diff --git a/.gitattributes b/.gitattributes
index 82ecfb28f18065f1e74571b41b1f05c9ba9b932e..b73ad4d1134d9fe3b743a9f3e0cb26d119af8f54 100644
--- a/.gitattributes
+++ b/.gitattributes
@@ -767,3 +767,10 @@ checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoin
 checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2736/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/tokenizer.json filter=lfs diff=lfs merge=lfs -text
diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1100/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1100/model-00001-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..ad0499a2dadbfb54d0f2be1bd1a82ac185b87a70
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1100/model-00001-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5ffc5ce6c42a363e63f1bf009b7e26493e6fa0a390407927f7ebe9e3b4ea05d
+size 4965798912
diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1200/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1200/model-00001-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..0693765a10867d021b1e21d4fd6702b810fbeb62
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1200/model-00001-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:02f302bcead71f8c5c4b816d585d501973ef83bb171ecf8e0ac202ae46539c04
+size 4965798912
diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1500/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1500/model-00001-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..57a4d99553e3bdd4d78239e1f1f2fa381d5ebc93
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1500/model-00001-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:81ab2c6922c2adf287ac6ddeb165bdca6d9e37c9637740c5b6f6d23cdda8b917
+size 4965798912
diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1700/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1700/model-00001-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..c3a22effc6f88f2e540add5aa5c5f6f856e44ee9
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-1700/model-00001-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd7a90faa91e1e485f69add60811c46b338fa7168c8c666b9acc817e3ac6dd47
+size 4965798912
diff --git a/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-600/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-600/model-00001-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..db3fdf4fc07b6f1da483e3e27d0870ce42508357
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_hop_tokens4_10M_seed0/runs/checkpoint-600/model-00001-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96ed7fbd257ca1caf26d32be423e7388ea50d9f4b95dec26e3de690709374fba
+size 4965798912
diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_partial_10M_seed0/runs/checkpoint-2500/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_reverse_partial_10M_seed0/runs/checkpoint-2500/model-00001-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..1e8a4823f3d7e6505a4e431876c973b705094d4e
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_reverse_partial_10M_seed0/runs/checkpoint-2500/model-00001-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82ff2f9bb29b7bfe764ca4f6ff30a7e88feffed064d7eb26b340eb7db979a819
+size 4965798912
diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_partial_10M_seed0/runs/checkpoint-2600/model-00001-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_reverse_partial_10M_seed0/runs/checkpoint-2600/model-00001-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..1044596c0cf3735b384bdbd7913ea0fe1d4ac9d4
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_reverse_partial_10M_seed0/runs/checkpoint-2600/model-00001-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6902bbe1baae896779c8f3b0a04742b8c3cced9339acda3f5a86208a2019b50
+size 4965798912
diff --git a/checkpoints/Llama-3.2-3B/babylm_reverse_partial_10M_seed0/runs/checkpoint-2900/global_step2900/mp_rank_00_model_states.pt b/checkpoints/Llama-3.2-3B/babylm_reverse_partial_10M_seed0/runs/checkpoint-2900/global_step2900/mp_rank_00_model_states.pt
new file mode 100644
index 0000000000000000000000000000000000000000..b8bcb6cb12869cd0e03dad87ea574bf8a91eab32
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_reverse_partial_10M_seed0/runs/checkpoint-2900/global_step2900/mp_rank_00_model_states.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2808d9c1c320add8f1fb29822244597874fe286f62a77ecf2a6eff4893ab4442
+size 6425576219
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/latest b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/latest
new file mode 100644
index 0000000000000000000000000000000000000000..744ae7dbad571b6f37ec6c7066549494261bb59e
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/latest
@@ -0,0 +1 @@
+global_step100
\ No newline at end of file
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/rng_state_0.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..44995716d9133cdde1b0eafab9b2cbbb3010cad2
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01b934e6ec558bac6fe169663502cde3d438344e9336e1e7d8afe651e9135428
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/rng_state_2.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..70cef6545138ddc2cc4b0c0ab18344cceb230f38
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:74c577a7e7176d95ccd3d6ec57d9ef0875615f3974af670a48262ddf57ea767d
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-100/zero_to_fp32.py
@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/config.json
@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "meta-llama/Llama-3.2-3B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "torch_dtype": "float16",
+  "transformers_version": "4.45.1",
+  "use_cache": true,
+  "vocab_size": 128256
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/generation_config.json
@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": 128001,
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.45.1"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/latest b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/latest
new file mode 100644
index 0000000000000000000000000000000000000000..4b10acccf3e8395339ff8799cea202bbc54d7f7d
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/latest
@@ -0,0 +1 @@
+global_step900
\ No newline at end of file
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/model.safetensors.index.json
new file mode 100644
index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/model.safetensors.index.json
@@ -0,0 +1,262 @@
+{
+  "metadata": {
+    "total_size": 7213504512
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00002-of-00002.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/rng_state_0.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..5a018a68bd5303fcda57dcb6093948dc236b697a
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34bfd29981738a04d7cb9b0b31f032749b5bd5fa1b120104d856a6f5b060af88
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/rng_state_1.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..85d20c3dd29d7f71371d2ce18d57829a2ebd22bc
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9e7ca9eed1d152be37852a8850dde9294edfaf37097b42c75c54b278336ef61
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/rng_state_2.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..137d87c143da2b66811890c8cf1e8b013ef6eacf
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62c45c6121ed4cfe9395ee5153d26d376e4f02f8dfa1bbceafb8a5416efcdd14
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/scheduler.pt b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..86063610918de03fe10ff5915f245b21a80f8845
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:977be84d836c6e1ba1cd1d60ad9c4b8ba9f4c0a63264746f477be2406b5629d7
+size 627
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/tokenizer.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb
+size 17210374
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128256": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|end_of_text|>",
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "[PAD]",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..febec8377abd19ba9f9f43a68f028fc1d9617e5f
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/trainer_state.json
@@ -0,0 +1,7053 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9863013698630136,
+  "eval_steps": 10,
+  "global_step": 900,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001095890410958904,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0566,
+      "step": 1
+    },
+    {
+      "epoch": 0.002191780821917808,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.045,
+      "step": 2
+    },
+    {
+      "epoch": 0.0032876712328767125,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0231,
+      "step": 3
+    },
+    {
+      "epoch": 0.004383561643835616,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0369,
+      "step": 4
+    },
+    {
+      "epoch": 0.005479452054794521,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0178,
+      "step": 5
+    },
+    {
+      "epoch": 0.006575342465753425,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.1366,
+      "step": 6
+    },
+    {
+      "epoch": 0.007671232876712329,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0555,
+      "step": 7
+    },
+    {
+      "epoch": 0.008767123287671232,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.094,
+      "step": 8
+    },
+    {
+      "epoch": 0.009863013698630137,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0471,
+      "step": 9
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.1006,
+      "step": 10
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "eval_loss": 3.0979692935943604,
+      "eval_runtime": 37.6015,
+      "eval_samples_per_second": 26.595,
+      "eval_steps_per_second": 1.117,
+      "step": 10
+    },
+    {
+      "epoch": 0.012054794520547946,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0849,
+      "step": 11
+    },
+    {
+      "epoch": 0.01315068493150685,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 2.9997,
+      "step": 12
+    },
+    {
+      "epoch": 0.014246575342465753,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0669,
+      "step": 13
+    },
+    {
+      "epoch": 0.015342465753424657,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0588,
+      "step": 14
+    },
+    {
+      "epoch": 0.01643835616438356,
+      "grad_norm": 5.574987411499023,
+      "learning_rate": 1.8248175182481753e-08,
+      "loss": 3.0983,
+      "step": 15
+    },
+    {
+      "epoch": 0.017534246575342465,
+      "grad_norm": 6.172506809234619,
+      "learning_rate": 3.6496350364963505e-08,
+      "loss": 3.0508,
+      "step": 16
+    },
+    {
+      "epoch": 0.01863013698630137,
+      "grad_norm": 6.234955787658691,
+      "learning_rate": 5.4744525547445264e-08,
+      "loss": 3.0467,
+      "step": 17
+    },
+    {
+      "epoch": 0.019726027397260273,
+      "grad_norm": 6.437706470489502,
+      "learning_rate": 7.299270072992701e-08,
+      "loss": 3.0846,
+      "step": 18
+    },
+    {
+      "epoch": 0.020821917808219178,
+      "grad_norm": 6.482471466064453,
+      "learning_rate": 9.124087591240876e-08,
+      "loss": 3.0289,
+      "step": 19
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "grad_norm": 5.701765060424805,
+      "learning_rate": 1.0948905109489053e-07,
+      "loss": 3.1008,
+      "step": 20
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "eval_loss": 3.096595287322998,
+      "eval_runtime": 37.2899,
+      "eval_samples_per_second": 26.817,
+      "eval_steps_per_second": 1.126,
+      "step": 20
+    },
+    {
+      "epoch": 0.023013698630136987,
+      "grad_norm": 6.222895622253418,
+      "learning_rate": 1.2773722627737227e-07,
+      "loss": 3.0819,
+      "step": 21
+    },
+    {
+      "epoch": 0.02410958904109589,
+      "grad_norm": 6.059528827667236,
+      "learning_rate": 1.4598540145985402e-07,
+      "loss": 3.0046,
+      "step": 22
+    },
+    {
+      "epoch": 0.025205479452054796,
+      "grad_norm": 5.951301574707031,
+      "learning_rate": 1.642335766423358e-07,
+      "loss": 3.0459,
+      "step": 23
+    },
+    {
+      "epoch": 0.0263013698630137,
+      "grad_norm": 5.824173927307129,
+      "learning_rate": 1.8248175182481753e-07,
+      "loss": 3.0823,
+      "step": 24
+    },
+    {
+      "epoch": 0.0273972602739726,
+      "grad_norm": 5.564741134643555,
+      "learning_rate": 2.0072992700729928e-07,
+      "loss": 3.0828,
+      "step": 25
+    },
+    {
+      "epoch": 0.028493150684931506,
+      "grad_norm": 4.622601509094238,
+      "learning_rate": 2.1897810218978106e-07,
+      "loss": 3.0587,
+      "step": 26
+    },
+    {
+      "epoch": 0.02958904109589041,
+      "grad_norm": 4.757779121398926,
+      "learning_rate": 2.3722627737226278e-07,
+      "loss": 3.0741,
+      "step": 27
+    },
+    {
+      "epoch": 0.030684931506849315,
+      "grad_norm": 4.821556568145752,
+      "learning_rate": 2.5547445255474454e-07,
+      "loss": 3.0245,
+      "step": 28
+    },
+    {
+      "epoch": 0.03178082191780822,
+      "grad_norm": 4.55632209777832,
+      "learning_rate": 2.737226277372263e-07,
+      "loss": 3.0195,
+      "step": 29
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "grad_norm": 4.604625701904297,
+      "learning_rate": 2.9197080291970804e-07,
+      "loss": 2.9946,
+      "step": 30
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "eval_loss": 3.0536928176879883,
+      "eval_runtime": 36.8045,
+      "eval_samples_per_second": 27.171,
+      "eval_steps_per_second": 1.141,
+      "step": 30
+    },
+    {
+      "epoch": 0.03397260273972603,
+      "grad_norm": 4.086721897125244,
+      "learning_rate": 3.102189781021898e-07,
+      "loss": 3.0171,
+      "step": 31
+    },
+    {
+      "epoch": 0.03506849315068493,
+      "grad_norm": 4.214878559112549,
+      "learning_rate": 3.284671532846716e-07,
+      "loss": 2.9961,
+      "step": 32
+    },
+    {
+      "epoch": 0.03616438356164384,
+      "grad_norm": 4.155057907104492,
+      "learning_rate": 3.467153284671533e-07,
+      "loss": 2.9751,
+      "step": 33
+    },
+    {
+      "epoch": 0.03726027397260274,
+      "grad_norm": 3.9290964603424072,
+      "learning_rate": 3.6496350364963505e-07,
+      "loss": 2.9689,
+      "step": 34
+    },
+    {
+      "epoch": 0.038356164383561646,
+      "grad_norm": 4.033408164978027,
+      "learning_rate": 3.8321167883211686e-07,
+      "loss": 2.9864,
+      "step": 35
+    },
+    {
+      "epoch": 0.03945205479452055,
+      "grad_norm": 3.602386474609375,
+      "learning_rate": 4.0145985401459856e-07,
+      "loss": 3.0125,
+      "step": 36
+    },
+    {
+      "epoch": 0.040547945205479455,
+      "grad_norm": 3.1355831623077393,
+      "learning_rate": 4.197080291970803e-07,
+      "loss": 2.9467,
+      "step": 37
+    },
+    {
+      "epoch": 0.041643835616438356,
+      "grad_norm": 3.539684295654297,
+      "learning_rate": 4.379562043795621e-07,
+      "loss": 2.9117,
+      "step": 38
+    },
+    {
+      "epoch": 0.042739726027397264,
+      "grad_norm": 3.389564037322998,
+      "learning_rate": 4.562043795620438e-07,
+      "loss": 3.0009,
+      "step": 39
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "grad_norm": 3.539494752883911,
+      "learning_rate": 4.7445255474452557e-07,
+      "loss": 2.9096,
+      "step": 40
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "eval_loss": 2.9791505336761475,
+      "eval_runtime": 36.749,
+      "eval_samples_per_second": 27.212,
+      "eval_steps_per_second": 1.143,
+      "step": 40
+    },
+    {
+      "epoch": 0.044931506849315066,
+      "grad_norm": 3.2061192989349365,
+      "learning_rate": 4.927007299270073e-07,
+      "loss": 2.9378,
+      "step": 41
+    },
+    {
+      "epoch": 0.046027397260273974,
+      "grad_norm": 2.979443311691284,
+      "learning_rate": 5.109489051094891e-07,
+      "loss": 2.9661,
+      "step": 42
+    },
+    {
+      "epoch": 0.047123287671232875,
+      "grad_norm": 3.0862207412719727,
+      "learning_rate": 5.291970802919708e-07,
+      "loss": 2.9306,
+      "step": 43
+    },
+    {
+      "epoch": 0.04821917808219178,
+      "grad_norm": 3.0377442836761475,
+      "learning_rate": 5.474452554744526e-07,
+      "loss": 2.9484,
+      "step": 44
+    },
+    {
+      "epoch": 0.049315068493150684,
+      "grad_norm": 2.934086799621582,
+      "learning_rate": 5.656934306569343e-07,
+      "loss": 2.9258,
+      "step": 45
+    },
+    {
+      "epoch": 0.05041095890410959,
+      "grad_norm": 2.9917893409729004,
+      "learning_rate": 5.839416058394161e-07,
+      "loss": 2.8798,
+      "step": 46
+    },
+    {
+      "epoch": 0.05150684931506849,
+      "grad_norm": 2.9639346599578857,
+      "learning_rate": 6.021897810218979e-07,
+      "loss": 2.8665,
+      "step": 47
+    },
+    {
+      "epoch": 0.0526027397260274,
+      "grad_norm": 2.7003400325775146,
+      "learning_rate": 6.204379562043796e-07,
+      "loss": 2.8879,
+      "step": 48
+    },
+    {
+      "epoch": 0.0536986301369863,
+      "grad_norm": 2.8495047092437744,
+      "learning_rate": 6.386861313868614e-07,
+      "loss": 2.8109,
+      "step": 49
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "grad_norm": 2.1759133338928223,
+      "learning_rate": 6.569343065693432e-07,
+      "loss": 2.929,
+      "step": 50
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "eval_loss": 2.891357421875,
+      "eval_runtime": 36.7855,
+      "eval_samples_per_second": 27.185,
+      "eval_steps_per_second": 1.142,
+      "step": 50
+    },
+    {
+      "epoch": 0.05589041095890411,
+      "grad_norm": 2.7880029678344727,
+      "learning_rate": 6.751824817518248e-07,
+      "loss": 2.8279,
+      "step": 51
+    },
+    {
+      "epoch": 0.05698630136986301,
+      "grad_norm": 2.3388028144836426,
+      "learning_rate": 6.934306569343066e-07,
+      "loss": 2.8474,
+      "step": 52
+    },
+    {
+      "epoch": 0.05808219178082192,
+      "grad_norm": 2.34242844581604,
+      "learning_rate": 7.116788321167883e-07,
+      "loss": 2.7775,
+      "step": 53
+    },
+    {
+      "epoch": 0.05917808219178082,
+      "grad_norm": 2.2962350845336914,
+      "learning_rate": 7.299270072992701e-07,
+      "loss": 2.7831,
+      "step": 54
+    },
+    {
+      "epoch": 0.06027397260273973,
+      "grad_norm": 2.353131055831909,
+      "learning_rate": 7.48175182481752e-07,
+      "loss": 2.7994,
+      "step": 55
+    },
+    {
+      "epoch": 0.06136986301369863,
+      "grad_norm": 2.1646981239318848,
+      "learning_rate": 7.664233576642337e-07,
+      "loss": 2.8155,
+      "step": 56
+    },
+    {
+      "epoch": 0.06246575342465754,
+      "grad_norm": 2.34967303276062,
+      "learning_rate": 7.846715328467154e-07,
+      "loss": 2.8154,
+      "step": 57
+    },
+    {
+      "epoch": 0.06356164383561644,
+      "grad_norm": 2.1275205612182617,
+      "learning_rate": 8.029197080291971e-07,
+      "loss": 2.7908,
+      "step": 58
+    },
+    {
+      "epoch": 0.06465753424657535,
+      "grad_norm": 1.89387047290802,
+      "learning_rate": 8.211678832116789e-07,
+      "loss": 2.7442,
+      "step": 59
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "grad_norm": 1.8902339935302734,
+      "learning_rate": 8.394160583941606e-07,
+      "loss": 2.7103,
+      "step": 60
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "eval_loss": 2.8224520683288574,
+      "eval_runtime": 36.7623,
+      "eval_samples_per_second": 27.202,
+      "eval_steps_per_second": 1.142,
+      "step": 60
+    },
+    {
+      "epoch": 0.06684931506849315,
+      "grad_norm": 2.1961512565612793,
+      "learning_rate": 8.576642335766425e-07,
+      "loss": 2.7477,
+      "step": 61
+    },
+    {
+      "epoch": 0.06794520547945206,
+      "grad_norm": 1.8918277025222778,
+      "learning_rate": 8.759124087591242e-07,
+      "loss": 2.7376,
+      "step": 62
+    },
+    {
+      "epoch": 0.06904109589041096,
+      "grad_norm": 1.9290616512298584,
+      "learning_rate": 8.94160583941606e-07,
+      "loss": 2.7539,
+      "step": 63
+    },
+    {
+      "epoch": 0.07013698630136986,
+      "grad_norm": 1.7825456857681274,
+      "learning_rate": 9.124087591240876e-07,
+      "loss": 2.7609,
+      "step": 64
+    },
+    {
+      "epoch": 0.07123287671232877,
+      "grad_norm": 2.140242576599121,
+      "learning_rate": 9.306569343065694e-07,
+      "loss": 2.7161,
+      "step": 65
+    },
+    {
+      "epoch": 0.07232876712328767,
+      "grad_norm": 2.2971854209899902,
+      "learning_rate": 9.489051094890511e-07,
+      "loss": 2.7358,
+      "step": 66
+    },
+    {
+      "epoch": 0.07342465753424658,
+      "grad_norm": 1.9837063550949097,
+      "learning_rate": 9.67153284671533e-07,
+      "loss": 2.7235,
+      "step": 67
+    },
+    {
+      "epoch": 0.07452054794520548,
+      "grad_norm": 2.1990857124328613,
+      "learning_rate": 9.854014598540146e-07,
+      "loss": 2.771,
+      "step": 68
+    },
+    {
+      "epoch": 0.07561643835616438,
+      "grad_norm": 2.3288135528564453,
+      "learning_rate": 1.0036496350364965e-06,
+      "loss": 2.6975,
+      "step": 69
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "grad_norm": 2.0509581565856934,
+      "learning_rate": 1.0218978102189781e-06,
+      "loss": 2.6627,
+      "step": 70
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "eval_loss": 2.7516250610351562,
+      "eval_runtime": 36.9078,
+      "eval_samples_per_second": 27.095,
+      "eval_steps_per_second": 1.138,
+      "step": 70
+    },
+    {
+      "epoch": 0.07780821917808219,
+      "grad_norm": 2.070568561553955,
+      "learning_rate": 1.04014598540146e-06,
+      "loss": 2.7297,
+      "step": 71
+    },
+    {
+      "epoch": 0.0789041095890411,
+      "grad_norm": 2.5551228523254395,
+      "learning_rate": 1.0583941605839416e-06,
+      "loss": 2.6152,
+      "step": 72
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 2.060121536254883,
+      "learning_rate": 1.0766423357664233e-06,
+      "loss": 2.6144,
+      "step": 73
+    },
+    {
+      "epoch": 0.08109589041095891,
+      "grad_norm": 2.783064365386963,
+      "learning_rate": 1.0948905109489052e-06,
+      "loss": 2.6374,
+      "step": 74
+    },
+    {
+      "epoch": 0.0821917808219178,
+      "grad_norm": 2.765652894973755,
+      "learning_rate": 1.113138686131387e-06,
+      "loss": 2.661,
+      "step": 75
+    },
+    {
+      "epoch": 0.08328767123287671,
+      "grad_norm": 2.0037786960601807,
+      "learning_rate": 1.1313868613138687e-06,
+      "loss": 2.666,
+      "step": 76
+    },
+    {
+      "epoch": 0.08438356164383562,
+      "grad_norm": 2.9089527130126953,
+      "learning_rate": 1.1496350364963505e-06,
+      "loss": 2.5964,
+      "step": 77
+    },
+    {
+      "epoch": 0.08547945205479453,
+      "grad_norm": 2.1275577545166016,
+      "learning_rate": 1.1678832116788322e-06,
+      "loss": 2.5754,
+      "step": 78
+    },
+    {
+      "epoch": 0.08657534246575342,
+      "grad_norm": 2.3247408866882324,
+      "learning_rate": 1.1861313868613138e-06,
+      "loss": 2.585,
+      "step": 79
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "grad_norm": 2.5372109413146973,
+      "learning_rate": 1.2043795620437959e-06,
+      "loss": 2.5757,
+      "step": 80
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "eval_loss": 2.67269229888916,
+      "eval_runtime": 36.9427,
+      "eval_samples_per_second": 27.069,
+      "eval_steps_per_second": 1.137,
+      "step": 80
+    },
+    {
+      "epoch": 0.08876712328767124,
+      "grad_norm": 2.0754737854003906,
+      "learning_rate": 1.2226277372262775e-06,
+      "loss": 2.581,
+      "step": 81
+    },
+    {
+      "epoch": 0.08986301369863013,
+      "grad_norm": 2.7675907611846924,
+      "learning_rate": 1.2408759124087592e-06,
+      "loss": 2.6194,
+      "step": 82
+    },
+    {
+      "epoch": 0.09095890410958904,
+      "grad_norm": 2.2297277450561523,
+      "learning_rate": 1.259124087591241e-06,
+      "loss": 2.5604,
+      "step": 83
+    },
+    {
+      "epoch": 0.09205479452054795,
+      "grad_norm": 2.3006038665771484,
+      "learning_rate": 1.2773722627737229e-06,
+      "loss": 2.5737,
+      "step": 84
+    },
+    {
+      "epoch": 0.09315068493150686,
+      "grad_norm": 2.127249002456665,
+      "learning_rate": 1.2956204379562045e-06,
+      "loss": 2.5599,
+      "step": 85
+    },
+    {
+      "epoch": 0.09424657534246575,
+      "grad_norm": 2.4705958366394043,
+      "learning_rate": 1.3138686131386864e-06,
+      "loss": 2.5135,
+      "step": 86
+    },
+    {
+      "epoch": 0.09534246575342466,
+      "grad_norm": 2.6609578132629395,
+      "learning_rate": 1.332116788321168e-06,
+      "loss": 2.5267,
+      "step": 87
+    },
+    {
+      "epoch": 0.09643835616438357,
+      "grad_norm": 2.690622091293335,
+      "learning_rate": 1.3503649635036497e-06,
+      "loss": 2.4526,
+      "step": 88
+    },
+    {
+      "epoch": 0.09753424657534246,
+      "grad_norm": 2.703291893005371,
+      "learning_rate": 1.3686131386861315e-06,
+      "loss": 2.4935,
+      "step": 89
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "grad_norm": 2.3661017417907715,
+      "learning_rate": 1.3868613138686132e-06,
+      "loss": 2.425,
+      "step": 90
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "eval_loss": 2.5889577865600586,
+      "eval_runtime": 36.912,
+      "eval_samples_per_second": 27.091,
+      "eval_steps_per_second": 1.138,
+      "step": 90
+    },
+    {
+      "epoch": 0.09972602739726028,
+      "grad_norm": 2.596808910369873,
+      "learning_rate": 1.405109489051095e-06,
+      "loss": 2.5244,
+      "step": 91
+    },
+    {
+      "epoch": 0.10082191780821918,
+      "grad_norm": 2.5526316165924072,
+      "learning_rate": 1.4233576642335767e-06,
+      "loss": 2.4803,
+      "step": 92
+    },
+    {
+      "epoch": 0.10191780821917808,
+      "grad_norm": 2.4569883346557617,
+      "learning_rate": 1.4416058394160583e-06,
+      "loss": 2.5149,
+      "step": 93
+    },
+    {
+      "epoch": 0.10301369863013699,
+      "grad_norm": 2.9641432762145996,
+      "learning_rate": 1.4598540145985402e-06,
+      "loss": 2.4616,
+      "step": 94
+    },
+    {
+      "epoch": 0.10410958904109589,
+      "grad_norm": 2.821410894393921,
+      "learning_rate": 1.478102189781022e-06,
+      "loss": 2.429,
+      "step": 95
+    },
+    {
+      "epoch": 0.1052054794520548,
+      "grad_norm": 3.8620495796203613,
+      "learning_rate": 1.496350364963504e-06,
+      "loss": 2.5028,
+      "step": 96
+    },
+    {
+      "epoch": 0.1063013698630137,
+      "grad_norm": 3.0422098636627197,
+      "learning_rate": 1.5145985401459856e-06,
+      "loss": 2.4631,
+      "step": 97
+    },
+    {
+      "epoch": 0.1073972602739726,
+      "grad_norm": 2.8717942237854004,
+      "learning_rate": 1.5328467153284674e-06,
+      "loss": 2.4786,
+      "step": 98
+    },
+    {
+      "epoch": 0.10849315068493151,
+      "grad_norm": 2.6856589317321777,
+      "learning_rate": 1.551094890510949e-06,
+      "loss": 2.4331,
+      "step": 99
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "grad_norm": 4.284306526184082,
+      "learning_rate": 1.5693430656934307e-06,
+      "loss": 2.3742,
+      "step": 100
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "eval_loss": 2.524946689605713,
+      "eval_runtime": 37.1044,
+      "eval_samples_per_second": 26.951,
+      "eval_steps_per_second": 1.132,
+      "step": 100
+    },
+    {
+      "epoch": 0.11068493150684931,
+      "grad_norm": 4.973050117492676,
+      "learning_rate": 1.5875912408759126e-06,
+      "loss": 2.3809,
+      "step": 101
+    },
+    {
+      "epoch": 0.11178082191780822,
+      "grad_norm": 2.808378219604492,
+      "learning_rate": 1.6058394160583942e-06,
+      "loss": 2.4314,
+      "step": 102
+    },
+    {
+      "epoch": 0.11287671232876713,
+      "grad_norm": 3.9805150032043457,
+      "learning_rate": 1.624087591240876e-06,
+      "loss": 2.4848,
+      "step": 103
+    },
+    {
+      "epoch": 0.11397260273972602,
+      "grad_norm": 2.7861433029174805,
+      "learning_rate": 1.6423357664233577e-06,
+      "loss": 2.3981,
+      "step": 104
+    },
+    {
+      "epoch": 0.11506849315068493,
+      "grad_norm": 3.4071974754333496,
+      "learning_rate": 1.6605839416058394e-06,
+      "loss": 2.3422,
+      "step": 105
+    },
+    {
+      "epoch": 0.11616438356164384,
+      "grad_norm": 2.4880754947662354,
+      "learning_rate": 1.6788321167883212e-06,
+      "loss": 2.3764,
+      "step": 106
+    },
+    {
+      "epoch": 0.11726027397260275,
+      "grad_norm": 3.5724127292633057,
+      "learning_rate": 1.6970802919708029e-06,
+      "loss": 2.3969,
+      "step": 107
+    },
+    {
+      "epoch": 0.11835616438356164,
+      "grad_norm": 2.5322089195251465,
+      "learning_rate": 1.715328467153285e-06,
+      "loss": 2.3188,
+      "step": 108
+    },
+    {
+      "epoch": 0.11945205479452055,
+      "grad_norm": 3.194002628326416,
+      "learning_rate": 1.7335766423357666e-06,
+      "loss": 2.3672,
+      "step": 109
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "grad_norm": 2.489720344543457,
+      "learning_rate": 1.7518248175182485e-06,
+      "loss": 2.3749,
+      "step": 110
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "eval_loss": 2.4652323722839355,
+      "eval_runtime": 36.9551,
+      "eval_samples_per_second": 27.06,
+      "eval_steps_per_second": 1.137,
+      "step": 110
+    },
+    {
+      "epoch": 0.12164383561643835,
+      "grad_norm": 2.7961671352386475,
+      "learning_rate": 1.77007299270073e-06,
+      "loss": 2.4069,
+      "step": 111
+    },
+    {
+      "epoch": 0.12273972602739726,
+      "grad_norm": 2.6191494464874268,
+      "learning_rate": 1.788321167883212e-06,
+      "loss": 2.3764,
+      "step": 112
+    },
+    {
+      "epoch": 0.12383561643835617,
+      "grad_norm": 2.5566232204437256,
+      "learning_rate": 1.8065693430656936e-06,
+      "loss": 2.4182,
+      "step": 113
+    },
+    {
+      "epoch": 0.12493150684931507,
+      "grad_norm": 2.5462262630462646,
+      "learning_rate": 1.8248175182481753e-06,
+      "loss": 2.3665,
+      "step": 114
+    },
+    {
+      "epoch": 0.12602739726027398,
+      "grad_norm": 2.6968109607696533,
+      "learning_rate": 1.8430656934306571e-06,
+      "loss": 2.3045,
+      "step": 115
+    },
+    {
+      "epoch": 0.12712328767123288,
+      "grad_norm": 2.838625192642212,
+      "learning_rate": 1.8613138686131388e-06,
+      "loss": 2.3332,
+      "step": 116
+    },
+    {
+      "epoch": 0.12821917808219177,
+      "grad_norm": 2.584407091140747,
+      "learning_rate": 1.8795620437956206e-06,
+      "loss": 2.3204,
+      "step": 117
+    },
+    {
+      "epoch": 0.1293150684931507,
+      "grad_norm": 2.5297932624816895,
+      "learning_rate": 1.8978102189781023e-06,
+      "loss": 2.312,
+      "step": 118
+    },
+    {
+      "epoch": 0.1304109589041096,
+      "grad_norm": 2.6709144115448,
+      "learning_rate": 1.916058394160584e-06,
+      "loss": 2.3271,
+      "step": 119
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "grad_norm": 2.6457643508911133,
+      "learning_rate": 1.934306569343066e-06,
+      "loss": 2.3415,
+      "step": 120
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "eval_loss": 2.4270987510681152,
+      "eval_runtime": 36.969,
+      "eval_samples_per_second": 27.05,
+      "eval_steps_per_second": 1.136,
+      "step": 120
+    },
+    {
+      "epoch": 0.1326027397260274,
+      "grad_norm": 3.8355023860931396,
+      "learning_rate": 1.952554744525548e-06,
+      "loss": 2.2405,
+      "step": 121
+    },
+    {
+      "epoch": 0.1336986301369863,
+      "grad_norm": 2.6029245853424072,
+      "learning_rate": 1.9708029197080293e-06,
+      "loss": 2.2841,
+      "step": 122
+    },
+    {
+      "epoch": 0.13479452054794522,
+      "grad_norm": 3.9318411350250244,
+      "learning_rate": 1.989051094890511e-06,
+      "loss": 2.2975,
+      "step": 123
+    },
+    {
+      "epoch": 0.1358904109589041,
+      "grad_norm": 4.0954461097717285,
+      "learning_rate": 2.007299270072993e-06,
+      "loss": 2.2963,
+      "step": 124
+    },
+    {
+      "epoch": 0.136986301369863,
+      "grad_norm": 2.4977500438690186,
+      "learning_rate": 2.025547445255475e-06,
+      "loss": 2.3526,
+      "step": 125
+    },
+    {
+      "epoch": 0.13808219178082193,
+      "grad_norm": 4.252346992492676,
+      "learning_rate": 2.0437956204379563e-06,
+      "loss": 2.2856,
+      "step": 126
+    },
+    {
+      "epoch": 0.13917808219178082,
+      "grad_norm": 2.8565123081207275,
+      "learning_rate": 2.062043795620438e-06,
+      "loss": 2.294,
+      "step": 127
+    },
+    {
+      "epoch": 0.14027397260273972,
+      "grad_norm": 3.1502203941345215,
+      "learning_rate": 2.08029197080292e-06,
+      "loss": 2.2817,
+      "step": 128
+    },
+    {
+      "epoch": 0.14136986301369864,
+      "grad_norm": 3.80749249458313,
+      "learning_rate": 2.0985401459854014e-06,
+      "loss": 2.3483,
+      "step": 129
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "grad_norm": 3.1611809730529785,
+      "learning_rate": 2.1167883211678833e-06,
+      "loss": 2.2368,
+      "step": 130
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "eval_loss": 2.3936448097229004,
+      "eval_runtime": 36.8674,
+      "eval_samples_per_second": 27.124,
+      "eval_steps_per_second": 1.139,
+      "step": 130
+    },
+    {
+      "epoch": 0.14356164383561643,
+      "grad_norm": 3.130007028579712,
+      "learning_rate": 2.135036496350365e-06,
+      "loss": 2.3034,
+      "step": 131
+    },
+    {
+      "epoch": 0.14465753424657535,
+      "grad_norm": 3.0841190814971924,
+      "learning_rate": 2.1532846715328466e-06,
+      "loss": 2.3314,
+      "step": 132
+    },
+    {
+      "epoch": 0.14575342465753424,
+      "grad_norm": 2.8934333324432373,
+      "learning_rate": 2.171532846715329e-06,
+      "loss": 2.2322,
+      "step": 133
+    },
+    {
+      "epoch": 0.14684931506849316,
+      "grad_norm": 2.8835387229919434,
+      "learning_rate": 2.1897810218978103e-06,
+      "loss": 2.2449,
+      "step": 134
+    },
+    {
+      "epoch": 0.14794520547945206,
+      "grad_norm": 3.428131103515625,
+      "learning_rate": 2.208029197080292e-06,
+      "loss": 2.2861,
+      "step": 135
+    },
+    {
+      "epoch": 0.14904109589041095,
+      "grad_norm": 2.794792652130127,
+      "learning_rate": 2.226277372262774e-06,
+      "loss": 2.2611,
+      "step": 136
+    },
+    {
+      "epoch": 0.15013698630136987,
+      "grad_norm": 3.3428287506103516,
+      "learning_rate": 2.244525547445256e-06,
+      "loss": 2.2668,
+      "step": 137
+    },
+    {
+      "epoch": 0.15123287671232877,
+      "grad_norm": 3.4636261463165283,
+      "learning_rate": 2.2627737226277373e-06,
+      "loss": 2.2943,
+      "step": 138
+    },
+    {
+      "epoch": 0.15232876712328766,
+      "grad_norm": 2.7653632164001465,
+      "learning_rate": 2.281021897810219e-06,
+      "loss": 2.2869,
+      "step": 139
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "grad_norm": 3.011944532394409,
+      "learning_rate": 2.299270072992701e-06,
+      "loss": 2.2447,
+      "step": 140
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "eval_loss": 2.370593547821045,
+      "eval_runtime": 36.8851,
+      "eval_samples_per_second": 27.111,
+      "eval_steps_per_second": 1.139,
+      "step": 140
+    },
+    {
+      "epoch": 0.15452054794520548,
+      "grad_norm": 3.198716640472412,
+      "learning_rate": 2.3175182481751825e-06,
+      "loss": 2.2391,
+      "step": 141
+    },
+    {
+      "epoch": 0.15561643835616437,
+      "grad_norm": 2.816667079925537,
+      "learning_rate": 2.3357664233576643e-06,
+      "loss": 2.2547,
+      "step": 142
+    },
+    {
+      "epoch": 0.1567123287671233,
+      "grad_norm": 2.6860835552215576,
+      "learning_rate": 2.354014598540146e-06,
+      "loss": 2.2036,
+      "step": 143
+    },
+    {
+      "epoch": 0.1578082191780822,
+      "grad_norm": 2.654677152633667,
+      "learning_rate": 2.3722627737226276e-06,
+      "loss": 2.2461,
+      "step": 144
+    },
+    {
+      "epoch": 0.1589041095890411,
+      "grad_norm": 2.617124080657959,
+      "learning_rate": 2.39051094890511e-06,
+      "loss": 2.2565,
+      "step": 145
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 3.1751718521118164,
+      "learning_rate": 2.4087591240875918e-06,
+      "loss": 2.1914,
+      "step": 146
+    },
+    {
+      "epoch": 0.1610958904109589,
+      "grad_norm": 2.6073708534240723,
+      "learning_rate": 2.427007299270073e-06,
+      "loss": 2.2179,
+      "step": 147
+    },
+    {
+      "epoch": 0.16219178082191782,
+      "grad_norm": 3.581951141357422,
+      "learning_rate": 2.445255474452555e-06,
+      "loss": 2.2254,
+      "step": 148
+    },
+    {
+      "epoch": 0.1632876712328767,
+      "grad_norm": 3.2456817626953125,
+      "learning_rate": 2.463503649635037e-06,
+      "loss": 2.2691,
+      "step": 149
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "grad_norm": 3.040400505065918,
+      "learning_rate": 2.4817518248175183e-06,
+      "loss": 2.1837,
+      "step": 150
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "eval_loss": 2.3496975898742676,
+      "eval_runtime": 36.8324,
+      "eval_samples_per_second": 27.15,
+      "eval_steps_per_second": 1.14,
+      "step": 150
+    },
+    {
+      "epoch": 0.16547945205479453,
+      "grad_norm": 3.313336133956909,
+      "learning_rate": 2.5e-06,
+      "loss": 2.1501,
+      "step": 151
+    },
+    {
+      "epoch": 0.16657534246575342,
+      "grad_norm": 3.2824361324310303,
+      "learning_rate": 2.518248175182482e-06,
+      "loss": 2.2836,
+      "step": 152
+    },
+    {
+      "epoch": 0.16767123287671232,
+      "grad_norm": 2.8979086875915527,
+      "learning_rate": 2.5364963503649635e-06,
+      "loss": 2.1755,
+      "step": 153
+    },
+    {
+      "epoch": 0.16876712328767124,
+      "grad_norm": 2.945603609085083,
+      "learning_rate": 2.5547445255474458e-06,
+      "loss": 2.1791,
+      "step": 154
+    },
+    {
+      "epoch": 0.16986301369863013,
+      "grad_norm": 2.6949260234832764,
+      "learning_rate": 2.5729927007299272e-06,
+      "loss": 2.2338,
+      "step": 155
+    },
+    {
+      "epoch": 0.17095890410958905,
+      "grad_norm": 3.3816075325012207,
+      "learning_rate": 2.591240875912409e-06,
+      "loss": 2.2067,
+      "step": 156
+    },
+    {
+      "epoch": 0.17205479452054795,
+      "grad_norm": 2.516245126724243,
+      "learning_rate": 2.6094890510948905e-06,
+      "loss": 2.2064,
+      "step": 157
+    },
+    {
+      "epoch": 0.17315068493150684,
+      "grad_norm": 3.5714497566223145,
+      "learning_rate": 2.627737226277373e-06,
+      "loss": 2.2417,
+      "step": 158
+    },
+    {
+      "epoch": 0.17424657534246576,
+      "grad_norm": 2.8084025382995605,
+      "learning_rate": 2.6459854014598542e-06,
+      "loss": 2.244,
+      "step": 159
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "grad_norm": 2.94307541847229,
+      "learning_rate": 2.664233576642336e-06,
+      "loss": 2.2036,
+      "step": 160
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "eval_loss": 2.332892417907715,
+      "eval_runtime": 36.8624,
+      "eval_samples_per_second": 27.128,
+      "eval_steps_per_second": 1.139,
+      "step": 160
+    },
+    {
+      "epoch": 0.17643835616438355,
+      "grad_norm": 3.5338687896728516,
+      "learning_rate": 2.6824817518248175e-06,
+      "loss": 2.1825,
+      "step": 161
+    },
+    {
+      "epoch": 0.17753424657534247,
+      "grad_norm": 2.6816883087158203,
+      "learning_rate": 2.7007299270072994e-06,
+      "loss": 2.1991,
+      "step": 162
+    },
+    {
+      "epoch": 0.17863013698630137,
+      "grad_norm": 2.9373624324798584,
+      "learning_rate": 2.718978102189781e-06,
+      "loss": 2.1771,
+      "step": 163
+    },
+    {
+      "epoch": 0.17972602739726026,
+      "grad_norm": 3.060258388519287,
+      "learning_rate": 2.737226277372263e-06,
+      "loss": 2.2348,
+      "step": 164
+    },
+    {
+      "epoch": 0.18082191780821918,
+      "grad_norm": 2.9374494552612305,
+      "learning_rate": 2.755474452554745e-06,
+      "loss": 2.2295,
+      "step": 165
+    },
+    {
+      "epoch": 0.18191780821917808,
+      "grad_norm": 2.9218924045562744,
+      "learning_rate": 2.7737226277372264e-06,
+      "loss": 2.1743,
+      "step": 166
+    },
+    {
+      "epoch": 0.18301369863013697,
+      "grad_norm": 2.7671658992767334,
+      "learning_rate": 2.7919708029197082e-06,
+      "loss": 2.2821,
+      "step": 167
+    },
+    {
+      "epoch": 0.1841095890410959,
+      "grad_norm": 2.830994129180908,
+      "learning_rate": 2.81021897810219e-06,
+      "loss": 2.2369,
+      "step": 168
+    },
+    {
+      "epoch": 0.1852054794520548,
+      "grad_norm": 2.4539380073547363,
+      "learning_rate": 2.828467153284672e-06,
+      "loss": 2.1718,
+      "step": 169
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "grad_norm": 2.529789686203003,
+      "learning_rate": 2.8467153284671534e-06,
+      "loss": 2.2034,
+      "step": 170
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "eval_loss": 2.3183443546295166,
+      "eval_runtime": 36.9034,
+      "eval_samples_per_second": 27.098,
+      "eval_steps_per_second": 1.138,
+      "step": 170
+    },
+    {
+      "epoch": 0.1873972602739726,
+      "grad_norm": 2.776603937149048,
+      "learning_rate": 2.8649635036496353e-06,
+      "loss": 2.1205,
+      "step": 171
+    },
+    {
+      "epoch": 0.1884931506849315,
+      "grad_norm": 2.6126351356506348,
+      "learning_rate": 2.8832116788321167e-06,
+      "loss": 2.1788,
+      "step": 172
+    },
+    {
+      "epoch": 0.18958904109589042,
+      "grad_norm": 2.5450496673583984,
+      "learning_rate": 2.901459854014599e-06,
+      "loss": 2.1306,
+      "step": 173
+    },
+    {
+      "epoch": 0.19068493150684931,
+      "grad_norm": 2.5038654804229736,
+      "learning_rate": 2.9197080291970804e-06,
+      "loss": 2.2277,
+      "step": 174
+    },
+    {
+      "epoch": 0.1917808219178082,
+      "grad_norm": 2.7198233604431152,
+      "learning_rate": 2.9379562043795623e-06,
+      "loss": 2.2078,
+      "step": 175
+    },
+    {
+      "epoch": 0.19287671232876713,
+      "grad_norm": 3.0290682315826416,
+      "learning_rate": 2.956204379562044e-06,
+      "loss": 2.1902,
+      "step": 176
+    },
+    {
+      "epoch": 0.19397260273972602,
+      "grad_norm": 2.661337375640869,
+      "learning_rate": 2.9744525547445256e-06,
+      "loss": 2.236,
+      "step": 177
+    },
+    {
+      "epoch": 0.19506849315068492,
+      "grad_norm": 2.7548766136169434,
+      "learning_rate": 2.992700729927008e-06,
+      "loss": 2.1875,
+      "step": 178
+    },
+    {
+      "epoch": 0.19616438356164384,
+      "grad_norm": 3.7520737648010254,
+      "learning_rate": 3.0109489051094893e-06,
+      "loss": 2.1447,
+      "step": 179
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "grad_norm": 3.4713258743286133,
+      "learning_rate": 3.029197080291971e-06,
+      "loss": 2.1503,
+      "step": 180
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "eval_loss": 2.3034958839416504,
+      "eval_runtime": 36.9709,
+      "eval_samples_per_second": 27.048,
+      "eval_steps_per_second": 1.136,
+      "step": 180
+    },
+    {
+      "epoch": 0.19835616438356166,
+      "grad_norm": 2.3800981044769287,
+      "learning_rate": 3.0474452554744526e-06,
+      "loss": 2.2199,
+      "step": 181
+    },
+    {
+      "epoch": 0.19945205479452055,
+      "grad_norm": 2.623415946960449,
+      "learning_rate": 3.065693430656935e-06,
+      "loss": 2.2165,
+      "step": 182
+    },
+    {
+      "epoch": 0.20054794520547944,
+      "grad_norm": 2.571760416030884,
+      "learning_rate": 3.0839416058394163e-06,
+      "loss": 2.1962,
+      "step": 183
+    },
+    {
+      "epoch": 0.20164383561643837,
+      "grad_norm": 2.6511805057525635,
+      "learning_rate": 3.102189781021898e-06,
+      "loss": 2.2164,
+      "step": 184
+    },
+    {
+      "epoch": 0.20273972602739726,
+      "grad_norm": 2.8188459873199463,
+      "learning_rate": 3.1204379562043796e-06,
+      "loss": 2.1988,
+      "step": 185
+    },
+    {
+      "epoch": 0.20383561643835615,
+      "grad_norm": 2.6241681575775146,
+      "learning_rate": 3.1386861313868614e-06,
+      "loss": 2.1246,
+      "step": 186
+    },
+    {
+      "epoch": 0.20493150684931508,
+      "grad_norm": 2.674527645111084,
+      "learning_rate": 3.156934306569343e-06,
+      "loss": 2.1162,
+      "step": 187
+    },
+    {
+      "epoch": 0.20602739726027397,
+      "grad_norm": 2.8179783821105957,
+      "learning_rate": 3.175182481751825e-06,
+      "loss": 2.2266,
+      "step": 188
+    },
+    {
+      "epoch": 0.20712328767123286,
+      "grad_norm": 2.5623691082000732,
+      "learning_rate": 3.193430656934307e-06,
+      "loss": 2.1767,
+      "step": 189
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "grad_norm": 2.6922426223754883,
+      "learning_rate": 3.2116788321167884e-06,
+      "loss": 2.1556,
+      "step": 190
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "eval_loss": 2.2934091091156006,
+      "eval_runtime": 36.9653,
+      "eval_samples_per_second": 27.052,
+      "eval_steps_per_second": 1.136,
+      "step": 190
+    },
+    {
+      "epoch": 0.20931506849315068,
+      "grad_norm": 2.670802116394043,
+      "learning_rate": 3.2299270072992707e-06,
+      "loss": 2.1701,
+      "step": 191
+    },
+    {
+      "epoch": 0.2104109589041096,
+      "grad_norm": 2.7782397270202637,
+      "learning_rate": 3.248175182481752e-06,
+      "loss": 2.1563,
+      "step": 192
+    },
+    {
+      "epoch": 0.2115068493150685,
+      "grad_norm": 2.993387460708618,
+      "learning_rate": 3.266423357664234e-06,
+      "loss": 2.0815,
+      "step": 193
+    },
+    {
+      "epoch": 0.2126027397260274,
+      "grad_norm": 2.964684009552002,
+      "learning_rate": 3.2846715328467155e-06,
+      "loss": 2.1791,
+      "step": 194
+    },
+    {
+      "epoch": 0.2136986301369863,
+      "grad_norm": 2.738447427749634,
+      "learning_rate": 3.3029197080291973e-06,
+      "loss": 2.1535,
+      "step": 195
+    },
+    {
+      "epoch": 0.2147945205479452,
+      "grad_norm": 3.247713804244995,
+      "learning_rate": 3.3211678832116788e-06,
+      "loss": 2.1879,
+      "step": 196
+    },
+    {
+      "epoch": 0.2158904109589041,
+      "grad_norm": 3.095205783843994,
+      "learning_rate": 3.339416058394161e-06,
+      "loss": 2.1647,
+      "step": 197
+    },
+    {
+      "epoch": 0.21698630136986302,
+      "grad_norm": 2.6435153484344482,
+      "learning_rate": 3.3576642335766425e-06,
+      "loss": 2.1583,
+      "step": 198
+    },
+    {
+      "epoch": 0.21808219178082192,
+      "grad_norm": 3.1337387561798096,
+      "learning_rate": 3.3759124087591243e-06,
+      "loss": 2.1569,
+      "step": 199
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "grad_norm": 3.172367572784424,
+      "learning_rate": 3.3941605839416058e-06,
+      "loss": 2.1633,
+      "step": 200
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "eval_loss": 2.279322862625122,
+      "eval_runtime": 37.1254,
+      "eval_samples_per_second": 26.936,
+      "eval_steps_per_second": 1.131,
+      "step": 200
+    },
+    {
+      "epoch": 0.22027397260273973,
+      "grad_norm": 2.533245801925659,
+      "learning_rate": 3.412408759124088e-06,
+      "loss": 2.1633,
+      "step": 201
+    },
+    {
+      "epoch": 0.22136986301369863,
+      "grad_norm": 3.0553905963897705,
+      "learning_rate": 3.43065693430657e-06,
+      "loss": 2.1687,
+      "step": 202
+    },
+    {
+      "epoch": 0.22246575342465755,
+      "grad_norm": 3.2229137420654297,
+      "learning_rate": 3.4489051094890513e-06,
+      "loss": 2.1064,
+      "step": 203
+    },
+    {
+      "epoch": 0.22356164383561644,
+      "grad_norm": 2.752973794937134,
+      "learning_rate": 3.467153284671533e-06,
+      "loss": 2.1584,
+      "step": 204
+    },
+    {
+      "epoch": 0.22465753424657534,
+      "grad_norm": 2.75563645362854,
+      "learning_rate": 3.4854014598540146e-06,
+      "loss": 2.1016,
+      "step": 205
+    },
+    {
+      "epoch": 0.22575342465753426,
+      "grad_norm": 2.480123281478882,
+      "learning_rate": 3.503649635036497e-06,
+      "loss": 2.143,
+      "step": 206
+    },
+    {
+      "epoch": 0.22684931506849315,
+      "grad_norm": 2.7256920337677,
+      "learning_rate": 3.5218978102189783e-06,
+      "loss": 2.1388,
+      "step": 207
+    },
+    {
+      "epoch": 0.22794520547945205,
+      "grad_norm": 2.680922746658325,
+      "learning_rate": 3.54014598540146e-06,
+      "loss": 2.1712,
+      "step": 208
+    },
+    {
+      "epoch": 0.22904109589041097,
+      "grad_norm": 3.1483004093170166,
+      "learning_rate": 3.5583941605839416e-06,
+      "loss": 2.1807,
+      "step": 209
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "grad_norm": 2.7536072731018066,
+      "learning_rate": 3.576642335766424e-06,
+      "loss": 2.23,
+      "step": 210
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "eval_loss": 2.2728517055511475,
+      "eval_runtime": 36.975,
+      "eval_samples_per_second": 27.045,
+      "eval_steps_per_second": 1.136,
+      "step": 210
+    },
+    {
+      "epoch": 0.23123287671232876,
+      "grad_norm": 2.9679646492004395,
+      "learning_rate": 3.5948905109489054e-06,
+      "loss": 2.211,
+      "step": 211
+    },
+    {
+      "epoch": 0.23232876712328768,
+      "grad_norm": 2.755133628845215,
+      "learning_rate": 3.6131386861313872e-06,
+      "loss": 2.0679,
+      "step": 212
+    },
+    {
+      "epoch": 0.23342465753424657,
+      "grad_norm": 3.1508214473724365,
+      "learning_rate": 3.6313868613138687e-06,
+      "loss": 2.064,
+      "step": 213
+    },
+    {
+      "epoch": 0.2345205479452055,
+      "grad_norm": 3.004370927810669,
+      "learning_rate": 3.6496350364963505e-06,
+      "loss": 2.0942,
+      "step": 214
+    },
+    {
+      "epoch": 0.2356164383561644,
+      "grad_norm": 2.5298097133636475,
+      "learning_rate": 3.667883211678833e-06,
+      "loss": 2.1168,
+      "step": 215
+    },
+    {
+      "epoch": 0.23671232876712328,
+      "grad_norm": 3.2273404598236084,
+      "learning_rate": 3.6861313868613142e-06,
+      "loss": 2.0983,
+      "step": 216
+    },
+    {
+      "epoch": 0.2378082191780822,
+      "grad_norm": 2.6995913982391357,
+      "learning_rate": 3.704379562043796e-06,
+      "loss": 2.0932,
+      "step": 217
+    },
+    {
+      "epoch": 0.2389041095890411,
+      "grad_norm": 2.9947187900543213,
+      "learning_rate": 3.7226277372262775e-06,
+      "loss": 2.1848,
+      "step": 218
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 2.8766634464263916,
+      "learning_rate": 3.7408759124087594e-06,
+      "loss": 2.1533,
+      "step": 219
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "grad_norm": 3.011009454727173,
+      "learning_rate": 3.7591240875912412e-06,
+      "loss": 2.0936,
+      "step": 220
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "eval_loss": 2.259504795074463,
+      "eval_runtime": 36.827,
+      "eval_samples_per_second": 27.154,
+      "eval_steps_per_second": 1.14,
+      "step": 220
+    },
+    {
+      "epoch": 0.2421917808219178,
+      "grad_norm": 2.5399389266967773,
+      "learning_rate": 3.777372262773723e-06,
+      "loss": 2.1338,
+      "step": 221
+    },
+    {
+      "epoch": 0.2432876712328767,
+      "grad_norm": 2.793663263320923,
+      "learning_rate": 3.7956204379562045e-06,
+      "loss": 2.1216,
+      "step": 222
+    },
+    {
+      "epoch": 0.24438356164383562,
+      "grad_norm": 2.8745079040527344,
+      "learning_rate": 3.8138686131386864e-06,
+      "loss": 2.0957,
+      "step": 223
+    },
+    {
+      "epoch": 0.24547945205479452,
+      "grad_norm": 2.521819591522217,
+      "learning_rate": 3.832116788321168e-06,
+      "loss": 2.1489,
+      "step": 224
+    },
+    {
+      "epoch": 0.2465753424657534,
+      "grad_norm": 2.724716901779175,
+      "learning_rate": 3.85036496350365e-06,
+      "loss": 2.11,
+      "step": 225
+    },
+    {
+      "epoch": 0.24767123287671233,
+      "grad_norm": 2.4324700832366943,
+      "learning_rate": 3.868613138686132e-06,
+      "loss": 2.0759,
+      "step": 226
+    },
+    {
+      "epoch": 0.24876712328767123,
+      "grad_norm": 3.0802838802337646,
+      "learning_rate": 3.886861313868613e-06,
+      "loss": 2.1131,
+      "step": 227
+    },
+    {
+      "epoch": 0.24986301369863015,
+      "grad_norm": 2.79115629196167,
+      "learning_rate": 3.905109489051096e-06,
+      "loss": 2.125,
+      "step": 228
+    },
+    {
+      "epoch": 0.25095890410958904,
+      "grad_norm": 2.475189685821533,
+      "learning_rate": 3.923357664233577e-06,
+      "loss": 2.1606,
+      "step": 229
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "grad_norm": 2.7638697624206543,
+      "learning_rate": 3.9416058394160585e-06,
+      "loss": 2.1165,
+      "step": 230
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "eval_loss": 2.24651837348938,
+      "eval_runtime": 36.8196,
+      "eval_samples_per_second": 27.159,
+      "eval_steps_per_second": 1.141,
+      "step": 230
+    },
+    {
+      "epoch": 0.25315068493150683,
+      "grad_norm": 2.5692026615142822,
+      "learning_rate": 3.95985401459854e-06,
+      "loss": 2.1662,
+      "step": 231
+    },
+    {
+      "epoch": 0.25424657534246575,
+      "grad_norm": 2.398109197616577,
+      "learning_rate": 3.978102189781022e-06,
+      "loss": 2.1864,
+      "step": 232
+    },
+    {
+      "epoch": 0.2553424657534247,
+      "grad_norm": 2.628312826156616,
+      "learning_rate": 3.996350364963504e-06,
+      "loss": 2.0537,
+      "step": 233
+    },
+    {
+      "epoch": 0.25643835616438354,
+      "grad_norm": 2.615253448486328,
+      "learning_rate": 4.014598540145986e-06,
+      "loss": 2.1348,
+      "step": 234
+    },
+    {
+      "epoch": 0.25753424657534246,
+      "grad_norm": 2.499142646789551,
+      "learning_rate": 4.032846715328467e-06,
+      "loss": 2.0396,
+      "step": 235
+    },
+    {
+      "epoch": 0.2586301369863014,
+      "grad_norm": 2.7101237773895264,
+      "learning_rate": 4.05109489051095e-06,
+      "loss": 2.0969,
+      "step": 236
+    },
+    {
+      "epoch": 0.25972602739726025,
+      "grad_norm": 2.574430227279663,
+      "learning_rate": 4.069343065693431e-06,
+      "loss": 2.1229,
+      "step": 237
+    },
+    {
+      "epoch": 0.2608219178082192,
+      "grad_norm": 2.6702752113342285,
+      "learning_rate": 4.0875912408759126e-06,
+      "loss": 2.0845,
+      "step": 238
+    },
+    {
+      "epoch": 0.2619178082191781,
+      "grad_norm": 3.2056877613067627,
+      "learning_rate": 4.105839416058395e-06,
+      "loss": 2.1096,
+      "step": 239
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "grad_norm": 2.633591413497925,
+      "learning_rate": 4.124087591240876e-06,
+      "loss": 2.1783,
+      "step": 240
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "eval_loss": 2.2385339736938477,
+      "eval_runtime": 36.8527,
+      "eval_samples_per_second": 27.135,
+      "eval_steps_per_second": 1.14,
+      "step": 240
+    },
+    {
+      "epoch": 0.2641095890410959,
+      "grad_norm": 2.780073642730713,
+      "learning_rate": 4.1423357664233586e-06,
+      "loss": 2.0974,
+      "step": 241
+    },
+    {
+      "epoch": 0.2652054794520548,
+      "grad_norm": 2.7121286392211914,
+      "learning_rate": 4.16058394160584e-06,
+      "loss": 2.0772,
+      "step": 242
+    },
+    {
+      "epoch": 0.2663013698630137,
+      "grad_norm": 3.238070011138916,
+      "learning_rate": 4.1788321167883214e-06,
+      "loss": 2.1567,
+      "step": 243
+    },
+    {
+      "epoch": 0.2673972602739726,
+      "grad_norm": 3.9809019565582275,
+      "learning_rate": 4.197080291970803e-06,
+      "loss": 2.0467,
+      "step": 244
+    },
+    {
+      "epoch": 0.2684931506849315,
+      "grad_norm": 2.440645217895508,
+      "learning_rate": 4.215328467153285e-06,
+      "loss": 2.1199,
+      "step": 245
+    },
+    {
+      "epoch": 0.26958904109589044,
+      "grad_norm": 3.84128999710083,
+      "learning_rate": 4.233576642335767e-06,
+      "loss": 2.0787,
+      "step": 246
+    },
+    {
+      "epoch": 0.2706849315068493,
+      "grad_norm": 2.7698514461517334,
+      "learning_rate": 4.251824817518249e-06,
+      "loss": 2.0756,
+      "step": 247
+    },
+    {
+      "epoch": 0.2717808219178082,
+      "grad_norm": 2.625366687774658,
+      "learning_rate": 4.27007299270073e-06,
+      "loss": 2.1093,
+      "step": 248
+    },
+    {
+      "epoch": 0.27287671232876715,
+      "grad_norm": 2.6527903079986572,
+      "learning_rate": 4.288321167883212e-06,
+      "loss": 2.0747,
+      "step": 249
+    },
+    {
+      "epoch": 0.273972602739726,
+      "grad_norm": 2.7589945793151855,
+      "learning_rate": 4.306569343065693e-06,
+      "loss": 2.0708,
+      "step": 250
+    },
+    {
+      "epoch": 0.273972602739726,
+      "eval_loss": 2.2307686805725098,
+      "eval_runtime": 36.8555,
+      "eval_samples_per_second": 27.133,
+      "eval_steps_per_second": 1.14,
+      "step": 250
+    },
+    {
+      "epoch": 0.27506849315068493,
+      "grad_norm": 2.598238468170166,
+      "learning_rate": 4.3248175182481755e-06,
+      "loss": 2.0421,
+      "step": 251
+    },
+    {
+      "epoch": 0.27616438356164386,
+      "grad_norm": 2.893688440322876,
+      "learning_rate": 4.343065693430658e-06,
+      "loss": 2.0997,
+      "step": 252
+    },
+    {
+      "epoch": 0.2772602739726027,
+      "grad_norm": 2.436887741088867,
+      "learning_rate": 4.361313868613139e-06,
+      "loss": 2.0941,
+      "step": 253
+    },
+    {
+      "epoch": 0.27835616438356164,
+      "grad_norm": 2.671720504760742,
+      "learning_rate": 4.379562043795621e-06,
+      "loss": 2.0818,
+      "step": 254
+    },
+    {
+      "epoch": 0.27945205479452057,
+      "grad_norm": 2.575180768966675,
+      "learning_rate": 4.397810218978102e-06,
+      "loss": 2.1164,
+      "step": 255
+    },
+    {
+      "epoch": 0.28054794520547943,
+      "grad_norm": 2.7255051136016846,
+      "learning_rate": 4.416058394160584e-06,
+      "loss": 2.1104,
+      "step": 256
+    },
+    {
+      "epoch": 0.28164383561643835,
+      "grad_norm": 2.75199031829834,
+      "learning_rate": 4.434306569343066e-06,
+      "loss": 2.0887,
+      "step": 257
+    },
+    {
+      "epoch": 0.2827397260273973,
+      "grad_norm": 2.5259952545166016,
+      "learning_rate": 4.452554744525548e-06,
+      "loss": 2.0612,
+      "step": 258
+    },
+    {
+      "epoch": 0.28383561643835614,
+      "grad_norm": 3.1171822547912598,
+      "learning_rate": 4.4708029197080295e-06,
+      "loss": 2.1114,
+      "step": 259
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "grad_norm": 2.5662221908569336,
+      "learning_rate": 4.489051094890512e-06,
+      "loss": 2.1016,
+      "step": 260
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "eval_loss": 2.225515604019165,
+      "eval_runtime": 36.8548,
+      "eval_samples_per_second": 27.134,
+      "eval_steps_per_second": 1.14,
+      "step": 260
+    },
+    {
+      "epoch": 0.286027397260274,
+      "grad_norm": 2.3760831356048584,
+      "learning_rate": 4.507299270072993e-06,
+      "loss": 2.1404,
+      "step": 261
+    },
+    {
+      "epoch": 0.28712328767123285,
+      "grad_norm": 3.341439723968506,
+      "learning_rate": 4.525547445255475e-06,
+      "loss": 2.1037,
+      "step": 262
+    },
+    {
+      "epoch": 0.2882191780821918,
+      "grad_norm": 2.7790281772613525,
+      "learning_rate": 4.543795620437956e-06,
+      "loss": 2.0863,
+      "step": 263
+    },
+    {
+      "epoch": 0.2893150684931507,
+      "grad_norm": 2.9859230518341064,
+      "learning_rate": 4.562043795620438e-06,
+      "loss": 2.0555,
+      "step": 264
+    },
+    {
+      "epoch": 0.29041095890410956,
+      "grad_norm": 2.668275833129883,
+      "learning_rate": 4.580291970802921e-06,
+      "loss": 2.1464,
+      "step": 265
+    },
+    {
+      "epoch": 0.2915068493150685,
+      "grad_norm": 2.858091115951538,
+      "learning_rate": 4.598540145985402e-06,
+      "loss": 2.1494,
+      "step": 266
+    },
+    {
+      "epoch": 0.2926027397260274,
+      "grad_norm": 2.5387940406799316,
+      "learning_rate": 4.6167883211678835e-06,
+      "loss": 2.109,
+      "step": 267
+    },
+    {
+      "epoch": 0.2936986301369863,
+      "grad_norm": 2.4841413497924805,
+      "learning_rate": 4.635036496350365e-06,
+      "loss": 2.0626,
+      "step": 268
+    },
+    {
+      "epoch": 0.2947945205479452,
+      "grad_norm": 2.696976661682129,
+      "learning_rate": 4.653284671532847e-06,
+      "loss": 2.0635,
+      "step": 269
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "grad_norm": 2.710571527481079,
+      "learning_rate": 4.671532846715329e-06,
+      "loss": 2.1056,
+      "step": 270
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "eval_loss": 2.218646764755249,
+      "eval_runtime": 36.8533,
+      "eval_samples_per_second": 27.135,
+      "eval_steps_per_second": 1.14,
+      "step": 270
+    },
+    {
+      "epoch": 0.29698630136986304,
+      "grad_norm": 2.6863739490509033,
+      "learning_rate": 4.689781021897811e-06,
+      "loss": 2.1186,
+      "step": 271
+    },
+    {
+      "epoch": 0.2980821917808219,
+      "grad_norm": 2.546550989151001,
+      "learning_rate": 4.708029197080292e-06,
+      "loss": 2.0856,
+      "step": 272
+    },
+    {
+      "epoch": 0.2991780821917808,
+      "grad_norm": 2.949937582015991,
+      "learning_rate": 4.726277372262774e-06,
+      "loss": 2.1301,
+      "step": 273
+    },
+    {
+      "epoch": 0.30027397260273975,
+      "grad_norm": 2.292760133743286,
+      "learning_rate": 4.744525547445255e-06,
+      "loss": 2.0987,
+      "step": 274
+    },
+    {
+      "epoch": 0.3013698630136986,
+      "grad_norm": 2.5523879528045654,
+      "learning_rate": 4.7627737226277375e-06,
+      "loss": 2.0296,
+      "step": 275
+    },
+    {
+      "epoch": 0.30246575342465754,
+      "grad_norm": 2.2079036235809326,
+      "learning_rate": 4.78102189781022e-06,
+      "loss": 2.1018,
+      "step": 276
+    },
+    {
+      "epoch": 0.30356164383561646,
+      "grad_norm": 2.5277910232543945,
+      "learning_rate": 4.799270072992701e-06,
+      "loss": 2.0953,
+      "step": 277
+    },
+    {
+      "epoch": 0.3046575342465753,
+      "grad_norm": 2.5521962642669678,
+      "learning_rate": 4.8175182481751835e-06,
+      "loss": 2.0709,
+      "step": 278
+    },
+    {
+      "epoch": 0.30575342465753425,
+      "grad_norm": 2.2348296642303467,
+      "learning_rate": 4.835766423357665e-06,
+      "loss": 2.0961,
+      "step": 279
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "grad_norm": 2.782870292663574,
+      "learning_rate": 4.854014598540146e-06,
+      "loss": 2.0775,
+      "step": 280
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "eval_loss": 2.2204668521881104,
+      "eval_runtime": 37.1159,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 280
+    },
+    {
+      "epoch": 0.30794520547945203,
+      "grad_norm": 2.6238420009613037,
+      "learning_rate": 4.872262773722628e-06,
+      "loss": 2.03,
+      "step": 281
+    },
+    {
+      "epoch": 0.30904109589041096,
+      "grad_norm": 2.29125714302063,
+      "learning_rate": 4.89051094890511e-06,
+      "loss": 2.0434,
+      "step": 282
+    },
+    {
+      "epoch": 0.3101369863013699,
+      "grad_norm": 2.3889894485473633,
+      "learning_rate": 4.9087591240875915e-06,
+      "loss": 2.0284,
+      "step": 283
+    },
+    {
+      "epoch": 0.31123287671232874,
+      "grad_norm": 2.3177151679992676,
+      "learning_rate": 4.927007299270074e-06,
+      "loss": 2.0695,
+      "step": 284
+    },
+    {
+      "epoch": 0.31232876712328766,
+      "grad_norm": 2.437307596206665,
+      "learning_rate": 4.945255474452555e-06,
+      "loss": 2.0729,
+      "step": 285
+    },
+    {
+      "epoch": 0.3134246575342466,
+      "grad_norm": 2.3625094890594482,
+      "learning_rate": 4.963503649635037e-06,
+      "loss": 2.0463,
+      "step": 286
+    },
+    {
+      "epoch": 0.31452054794520545,
+      "grad_norm": 2.32767391204834,
+      "learning_rate": 4.981751824817518e-06,
+      "loss": 2.1376,
+      "step": 287
+    },
+    {
+      "epoch": 0.3156164383561644,
+      "grad_norm": 2.291435956954956,
+      "learning_rate": 5e-06,
+      "loss": 2.132,
+      "step": 288
+    },
+    {
+      "epoch": 0.3167123287671233,
+      "grad_norm": 2.226759910583496,
+      "learning_rate": 4.997969130787978e-06,
+      "loss": 2.0759,
+      "step": 289
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "grad_norm": 2.2493679523468018,
+      "learning_rate": 4.995938261575955e-06,
+      "loss": 2.1042,
+      "step": 290
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "eval_loss": 2.201479434967041,
+      "eval_runtime": 37.0713,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 290
+    },
+    {
+      "epoch": 0.3189041095890411,
+      "grad_norm": 2.4757614135742188,
+      "learning_rate": 4.993907392363932e-06,
+      "loss": 2.012,
+      "step": 291
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 2.420912981033325,
+      "learning_rate": 4.9918765231519095e-06,
+      "loss": 2.0936,
+      "step": 292
+    },
+    {
+      "epoch": 0.32109589041095893,
+      "grad_norm": 2.3863155841827393,
+      "learning_rate": 4.989845653939887e-06,
+      "loss": 2.0119,
+      "step": 293
+    },
+    {
+      "epoch": 0.3221917808219178,
+      "grad_norm": 2.4652020931243896,
+      "learning_rate": 4.987814784727864e-06,
+      "loss": 2.0625,
+      "step": 294
+    },
+    {
+      "epoch": 0.3232876712328767,
+      "grad_norm": 2.2076728343963623,
+      "learning_rate": 4.985783915515841e-06,
+      "loss": 2.0591,
+      "step": 295
+    },
+    {
+      "epoch": 0.32438356164383564,
+      "grad_norm": 2.4880995750427246,
+      "learning_rate": 4.983753046303819e-06,
+      "loss": 2.0548,
+      "step": 296
+    },
+    {
+      "epoch": 0.3254794520547945,
+      "grad_norm": 2.396613359451294,
+      "learning_rate": 4.981722177091796e-06,
+      "loss": 2.0678,
+      "step": 297
+    },
+    {
+      "epoch": 0.3265753424657534,
+      "grad_norm": 2.6646151542663574,
+      "learning_rate": 4.979691307879773e-06,
+      "loss": 2.0594,
+      "step": 298
+    },
+    {
+      "epoch": 0.32767123287671235,
+      "grad_norm": 2.393214464187622,
+      "learning_rate": 4.9776604386677504e-06,
+      "loss": 2.0862,
+      "step": 299
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "grad_norm": 2.651883840560913,
+      "learning_rate": 4.975629569455728e-06,
+      "loss": 2.0659,
+      "step": 300
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "eval_loss": 2.198840856552124,
+      "eval_runtime": 37.072,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 300
+    },
+    {
+      "epoch": 0.32986301369863014,
+      "grad_norm": 2.3431386947631836,
+      "learning_rate": 4.973598700243705e-06,
+      "loss": 2.1162,
+      "step": 301
+    },
+    {
+      "epoch": 0.33095890410958906,
+      "grad_norm": 2.754340171813965,
+      "learning_rate": 4.971567831031681e-06,
+      "loss": 2.0276,
+      "step": 302
+    },
+    {
+      "epoch": 0.3320547945205479,
+      "grad_norm": 3.2804505825042725,
+      "learning_rate": 4.9695369618196595e-06,
+      "loss": 2.038,
+      "step": 303
+    },
+    {
+      "epoch": 0.33315068493150685,
+      "grad_norm": 2.722909927368164,
+      "learning_rate": 4.967506092607636e-06,
+      "loss": 1.9115,
+      "step": 304
+    },
+    {
+      "epoch": 0.33424657534246577,
+      "grad_norm": 2.82663893699646,
+      "learning_rate": 4.965475223395614e-06,
+      "loss": 2.0034,
+      "step": 305
+    },
+    {
+      "epoch": 0.33534246575342463,
+      "grad_norm": 2.743824005126953,
+      "learning_rate": 4.9634443541835905e-06,
+      "loss": 2.0251,
+      "step": 306
+    },
+    {
+      "epoch": 0.33643835616438356,
+      "grad_norm": 2.371054172515869,
+      "learning_rate": 4.961413484971569e-06,
+      "loss": 2.0359,
+      "step": 307
+    },
+    {
+      "epoch": 0.3375342465753425,
+      "grad_norm": 2.5377180576324463,
+      "learning_rate": 4.959382615759545e-06,
+      "loss": 2.0489,
+      "step": 308
+    },
+    {
+      "epoch": 0.33863013698630134,
+      "grad_norm": 2.7547554969787598,
+      "learning_rate": 4.957351746547523e-06,
+      "loss": 2.0567,
+      "step": 309
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "grad_norm": 2.832305669784546,
+      "learning_rate": 4.9553208773355e-06,
+      "loss": 2.032,
+      "step": 310
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "eval_loss": 2.2051939964294434,
+      "eval_runtime": 36.8826,
+      "eval_samples_per_second": 27.113,
+      "eval_steps_per_second": 1.139,
+      "step": 310
+    },
+    {
+      "epoch": 0.3408219178082192,
+      "grad_norm": 2.46624493598938,
+      "learning_rate": 4.953290008123478e-06,
+      "loss": 2.1559,
+      "step": 311
+    },
+    {
+      "epoch": 0.3419178082191781,
+      "grad_norm": 2.380703926086426,
+      "learning_rate": 4.951259138911454e-06,
+      "loss": 2.0618,
+      "step": 312
+    },
+    {
+      "epoch": 0.343013698630137,
+      "grad_norm": 2.3415584564208984,
+      "learning_rate": 4.9492282696994315e-06,
+      "loss": 1.9843,
+      "step": 313
+    },
+    {
+      "epoch": 0.3441095890410959,
+      "grad_norm": 2.4721949100494385,
+      "learning_rate": 4.947197400487409e-06,
+      "loss": 2.1236,
+      "step": 314
+    },
+    {
+      "epoch": 0.3452054794520548,
+      "grad_norm": 2.5910539627075195,
+      "learning_rate": 4.945166531275386e-06,
+      "loss": 2.0454,
+      "step": 315
+    },
+    {
+      "epoch": 0.3463013698630137,
+      "grad_norm": 2.761718511581421,
+      "learning_rate": 4.943135662063363e-06,
+      "loss": 2.0419,
+      "step": 316
+    },
+    {
+      "epoch": 0.3473972602739726,
+      "grad_norm": 3.101546287536621,
+      "learning_rate": 4.9411047928513406e-06,
+      "loss": 2.0471,
+      "step": 317
+    },
+    {
+      "epoch": 0.34849315068493153,
+      "grad_norm": 2.7173798084259033,
+      "learning_rate": 4.939073923639318e-06,
+      "loss": 2.0834,
+      "step": 318
+    },
+    {
+      "epoch": 0.3495890410958904,
+      "grad_norm": 2.716454267501831,
+      "learning_rate": 4.937043054427295e-06,
+      "loss": 2.0889,
+      "step": 319
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "grad_norm": 2.755162239074707,
+      "learning_rate": 4.935012185215272e-06,
+      "loss": 2.0245,
+      "step": 320
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "eval_loss": 2.1907594203948975,
+      "eval_runtime": 36.8224,
+      "eval_samples_per_second": 27.157,
+      "eval_steps_per_second": 1.141,
+      "step": 320
+    },
+    {
+      "epoch": 0.35178082191780824,
+      "grad_norm": 2.6325459480285645,
+      "learning_rate": 4.93298131600325e-06,
+      "loss": 2.0095,
+      "step": 321
+    },
+    {
+      "epoch": 0.3528767123287671,
+      "grad_norm": 2.909247398376465,
+      "learning_rate": 4.930950446791227e-06,
+      "loss": 2.0745,
+      "step": 322
+    },
+    {
+      "epoch": 0.353972602739726,
+      "grad_norm": 3.1001574993133545,
+      "learning_rate": 4.928919577579204e-06,
+      "loss": 2.0531,
+      "step": 323
+    },
+    {
+      "epoch": 0.35506849315068495,
+      "grad_norm": 2.546571731567383,
+      "learning_rate": 4.9268887083671815e-06,
+      "loss": 2.0562,
+      "step": 324
+    },
+    {
+      "epoch": 0.3561643835616438,
+      "grad_norm": 3.4734079837799072,
+      "learning_rate": 4.924857839155159e-06,
+      "loss": 2.0396,
+      "step": 325
+    },
+    {
+      "epoch": 0.35726027397260274,
+      "grad_norm": 2.253105878829956,
+      "learning_rate": 4.922826969943136e-06,
+      "loss": 2.0418,
+      "step": 326
+    },
+    {
+      "epoch": 0.35835616438356166,
+      "grad_norm": 3.5297048091888428,
+      "learning_rate": 4.920796100731113e-06,
+      "loss": 2.0683,
+      "step": 327
+    },
+    {
+      "epoch": 0.3594520547945205,
+      "grad_norm": 2.448322296142578,
+      "learning_rate": 4.918765231519091e-06,
+      "loss": 1.9887,
+      "step": 328
+    },
+    {
+      "epoch": 0.36054794520547945,
+      "grad_norm": 2.4430301189422607,
+      "learning_rate": 4.916734362307068e-06,
+      "loss": 2.0489,
+      "step": 329
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "grad_norm": 2.4111526012420654,
+      "learning_rate": 4.914703493095045e-06,
+      "loss": 1.9955,
+      "step": 330
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "eval_loss": 2.191542863845825,
+      "eval_runtime": 36.8305,
+      "eval_samples_per_second": 27.151,
+      "eval_steps_per_second": 1.14,
+      "step": 330
+    },
+    {
+      "epoch": 0.36273972602739724,
+      "grad_norm": 2.182343006134033,
+      "learning_rate": 4.912672623883022e-06,
+      "loss": 2.0105,
+      "step": 331
+    },
+    {
+      "epoch": 0.36383561643835616,
+      "grad_norm": 2.443370819091797,
+      "learning_rate": 4.910641754671e-06,
+      "loss": 2.0024,
+      "step": 332
+    },
+    {
+      "epoch": 0.3649315068493151,
+      "grad_norm": 2.509678363800049,
+      "learning_rate": 4.908610885458977e-06,
+      "loss": 2.0316,
+      "step": 333
+    },
+    {
+      "epoch": 0.36602739726027395,
+      "grad_norm": 2.4028046131134033,
+      "learning_rate": 4.906580016246954e-06,
+      "loss": 2.0206,
+      "step": 334
+    },
+    {
+      "epoch": 0.36712328767123287,
+      "grad_norm": 2.22725772857666,
+      "learning_rate": 4.9045491470349315e-06,
+      "loss": 2.0749,
+      "step": 335
+    },
+    {
+      "epoch": 0.3682191780821918,
+      "grad_norm": 2.4052486419677734,
+      "learning_rate": 4.902518277822909e-06,
+      "loss": 2.0441,
+      "step": 336
+    },
+    {
+      "epoch": 0.3693150684931507,
+      "grad_norm": 2.509493589401245,
+      "learning_rate": 4.900487408610886e-06,
+      "loss": 2.0522,
+      "step": 337
+    },
+    {
+      "epoch": 0.3704109589041096,
+      "grad_norm": 2.263730764389038,
+      "learning_rate": 4.898456539398863e-06,
+      "loss": 2.0579,
+      "step": 338
+    },
+    {
+      "epoch": 0.3715068493150685,
+      "grad_norm": 2.6925342082977295,
+      "learning_rate": 4.89642567018684e-06,
+      "loss": 2.0712,
+      "step": 339
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "grad_norm": 2.517216920852661,
+      "learning_rate": 4.894394800974818e-06,
+      "loss": 2.1426,
+      "step": 340
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "eval_loss": 2.1843371391296387,
+      "eval_runtime": 36.8711,
+      "eval_samples_per_second": 27.122,
+      "eval_steps_per_second": 1.139,
+      "step": 340
+    },
+    {
+      "epoch": 0.3736986301369863,
+      "grad_norm": 2.222014904022217,
+      "learning_rate": 4.892363931762794e-06,
+      "loss": 1.9838,
+      "step": 341
+    },
+    {
+      "epoch": 0.3747945205479452,
+      "grad_norm": 2.2543673515319824,
+      "learning_rate": 4.8903330625507725e-06,
+      "loss": 2.0666,
+      "step": 342
+    },
+    {
+      "epoch": 0.37589041095890413,
+      "grad_norm": 2.496208906173706,
+      "learning_rate": 4.888302193338749e-06,
+      "loss": 2.0072,
+      "step": 343
+    },
+    {
+      "epoch": 0.376986301369863,
+      "grad_norm": 2.194899559020996,
+      "learning_rate": 4.886271324126727e-06,
+      "loss": 1.9863,
+      "step": 344
+    },
+    {
+      "epoch": 0.3780821917808219,
+      "grad_norm": 2.666233539581299,
+      "learning_rate": 4.8842404549147034e-06,
+      "loss": 2.0828,
+      "step": 345
+    },
+    {
+      "epoch": 0.37917808219178084,
+      "grad_norm": 2.239429473876953,
+      "learning_rate": 4.8822095857026816e-06,
+      "loss": 2.0716,
+      "step": 346
+    },
+    {
+      "epoch": 0.3802739726027397,
+      "grad_norm": 2.4315106868743896,
+      "learning_rate": 4.880178716490658e-06,
+      "loss": 2.0361,
+      "step": 347
+    },
+    {
+      "epoch": 0.38136986301369863,
+      "grad_norm": 2.616281509399414,
+      "learning_rate": 4.878147847278636e-06,
+      "loss": 1.987,
+      "step": 348
+    },
+    {
+      "epoch": 0.38246575342465755,
+      "grad_norm": 2.1155405044555664,
+      "learning_rate": 4.8761169780666125e-06,
+      "loss": 1.9996,
+      "step": 349
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "grad_norm": 2.784736156463623,
+      "learning_rate": 4.874086108854591e-06,
+      "loss": 1.9723,
+      "step": 350
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "eval_loss": 2.1786441802978516,
+      "eval_runtime": 36.8481,
+      "eval_samples_per_second": 27.138,
+      "eval_steps_per_second": 1.14,
+      "step": 350
+    },
+    {
+      "epoch": 0.38465753424657534,
+      "grad_norm": 2.425434112548828,
+      "learning_rate": 4.872055239642567e-06,
+      "loss": 1.9823,
+      "step": 351
+    },
+    {
+      "epoch": 0.38575342465753426,
+      "grad_norm": 2.474745512008667,
+      "learning_rate": 4.870024370430544e-06,
+      "loss": 1.9657,
+      "step": 352
+    },
+    {
+      "epoch": 0.3868493150684931,
+      "grad_norm": 2.5412509441375732,
+      "learning_rate": 4.867993501218522e-06,
+      "loss": 2.017,
+      "step": 353
+    },
+    {
+      "epoch": 0.38794520547945205,
+      "grad_norm": 2.587143659591675,
+      "learning_rate": 4.865962632006499e-06,
+      "loss": 2.0564,
+      "step": 354
+    },
+    {
+      "epoch": 0.38904109589041097,
+      "grad_norm": 2.4585652351379395,
+      "learning_rate": 4.863931762794476e-06,
+      "loss": 1.9895,
+      "step": 355
+    },
+    {
+      "epoch": 0.39013698630136984,
+      "grad_norm": 3.121283769607544,
+      "learning_rate": 4.8619008935824535e-06,
+      "loss": 2.0179,
+      "step": 356
+    },
+    {
+      "epoch": 0.39123287671232876,
+      "grad_norm": 2.432809352874756,
+      "learning_rate": 4.859870024370431e-06,
+      "loss": 2.049,
+      "step": 357
+    },
+    {
+      "epoch": 0.3923287671232877,
+      "grad_norm": 2.6077938079833984,
+      "learning_rate": 4.857839155158408e-06,
+      "loss": 2.0667,
+      "step": 358
+    },
+    {
+      "epoch": 0.3934246575342466,
+      "grad_norm": 2.636944055557251,
+      "learning_rate": 4.855808285946385e-06,
+      "loss": 2.0776,
+      "step": 359
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "grad_norm": 2.3640847206115723,
+      "learning_rate": 4.8537774167343626e-06,
+      "loss": 2.0533,
+      "step": 360
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "eval_loss": 2.1743040084838867,
+      "eval_runtime": 36.8418,
+      "eval_samples_per_second": 27.143,
+      "eval_steps_per_second": 1.14,
+      "step": 360
+    },
+    {
+      "epoch": 0.3956164383561644,
+      "grad_norm": 2.7718842029571533,
+      "learning_rate": 4.85174654752234e-06,
+      "loss": 2.0374,
+      "step": 361
+    },
+    {
+      "epoch": 0.3967123287671233,
+      "grad_norm": 2.86139178276062,
+      "learning_rate": 4.849715678310317e-06,
+      "loss": 2.035,
+      "step": 362
+    },
+    {
+      "epoch": 0.3978082191780822,
+      "grad_norm": 2.1344094276428223,
+      "learning_rate": 4.847684809098294e-06,
+      "loss": 2.0307,
+      "step": 363
+    },
+    {
+      "epoch": 0.3989041095890411,
+      "grad_norm": 2.383131742477417,
+      "learning_rate": 4.845653939886272e-06,
+      "loss": 2.0346,
+      "step": 364
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 2.606567144393921,
+      "learning_rate": 4.843623070674249e-06,
+      "loss": 2.0325,
+      "step": 365
+    },
+    {
+      "epoch": 0.4010958904109589,
+      "grad_norm": 2.1096670627593994,
+      "learning_rate": 4.841592201462226e-06,
+      "loss": 2.0065,
+      "step": 366
+    },
+    {
+      "epoch": 0.4021917808219178,
+      "grad_norm": 2.513556480407715,
+      "learning_rate": 4.8395613322502035e-06,
+      "loss": 2.0643,
+      "step": 367
+    },
+    {
+      "epoch": 0.40328767123287673,
+      "grad_norm": 2.236032485961914,
+      "learning_rate": 4.837530463038181e-06,
+      "loss": 1.9637,
+      "step": 368
+    },
+    {
+      "epoch": 0.4043835616438356,
+      "grad_norm": 2.1936111450195312,
+      "learning_rate": 4.835499593826158e-06,
+      "loss": 2.0217,
+      "step": 369
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "grad_norm": 2.11141037940979,
+      "learning_rate": 4.833468724614135e-06,
+      "loss": 2.049,
+      "step": 370
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "eval_loss": 2.1685962677001953,
+      "eval_runtime": 36.9866,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 370
+    },
+    {
+      "epoch": 0.40657534246575344,
+      "grad_norm": 2.2242090702056885,
+      "learning_rate": 4.831437855402113e-06,
+      "loss": 2.0267,
+      "step": 371
+    },
+    {
+      "epoch": 0.4076712328767123,
+      "grad_norm": 2.1799142360687256,
+      "learning_rate": 4.82940698619009e-06,
+      "loss": 1.9266,
+      "step": 372
+    },
+    {
+      "epoch": 0.40876712328767123,
+      "grad_norm": 2.0943312644958496,
+      "learning_rate": 4.827376116978067e-06,
+      "loss": 1.9934,
+      "step": 373
+    },
+    {
+      "epoch": 0.40986301369863015,
+      "grad_norm": 2.386090040206909,
+      "learning_rate": 4.8253452477660444e-06,
+      "loss": 2.0524,
+      "step": 374
+    },
+    {
+      "epoch": 0.410958904109589,
+      "grad_norm": 2.259557008743286,
+      "learning_rate": 4.823314378554022e-06,
+      "loss": 2.0335,
+      "step": 375
+    },
+    {
+      "epoch": 0.41205479452054794,
+      "grad_norm": 2.342690944671631,
+      "learning_rate": 4.821283509341998e-06,
+      "loss": 2.0234,
+      "step": 376
+    },
+    {
+      "epoch": 0.41315068493150686,
+      "grad_norm": 2.207160472869873,
+      "learning_rate": 4.819252640129976e-06,
+      "loss": 1.9935,
+      "step": 377
+    },
+    {
+      "epoch": 0.41424657534246573,
+      "grad_norm": 2.3230350017547607,
+      "learning_rate": 4.817221770917953e-06,
+      "loss": 2.0111,
+      "step": 378
+    },
+    {
+      "epoch": 0.41534246575342465,
+      "grad_norm": 2.3750221729278564,
+      "learning_rate": 4.815190901705931e-06,
+      "loss": 1.9941,
+      "step": 379
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "grad_norm": 2.139505624771118,
+      "learning_rate": 4.813160032493907e-06,
+      "loss": 2.0216,
+      "step": 380
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "eval_loss": 2.1608307361602783,
+      "eval_runtime": 37.0997,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 380
+    },
+    {
+      "epoch": 0.41753424657534244,
+      "grad_norm": 2.3400697708129883,
+      "learning_rate": 4.811129163281885e-06,
+      "loss": 2.0082,
+      "step": 381
+    },
+    {
+      "epoch": 0.41863013698630136,
+      "grad_norm": 2.4719579219818115,
+      "learning_rate": 4.809098294069862e-06,
+      "loss": 2.0431,
+      "step": 382
+    },
+    {
+      "epoch": 0.4197260273972603,
+      "grad_norm": 2.1330277919769287,
+      "learning_rate": 4.80706742485784e-06,
+      "loss": 2.0241,
+      "step": 383
+    },
+    {
+      "epoch": 0.4208219178082192,
+      "grad_norm": 2.3635149002075195,
+      "learning_rate": 4.805036555645816e-06,
+      "loss": 2.0795,
+      "step": 384
+    },
+    {
+      "epoch": 0.42191780821917807,
+      "grad_norm": 2.331961154937744,
+      "learning_rate": 4.8030056864337945e-06,
+      "loss": 2.001,
+      "step": 385
+    },
+    {
+      "epoch": 0.423013698630137,
+      "grad_norm": 2.593552589416504,
+      "learning_rate": 4.800974817221771e-06,
+      "loss": 2.0552,
+      "step": 386
+    },
+    {
+      "epoch": 0.4241095890410959,
+      "grad_norm": 2.3201818466186523,
+      "learning_rate": 4.798943948009749e-06,
+      "loss": 2.0069,
+      "step": 387
+    },
+    {
+      "epoch": 0.4252054794520548,
+      "grad_norm": 2.2886810302734375,
+      "learning_rate": 4.7969130787977254e-06,
+      "loss": 2.0232,
+      "step": 388
+    },
+    {
+      "epoch": 0.4263013698630137,
+      "grad_norm": 2.208707809448242,
+      "learning_rate": 4.794882209585703e-06,
+      "loss": 2.0149,
+      "step": 389
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "grad_norm": 2.287991523742676,
+      "learning_rate": 4.79285134037368e-06,
+      "loss": 2.0436,
+      "step": 390
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "eval_loss": 2.1550230979919434,
+      "eval_runtime": 36.9869,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 390
+    },
+    {
+      "epoch": 0.4284931506849315,
+      "grad_norm": 2.0982491970062256,
+      "learning_rate": 4.790820471161657e-06,
+      "loss": 1.9824,
+      "step": 391
+    },
+    {
+      "epoch": 0.4295890410958904,
+      "grad_norm": 2.2861428260803223,
+      "learning_rate": 4.7887896019496345e-06,
+      "loss": 1.9914,
+      "step": 392
+    },
+    {
+      "epoch": 0.43068493150684933,
+      "grad_norm": 2.2473807334899902,
+      "learning_rate": 4.786758732737612e-06,
+      "loss": 2.0057,
+      "step": 393
+    },
+    {
+      "epoch": 0.4317808219178082,
+      "grad_norm": 2.1168124675750732,
+      "learning_rate": 4.784727863525589e-06,
+      "loss": 2.0457,
+      "step": 394
+    },
+    {
+      "epoch": 0.4328767123287671,
+      "grad_norm": 2.2538185119628906,
+      "learning_rate": 4.782696994313566e-06,
+      "loss": 1.9618,
+      "step": 395
+    },
+    {
+      "epoch": 0.43397260273972604,
+      "grad_norm": 2.1803009510040283,
+      "learning_rate": 4.780666125101544e-06,
+      "loss": 2.0172,
+      "step": 396
+    },
+    {
+      "epoch": 0.4350684931506849,
+      "grad_norm": 2.3871731758117676,
+      "learning_rate": 4.778635255889521e-06,
+      "loss": 1.9548,
+      "step": 397
+    },
+    {
+      "epoch": 0.43616438356164383,
+      "grad_norm": 2.3013761043548584,
+      "learning_rate": 4.776604386677498e-06,
+      "loss": 1.978,
+      "step": 398
+    },
+    {
+      "epoch": 0.43726027397260275,
+      "grad_norm": 2.259394645690918,
+      "learning_rate": 4.7745735174654755e-06,
+      "loss": 1.9895,
+      "step": 399
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "grad_norm": 2.233656406402588,
+      "learning_rate": 4.772542648253453e-06,
+      "loss": 2.0485,
+      "step": 400
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "eval_loss": 2.1557083129882812,
+      "eval_runtime": 37.004,
+      "eval_samples_per_second": 27.024,
+      "eval_steps_per_second": 1.135,
+      "step": 400
+    },
+    {
+      "epoch": 0.43945205479452054,
+      "grad_norm": 2.131877899169922,
+      "learning_rate": 4.77051177904143e-06,
+      "loss": 2.051,
+      "step": 401
+    },
+    {
+      "epoch": 0.44054794520547946,
+      "grad_norm": 2.1987686157226562,
+      "learning_rate": 4.768480909829407e-06,
+      "loss": 1.977,
+      "step": 402
+    },
+    {
+      "epoch": 0.44164383561643833,
+      "grad_norm": 2.26301646232605,
+      "learning_rate": 4.766450040617385e-06,
+      "loss": 2.0238,
+      "step": 403
+    },
+    {
+      "epoch": 0.44273972602739725,
+      "grad_norm": 2.288947105407715,
+      "learning_rate": 4.764419171405362e-06,
+      "loss": 1.9943,
+      "step": 404
+    },
+    {
+      "epoch": 0.4438356164383562,
+      "grad_norm": 2.364522695541382,
+      "learning_rate": 4.762388302193339e-06,
+      "loss": 2.0376,
+      "step": 405
+    },
+    {
+      "epoch": 0.4449315068493151,
+      "grad_norm": 2.1967415809631348,
+      "learning_rate": 4.760357432981316e-06,
+      "loss": 2.0184,
+      "step": 406
+    },
+    {
+      "epoch": 0.44602739726027396,
+      "grad_norm": 2.0930731296539307,
+      "learning_rate": 4.758326563769294e-06,
+      "loss": 1.9862,
+      "step": 407
+    },
+    {
+      "epoch": 0.4471232876712329,
+      "grad_norm": 2.2198588848114014,
+      "learning_rate": 4.756295694557271e-06,
+      "loss": 2.022,
+      "step": 408
+    },
+    {
+      "epoch": 0.4482191780821918,
+      "grad_norm": 2.0918731689453125,
+      "learning_rate": 4.754264825345248e-06,
+      "loss": 2.0166,
+      "step": 409
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "grad_norm": 2.40588116645813,
+      "learning_rate": 4.7522339561332255e-06,
+      "loss": 1.9502,
+      "step": 410
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "eval_loss": 2.1560990810394287,
+      "eval_runtime": 36.9649,
+      "eval_samples_per_second": 27.053,
+      "eval_steps_per_second": 1.136,
+      "step": 410
+    },
+    {
+      "epoch": 0.4504109589041096,
+      "grad_norm": 2.228062629699707,
+      "learning_rate": 4.750203086921203e-06,
+      "loss": 1.9715,
+      "step": 411
+    },
+    {
+      "epoch": 0.4515068493150685,
+      "grad_norm": 2.216545820236206,
+      "learning_rate": 4.74817221770918e-06,
+      "loss": 1.9359,
+      "step": 412
+    },
+    {
+      "epoch": 0.4526027397260274,
+      "grad_norm": 2.3648414611816406,
+      "learning_rate": 4.7461413484971565e-06,
+      "loss": 2.0099,
+      "step": 413
+    },
+    {
+      "epoch": 0.4536986301369863,
+      "grad_norm": 2.2856826782226562,
+      "learning_rate": 4.744110479285135e-06,
+      "loss": 2.0623,
+      "step": 414
+    },
+    {
+      "epoch": 0.4547945205479452,
+      "grad_norm": 2.5282676219940186,
+      "learning_rate": 4.742079610073111e-06,
+      "loss": 1.9527,
+      "step": 415
+    },
+    {
+      "epoch": 0.4558904109589041,
+      "grad_norm": 2.1065621376037598,
+      "learning_rate": 4.740048740861089e-06,
+      "loss": 1.9769,
+      "step": 416
+    },
+    {
+      "epoch": 0.456986301369863,
+      "grad_norm": 2.3433425426483154,
+      "learning_rate": 4.738017871649066e-06,
+      "loss": 1.9881,
+      "step": 417
+    },
+    {
+      "epoch": 0.45808219178082193,
+      "grad_norm": 2.2476186752319336,
+      "learning_rate": 4.735987002437044e-06,
+      "loss": 1.9982,
+      "step": 418
+    },
+    {
+      "epoch": 0.4591780821917808,
+      "grad_norm": 2.3849618434906006,
+      "learning_rate": 4.73395613322502e-06,
+      "loss": 2.0616,
+      "step": 419
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "grad_norm": 2.398130416870117,
+      "learning_rate": 4.731925264012998e-06,
+      "loss": 2.0686,
+      "step": 420
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "eval_loss": 2.1495046615600586,
+      "eval_runtime": 36.8294,
+      "eval_samples_per_second": 27.152,
+      "eval_steps_per_second": 1.14,
+      "step": 420
+    },
+    {
+      "epoch": 0.46136986301369864,
+      "grad_norm": 2.224095106124878,
+      "learning_rate": 4.729894394800975e-06,
+      "loss": 2.0278,
+      "step": 421
+    },
+    {
+      "epoch": 0.4624657534246575,
+      "grad_norm": 2.177316665649414,
+      "learning_rate": 4.727863525588953e-06,
+      "loss": 1.9633,
+      "step": 422
+    },
+    {
+      "epoch": 0.46356164383561643,
+      "grad_norm": 2.212502956390381,
+      "learning_rate": 4.725832656376929e-06,
+      "loss": 1.9211,
+      "step": 423
+    },
+    {
+      "epoch": 0.46465753424657535,
+      "grad_norm": 2.329396963119507,
+      "learning_rate": 4.723801787164907e-06,
+      "loss": 1.9822,
+      "step": 424
+    },
+    {
+      "epoch": 0.4657534246575342,
+      "grad_norm": 2.351280689239502,
+      "learning_rate": 4.721770917952885e-06,
+      "loss": 2.0233,
+      "step": 425
+    },
+    {
+      "epoch": 0.46684931506849314,
+      "grad_norm": 2.1245787143707275,
+      "learning_rate": 4.719740048740861e-06,
+      "loss": 1.9609,
+      "step": 426
+    },
+    {
+      "epoch": 0.46794520547945206,
+      "grad_norm": 2.1225812435150146,
+      "learning_rate": 4.717709179528839e-06,
+      "loss": 1.9905,
+      "step": 427
+    },
+    {
+      "epoch": 0.469041095890411,
+      "grad_norm": 2.0605881214141846,
+      "learning_rate": 4.715678310316816e-06,
+      "loss": 2.0114,
+      "step": 428
+    },
+    {
+      "epoch": 0.47013698630136985,
+      "grad_norm": 2.241417169570923,
+      "learning_rate": 4.713647441104794e-06,
+      "loss": 2.0146,
+      "step": 429
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "grad_norm": 2.2796316146850586,
+      "learning_rate": 4.71161657189277e-06,
+      "loss": 1.9548,
+      "step": 430
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "eval_loss": 2.142974615097046,
+      "eval_runtime": 36.8621,
+      "eval_samples_per_second": 27.128,
+      "eval_steps_per_second": 1.139,
+      "step": 430
+    },
+    {
+      "epoch": 0.4723287671232877,
+      "grad_norm": 2.4043936729431152,
+      "learning_rate": 4.709585702680748e-06,
+      "loss": 1.9465,
+      "step": 431
+    },
+    {
+      "epoch": 0.47342465753424656,
+      "grad_norm": 2.176814317703247,
+      "learning_rate": 4.707554833468725e-06,
+      "loss": 2.0234,
+      "step": 432
+    },
+    {
+      "epoch": 0.4745205479452055,
+      "grad_norm": 2.3174710273742676,
+      "learning_rate": 4.705523964256703e-06,
+      "loss": 2.0795,
+      "step": 433
+    },
+    {
+      "epoch": 0.4756164383561644,
+      "grad_norm": 2.460334539413452,
+      "learning_rate": 4.703493095044679e-06,
+      "loss": 1.9687,
+      "step": 434
+    },
+    {
+      "epoch": 0.4767123287671233,
+      "grad_norm": 2.4824578762054443,
+      "learning_rate": 4.701462225832657e-06,
+      "loss": 1.9911,
+      "step": 435
+    },
+    {
+      "epoch": 0.4778082191780822,
+      "grad_norm": 2.5180532932281494,
+      "learning_rate": 4.699431356620634e-06,
+      "loss": 1.9128,
+      "step": 436
+    },
+    {
+      "epoch": 0.4789041095890411,
+      "grad_norm": 2.380042791366577,
+      "learning_rate": 4.697400487408612e-06,
+      "loss": 2.071,
+      "step": 437
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 2.441718816757202,
+      "learning_rate": 4.695369618196588e-06,
+      "loss": 2.0143,
+      "step": 438
+    },
+    {
+      "epoch": 0.4810958904109589,
+      "grad_norm": 2.9212632179260254,
+      "learning_rate": 4.693338748984566e-06,
+      "loss": 1.9867,
+      "step": 439
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "grad_norm": 2.3351190090179443,
+      "learning_rate": 4.691307879772543e-06,
+      "loss": 1.9852,
+      "step": 440
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "eval_loss": 2.1418583393096924,
+      "eval_runtime": 36.8638,
+      "eval_samples_per_second": 27.127,
+      "eval_steps_per_second": 1.139,
+      "step": 440
+    },
+    {
+      "epoch": 0.4832876712328767,
+      "grad_norm": 3.4289700984954834,
+      "learning_rate": 4.68927701056052e-06,
+      "loss": 1.9353,
+      "step": 441
+    },
+    {
+      "epoch": 0.4843835616438356,
+      "grad_norm": 2.154865026473999,
+      "learning_rate": 4.6872461413484975e-06,
+      "loss": 2.0259,
+      "step": 442
+    },
+    {
+      "epoch": 0.48547945205479454,
+      "grad_norm": 2.4053304195404053,
+      "learning_rate": 4.685215272136475e-06,
+      "loss": 2.0101,
+      "step": 443
+    },
+    {
+      "epoch": 0.4865753424657534,
+      "grad_norm": 2.250504493713379,
+      "learning_rate": 4.683184402924452e-06,
+      "loss": 1.9976,
+      "step": 444
+    },
+    {
+      "epoch": 0.4876712328767123,
+      "grad_norm": 2.9280881881713867,
+      "learning_rate": 4.681153533712429e-06,
+      "loss": 1.9925,
+      "step": 445
+    },
+    {
+      "epoch": 0.48876712328767125,
+      "grad_norm": 2.3568203449249268,
+      "learning_rate": 4.679122664500407e-06,
+      "loss": 2.0534,
+      "step": 446
+    },
+    {
+      "epoch": 0.4898630136986301,
+      "grad_norm": 3.101870536804199,
+      "learning_rate": 4.677091795288384e-06,
+      "loss": 1.9486,
+      "step": 447
+    },
+    {
+      "epoch": 0.49095890410958903,
+      "grad_norm": 2.2512426376342773,
+      "learning_rate": 4.675060926076361e-06,
+      "loss": 1.9334,
+      "step": 448
+    },
+    {
+      "epoch": 0.49205479452054796,
+      "grad_norm": 2.2165493965148926,
+      "learning_rate": 4.673030056864338e-06,
+      "loss": 2.0329,
+      "step": 449
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "grad_norm": 2.2993414402008057,
+      "learning_rate": 4.670999187652316e-06,
+      "loss": 1.9409,
+      "step": 450
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "eval_loss": 2.1386303901672363,
+      "eval_runtime": 36.8687,
+      "eval_samples_per_second": 27.123,
+      "eval_steps_per_second": 1.139,
+      "step": 450
+    },
+    {
+      "epoch": 0.49424657534246574,
+      "grad_norm": 2.196807622909546,
+      "learning_rate": 4.668968318440293e-06,
+      "loss": 1.9416,
+      "step": 451
+    },
+    {
+      "epoch": 0.49534246575342467,
+      "grad_norm": 2.04422926902771,
+      "learning_rate": 4.66693744922827e-06,
+      "loss": 1.9426,
+      "step": 452
+    },
+    {
+      "epoch": 0.4964383561643836,
+      "grad_norm": 2.2703120708465576,
+      "learning_rate": 4.6649065800162475e-06,
+      "loss": 2.0223,
+      "step": 453
+    },
+    {
+      "epoch": 0.49753424657534245,
+      "grad_norm": 2.350550889968872,
+      "learning_rate": 4.662875710804225e-06,
+      "loss": 1.9215,
+      "step": 454
+    },
+    {
+      "epoch": 0.4986301369863014,
+      "grad_norm": 2.192880868911743,
+      "learning_rate": 4.660844841592202e-06,
+      "loss": 1.9945,
+      "step": 455
+    },
+    {
+      "epoch": 0.4997260273972603,
+      "grad_norm": 2.3187272548675537,
+      "learning_rate": 4.658813972380179e-06,
+      "loss": 1.9014,
+      "step": 456
+    },
+    {
+      "epoch": 0.5008219178082192,
+      "grad_norm": 2.3140294551849365,
+      "learning_rate": 4.656783103168157e-06,
+      "loss": 2.0314,
+      "step": 457
+    },
+    {
+      "epoch": 0.5019178082191781,
+      "grad_norm": 2.1471943855285645,
+      "learning_rate": 4.654752233956134e-06,
+      "loss": 1.896,
+      "step": 458
+    },
+    {
+      "epoch": 0.503013698630137,
+      "grad_norm": 2.2951014041900635,
+      "learning_rate": 4.652721364744111e-06,
+      "loss": 1.9878,
+      "step": 459
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "grad_norm": 2.2135801315307617,
+      "learning_rate": 4.6506904955320885e-06,
+      "loss": 1.9573,
+      "step": 460
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "eval_loss": 2.1433045864105225,
+      "eval_runtime": 36.8723,
+      "eval_samples_per_second": 27.121,
+      "eval_steps_per_second": 1.139,
+      "step": 460
+    },
+    {
+      "epoch": 0.5052054794520547,
+      "grad_norm": 2.1129870414733887,
+      "learning_rate": 4.648659626320066e-06,
+      "loss": 2.0127,
+      "step": 461
+    },
+    {
+      "epoch": 0.5063013698630137,
+      "grad_norm": 2.299039602279663,
+      "learning_rate": 4.646628757108043e-06,
+      "loss": 2.0106,
+      "step": 462
+    },
+    {
+      "epoch": 0.5073972602739726,
+      "grad_norm": 2.299088716506958,
+      "learning_rate": 4.6445978878960194e-06,
+      "loss": 1.9812,
+      "step": 463
+    },
+    {
+      "epoch": 0.5084931506849315,
+      "grad_norm": 2.0711328983306885,
+      "learning_rate": 4.6425670186839976e-06,
+      "loss": 1.9435,
+      "step": 464
+    },
+    {
+      "epoch": 0.5095890410958904,
+      "grad_norm": 2.1749722957611084,
+      "learning_rate": 4.640536149471974e-06,
+      "loss": 1.9793,
+      "step": 465
+    },
+    {
+      "epoch": 0.5106849315068493,
+      "grad_norm": 2.127612352371216,
+      "learning_rate": 4.638505280259952e-06,
+      "loss": 2.0117,
+      "step": 466
+    },
+    {
+      "epoch": 0.5117808219178083,
+      "grad_norm": 2.1723556518554688,
+      "learning_rate": 4.6364744110479285e-06,
+      "loss": 2.0336,
+      "step": 467
+    },
+    {
+      "epoch": 0.5128767123287671,
+      "grad_norm": 2.1473746299743652,
+      "learning_rate": 4.634443541835907e-06,
+      "loss": 1.9861,
+      "step": 468
+    },
+    {
+      "epoch": 0.513972602739726,
+      "grad_norm": 2.2084238529205322,
+      "learning_rate": 4.632412672623883e-06,
+      "loss": 2.0428,
+      "step": 469
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "grad_norm": 2.3061389923095703,
+      "learning_rate": 4.630381803411861e-06,
+      "loss": 1.9594,
+      "step": 470
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "eval_loss": 2.1333870887756348,
+      "eval_runtime": 37.1096,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 470
+    },
+    {
+      "epoch": 0.5161643835616438,
+      "grad_norm": 2.340785264968872,
+      "learning_rate": 4.628350934199838e-06,
+      "loss": 1.8626,
+      "step": 471
+    },
+    {
+      "epoch": 0.5172602739726028,
+      "grad_norm": 2.1136975288391113,
+      "learning_rate": 4.626320064987816e-06,
+      "loss": 1.9679,
+      "step": 472
+    },
+    {
+      "epoch": 0.5183561643835617,
+      "grad_norm": 2.311692714691162,
+      "learning_rate": 4.624289195775792e-06,
+      "loss": 2.001,
+      "step": 473
+    },
+    {
+      "epoch": 0.5194520547945205,
+      "grad_norm": 2.2172560691833496,
+      "learning_rate": 4.62225832656377e-06,
+      "loss": 1.9731,
+      "step": 474
+    },
+    {
+      "epoch": 0.5205479452054794,
+      "grad_norm": 2.2663748264312744,
+      "learning_rate": 4.620227457351747e-06,
+      "loss": 1.9804,
+      "step": 475
+    },
+    {
+      "epoch": 0.5216438356164383,
+      "grad_norm": 2.2165415287017822,
+      "learning_rate": 4.618196588139724e-06,
+      "loss": 1.972,
+      "step": 476
+    },
+    {
+      "epoch": 0.5227397260273973,
+      "grad_norm": 2.1509106159210205,
+      "learning_rate": 4.616165718927701e-06,
+      "loss": 1.9886,
+      "step": 477
+    },
+    {
+      "epoch": 0.5238356164383562,
+      "grad_norm": 2.096590280532837,
+      "learning_rate": 4.6141348497156786e-06,
+      "loss": 1.9412,
+      "step": 478
+    },
+    {
+      "epoch": 0.5249315068493151,
+      "grad_norm": 2.123234987258911,
+      "learning_rate": 4.612103980503656e-06,
+      "loss": 2.0331,
+      "step": 479
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "grad_norm": 2.3646109104156494,
+      "learning_rate": 4.610073111291633e-06,
+      "loss": 2.0127,
+      "step": 480
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "eval_loss": 2.1293928623199463,
+      "eval_runtime": 37.1286,
+      "eval_samples_per_second": 26.933,
+      "eval_steps_per_second": 1.131,
+      "step": 480
+    },
+    {
+      "epoch": 0.5271232876712328,
+      "grad_norm": 2.15993070602417,
+      "learning_rate": 4.60804224207961e-06,
+      "loss": 1.9867,
+      "step": 481
+    },
+    {
+      "epoch": 0.5282191780821918,
+      "grad_norm": 2.0517752170562744,
+      "learning_rate": 4.606011372867588e-06,
+      "loss": 1.9611,
+      "step": 482
+    },
+    {
+      "epoch": 0.5293150684931507,
+      "grad_norm": 2.262526035308838,
+      "learning_rate": 4.603980503655565e-06,
+      "loss": 1.9939,
+      "step": 483
+    },
+    {
+      "epoch": 0.5304109589041096,
+      "grad_norm": 2.203848361968994,
+      "learning_rate": 4.601949634443542e-06,
+      "loss": 1.9674,
+      "step": 484
+    },
+    {
+      "epoch": 0.5315068493150685,
+      "grad_norm": 2.0661096572875977,
+      "learning_rate": 4.5999187652315195e-06,
+      "loss": 1.902,
+      "step": 485
+    },
+    {
+      "epoch": 0.5326027397260275,
+      "grad_norm": 2.0441699028015137,
+      "learning_rate": 4.597887896019497e-06,
+      "loss": 1.9026,
+      "step": 486
+    },
+    {
+      "epoch": 0.5336986301369863,
+      "grad_norm": 2.330298900604248,
+      "learning_rate": 4.595857026807474e-06,
+      "loss": 1.9594,
+      "step": 487
+    },
+    {
+      "epoch": 0.5347945205479452,
+      "grad_norm": 2.0567939281463623,
+      "learning_rate": 4.593826157595451e-06,
+      "loss": 1.969,
+      "step": 488
+    },
+    {
+      "epoch": 0.5358904109589041,
+      "grad_norm": 2.2553024291992188,
+      "learning_rate": 4.591795288383429e-06,
+      "loss": 1.9501,
+      "step": 489
+    },
+    {
+      "epoch": 0.536986301369863,
+      "grad_norm": 2.525721549987793,
+      "learning_rate": 4.589764419171406e-06,
+      "loss": 2.0365,
+      "step": 490
+    },
+    {
+      "epoch": 0.536986301369863,
+      "eval_loss": 2.124099016189575,
+      "eval_runtime": 37.1141,
+      "eval_samples_per_second": 26.944,
+      "eval_steps_per_second": 1.132,
+      "step": 490
+    },
+    {
+      "epoch": 0.538082191780822,
+      "grad_norm": 2.017068386077881,
+      "learning_rate": 4.587733549959383e-06,
+      "loss": 1.9768,
+      "step": 491
+    },
+    {
+      "epoch": 0.5391780821917809,
+      "grad_norm": 2.3081953525543213,
+      "learning_rate": 4.5857026807473604e-06,
+      "loss": 1.9499,
+      "step": 492
+    },
+    {
+      "epoch": 0.5402739726027397,
+      "grad_norm": 2.317981243133545,
+      "learning_rate": 4.583671811535338e-06,
+      "loss": 1.9289,
+      "step": 493
+    },
+    {
+      "epoch": 0.5413698630136986,
+      "grad_norm": 2.2007813453674316,
+      "learning_rate": 4.581640942323315e-06,
+      "loss": 1.9095,
+      "step": 494
+    },
+    {
+      "epoch": 0.5424657534246575,
+      "grad_norm": 2.251990795135498,
+      "learning_rate": 4.579610073111292e-06,
+      "loss": 2.0063,
+      "step": 495
+    },
+    {
+      "epoch": 0.5435616438356164,
+      "grad_norm": 2.774461507797241,
+      "learning_rate": 4.5775792038992695e-06,
+      "loss": 2.0236,
+      "step": 496
+    },
+    {
+      "epoch": 0.5446575342465754,
+      "grad_norm": 2.346486806869507,
+      "learning_rate": 4.575548334687247e-06,
+      "loss": 2.0146,
+      "step": 497
+    },
+    {
+      "epoch": 0.5457534246575343,
+      "grad_norm": 2.2475638389587402,
+      "learning_rate": 4.573517465475224e-06,
+      "loss": 2.0431,
+      "step": 498
+    },
+    {
+      "epoch": 0.5468493150684931,
+      "grad_norm": 2.385378122329712,
+      "learning_rate": 4.571486596263201e-06,
+      "loss": 1.943,
+      "step": 499
+    },
+    {
+      "epoch": 0.547945205479452,
+      "grad_norm": 2.4229276180267334,
+      "learning_rate": 4.569455727051178e-06,
+      "loss": 1.9989,
+      "step": 500
+    },
+    {
+      "epoch": 0.547945205479452,
+      "eval_loss": 2.1249091625213623,
+      "eval_runtime": 37.1158,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 500
+    },
+    {
+      "epoch": 0.549041095890411,
+      "grad_norm": 2.272860527038574,
+      "learning_rate": 4.567424857839156e-06,
+      "loss": 1.9964,
+      "step": 501
+    },
+    {
+      "epoch": 0.5501369863013699,
+      "grad_norm": 2.585275888442993,
+      "learning_rate": 4.565393988627132e-06,
+      "loss": 1.9761,
+      "step": 502
+    },
+    {
+      "epoch": 0.5512328767123288,
+      "grad_norm": 2.2231333255767822,
+      "learning_rate": 4.5633631194151105e-06,
+      "loss": 2.0429,
+      "step": 503
+    },
+    {
+      "epoch": 0.5523287671232877,
+      "grad_norm": 2.550520181655884,
+      "learning_rate": 4.561332250203087e-06,
+      "loss": 1.9702,
+      "step": 504
+    },
+    {
+      "epoch": 0.5534246575342465,
+      "grad_norm": 2.5849199295043945,
+      "learning_rate": 4.559301380991065e-06,
+      "loss": 1.9705,
+      "step": 505
+    },
+    {
+      "epoch": 0.5545205479452054,
+      "grad_norm": 2.135770559310913,
+      "learning_rate": 4.5572705117790414e-06,
+      "loss": 1.9109,
+      "step": 506
+    },
+    {
+      "epoch": 0.5556164383561644,
+      "grad_norm": 2.49159574508667,
+      "learning_rate": 4.5552396425670196e-06,
+      "loss": 1.9574,
+      "step": 507
+    },
+    {
+      "epoch": 0.5567123287671233,
+      "grad_norm": 2.4807381629943848,
+      "learning_rate": 4.553208773354996e-06,
+      "loss": 2.0113,
+      "step": 508
+    },
+    {
+      "epoch": 0.5578082191780822,
+      "grad_norm": 2.373417854309082,
+      "learning_rate": 4.551177904142974e-06,
+      "loss": 1.9272,
+      "step": 509
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "grad_norm": 2.057331085205078,
+      "learning_rate": 4.5491470349309505e-06,
+      "loss": 2.0213,
+      "step": 510
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "eval_loss": 2.1239938735961914,
+      "eval_runtime": 36.8487,
+      "eval_samples_per_second": 27.138,
+      "eval_steps_per_second": 1.14,
+      "step": 510
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 2.352627754211426,
+      "learning_rate": 4.547116165718929e-06,
+      "loss": 1.9467,
+      "step": 511
+    },
+    {
+      "epoch": 0.5610958904109589,
+      "grad_norm": 2.4589340686798096,
+      "learning_rate": 4.545085296506905e-06,
+      "loss": 1.9719,
+      "step": 512
+    },
+    {
+      "epoch": 0.5621917808219178,
+      "grad_norm": 2.0117125511169434,
+      "learning_rate": 4.543054427294882e-06,
+      "loss": 1.9388,
+      "step": 513
+    },
+    {
+      "epoch": 0.5632876712328767,
+      "grad_norm": 2.0877344608306885,
+      "learning_rate": 4.54102355808286e-06,
+      "loss": 1.9214,
+      "step": 514
+    },
+    {
+      "epoch": 0.5643835616438356,
+      "grad_norm": 2.4225330352783203,
+      "learning_rate": 4.538992688870837e-06,
+      "loss": 1.9312,
+      "step": 515
+    },
+    {
+      "epoch": 0.5654794520547946,
+      "grad_norm": 2.2232658863067627,
+      "learning_rate": 4.536961819658814e-06,
+      "loss": 1.9588,
+      "step": 516
+    },
+    {
+      "epoch": 0.5665753424657535,
+      "grad_norm": 2.0866599082946777,
+      "learning_rate": 4.5349309504467915e-06,
+      "loss": 1.9528,
+      "step": 517
+    },
+    {
+      "epoch": 0.5676712328767123,
+      "grad_norm": 2.272505760192871,
+      "learning_rate": 4.532900081234769e-06,
+      "loss": 1.9834,
+      "step": 518
+    },
+    {
+      "epoch": 0.5687671232876712,
+      "grad_norm": 2.2399017810821533,
+      "learning_rate": 4.530869212022746e-06,
+      "loss": 1.9479,
+      "step": 519
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "grad_norm": 2.8384666442871094,
+      "learning_rate": 4.528838342810723e-06,
+      "loss": 1.9756,
+      "step": 520
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "eval_loss": 2.1196842193603516,
+      "eval_runtime": 36.9011,
+      "eval_samples_per_second": 27.099,
+      "eval_steps_per_second": 1.138,
+      "step": 520
+    },
+    {
+      "epoch": 0.570958904109589,
+      "grad_norm": 2.231113910675049,
+      "learning_rate": 4.526807473598701e-06,
+      "loss": 1.9594,
+      "step": 521
+    },
+    {
+      "epoch": 0.572054794520548,
+      "grad_norm": 2.4858415126800537,
+      "learning_rate": 4.524776604386678e-06,
+      "loss": 1.9669,
+      "step": 522
+    },
+    {
+      "epoch": 0.5731506849315069,
+      "grad_norm": 2.2841124534606934,
+      "learning_rate": 4.522745735174655e-06,
+      "loss": 1.9429,
+      "step": 523
+    },
+    {
+      "epoch": 0.5742465753424657,
+      "grad_norm": 2.0490520000457764,
+      "learning_rate": 4.520714865962632e-06,
+      "loss": 1.9953,
+      "step": 524
+    },
+    {
+      "epoch": 0.5753424657534246,
+      "grad_norm": 2.2872908115386963,
+      "learning_rate": 4.51868399675061e-06,
+      "loss": 1.9729,
+      "step": 525
+    },
+    {
+      "epoch": 0.5764383561643835,
+      "grad_norm": 2.2335281372070312,
+      "learning_rate": 4.516653127538587e-06,
+      "loss": 1.9891,
+      "step": 526
+    },
+    {
+      "epoch": 0.5775342465753425,
+      "grad_norm": 2.0251572132110596,
+      "learning_rate": 4.514622258326564e-06,
+      "loss": 1.9387,
+      "step": 527
+    },
+    {
+      "epoch": 0.5786301369863014,
+      "grad_norm": 2.3072011470794678,
+      "learning_rate": 4.5125913891145415e-06,
+      "loss": 1.9275,
+      "step": 528
+    },
+    {
+      "epoch": 0.5797260273972603,
+      "grad_norm": 2.4449033737182617,
+      "learning_rate": 4.510560519902519e-06,
+      "loss": 1.9524,
+      "step": 529
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "grad_norm": 2.2075350284576416,
+      "learning_rate": 4.508529650690496e-06,
+      "loss": 1.99,
+      "step": 530
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "eval_loss": 2.1205594539642334,
+      "eval_runtime": 36.8406,
+      "eval_samples_per_second": 27.144,
+      "eval_steps_per_second": 1.14,
+      "step": 530
+    },
+    {
+      "epoch": 0.581917808219178,
+      "grad_norm": 2.148533821105957,
+      "learning_rate": 4.506498781478473e-06,
+      "loss": 1.9133,
+      "step": 531
+    },
+    {
+      "epoch": 0.583013698630137,
+      "grad_norm": 2.2660858631134033,
+      "learning_rate": 4.504467912266451e-06,
+      "loss": 1.9131,
+      "step": 532
+    },
+    {
+      "epoch": 0.5841095890410959,
+      "grad_norm": 2.31901478767395,
+      "learning_rate": 4.502437043054428e-06,
+      "loss": 1.955,
+      "step": 533
+    },
+    {
+      "epoch": 0.5852054794520548,
+      "grad_norm": 2.0394129753112793,
+      "learning_rate": 4.500406173842405e-06,
+      "loss": 2.0227,
+      "step": 534
+    },
+    {
+      "epoch": 0.5863013698630137,
+      "grad_norm": 2.2833235263824463,
+      "learning_rate": 4.4983753046303824e-06,
+      "loss": 1.9819,
+      "step": 535
+    },
+    {
+      "epoch": 0.5873972602739727,
+      "grad_norm": 2.2009031772613525,
+      "learning_rate": 4.49634443541836e-06,
+      "loss": 1.9587,
+      "step": 536
+    },
+    {
+      "epoch": 0.5884931506849315,
+      "grad_norm": 2.1384377479553223,
+      "learning_rate": 4.494313566206336e-06,
+      "loss": 1.9458,
+      "step": 537
+    },
+    {
+      "epoch": 0.5895890410958904,
+      "grad_norm": 2.2180237770080566,
+      "learning_rate": 4.492282696994314e-06,
+      "loss": 2.0082,
+      "step": 538
+    },
+    {
+      "epoch": 0.5906849315068493,
+      "grad_norm": 2.1079695224761963,
+      "learning_rate": 4.490251827782291e-06,
+      "loss": 1.9423,
+      "step": 539
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "grad_norm": 2.3531908988952637,
+      "learning_rate": 4.488220958570269e-06,
+      "loss": 1.911,
+      "step": 540
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "eval_loss": 2.11873197555542,
+      "eval_runtime": 36.7998,
+      "eval_samples_per_second": 27.174,
+      "eval_steps_per_second": 1.141,
+      "step": 540
+    },
+    {
+      "epoch": 0.5928767123287672,
+      "grad_norm": 2.173271417617798,
+      "learning_rate": 4.486190089358245e-06,
+      "loss": 1.942,
+      "step": 541
+    },
+    {
+      "epoch": 0.5939726027397261,
+      "grad_norm": 2.161085367202759,
+      "learning_rate": 4.484159220146223e-06,
+      "loss": 1.8837,
+      "step": 542
+    },
+    {
+      "epoch": 0.5950684931506849,
+      "grad_norm": 2.1999270915985107,
+      "learning_rate": 4.4821283509342e-06,
+      "loss": 1.9926,
+      "step": 543
+    },
+    {
+      "epoch": 0.5961643835616438,
+      "grad_norm": 2.0908172130584717,
+      "learning_rate": 4.480097481722178e-06,
+      "loss": 1.9738,
+      "step": 544
+    },
+    {
+      "epoch": 0.5972602739726027,
+      "grad_norm": 2.1115596294403076,
+      "learning_rate": 4.478066612510154e-06,
+      "loss": 1.9485,
+      "step": 545
+    },
+    {
+      "epoch": 0.5983561643835617,
+      "grad_norm": 2.0752980709075928,
+      "learning_rate": 4.4760357432981325e-06,
+      "loss": 1.9171,
+      "step": 546
+    },
+    {
+      "epoch": 0.5994520547945206,
+      "grad_norm": 2.1090850830078125,
+      "learning_rate": 4.474004874086109e-06,
+      "loss": 1.9363,
+      "step": 547
+    },
+    {
+      "epoch": 0.6005479452054795,
+      "grad_norm": 2.2388253211975098,
+      "learning_rate": 4.471974004874087e-06,
+      "loss": 2.079,
+      "step": 548
+    },
+    {
+      "epoch": 0.6016438356164383,
+      "grad_norm": 2.13826584815979,
+      "learning_rate": 4.4699431356620635e-06,
+      "loss": 1.9066,
+      "step": 549
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "grad_norm": 2.139592409133911,
+      "learning_rate": 4.467912266450041e-06,
+      "loss": 2.0115,
+      "step": 550
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "eval_loss": 2.1150012016296387,
+      "eval_runtime": 36.8293,
+      "eval_samples_per_second": 27.152,
+      "eval_steps_per_second": 1.14,
+      "step": 550
+    },
+    {
+      "epoch": 0.6038356164383561,
+      "grad_norm": 2.2159054279327393,
+      "learning_rate": 4.465881397238018e-06,
+      "loss": 1.953,
+      "step": 551
+    },
+    {
+      "epoch": 0.6049315068493151,
+      "grad_norm": 2.0024075508117676,
+      "learning_rate": 4.463850528025995e-06,
+      "loss": 1.9418,
+      "step": 552
+    },
+    {
+      "epoch": 0.606027397260274,
+      "grad_norm": 2.0435662269592285,
+      "learning_rate": 4.4618196588139726e-06,
+      "loss": 1.959,
+      "step": 553
+    },
+    {
+      "epoch": 0.6071232876712329,
+      "grad_norm": 2.204221487045288,
+      "learning_rate": 4.45978878960195e-06,
+      "loss": 1.9876,
+      "step": 554
+    },
+    {
+      "epoch": 0.6082191780821918,
+      "grad_norm": 2.1536989212036133,
+      "learning_rate": 4.457757920389927e-06,
+      "loss": 1.976,
+      "step": 555
+    },
+    {
+      "epoch": 0.6093150684931506,
+      "grad_norm": 2.1285719871520996,
+      "learning_rate": 4.455727051177904e-06,
+      "loss": 1.9185,
+      "step": 556
+    },
+    {
+      "epoch": 0.6104109589041096,
+      "grad_norm": 2.1217541694641113,
+      "learning_rate": 4.453696181965882e-06,
+      "loss": 1.9482,
+      "step": 557
+    },
+    {
+      "epoch": 0.6115068493150685,
+      "grad_norm": 2.1897120475769043,
+      "learning_rate": 4.451665312753859e-06,
+      "loss": 2.0028,
+      "step": 558
+    },
+    {
+      "epoch": 0.6126027397260274,
+      "grad_norm": 2.0617103576660156,
+      "learning_rate": 4.449634443541836e-06,
+      "loss": 1.9683,
+      "step": 559
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "grad_norm": 2.091261863708496,
+      "learning_rate": 4.4476035743298135e-06,
+      "loss": 2.0174,
+      "step": 560
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "eval_loss": 2.11283540725708,
+      "eval_runtime": 36.8348,
+      "eval_samples_per_second": 27.148,
+      "eval_steps_per_second": 1.14,
+      "step": 560
+    },
+    {
+      "epoch": 0.6147945205479453,
+      "grad_norm": 2.2024059295654297,
+      "learning_rate": 4.445572705117791e-06,
+      "loss": 1.9247,
+      "step": 561
+    },
+    {
+      "epoch": 0.6158904109589041,
+      "grad_norm": 2.234591484069824,
+      "learning_rate": 4.443541835905768e-06,
+      "loss": 1.9419,
+      "step": 562
+    },
+    {
+      "epoch": 0.616986301369863,
+      "grad_norm": 2.073786497116089,
+      "learning_rate": 4.441510966693745e-06,
+      "loss": 2.0076,
+      "step": 563
+    },
+    {
+      "epoch": 0.6180821917808219,
+      "grad_norm": 2.250511407852173,
+      "learning_rate": 4.439480097481723e-06,
+      "loss": 1.9475,
+      "step": 564
+    },
+    {
+      "epoch": 0.6191780821917808,
+      "grad_norm": 2.1097965240478516,
+      "learning_rate": 4.4374492282697e-06,
+      "loss": 2.0042,
+      "step": 565
+    },
+    {
+      "epoch": 0.6202739726027398,
+      "grad_norm": 2.1740925312042236,
+      "learning_rate": 4.435418359057677e-06,
+      "loss": 1.8782,
+      "step": 566
+    },
+    {
+      "epoch": 0.6213698630136987,
+      "grad_norm": 2.072244644165039,
+      "learning_rate": 4.433387489845654e-06,
+      "loss": 1.9604,
+      "step": 567
+    },
+    {
+      "epoch": 0.6224657534246575,
+      "grad_norm": 2.042480707168579,
+      "learning_rate": 4.431356620633632e-06,
+      "loss": 1.9673,
+      "step": 568
+    },
+    {
+      "epoch": 0.6235616438356164,
+      "grad_norm": 2.0622525215148926,
+      "learning_rate": 4.429325751421609e-06,
+      "loss": 1.9208,
+      "step": 569
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "grad_norm": 2.120314836502075,
+      "learning_rate": 4.427294882209586e-06,
+      "loss": 1.9273,
+      "step": 570
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "eval_loss": 2.1131231784820557,
+      "eval_runtime": 36.8113,
+      "eval_samples_per_second": 27.166,
+      "eval_steps_per_second": 1.141,
+      "step": 570
+    },
+    {
+      "epoch": 0.6257534246575343,
+      "grad_norm": 2.1246325969696045,
+      "learning_rate": 4.4252640129975635e-06,
+      "loss": 1.9538,
+      "step": 571
+    },
+    {
+      "epoch": 0.6268493150684932,
+      "grad_norm": 2.1085169315338135,
+      "learning_rate": 4.423233143785541e-06,
+      "loss": 1.8853,
+      "step": 572
+    },
+    {
+      "epoch": 0.6279452054794521,
+      "grad_norm": 2.0212948322296143,
+      "learning_rate": 4.421202274573518e-06,
+      "loss": 1.9515,
+      "step": 573
+    },
+    {
+      "epoch": 0.6290410958904109,
+      "grad_norm": 2.0736632347106934,
+      "learning_rate": 4.4191714053614945e-06,
+      "loss": 1.9543,
+      "step": 574
+    },
+    {
+      "epoch": 0.6301369863013698,
+      "grad_norm": 2.0515239238739014,
+      "learning_rate": 4.417140536149473e-06,
+      "loss": 1.854,
+      "step": 575
+    },
+    {
+      "epoch": 0.6312328767123287,
+      "grad_norm": 2.026869058609009,
+      "learning_rate": 4.415109666937449e-06,
+      "loss": 1.8974,
+      "step": 576
+    },
+    {
+      "epoch": 0.6323287671232877,
+      "grad_norm": 2.0393314361572266,
+      "learning_rate": 4.413078797725427e-06,
+      "loss": 1.9686,
+      "step": 577
+    },
+    {
+      "epoch": 0.6334246575342466,
+      "grad_norm": 2.063215494155884,
+      "learning_rate": 4.411047928513404e-06,
+      "loss": 1.9154,
+      "step": 578
+    },
+    {
+      "epoch": 0.6345205479452055,
+      "grad_norm": 2.167123556137085,
+      "learning_rate": 4.409017059301382e-06,
+      "loss": 1.9585,
+      "step": 579
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "grad_norm": 2.155423879623413,
+      "learning_rate": 4.406986190089358e-06,
+      "loss": 1.9644,
+      "step": 580
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "eval_loss": 2.1095986366271973,
+      "eval_runtime": 37.0543,
+      "eval_samples_per_second": 26.987,
+      "eval_steps_per_second": 1.133,
+      "step": 580
+    },
+    {
+      "epoch": 0.6367123287671232,
+      "grad_norm": 2.0016093254089355,
+      "learning_rate": 4.404955320877336e-06,
+      "loss": 1.9824,
+      "step": 581
+    },
+    {
+      "epoch": 0.6378082191780822,
+      "grad_norm": 2.0963051319122314,
+      "learning_rate": 4.402924451665313e-06,
+      "loss": 1.9713,
+      "step": 582
+    },
+    {
+      "epoch": 0.6389041095890411,
+      "grad_norm": 1.9784541130065918,
+      "learning_rate": 4.400893582453291e-06,
+      "loss": 1.9263,
+      "step": 583
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 1.9902760982513428,
+      "learning_rate": 4.398862713241267e-06,
+      "loss": 1.9,
+      "step": 584
+    },
+    {
+      "epoch": 0.6410958904109589,
+      "grad_norm": 2.1675639152526855,
+      "learning_rate": 4.396831844029245e-06,
+      "loss": 1.9422,
+      "step": 585
+    },
+    {
+      "epoch": 0.6421917808219179,
+      "grad_norm": 2.028651237487793,
+      "learning_rate": 4.394800974817222e-06,
+      "loss": 1.9559,
+      "step": 586
+    },
+    {
+      "epoch": 0.6432876712328767,
+      "grad_norm": 2.1689412593841553,
+      "learning_rate": 4.392770105605199e-06,
+      "loss": 1.945,
+      "step": 587
+    },
+    {
+      "epoch": 0.6443835616438356,
+      "grad_norm": 2.1114044189453125,
+      "learning_rate": 4.390739236393176e-06,
+      "loss": 1.9596,
+      "step": 588
+    },
+    {
+      "epoch": 0.6454794520547945,
+      "grad_norm": 2.117853879928589,
+      "learning_rate": 4.388708367181154e-06,
+      "loss": 1.9712,
+      "step": 589
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "grad_norm": 1.9970754384994507,
+      "learning_rate": 4.386677497969131e-06,
+      "loss": 1.9746,
+      "step": 590
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "eval_loss": 2.105872631072998,
+      "eval_runtime": 37.1042,
+      "eval_samples_per_second": 26.951,
+      "eval_steps_per_second": 1.132,
+      "step": 590
+    },
+    {
+      "epoch": 0.6476712328767124,
+      "grad_norm": 2.060075044631958,
+      "learning_rate": 4.384646628757108e-06,
+      "loss": 1.9602,
+      "step": 591
+    },
+    {
+      "epoch": 0.6487671232876713,
+      "grad_norm": 1.9594982862472534,
+      "learning_rate": 4.3826157595450855e-06,
+      "loss": 2.0127,
+      "step": 592
+    },
+    {
+      "epoch": 0.6498630136986301,
+      "grad_norm": 2.054703950881958,
+      "learning_rate": 4.380584890333063e-06,
+      "loss": 1.9324,
+      "step": 593
+    },
+    {
+      "epoch": 0.650958904109589,
+      "grad_norm": 2.041199207305908,
+      "learning_rate": 4.37855402112104e-06,
+      "loss": 1.8969,
+      "step": 594
+    },
+    {
+      "epoch": 0.6520547945205479,
+      "grad_norm": 2.0088963508605957,
+      "learning_rate": 4.376523151909017e-06,
+      "loss": 1.8449,
+      "step": 595
+    },
+    {
+      "epoch": 0.6531506849315069,
+      "grad_norm": 2.000422477722168,
+      "learning_rate": 4.3744922826969946e-06,
+      "loss": 2.0027,
+      "step": 596
+    },
+    {
+      "epoch": 0.6542465753424658,
+      "grad_norm": 2.008295774459839,
+      "learning_rate": 4.372461413484972e-06,
+      "loss": 1.9225,
+      "step": 597
+    },
+    {
+      "epoch": 0.6553424657534247,
+      "grad_norm": 2.252830982208252,
+      "learning_rate": 4.370430544272949e-06,
+      "loss": 1.9575,
+      "step": 598
+    },
+    {
+      "epoch": 0.6564383561643835,
+      "grad_norm": 2.0865542888641357,
+      "learning_rate": 4.368399675060926e-06,
+      "loss": 1.9791,
+      "step": 599
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "grad_norm": 2.2093706130981445,
+      "learning_rate": 4.366368805848904e-06,
+      "loss": 1.993,
+      "step": 600
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "eval_loss": 2.1032118797302246,
+      "eval_runtime": 37.1491,
+      "eval_samples_per_second": 26.919,
+      "eval_steps_per_second": 1.131,
+      "step": 600
+    },
+    {
+      "epoch": 0.6586301369863014,
+      "grad_norm": 2.096261739730835,
+      "learning_rate": 4.364337936636881e-06,
+      "loss": 1.9364,
+      "step": 601
+    },
+    {
+      "epoch": 0.6597260273972603,
+      "grad_norm": 2.3028810024261475,
+      "learning_rate": 4.362307067424858e-06,
+      "loss": 1.8922,
+      "step": 602
+    },
+    {
+      "epoch": 0.6608219178082192,
+      "grad_norm": 2.106886863708496,
+      "learning_rate": 4.3602761982128355e-06,
+      "loss": 1.9572,
+      "step": 603
+    },
+    {
+      "epoch": 0.6619178082191781,
+      "grad_norm": 2.1422243118286133,
+      "learning_rate": 4.358245329000813e-06,
+      "loss": 1.9132,
+      "step": 604
+    },
+    {
+      "epoch": 0.663013698630137,
+      "grad_norm": 2.0614800453186035,
+      "learning_rate": 4.35621445978879e-06,
+      "loss": 1.9356,
+      "step": 605
+    },
+    {
+      "epoch": 0.6641095890410958,
+      "grad_norm": 2.013725996017456,
+      "learning_rate": 4.354183590576767e-06,
+      "loss": 1.9516,
+      "step": 606
+    },
+    {
+      "epoch": 0.6652054794520548,
+      "grad_norm": 2.1779627799987793,
+      "learning_rate": 4.352152721364745e-06,
+      "loss": 1.9341,
+      "step": 607
+    },
+    {
+      "epoch": 0.6663013698630137,
+      "grad_norm": 2.1698031425476074,
+      "learning_rate": 4.350121852152722e-06,
+      "loss": 1.8796,
+      "step": 608
+    },
+    {
+      "epoch": 0.6673972602739726,
+      "grad_norm": 2.1617958545684814,
+      "learning_rate": 4.348090982940699e-06,
+      "loss": 1.9606,
+      "step": 609
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "grad_norm": 2.189724922180176,
+      "learning_rate": 4.3460601137286764e-06,
+      "loss": 1.9491,
+      "step": 610
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "eval_loss": 2.101118564605713,
+      "eval_runtime": 36.7937,
+      "eval_samples_per_second": 27.179,
+      "eval_steps_per_second": 1.141,
+      "step": 610
+    },
+    {
+      "epoch": 0.6695890410958905,
+      "grad_norm": 2.300693988800049,
+      "learning_rate": 4.344029244516654e-06,
+      "loss": 1.971,
+      "step": 611
+    },
+    {
+      "epoch": 0.6706849315068493,
+      "grad_norm": 2.2398195266723633,
+      "learning_rate": 4.341998375304631e-06,
+      "loss": 1.9289,
+      "step": 612
+    },
+    {
+      "epoch": 0.6717808219178082,
+      "grad_norm": 2.0529696941375732,
+      "learning_rate": 4.339967506092607e-06,
+      "loss": 2.0109,
+      "step": 613
+    },
+    {
+      "epoch": 0.6728767123287671,
+      "grad_norm": 2.0417604446411133,
+      "learning_rate": 4.3379366368805855e-06,
+      "loss": 1.9443,
+      "step": 614
+    },
+    {
+      "epoch": 0.673972602739726,
+      "grad_norm": 2.1295909881591797,
+      "learning_rate": 4.335905767668562e-06,
+      "loss": 1.942,
+      "step": 615
+    },
+    {
+      "epoch": 0.675068493150685,
+      "grad_norm": 2.1724905967712402,
+      "learning_rate": 4.33387489845654e-06,
+      "loss": 1.9272,
+      "step": 616
+    },
+    {
+      "epoch": 0.6761643835616439,
+      "grad_norm": 2.078840494155884,
+      "learning_rate": 4.3318440292445165e-06,
+      "loss": 1.9628,
+      "step": 617
+    },
+    {
+      "epoch": 0.6772602739726027,
+      "grad_norm": 2.073406934738159,
+      "learning_rate": 4.329813160032495e-06,
+      "loss": 1.9398,
+      "step": 618
+    },
+    {
+      "epoch": 0.6783561643835616,
+      "grad_norm": 2.3771591186523438,
+      "learning_rate": 4.327782290820471e-06,
+      "loss": 1.911,
+      "step": 619
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "grad_norm": 2.0965006351470947,
+      "learning_rate": 4.325751421608449e-06,
+      "loss": 1.9037,
+      "step": 620
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "eval_loss": 2.103893995285034,
+      "eval_runtime": 36.7441,
+      "eval_samples_per_second": 27.215,
+      "eval_steps_per_second": 1.143,
+      "step": 620
+    },
+    {
+      "epoch": 0.6805479452054795,
+      "grad_norm": 2.189476490020752,
+      "learning_rate": 4.323720552396426e-06,
+      "loss": 1.9794,
+      "step": 621
+    },
+    {
+      "epoch": 0.6816438356164384,
+      "grad_norm": 2.453953504562378,
+      "learning_rate": 4.321689683184404e-06,
+      "loss": 1.9224,
+      "step": 622
+    },
+    {
+      "epoch": 0.6827397260273973,
+      "grad_norm": 2.0869131088256836,
+      "learning_rate": 4.31965881397238e-06,
+      "loss": 1.9195,
+      "step": 623
+    },
+    {
+      "epoch": 0.6838356164383562,
+      "grad_norm": 2.181950807571411,
+      "learning_rate": 4.3176279447603574e-06,
+      "loss": 1.9214,
+      "step": 624
+    },
+    {
+      "epoch": 0.684931506849315,
+      "grad_norm": 2.0392794609069824,
+      "learning_rate": 4.315597075548335e-06,
+      "loss": 1.9694,
+      "step": 625
+    },
+    {
+      "epoch": 0.686027397260274,
+      "grad_norm": 2.268585205078125,
+      "learning_rate": 4.313566206336312e-06,
+      "loss": 1.9623,
+      "step": 626
+    },
+    {
+      "epoch": 0.6871232876712329,
+      "grad_norm": 2.0459189414978027,
+      "learning_rate": 4.311535337124289e-06,
+      "loss": 1.9815,
+      "step": 627
+    },
+    {
+      "epoch": 0.6882191780821918,
+      "grad_norm": 2.1525423526763916,
+      "learning_rate": 4.3095044679122665e-06,
+      "loss": 1.9296,
+      "step": 628
+    },
+    {
+      "epoch": 0.6893150684931507,
+      "grad_norm": 2.2998454570770264,
+      "learning_rate": 4.307473598700244e-06,
+      "loss": 1.9361,
+      "step": 629
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "grad_norm": 2.057570695877075,
+      "learning_rate": 4.305442729488221e-06,
+      "loss": 1.9503,
+      "step": 630
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "eval_loss": 2.1017305850982666,
+      "eval_runtime": 36.7458,
+      "eval_samples_per_second": 27.214,
+      "eval_steps_per_second": 1.143,
+      "step": 630
+    },
+    {
+      "epoch": 0.6915068493150685,
+      "grad_norm": 2.1565167903900146,
+      "learning_rate": 4.303411860276198e-06,
+      "loss": 1.8801,
+      "step": 631
+    },
+    {
+      "epoch": 0.6926027397260274,
+      "grad_norm": 2.100191593170166,
+      "learning_rate": 4.301380991064176e-06,
+      "loss": 1.9861,
+      "step": 632
+    },
+    {
+      "epoch": 0.6936986301369863,
+      "grad_norm": 2.0314157009124756,
+      "learning_rate": 4.299350121852153e-06,
+      "loss": 1.9588,
+      "step": 633
+    },
+    {
+      "epoch": 0.6947945205479452,
+      "grad_norm": 2.270892381668091,
+      "learning_rate": 4.29731925264013e-06,
+      "loss": 1.9797,
+      "step": 634
+    },
+    {
+      "epoch": 0.6958904109589041,
+      "grad_norm": 1.9957224130630493,
+      "learning_rate": 4.2952883834281075e-06,
+      "loss": 1.9834,
+      "step": 635
+    },
+    {
+      "epoch": 0.6969863013698631,
+      "grad_norm": 2.145697593688965,
+      "learning_rate": 4.293257514216085e-06,
+      "loss": 1.9332,
+      "step": 636
+    },
+    {
+      "epoch": 0.6980821917808219,
+      "grad_norm": 2.3837900161743164,
+      "learning_rate": 4.291226645004062e-06,
+      "loss": 1.8873,
+      "step": 637
+    },
+    {
+      "epoch": 0.6991780821917808,
+      "grad_norm": 1.9581835269927979,
+      "learning_rate": 4.289195775792039e-06,
+      "loss": 1.965,
+      "step": 638
+    },
+    {
+      "epoch": 0.7002739726027397,
+      "grad_norm": 2.091428518295288,
+      "learning_rate": 4.287164906580017e-06,
+      "loss": 1.9812,
+      "step": 639
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "grad_norm": 2.1069414615631104,
+      "learning_rate": 4.285134037367994e-06,
+      "loss": 1.9201,
+      "step": 640
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "eval_loss": 2.0906243324279785,
+      "eval_runtime": 36.8157,
+      "eval_samples_per_second": 27.162,
+      "eval_steps_per_second": 1.141,
+      "step": 640
+    },
+    {
+      "epoch": 0.7024657534246576,
+      "grad_norm": 1.9328324794769287,
+      "learning_rate": 4.283103168155971e-06,
+      "loss": 1.9108,
+      "step": 641
+    },
+    {
+      "epoch": 0.7035616438356165,
+      "grad_norm": 2.0553345680236816,
+      "learning_rate": 4.281072298943948e-06,
+      "loss": 1.888,
+      "step": 642
+    },
+    {
+      "epoch": 0.7046575342465753,
+      "grad_norm": 2.153369426727295,
+      "learning_rate": 4.279041429731926e-06,
+      "loss": 1.885,
+      "step": 643
+    },
+    {
+      "epoch": 0.7057534246575342,
+      "grad_norm": 2.1366844177246094,
+      "learning_rate": 4.277010560519903e-06,
+      "loss": 1.9395,
+      "step": 644
+    },
+    {
+      "epoch": 0.7068493150684931,
+      "grad_norm": 2.0654947757720947,
+      "learning_rate": 4.27497969130788e-06,
+      "loss": 1.9104,
+      "step": 645
+    },
+    {
+      "epoch": 0.707945205479452,
+      "grad_norm": 2.058098077774048,
+      "learning_rate": 4.2729488220958575e-06,
+      "loss": 1.9137,
+      "step": 646
+    },
+    {
+      "epoch": 0.709041095890411,
+      "grad_norm": 2.2112984657287598,
+      "learning_rate": 4.270917952883835e-06,
+      "loss": 1.956,
+      "step": 647
+    },
+    {
+      "epoch": 0.7101369863013699,
+      "grad_norm": 2.1838035583496094,
+      "learning_rate": 4.268887083671812e-06,
+      "loss": 1.9758,
+      "step": 648
+    },
+    {
+      "epoch": 0.7112328767123288,
+      "grad_norm": 2.142172336578369,
+      "learning_rate": 4.266856214459789e-06,
+      "loss": 1.8933,
+      "step": 649
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "grad_norm": 2.149646520614624,
+      "learning_rate": 4.264825345247766e-06,
+      "loss": 2.0118,
+      "step": 650
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "eval_loss": 2.0873661041259766,
+      "eval_runtime": 36.8064,
+      "eval_samples_per_second": 27.169,
+      "eval_steps_per_second": 1.141,
+      "step": 650
+    },
+    {
+      "epoch": 0.7134246575342466,
+      "grad_norm": 2.117799758911133,
+      "learning_rate": 4.262794476035744e-06,
+      "loss": 1.8852,
+      "step": 651
+    },
+    {
+      "epoch": 0.7145205479452055,
+      "grad_norm": 2.248161554336548,
+      "learning_rate": 4.26076360682372e-06,
+      "loss": 1.9385,
+      "step": 652
+    },
+    {
+      "epoch": 0.7156164383561644,
+      "grad_norm": 2.224853038787842,
+      "learning_rate": 4.2587327376116984e-06,
+      "loss": 1.8479,
+      "step": 653
+    },
+    {
+      "epoch": 0.7167123287671233,
+      "grad_norm": 2.2025816440582275,
+      "learning_rate": 4.256701868399675e-06,
+      "loss": 1.9087,
+      "step": 654
+    },
+    {
+      "epoch": 0.7178082191780822,
+      "grad_norm": 2.158379554748535,
+      "learning_rate": 4.254670999187653e-06,
+      "loss": 1.9438,
+      "step": 655
+    },
+    {
+      "epoch": 0.718904109589041,
+      "grad_norm": 2.09287428855896,
+      "learning_rate": 4.252640129975629e-06,
+      "loss": 1.9343,
+      "step": 656
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 2.166685104370117,
+      "learning_rate": 4.2506092607636075e-06,
+      "loss": 1.9391,
+      "step": 657
+    },
+    {
+      "epoch": 0.7210958904109589,
+      "grad_norm": 2.125105142593384,
+      "learning_rate": 4.248578391551584e-06,
+      "loss": 1.9191,
+      "step": 658
+    },
+    {
+      "epoch": 0.7221917808219178,
+      "grad_norm": 2.100684642791748,
+      "learning_rate": 4.246547522339562e-06,
+      "loss": 1.862,
+      "step": 659
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "grad_norm": 2.2361960411071777,
+      "learning_rate": 4.2445166531275385e-06,
+      "loss": 1.9038,
+      "step": 660
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "eval_loss": 2.0889461040496826,
+      "eval_runtime": 36.823,
+      "eval_samples_per_second": 27.157,
+      "eval_steps_per_second": 1.141,
+      "step": 660
+    },
+    {
+      "epoch": 0.7243835616438357,
+      "grad_norm": 2.309504270553589,
+      "learning_rate": 4.242485783915517e-06,
+      "loss": 1.9218,
+      "step": 661
+    },
+    {
+      "epoch": 0.7254794520547945,
+      "grad_norm": 2.102313280105591,
+      "learning_rate": 4.240454914703493e-06,
+      "loss": 1.8704,
+      "step": 662
+    },
+    {
+      "epoch": 0.7265753424657534,
+      "grad_norm": 2.2136611938476562,
+      "learning_rate": 4.23842404549147e-06,
+      "loss": 1.8814,
+      "step": 663
+    },
+    {
+      "epoch": 0.7276712328767123,
+      "grad_norm": 2.0127627849578857,
+      "learning_rate": 4.236393176279448e-06,
+      "loss": 1.8775,
+      "step": 664
+    },
+    {
+      "epoch": 0.7287671232876712,
+      "grad_norm": 2.192124366760254,
+      "learning_rate": 4.234362307067425e-06,
+      "loss": 1.8682,
+      "step": 665
+    },
+    {
+      "epoch": 0.7298630136986302,
+      "grad_norm": 2.4385814666748047,
+      "learning_rate": 4.232331437855402e-06,
+      "loss": 1.9512,
+      "step": 666
+    },
+    {
+      "epoch": 0.7309589041095891,
+      "grad_norm": 2.0868494510650635,
+      "learning_rate": 4.2303005686433795e-06,
+      "loss": 1.9669,
+      "step": 667
+    },
+    {
+      "epoch": 0.7320547945205479,
+      "grad_norm": 2.2664904594421387,
+      "learning_rate": 4.228269699431357e-06,
+      "loss": 1.971,
+      "step": 668
+    },
+    {
+      "epoch": 0.7331506849315068,
+      "grad_norm": 2.043488025665283,
+      "learning_rate": 4.226238830219334e-06,
+      "loss": 1.9259,
+      "step": 669
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "grad_norm": 2.0816924571990967,
+      "learning_rate": 4.224207961007311e-06,
+      "loss": 1.9357,
+      "step": 670
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "eval_loss": 2.093440294265747,
+      "eval_runtime": 36.8907,
+      "eval_samples_per_second": 27.107,
+      "eval_steps_per_second": 1.138,
+      "step": 670
+    },
+    {
+      "epoch": 0.7353424657534247,
+      "grad_norm": 2.134960889816284,
+      "learning_rate": 4.2221770917952886e-06,
+      "loss": 1.9234,
+      "step": 671
+    },
+    {
+      "epoch": 0.7364383561643836,
+      "grad_norm": 2.1693387031555176,
+      "learning_rate": 4.220146222583266e-06,
+      "loss": 1.9128,
+      "step": 672
+    },
+    {
+      "epoch": 0.7375342465753425,
+      "grad_norm": 2.2874763011932373,
+      "learning_rate": 4.218115353371243e-06,
+      "loss": 1.9364,
+      "step": 673
+    },
+    {
+      "epoch": 0.7386301369863014,
+      "grad_norm": 2.123231887817383,
+      "learning_rate": 4.21608448415922e-06,
+      "loss": 1.9281,
+      "step": 674
+    },
+    {
+      "epoch": 0.7397260273972602,
+      "grad_norm": 2.020890712738037,
+      "learning_rate": 4.214053614947198e-06,
+      "loss": 1.8954,
+      "step": 675
+    },
+    {
+      "epoch": 0.7408219178082192,
+      "grad_norm": 2.0817489624023438,
+      "learning_rate": 4.212022745735175e-06,
+      "loss": 1.9805,
+      "step": 676
+    },
+    {
+      "epoch": 0.7419178082191781,
+      "grad_norm": 2.266101360321045,
+      "learning_rate": 4.209991876523152e-06,
+      "loss": 1.8726,
+      "step": 677
+    },
+    {
+      "epoch": 0.743013698630137,
+      "grad_norm": 2.0021142959594727,
+      "learning_rate": 4.2079610073111295e-06,
+      "loss": 1.9528,
+      "step": 678
+    },
+    {
+      "epoch": 0.7441095890410959,
+      "grad_norm": 2.213418483734131,
+      "learning_rate": 4.205930138099107e-06,
+      "loss": 1.8749,
+      "step": 679
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "grad_norm": 2.2187159061431885,
+      "learning_rate": 4.203899268887084e-06,
+      "loss": 1.9787,
+      "step": 680
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "eval_loss": 2.0915141105651855,
+      "eval_runtime": 37.0523,
+      "eval_samples_per_second": 26.989,
+      "eval_steps_per_second": 1.134,
+      "step": 680
+    },
+    {
+      "epoch": 0.7463013698630137,
+      "grad_norm": 1.986974835395813,
+      "learning_rate": 4.201868399675061e-06,
+      "loss": 1.9257,
+      "step": 681
+    },
+    {
+      "epoch": 0.7473972602739726,
+      "grad_norm": 2.0920112133026123,
+      "learning_rate": 4.199837530463039e-06,
+      "loss": 1.9743,
+      "step": 682
+    },
+    {
+      "epoch": 0.7484931506849315,
+      "grad_norm": 2.2249104976654053,
+      "learning_rate": 4.197806661251016e-06,
+      "loss": 1.8572,
+      "step": 683
+    },
+    {
+      "epoch": 0.7495890410958904,
+      "grad_norm": 2.0482609272003174,
+      "learning_rate": 4.195775792038993e-06,
+      "loss": 1.9271,
+      "step": 684
+    },
+    {
+      "epoch": 0.7506849315068493,
+      "grad_norm": 2.1768155097961426,
+      "learning_rate": 4.19374492282697e-06,
+      "loss": 1.9097,
+      "step": 685
+    },
+    {
+      "epoch": 0.7517808219178083,
+      "grad_norm": 2.1729984283447266,
+      "learning_rate": 4.191714053614948e-06,
+      "loss": 1.8393,
+      "step": 686
+    },
+    {
+      "epoch": 0.7528767123287671,
+      "grad_norm": 2.2250611782073975,
+      "learning_rate": 4.189683184402924e-06,
+      "loss": 1.9426,
+      "step": 687
+    },
+    {
+      "epoch": 0.753972602739726,
+      "grad_norm": 2.1344521045684814,
+      "learning_rate": 4.187652315190902e-06,
+      "loss": 1.8935,
+      "step": 688
+    },
+    {
+      "epoch": 0.7550684931506849,
+      "grad_norm": 2.2652342319488525,
+      "learning_rate": 4.185621445978879e-06,
+      "loss": 1.9631,
+      "step": 689
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "grad_norm": 2.2338240146636963,
+      "learning_rate": 4.183590576766857e-06,
+      "loss": 1.8911,
+      "step": 690
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "eval_loss": 2.0889363288879395,
+      "eval_runtime": 37.1002,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 690
+    },
+    {
+      "epoch": 0.7572602739726028,
+      "grad_norm": 2.099368095397949,
+      "learning_rate": 4.181559707554833e-06,
+      "loss": 1.9124,
+      "step": 691
+    },
+    {
+      "epoch": 0.7583561643835617,
+      "grad_norm": 2.2774770259857178,
+      "learning_rate": 4.179528838342811e-06,
+      "loss": 1.9653,
+      "step": 692
+    },
+    {
+      "epoch": 0.7594520547945205,
+      "grad_norm": 2.1390492916107178,
+      "learning_rate": 4.177497969130788e-06,
+      "loss": 1.883,
+      "step": 693
+    },
+    {
+      "epoch": 0.7605479452054794,
+      "grad_norm": 2.1724212169647217,
+      "learning_rate": 4.175467099918766e-06,
+      "loss": 1.8933,
+      "step": 694
+    },
+    {
+      "epoch": 0.7616438356164383,
+      "grad_norm": 2.0856740474700928,
+      "learning_rate": 4.173436230706742e-06,
+      "loss": 1.9805,
+      "step": 695
+    },
+    {
+      "epoch": 0.7627397260273973,
+      "grad_norm": 2.1491377353668213,
+      "learning_rate": 4.1714053614947205e-06,
+      "loss": 1.9129,
+      "step": 696
+    },
+    {
+      "epoch": 0.7638356164383562,
+      "grad_norm": 2.207265615463257,
+      "learning_rate": 4.169374492282697e-06,
+      "loss": 1.8939,
+      "step": 697
+    },
+    {
+      "epoch": 0.7649315068493151,
+      "grad_norm": 2.091902494430542,
+      "learning_rate": 4.167343623070675e-06,
+      "loss": 1.9003,
+      "step": 698
+    },
+    {
+      "epoch": 0.766027397260274,
+      "grad_norm": 2.048015594482422,
+      "learning_rate": 4.165312753858652e-06,
+      "loss": 1.9565,
+      "step": 699
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "grad_norm": 2.1699411869049072,
+      "learning_rate": 4.163281884646629e-06,
+      "loss": 1.9066,
+      "step": 700
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "eval_loss": 2.0834600925445557,
+      "eval_runtime": 36.82,
+      "eval_samples_per_second": 27.159,
+      "eval_steps_per_second": 1.141,
+      "step": 700
+    },
+    {
+      "epoch": 0.7682191780821918,
+      "grad_norm": 2.1105830669403076,
+      "learning_rate": 4.161251015434607e-06,
+      "loss": 1.9024,
+      "step": 701
+    },
+    {
+      "epoch": 0.7693150684931507,
+      "grad_norm": 2.0144476890563965,
+      "learning_rate": 4.159220146222583e-06,
+      "loss": 1.9268,
+      "step": 702
+    },
+    {
+      "epoch": 0.7704109589041096,
+      "grad_norm": 2.060319662094116,
+      "learning_rate": 4.157189277010561e-06,
+      "loss": 1.879,
+      "step": 703
+    },
+    {
+      "epoch": 0.7715068493150685,
+      "grad_norm": 2.14819598197937,
+      "learning_rate": 4.155158407798538e-06,
+      "loss": 1.9073,
+      "step": 704
+    },
+    {
+      "epoch": 0.7726027397260274,
+      "grad_norm": 2.2959697246551514,
+      "learning_rate": 4.153127538586516e-06,
+      "loss": 1.947,
+      "step": 705
+    },
+    {
+      "epoch": 0.7736986301369863,
+      "grad_norm": 2.1901297569274902,
+      "learning_rate": 4.151096669374492e-06,
+      "loss": 1.9127,
+      "step": 706
+    },
+    {
+      "epoch": 0.7747945205479452,
+      "grad_norm": 2.264690637588501,
+      "learning_rate": 4.1490658001624705e-06,
+      "loss": 1.9376,
+      "step": 707
+    },
+    {
+      "epoch": 0.7758904109589041,
+      "grad_norm": 2.155043601989746,
+      "learning_rate": 4.147034930950447e-06,
+      "loss": 1.9048,
+      "step": 708
+    },
+    {
+      "epoch": 0.776986301369863,
+      "grad_norm": 2.059197425842285,
+      "learning_rate": 4.145004061738425e-06,
+      "loss": 1.9049,
+      "step": 709
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "grad_norm": 2.4279260635375977,
+      "learning_rate": 4.1429731925264015e-06,
+      "loss": 1.9038,
+      "step": 710
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "eval_loss": 2.0821099281311035,
+      "eval_runtime": 36.8421,
+      "eval_samples_per_second": 27.143,
+      "eval_steps_per_second": 1.14,
+      "step": 710
+    },
+    {
+      "epoch": 0.7791780821917809,
+      "grad_norm": 2.1462533473968506,
+      "learning_rate": 4.140942323314379e-06,
+      "loss": 1.8325,
+      "step": 711
+    },
+    {
+      "epoch": 0.7802739726027397,
+      "grad_norm": 2.3161439895629883,
+      "learning_rate": 4.138911454102356e-06,
+      "loss": 1.8739,
+      "step": 712
+    },
+    {
+      "epoch": 0.7813698630136986,
+      "grad_norm": 2.130906343460083,
+      "learning_rate": 4.136880584890333e-06,
+      "loss": 1.8985,
+      "step": 713
+    },
+    {
+      "epoch": 0.7824657534246575,
+      "grad_norm": 2.107179641723633,
+      "learning_rate": 4.1348497156783106e-06,
+      "loss": 1.9336,
+      "step": 714
+    },
+    {
+      "epoch": 0.7835616438356164,
+      "grad_norm": 2.4675867557525635,
+      "learning_rate": 4.132818846466288e-06,
+      "loss": 1.9318,
+      "step": 715
+    },
+    {
+      "epoch": 0.7846575342465754,
+      "grad_norm": 2.1309103965759277,
+      "learning_rate": 4.130787977254265e-06,
+      "loss": 1.9279,
+      "step": 716
+    },
+    {
+      "epoch": 0.7857534246575343,
+      "grad_norm": 2.2152087688446045,
+      "learning_rate": 4.128757108042242e-06,
+      "loss": 1.9428,
+      "step": 717
+    },
+    {
+      "epoch": 0.7868493150684932,
+      "grad_norm": 2.02030873298645,
+      "learning_rate": 4.12672623883022e-06,
+      "loss": 1.9456,
+      "step": 718
+    },
+    {
+      "epoch": 0.787945205479452,
+      "grad_norm": 2.2501220703125,
+      "learning_rate": 4.124695369618197e-06,
+      "loss": 1.9994,
+      "step": 719
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "grad_norm": 2.082099199295044,
+      "learning_rate": 4.122664500406174e-06,
+      "loss": 1.9181,
+      "step": 720
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "eval_loss": 2.0795793533325195,
+      "eval_runtime": 36.8419,
+      "eval_samples_per_second": 27.143,
+      "eval_steps_per_second": 1.14,
+      "step": 720
+    },
+    {
+      "epoch": 0.7901369863013699,
+      "grad_norm": 2.121677875518799,
+      "learning_rate": 4.1206336311941515e-06,
+      "loss": 1.9112,
+      "step": 721
+    },
+    {
+      "epoch": 0.7912328767123288,
+      "grad_norm": 2.154696226119995,
+      "learning_rate": 4.118602761982129e-06,
+      "loss": 1.9328,
+      "step": 722
+    },
+    {
+      "epoch": 0.7923287671232877,
+      "grad_norm": 2.141899347305298,
+      "learning_rate": 4.116571892770106e-06,
+      "loss": 1.8682,
+      "step": 723
+    },
+    {
+      "epoch": 0.7934246575342466,
+      "grad_norm": 2.1390345096588135,
+      "learning_rate": 4.114541023558083e-06,
+      "loss": 1.85,
+      "step": 724
+    },
+    {
+      "epoch": 0.7945205479452054,
+      "grad_norm": 2.22886061668396,
+      "learning_rate": 4.112510154346061e-06,
+      "loss": 1.9754,
+      "step": 725
+    },
+    {
+      "epoch": 0.7956164383561644,
+      "grad_norm": 2.2299511432647705,
+      "learning_rate": 4.110479285134038e-06,
+      "loss": 1.8985,
+      "step": 726
+    },
+    {
+      "epoch": 0.7967123287671233,
+      "grad_norm": 2.251913070678711,
+      "learning_rate": 4.108448415922015e-06,
+      "loss": 1.918,
+      "step": 727
+    },
+    {
+      "epoch": 0.7978082191780822,
+      "grad_norm": 2.2145745754241943,
+      "learning_rate": 4.1064175467099924e-06,
+      "loss": 1.9469,
+      "step": 728
+    },
+    {
+      "epoch": 0.7989041095890411,
+      "grad_norm": 2.0919570922851562,
+      "learning_rate": 4.10438667749797e-06,
+      "loss": 1.9395,
+      "step": 729
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 2.3385298252105713,
+      "learning_rate": 4.102355808285947e-06,
+      "loss": 1.7958,
+      "step": 730
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 2.081737756729126,
+      "eval_runtime": 36.7686,
+      "eval_samples_per_second": 27.197,
+      "eval_steps_per_second": 1.142,
+      "step": 730
+    },
+    {
+      "epoch": 0.8010958904109589,
+      "grad_norm": 2.3434054851531982,
+      "learning_rate": 4.100324939073924e-06,
+      "loss": 1.9948,
+      "step": 731
+    },
+    {
+      "epoch": 0.8021917808219178,
+      "grad_norm": 2.129086971282959,
+      "learning_rate": 4.0982940698619015e-06,
+      "loss": 1.8994,
+      "step": 732
+    },
+    {
+      "epoch": 0.8032876712328767,
+      "grad_norm": 2.2564284801483154,
+      "learning_rate": 4.096263200649879e-06,
+      "loss": 1.8951,
+      "step": 733
+    },
+    {
+      "epoch": 0.8043835616438356,
+      "grad_norm": 2.53267240524292,
+      "learning_rate": 4.094232331437856e-06,
+      "loss": 1.9083,
+      "step": 734
+    },
+    {
+      "epoch": 0.8054794520547945,
+      "grad_norm": 2.164807081222534,
+      "learning_rate": 4.092201462225833e-06,
+      "loss": 1.8775,
+      "step": 735
+    },
+    {
+      "epoch": 0.8065753424657535,
+      "grad_norm": 2.3593950271606445,
+      "learning_rate": 4.090170593013811e-06,
+      "loss": 1.8777,
+      "step": 736
+    },
+    {
+      "epoch": 0.8076712328767123,
+      "grad_norm": 2.470942258834839,
+      "learning_rate": 4.088139723801787e-06,
+      "loss": 1.9196,
+      "step": 737
+    },
+    {
+      "epoch": 0.8087671232876712,
+      "grad_norm": 2.0244338512420654,
+      "learning_rate": 4.086108854589765e-06,
+      "loss": 1.86,
+      "step": 738
+    },
+    {
+      "epoch": 0.8098630136986301,
+      "grad_norm": 2.2432093620300293,
+      "learning_rate": 4.084077985377742e-06,
+      "loss": 1.8832,
+      "step": 739
+    },
+    {
+      "epoch": 0.810958904109589,
+      "grad_norm": 2.087015390396118,
+      "learning_rate": 4.08204711616572e-06,
+      "loss": 1.882,
+      "step": 740
+    },
+    {
+      "epoch": 0.810958904109589,
+      "eval_loss": 2.0833325386047363,
+      "eval_runtime": 36.8099,
+      "eval_samples_per_second": 27.167,
+      "eval_steps_per_second": 1.141,
+      "step": 740
+    },
+    {
+      "epoch": 0.812054794520548,
+      "grad_norm": 2.038623571395874,
+      "learning_rate": 4.080016246953696e-06,
+      "loss": 1.9526,
+      "step": 741
+    },
+    {
+      "epoch": 0.8131506849315069,
+      "grad_norm": 2.1820836067199707,
+      "learning_rate": 4.077985377741674e-06,
+      "loss": 1.8739,
+      "step": 742
+    },
+    {
+      "epoch": 0.8142465753424658,
+      "grad_norm": 2.1047675609588623,
+      "learning_rate": 4.075954508529651e-06,
+      "loss": 1.941,
+      "step": 743
+    },
+    {
+      "epoch": 0.8153424657534246,
+      "grad_norm": 2.1180074214935303,
+      "learning_rate": 4.073923639317629e-06,
+      "loss": 1.9522,
+      "step": 744
+    },
+    {
+      "epoch": 0.8164383561643835,
+      "grad_norm": 2.1034464836120605,
+      "learning_rate": 4.071892770105605e-06,
+      "loss": 1.9644,
+      "step": 745
+    },
+    {
+      "epoch": 0.8175342465753425,
+      "grad_norm": 2.1570026874542236,
+      "learning_rate": 4.069861900893583e-06,
+      "loss": 1.8953,
+      "step": 746
+    },
+    {
+      "epoch": 0.8186301369863014,
+      "grad_norm": 2.041189432144165,
+      "learning_rate": 4.06783103168156e-06,
+      "loss": 1.8557,
+      "step": 747
+    },
+    {
+      "epoch": 0.8197260273972603,
+      "grad_norm": 2.1053237915039062,
+      "learning_rate": 4.065800162469538e-06,
+      "loss": 1.8746,
+      "step": 748
+    },
+    {
+      "epoch": 0.8208219178082192,
+      "grad_norm": 2.223841667175293,
+      "learning_rate": 4.063769293257514e-06,
+      "loss": 1.8773,
+      "step": 749
+    },
+    {
+      "epoch": 0.821917808219178,
+      "grad_norm": 2.2014336585998535,
+      "learning_rate": 4.061738424045492e-06,
+      "loss": 1.9193,
+      "step": 750
+    },
+    {
+      "epoch": 0.821917808219178,
+      "eval_loss": 2.0867912769317627,
+      "eval_runtime": 36.8009,
+      "eval_samples_per_second": 27.173,
+      "eval_steps_per_second": 1.141,
+      "step": 750
+    },
+    {
+      "epoch": 0.823013698630137,
+      "grad_norm": 2.1470754146575928,
+      "learning_rate": 4.059707554833469e-06,
+      "loss": 1.8899,
+      "step": 751
+    },
+    {
+      "epoch": 0.8241095890410959,
+      "grad_norm": 2.1802284717559814,
+      "learning_rate": 4.057676685621446e-06,
+      "loss": 1.9144,
+      "step": 752
+    },
+    {
+      "epoch": 0.8252054794520548,
+      "grad_norm": 2.139535665512085,
+      "learning_rate": 4.0556458164094235e-06,
+      "loss": 1.9432,
+      "step": 753
+    },
+    {
+      "epoch": 0.8263013698630137,
+      "grad_norm": 2.116802215576172,
+      "learning_rate": 4.053614947197401e-06,
+      "loss": 1.9108,
+      "step": 754
+    },
+    {
+      "epoch": 0.8273972602739726,
+      "grad_norm": 2.031766414642334,
+      "learning_rate": 4.051584077985378e-06,
+      "loss": 1.8275,
+      "step": 755
+    },
+    {
+      "epoch": 0.8284931506849315,
+      "grad_norm": 2.2119126319885254,
+      "learning_rate": 4.049553208773355e-06,
+      "loss": 1.9109,
+      "step": 756
+    },
+    {
+      "epoch": 0.8295890410958904,
+      "grad_norm": 2.1019320487976074,
+      "learning_rate": 4.047522339561333e-06,
+      "loss": 1.9453,
+      "step": 757
+    },
+    {
+      "epoch": 0.8306849315068493,
+      "grad_norm": 2.1405951976776123,
+      "learning_rate": 4.04549147034931e-06,
+      "loss": 1.8961,
+      "step": 758
+    },
+    {
+      "epoch": 0.8317808219178082,
+      "grad_norm": 2.067448616027832,
+      "learning_rate": 4.043460601137287e-06,
+      "loss": 1.896,
+      "step": 759
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "grad_norm": 2.099836587905884,
+      "learning_rate": 4.041429731925264e-06,
+      "loss": 2.0224,
+      "step": 760
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "eval_loss": 2.07991886138916,
+      "eval_runtime": 36.8178,
+      "eval_samples_per_second": 27.161,
+      "eval_steps_per_second": 1.141,
+      "step": 760
+    },
+    {
+      "epoch": 0.8339726027397261,
+      "grad_norm": 2.2045092582702637,
+      "learning_rate": 4.039398862713242e-06,
+      "loss": 1.9005,
+      "step": 761
+    },
+    {
+      "epoch": 0.8350684931506849,
+      "grad_norm": 2.096851348876953,
+      "learning_rate": 4.037367993501219e-06,
+      "loss": 1.901,
+      "step": 762
+    },
+    {
+      "epoch": 0.8361643835616438,
+      "grad_norm": 2.09584903717041,
+      "learning_rate": 4.035337124289196e-06,
+      "loss": 1.9619,
+      "step": 763
+    },
+    {
+      "epoch": 0.8372602739726027,
+      "grad_norm": 2.236417770385742,
+      "learning_rate": 4.0333062550771735e-06,
+      "loss": 1.8862,
+      "step": 764
+    },
+    {
+      "epoch": 0.8383561643835616,
+      "grad_norm": 2.298532009124756,
+      "learning_rate": 4.031275385865151e-06,
+      "loss": 1.8609,
+      "step": 765
+    },
+    {
+      "epoch": 0.8394520547945206,
+      "grad_norm": 2.2270593643188477,
+      "learning_rate": 4.029244516653128e-06,
+      "loss": 1.8942,
+      "step": 766
+    },
+    {
+      "epoch": 0.8405479452054795,
+      "grad_norm": 2.2111141681671143,
+      "learning_rate": 4.027213647441105e-06,
+      "loss": 1.9365,
+      "step": 767
+    },
+    {
+      "epoch": 0.8416438356164384,
+      "grad_norm": 2.1535582542419434,
+      "learning_rate": 4.025182778229083e-06,
+      "loss": 1.8393,
+      "step": 768
+    },
+    {
+      "epoch": 0.8427397260273972,
+      "grad_norm": 2.0704867839813232,
+      "learning_rate": 4.02315190901706e-06,
+      "loss": 1.8793,
+      "step": 769
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "grad_norm": 2.2168002128601074,
+      "learning_rate": 4.021121039805037e-06,
+      "loss": 1.9023,
+      "step": 770
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "eval_loss": 2.0816757678985596,
+      "eval_runtime": 36.8186,
+      "eval_samples_per_second": 27.16,
+      "eval_steps_per_second": 1.141,
+      "step": 770
+    },
+    {
+      "epoch": 0.8449315068493151,
+      "grad_norm": 2.4146242141723633,
+      "learning_rate": 4.0190901705930144e-06,
+      "loss": 1.9016,
+      "step": 771
+    },
+    {
+      "epoch": 0.846027397260274,
+      "grad_norm": 2.2301456928253174,
+      "learning_rate": 4.017059301380992e-06,
+      "loss": 1.9945,
+      "step": 772
+    },
+    {
+      "epoch": 0.8471232876712329,
+      "grad_norm": 2.299612283706665,
+      "learning_rate": 4.015028432168969e-06,
+      "loss": 1.8808,
+      "step": 773
+    },
+    {
+      "epoch": 0.8482191780821918,
+      "grad_norm": 2.161613941192627,
+      "learning_rate": 4.012997562956945e-06,
+      "loss": 1.9481,
+      "step": 774
+    },
+    {
+      "epoch": 0.8493150684931506,
+      "grad_norm": 2.0541019439697266,
+      "learning_rate": 4.0109666937449235e-06,
+      "loss": 1.8609,
+      "step": 775
+    },
+    {
+      "epoch": 0.8504109589041096,
+      "grad_norm": 2.0485072135925293,
+      "learning_rate": 4.0089358245329e-06,
+      "loss": 1.8789,
+      "step": 776
+    },
+    {
+      "epoch": 0.8515068493150685,
+      "grad_norm": 1.985097050666809,
+      "learning_rate": 4.006904955320878e-06,
+      "loss": 1.8845,
+      "step": 777
+    },
+    {
+      "epoch": 0.8526027397260274,
+      "grad_norm": 2.200528621673584,
+      "learning_rate": 4.0048740861088545e-06,
+      "loss": 1.8616,
+      "step": 778
+    },
+    {
+      "epoch": 0.8536986301369863,
+      "grad_norm": 2.003523349761963,
+      "learning_rate": 4.002843216896833e-06,
+      "loss": 1.9393,
+      "step": 779
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "grad_norm": 2.208388090133667,
+      "learning_rate": 4.000812347684809e-06,
+      "loss": 1.9272,
+      "step": 780
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "eval_loss": 2.074138641357422,
+      "eval_runtime": 37.0322,
+      "eval_samples_per_second": 27.004,
+      "eval_steps_per_second": 1.134,
+      "step": 780
+    },
+    {
+      "epoch": 0.8558904109589041,
+      "grad_norm": 2.384833812713623,
+      "learning_rate": 3.998781478472787e-06,
+      "loss": 1.8743,
+      "step": 781
+    },
+    {
+      "epoch": 0.856986301369863,
+      "grad_norm": 2.3035428524017334,
+      "learning_rate": 3.996750609260764e-06,
+      "loss": 1.8877,
+      "step": 782
+    },
+    {
+      "epoch": 0.8580821917808219,
+      "grad_norm": 2.2021212577819824,
+      "learning_rate": 3.994719740048742e-06,
+      "loss": 1.8697,
+      "step": 783
+    },
+    {
+      "epoch": 0.8591780821917808,
+      "grad_norm": 2.4477264881134033,
+      "learning_rate": 3.992688870836718e-06,
+      "loss": 2.004,
+      "step": 784
+    },
+    {
+      "epoch": 0.8602739726027397,
+      "grad_norm": 2.5018362998962402,
+      "learning_rate": 3.990658001624696e-06,
+      "loss": 1.911,
+      "step": 785
+    },
+    {
+      "epoch": 0.8613698630136987,
+      "grad_norm": 2.1495399475097656,
+      "learning_rate": 3.988627132412673e-06,
+      "loss": 1.8654,
+      "step": 786
+    },
+    {
+      "epoch": 0.8624657534246576,
+      "grad_norm": 2.249499559402466,
+      "learning_rate": 3.98659626320065e-06,
+      "loss": 1.9429,
+      "step": 787
+    },
+    {
+      "epoch": 0.8635616438356164,
+      "grad_norm": 2.622509002685547,
+      "learning_rate": 3.984565393988627e-06,
+      "loss": 1.8938,
+      "step": 788
+    },
+    {
+      "epoch": 0.8646575342465753,
+      "grad_norm": 2.1475276947021484,
+      "learning_rate": 3.9825345247766046e-06,
+      "loss": 1.9014,
+      "step": 789
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "grad_norm": 2.3051605224609375,
+      "learning_rate": 3.980503655564582e-06,
+      "loss": 1.906,
+      "step": 790
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "eval_loss": 2.0729031562805176,
+      "eval_runtime": 37.0595,
+      "eval_samples_per_second": 26.984,
+      "eval_steps_per_second": 1.133,
+      "step": 790
+    },
+    {
+      "epoch": 0.8668493150684932,
+      "grad_norm": 2.2738583087921143,
+      "learning_rate": 3.978472786352559e-06,
+      "loss": 1.9218,
+      "step": 791
+    },
+    {
+      "epoch": 0.8679452054794521,
+      "grad_norm": 2.2438387870788574,
+      "learning_rate": 3.976441917140536e-06,
+      "loss": 1.9311,
+      "step": 792
+    },
+    {
+      "epoch": 0.869041095890411,
+      "grad_norm": 2.1148440837860107,
+      "learning_rate": 3.974411047928514e-06,
+      "loss": 1.9059,
+      "step": 793
+    },
+    {
+      "epoch": 0.8701369863013698,
+      "grad_norm": 2.077681541442871,
+      "learning_rate": 3.972380178716491e-06,
+      "loss": 1.9682,
+      "step": 794
+    },
+    {
+      "epoch": 0.8712328767123287,
+      "grad_norm": 2.0375349521636963,
+      "learning_rate": 3.970349309504468e-06,
+      "loss": 1.9329,
+      "step": 795
+    },
+    {
+      "epoch": 0.8723287671232877,
+      "grad_norm": 2.135745048522949,
+      "learning_rate": 3.9683184402924455e-06,
+      "loss": 1.9393,
+      "step": 796
+    },
+    {
+      "epoch": 0.8734246575342466,
+      "grad_norm": 2.0594282150268555,
+      "learning_rate": 3.966287571080423e-06,
+      "loss": 1.8535,
+      "step": 797
+    },
+    {
+      "epoch": 0.8745205479452055,
+      "grad_norm": 2.181072950363159,
+      "learning_rate": 3.9642567018684e-06,
+      "loss": 1.8401,
+      "step": 798
+    },
+    {
+      "epoch": 0.8756164383561644,
+      "grad_norm": 2.0925304889678955,
+      "learning_rate": 3.962225832656377e-06,
+      "loss": 1.9216,
+      "step": 799
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "grad_norm": 2.0854146480560303,
+      "learning_rate": 3.960194963444355e-06,
+      "loss": 1.8973,
+      "step": 800
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "eval_loss": 2.06680965423584,
+      "eval_runtime": 36.8115,
+      "eval_samples_per_second": 27.165,
+      "eval_steps_per_second": 1.141,
+      "step": 800
+    },
+    {
+      "epoch": 0.8778082191780822,
+      "grad_norm": 2.068798542022705,
+      "learning_rate": 3.958164094232332e-06,
+      "loss": 1.8216,
+      "step": 801
+    },
+    {
+      "epoch": 0.8789041095890411,
+      "grad_norm": 2.124248504638672,
+      "learning_rate": 3.956133225020309e-06,
+      "loss": 1.9033,
+      "step": 802
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 2.0681240558624268,
+      "learning_rate": 3.954102355808286e-06,
+      "loss": 1.8997,
+      "step": 803
+    },
+    {
+      "epoch": 0.8810958904109589,
+      "grad_norm": 2.1278390884399414,
+      "learning_rate": 3.952071486596264e-06,
+      "loss": 1.9053,
+      "step": 804
+    },
+    {
+      "epoch": 0.8821917808219178,
+      "grad_norm": 2.20721435546875,
+      "learning_rate": 3.950040617384241e-06,
+      "loss": 1.9038,
+      "step": 805
+    },
+    {
+      "epoch": 0.8832876712328767,
+      "grad_norm": 2.221014976501465,
+      "learning_rate": 3.948009748172218e-06,
+      "loss": 1.8865,
+      "step": 806
+    },
+    {
+      "epoch": 0.8843835616438356,
+      "grad_norm": 2.0922741889953613,
+      "learning_rate": 3.9459788789601955e-06,
+      "loss": 1.8825,
+      "step": 807
+    },
+    {
+      "epoch": 0.8854794520547945,
+      "grad_norm": 2.087143898010254,
+      "learning_rate": 3.943948009748173e-06,
+      "loss": 1.9459,
+      "step": 808
+    },
+    {
+      "epoch": 0.8865753424657534,
+      "grad_norm": 2.2888102531433105,
+      "learning_rate": 3.94191714053615e-06,
+      "loss": 1.9057,
+      "step": 809
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "grad_norm": 2.1566672325134277,
+      "learning_rate": 3.939886271324127e-06,
+      "loss": 1.9013,
+      "step": 810
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "eval_loss": 2.062390089035034,
+      "eval_runtime": 36.8037,
+      "eval_samples_per_second": 27.171,
+      "eval_steps_per_second": 1.141,
+      "step": 810
+    },
+    {
+      "epoch": 0.8887671232876713,
+      "grad_norm": 2.1750905513763428,
+      "learning_rate": 3.937855402112104e-06,
+      "loss": 1.9257,
+      "step": 811
+    },
+    {
+      "epoch": 0.8898630136986302,
+      "grad_norm": 2.1565544605255127,
+      "learning_rate": 3.935824532900082e-06,
+      "loss": 1.8484,
+      "step": 812
+    },
+    {
+      "epoch": 0.890958904109589,
+      "grad_norm": 2.119025468826294,
+      "learning_rate": 3.933793663688058e-06,
+      "loss": 1.8503,
+      "step": 813
+    },
+    {
+      "epoch": 0.8920547945205479,
+      "grad_norm": 2.204523801803589,
+      "learning_rate": 3.9317627944760365e-06,
+      "loss": 1.9263,
+      "step": 814
+    },
+    {
+      "epoch": 0.8931506849315068,
+      "grad_norm": 2.2436130046844482,
+      "learning_rate": 3.929731925264013e-06,
+      "loss": 1.9419,
+      "step": 815
+    },
+    {
+      "epoch": 0.8942465753424658,
+      "grad_norm": 2.208177328109741,
+      "learning_rate": 3.927701056051991e-06,
+      "loss": 1.8433,
+      "step": 816
+    },
+    {
+      "epoch": 0.8953424657534247,
+      "grad_norm": 2.111579179763794,
+      "learning_rate": 3.9256701868399674e-06,
+      "loss": 1.8767,
+      "step": 817
+    },
+    {
+      "epoch": 0.8964383561643836,
+      "grad_norm": 2.1987080574035645,
+      "learning_rate": 3.9236393176279456e-06,
+      "loss": 1.8753,
+      "step": 818
+    },
+    {
+      "epoch": 0.8975342465753424,
+      "grad_norm": 2.0366690158843994,
+      "learning_rate": 3.921608448415922e-06,
+      "loss": 1.8663,
+      "step": 819
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "grad_norm": 2.3097612857818604,
+      "learning_rate": 3.9195775792039e-06,
+      "loss": 1.8574,
+      "step": 820
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "eval_loss": 2.0613417625427246,
+      "eval_runtime": 36.8049,
+      "eval_samples_per_second": 27.17,
+      "eval_steps_per_second": 1.141,
+      "step": 820
+    },
+    {
+      "epoch": 0.8997260273972603,
+      "grad_norm": 2.27079439163208,
+      "learning_rate": 3.9175467099918765e-06,
+      "loss": 1.8997,
+      "step": 821
+    },
+    {
+      "epoch": 0.9008219178082192,
+      "grad_norm": 2.053964138031006,
+      "learning_rate": 3.915515840779855e-06,
+      "loss": 1.8868,
+      "step": 822
+    },
+    {
+      "epoch": 0.9019178082191781,
+      "grad_norm": 2.2430622577667236,
+      "learning_rate": 3.913484971567831e-06,
+      "loss": 1.9368,
+      "step": 823
+    },
+    {
+      "epoch": 0.903013698630137,
+      "grad_norm": 1.9371908903121948,
+      "learning_rate": 3.911454102355808e-06,
+      "loss": 1.9162,
+      "step": 824
+    },
+    {
+      "epoch": 0.9041095890410958,
+      "grad_norm": 2.0140321254730225,
+      "learning_rate": 3.909423233143786e-06,
+      "loss": 1.8195,
+      "step": 825
+    },
+    {
+      "epoch": 0.9052054794520548,
+      "grad_norm": 2.2602760791778564,
+      "learning_rate": 3.907392363931763e-06,
+      "loss": 1.9084,
+      "step": 826
+    },
+    {
+      "epoch": 0.9063013698630137,
+      "grad_norm": 2.1431221961975098,
+      "learning_rate": 3.90536149471974e-06,
+      "loss": 1.9092,
+      "step": 827
+    },
+    {
+      "epoch": 0.9073972602739726,
+      "grad_norm": 1.9904584884643555,
+      "learning_rate": 3.9033306255077175e-06,
+      "loss": 1.9688,
+      "step": 828
+    },
+    {
+      "epoch": 0.9084931506849315,
+      "grad_norm": 2.1969170570373535,
+      "learning_rate": 3.901299756295695e-06,
+      "loss": 2.0016,
+      "step": 829
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "grad_norm": 2.2480008602142334,
+      "learning_rate": 3.899268887083672e-06,
+      "loss": 2.0181,
+      "step": 830
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "eval_loss": 2.0588274002075195,
+      "eval_runtime": 36.7503,
+      "eval_samples_per_second": 27.211,
+      "eval_steps_per_second": 1.143,
+      "step": 830
+    },
+    {
+      "epoch": 0.9106849315068493,
+      "grad_norm": 2.0574944019317627,
+      "learning_rate": 3.897238017871649e-06,
+      "loss": 1.8834,
+      "step": 831
+    },
+    {
+      "epoch": 0.9117808219178082,
+      "grad_norm": 2.327791213989258,
+      "learning_rate": 3.8952071486596266e-06,
+      "loss": 1.8947,
+      "step": 832
+    },
+    {
+      "epoch": 0.9128767123287671,
+      "grad_norm": 2.1172237396240234,
+      "learning_rate": 3.893176279447604e-06,
+      "loss": 1.9049,
+      "step": 833
+    },
+    {
+      "epoch": 0.913972602739726,
+      "grad_norm": 2.154917001724243,
+      "learning_rate": 3.891145410235581e-06,
+      "loss": 1.8571,
+      "step": 834
+    },
+    {
+      "epoch": 0.915068493150685,
+      "grad_norm": 2.1272647380828857,
+      "learning_rate": 3.889114541023558e-06,
+      "loss": 1.975,
+      "step": 835
+    },
+    {
+      "epoch": 0.9161643835616439,
+      "grad_norm": 2.2209606170654297,
+      "learning_rate": 3.887083671811536e-06,
+      "loss": 1.9077,
+      "step": 836
+    },
+    {
+      "epoch": 0.9172602739726028,
+      "grad_norm": 2.1849212646484375,
+      "learning_rate": 3.885052802599513e-06,
+      "loss": 1.8818,
+      "step": 837
+    },
+    {
+      "epoch": 0.9183561643835616,
+      "grad_norm": 2.01114821434021,
+      "learning_rate": 3.88302193338749e-06,
+      "loss": 1.8357,
+      "step": 838
+    },
+    {
+      "epoch": 0.9194520547945205,
+      "grad_norm": 2.1323206424713135,
+      "learning_rate": 3.8809910641754675e-06,
+      "loss": 1.9642,
+      "step": 839
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "grad_norm": 2.013662815093994,
+      "learning_rate": 3.878960194963445e-06,
+      "loss": 1.9073,
+      "step": 840
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "eval_loss": 2.0627999305725098,
+      "eval_runtime": 36.7873,
+      "eval_samples_per_second": 27.183,
+      "eval_steps_per_second": 1.142,
+      "step": 840
+    },
+    {
+      "epoch": 0.9216438356164384,
+      "grad_norm": 2.2091948986053467,
+      "learning_rate": 3.876929325751422e-06,
+      "loss": 1.892,
+      "step": 841
+    },
+    {
+      "epoch": 0.9227397260273973,
+      "grad_norm": 2.227769613265991,
+      "learning_rate": 3.874898456539399e-06,
+      "loss": 1.9428,
+      "step": 842
+    },
+    {
+      "epoch": 0.9238356164383562,
+      "grad_norm": 2.259666919708252,
+      "learning_rate": 3.872867587327377e-06,
+      "loss": 1.8448,
+      "step": 843
+    },
+    {
+      "epoch": 0.924931506849315,
+      "grad_norm": 2.1382651329040527,
+      "learning_rate": 3.870836718115354e-06,
+      "loss": 1.9144,
+      "step": 844
+    },
+    {
+      "epoch": 0.9260273972602739,
+      "grad_norm": 2.1010990142822266,
+      "learning_rate": 3.868805848903331e-06,
+      "loss": 1.8679,
+      "step": 845
+    },
+    {
+      "epoch": 0.9271232876712329,
+      "grad_norm": 2.3949990272521973,
+      "learning_rate": 3.8667749796913084e-06,
+      "loss": 1.8697,
+      "step": 846
+    },
+    {
+      "epoch": 0.9282191780821918,
+      "grad_norm": 2.2491044998168945,
+      "learning_rate": 3.864744110479286e-06,
+      "loss": 1.903,
+      "step": 847
+    },
+    {
+      "epoch": 0.9293150684931507,
+      "grad_norm": 2.144421339035034,
+      "learning_rate": 3.862713241267262e-06,
+      "loss": 1.9016,
+      "step": 848
+    },
+    {
+      "epoch": 0.9304109589041096,
+      "grad_norm": 2.15675687789917,
+      "learning_rate": 3.86068237205524e-06,
+      "loss": 1.941,
+      "step": 849
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "grad_norm": 2.315117120742798,
+      "learning_rate": 3.858651502843217e-06,
+      "loss": 1.95,
+      "step": 850
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "eval_loss": 2.0650341510772705,
+      "eval_runtime": 36.7951,
+      "eval_samples_per_second": 27.178,
+      "eval_steps_per_second": 1.141,
+      "step": 850
+    },
+    {
+      "epoch": 0.9326027397260274,
+      "grad_norm": 2.231395959854126,
+      "learning_rate": 3.856620633631195e-06,
+      "loss": 1.9035,
+      "step": 851
+    },
+    {
+      "epoch": 0.9336986301369863,
+      "grad_norm": 2.250143051147461,
+      "learning_rate": 3.854589764419171e-06,
+      "loss": 1.8223,
+      "step": 852
+    },
+    {
+      "epoch": 0.9347945205479452,
+      "grad_norm": 2.1404998302459717,
+      "learning_rate": 3.852558895207149e-06,
+      "loss": 1.8813,
+      "step": 853
+    },
+    {
+      "epoch": 0.9358904109589041,
+      "grad_norm": 2.271339178085327,
+      "learning_rate": 3.850528025995126e-06,
+      "loss": 1.9847,
+      "step": 854
+    },
+    {
+      "epoch": 0.936986301369863,
+      "grad_norm": 2.2669527530670166,
+      "learning_rate": 3.848497156783104e-06,
+      "loss": 1.9299,
+      "step": 855
+    },
+    {
+      "epoch": 0.938082191780822,
+      "grad_norm": 2.2487215995788574,
+      "learning_rate": 3.84646628757108e-06,
+      "loss": 1.8706,
+      "step": 856
+    },
+    {
+      "epoch": 0.9391780821917808,
+      "grad_norm": 2.1427817344665527,
+      "learning_rate": 3.8444354183590585e-06,
+      "loss": 1.8843,
+      "step": 857
+    },
+    {
+      "epoch": 0.9402739726027397,
+      "grad_norm": 2.2420499324798584,
+      "learning_rate": 3.842404549147035e-06,
+      "loss": 1.8762,
+      "step": 858
+    },
+    {
+      "epoch": 0.9413698630136986,
+      "grad_norm": 1.9591865539550781,
+      "learning_rate": 3.840373679935013e-06,
+      "loss": 1.8813,
+      "step": 859
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "grad_norm": 2.1901352405548096,
+      "learning_rate": 3.8383428107229894e-06,
+      "loss": 1.8455,
+      "step": 860
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "eval_loss": 2.0641214847564697,
+      "eval_runtime": 36.7899,
+      "eval_samples_per_second": 27.181,
+      "eval_steps_per_second": 1.142,
+      "step": 860
+    },
+    {
+      "epoch": 0.9435616438356165,
+      "grad_norm": 2.2325565814971924,
+      "learning_rate": 3.836311941510967e-06,
+      "loss": 1.8885,
+      "step": 861
+    },
+    {
+      "epoch": 0.9446575342465754,
+      "grad_norm": 2.1804277896881104,
+      "learning_rate": 3.834281072298944e-06,
+      "loss": 1.8595,
+      "step": 862
+    },
+    {
+      "epoch": 0.9457534246575342,
+      "grad_norm": 2.310577630996704,
+      "learning_rate": 3.832250203086921e-06,
+      "loss": 1.922,
+      "step": 863
+    },
+    {
+      "epoch": 0.9468493150684931,
+      "grad_norm": 2.2623233795166016,
+      "learning_rate": 3.8302193338748985e-06,
+      "loss": 1.8554,
+      "step": 864
+    },
+    {
+      "epoch": 0.947945205479452,
+      "grad_norm": 2.1515281200408936,
+      "learning_rate": 3.828188464662876e-06,
+      "loss": 1.9402,
+      "step": 865
+    },
+    {
+      "epoch": 0.949041095890411,
+      "grad_norm": 2.2868542671203613,
+      "learning_rate": 3.826157595450853e-06,
+      "loss": 1.9796,
+      "step": 866
+    },
+    {
+      "epoch": 0.9501369863013699,
+      "grad_norm": 2.1473448276519775,
+      "learning_rate": 3.82412672623883e-06,
+      "loss": 1.9088,
+      "step": 867
+    },
+    {
+      "epoch": 0.9512328767123288,
+      "grad_norm": 2.0565810203552246,
+      "learning_rate": 3.822095857026808e-06,
+      "loss": 1.8451,
+      "step": 868
+    },
+    {
+      "epoch": 0.9523287671232876,
+      "grad_norm": 2.0083553791046143,
+      "learning_rate": 3.820064987814785e-06,
+      "loss": 1.9164,
+      "step": 869
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "grad_norm": 2.011315107345581,
+      "learning_rate": 3.818034118602762e-06,
+      "loss": 1.8632,
+      "step": 870
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "eval_loss": 2.0615739822387695,
+      "eval_runtime": 36.7512,
+      "eval_samples_per_second": 27.21,
+      "eval_steps_per_second": 1.143,
+      "step": 870
+    },
+    {
+      "epoch": 0.9545205479452055,
+      "grad_norm": 2.116142749786377,
+      "learning_rate": 3.8160032493907395e-06,
+      "loss": 1.8591,
+      "step": 871
+    },
+    {
+      "epoch": 0.9556164383561644,
+      "grad_norm": 2.1953125,
+      "learning_rate": 3.813972380178717e-06,
+      "loss": 1.8719,
+      "step": 872
+    },
+    {
+      "epoch": 0.9567123287671233,
+      "grad_norm": 2.160209894180298,
+      "learning_rate": 3.811941510966694e-06,
+      "loss": 1.8716,
+      "step": 873
+    },
+    {
+      "epoch": 0.9578082191780822,
+      "grad_norm": 2.281137466430664,
+      "learning_rate": 3.809910641754671e-06,
+      "loss": 1.9214,
+      "step": 874
+    },
+    {
+      "epoch": 0.958904109589041,
+      "grad_norm": 2.1310486793518066,
+      "learning_rate": 3.8078797725426486e-06,
+      "loss": 1.8482,
+      "step": 875
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 2.139904022216797,
+      "learning_rate": 3.8058489033306254e-06,
+      "loss": 1.8024,
+      "step": 876
+    },
+    {
+      "epoch": 0.9610958904109589,
+      "grad_norm": 2.0222628116607666,
+      "learning_rate": 3.803818034118603e-06,
+      "loss": 1.8911,
+      "step": 877
+    },
+    {
+      "epoch": 0.9621917808219178,
+      "grad_norm": 2.1462645530700684,
+      "learning_rate": 3.80178716490658e-06,
+      "loss": 1.8707,
+      "step": 878
+    },
+    {
+      "epoch": 0.9632876712328767,
+      "grad_norm": 1.9841400384902954,
+      "learning_rate": 3.7997562956945577e-06,
+      "loss": 1.9406,
+      "step": 879
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "grad_norm": 2.0436203479766846,
+      "learning_rate": 3.7977254264825345e-06,
+      "loss": 1.9268,
+      "step": 880
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "eval_loss": 2.0551161766052246,
+      "eval_runtime": 37.0754,
+      "eval_samples_per_second": 26.972,
+      "eval_steps_per_second": 1.133,
+      "step": 880
+    },
+    {
+      "epoch": 0.9654794520547946,
+      "grad_norm": 2.1850926876068115,
+      "learning_rate": 3.7956945572705122e-06,
+      "loss": 1.8766,
+      "step": 881
+    },
+    {
+      "epoch": 0.9665753424657534,
+      "grad_norm": 2.1350739002227783,
+      "learning_rate": 3.793663688058489e-06,
+      "loss": 1.8582,
+      "step": 882
+    },
+    {
+      "epoch": 0.9676712328767123,
+      "grad_norm": 2.066235303878784,
+      "learning_rate": 3.791632818846467e-06,
+      "loss": 1.9486,
+      "step": 883
+    },
+    {
+      "epoch": 0.9687671232876712,
+      "grad_norm": 2.0621633529663086,
+      "learning_rate": 3.7896019496344436e-06,
+      "loss": 1.9465,
+      "step": 884
+    },
+    {
+      "epoch": 0.9698630136986301,
+      "grad_norm": 2.0093510150909424,
+      "learning_rate": 3.787571080422421e-06,
+      "loss": 1.9418,
+      "step": 885
+    },
+    {
+      "epoch": 0.9709589041095891,
+      "grad_norm": 2.1686882972717285,
+      "learning_rate": 3.785540211210398e-06,
+      "loss": 1.9396,
+      "step": 886
+    },
+    {
+      "epoch": 0.972054794520548,
+      "grad_norm": 2.0459682941436768,
+      "learning_rate": 3.7835093419983755e-06,
+      "loss": 1.9171,
+      "step": 887
+    },
+    {
+      "epoch": 0.9731506849315068,
+      "grad_norm": 2.034409523010254,
+      "learning_rate": 3.7814784727863527e-06,
+      "loss": 1.896,
+      "step": 888
+    },
+    {
+      "epoch": 0.9742465753424657,
+      "grad_norm": 2.1498303413391113,
+      "learning_rate": 3.77944760357433e-06,
+      "loss": 1.8738,
+      "step": 889
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "grad_norm": 2.1839663982391357,
+      "learning_rate": 3.7774167343623073e-06,
+      "loss": 1.9987,
+      "step": 890
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "eval_loss": 2.055088996887207,
+      "eval_runtime": 37.0749,
+      "eval_samples_per_second": 26.972,
+      "eval_steps_per_second": 1.133,
+      "step": 890
+    },
+    {
+      "epoch": 0.9764383561643836,
+      "grad_norm": 2.192547559738159,
+      "learning_rate": 3.7753858651502846e-06,
+      "loss": 1.9415,
+      "step": 891
+    },
+    {
+      "epoch": 0.9775342465753425,
+      "grad_norm": 2.0081186294555664,
+      "learning_rate": 3.773354995938262e-06,
+      "loss": 1.8653,
+      "step": 892
+    },
+    {
+      "epoch": 0.9786301369863014,
+      "grad_norm": 2.160459518432617,
+      "learning_rate": 3.771324126726239e-06,
+      "loss": 1.9353,
+      "step": 893
+    },
+    {
+      "epoch": 0.9797260273972602,
+      "grad_norm": 2.1357319355010986,
+      "learning_rate": 3.7692932575142164e-06,
+      "loss": 1.9078,
+      "step": 894
+    },
+    {
+      "epoch": 0.9808219178082191,
+      "grad_norm": 2.114255905151367,
+      "learning_rate": 3.7672623883021937e-06,
+      "loss": 1.9398,
+      "step": 895
+    },
+    {
+      "epoch": 0.9819178082191781,
+      "grad_norm": 2.1232616901397705,
+      "learning_rate": 3.765231519090171e-06,
+      "loss": 1.8687,
+      "step": 896
+    },
+    {
+      "epoch": 0.983013698630137,
+      "grad_norm": 2.227722644805908,
+      "learning_rate": 3.7632006498781482e-06,
+      "loss": 1.7913,
+      "step": 897
+    },
+    {
+      "epoch": 0.9841095890410959,
+      "grad_norm": 2.046050786972046,
+      "learning_rate": 3.761169780666125e-06,
+      "loss": 1.903,
+      "step": 898
+    },
+    {
+      "epoch": 0.9852054794520548,
+      "grad_norm": 2.0269315242767334,
+      "learning_rate": 3.7591389114541028e-06,
+      "loss": 1.8508,
+      "step": 899
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "grad_norm": 2.237417221069336,
+      "learning_rate": 3.7571080422420796e-06,
+      "loss": 1.8978,
+      "step": 900
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "eval_loss": 2.0542197227478027,
+      "eval_runtime": 36.7867,
+      "eval_samples_per_second": 27.184,
+      "eval_steps_per_second": 1.142,
+      "step": 900
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2736,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.805578468425728e+17,
+  "train_batch_size": 3,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/training_args.bin b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..4d0507e2f82944df028e22f11bd29f2ad6ae035a
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:792d444faf96ea7004659c6473540909c7b64212dbf0c341d4052967b42749eb
+size 6011
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic57_10M_seed0/runs/checkpoint-900/zero_to_fp32.py
@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/2d73a6863086ff9d491c28e49df9fb697cd92c2b.lock b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/2d73a6863086ff9d491c28e49df9fb697cd92c2b.lock
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/4719a04514ec2f060240711b7c33ab21187cac730ecaba3040b7a0fd95a9cefb.lock b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/4719a04514ec2f060240711b7c33ab21187cac730ecaba3040b7a0fd95a9cefb.lock
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/47d4a5aa69cdef91a53b77f5c5583647a578ca0e.lock b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/47d4a5aa69cdef91a53b77f5c5583647a578ca0e.lock
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/584d8d3e3f82f7964955174dfe5e3b1cf117a9d859f022cfdf7fcb884856e002.lock b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/584d8d3e3f82f7964955174dfe5e3b1cf117a9d859f022cfdf7fcb884856e002.lock
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58.lock b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/.locks/models--meta-llama--Llama-3.2-3B/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58.lock
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/adapter_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/.no_exist/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/2d73a6863086ff9d491c28e49df9fb697cd92c2b b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/2d73a6863086ff9d491c28e49df9fb697cd92c2b
new file mode 100644
index 0000000000000000000000000000000000000000..2d73a6863086ff9d491c28e49df9fb697cd92c2b
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/2d73a6863086ff9d491c28e49df9fb697cd92c2b
@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "transformers_version": "4.45.0.dev0",
+  "do_sample": true,
+  "temperature": 0.6,
+  "top_p": 0.9
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/47d4a5aa69cdef91a53b77f5c5583647a578ca0e b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/47d4a5aa69cdef91a53b77f5c5583647a578ca0e
new file mode 100644
index 0000000000000000000000000000000000000000..47d4a5aa69cdef91a53b77f5c5583647a578ca0e
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/47d4a5aa69cdef91a53b77f5c5583647a578ca0e
@@ -0,0 +1,35 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.45.0.dev0",
+  "use_cache": true,
+  "vocab_size": 128256
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58 b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58
new file mode 100644
index 0000000000000000000000000000000000000000..d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/blobs/d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58
@@ -0,0 +1,261 @@
+{
+  "metadata": {
+    "total_size": 6425499648
+  },
+  "weight_map": {
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/refs/main b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/refs/main
new file mode 100644
index 0000000000000000000000000000000000000000..940c46e96db631910d2cc5186ab76cd83e32b124
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/refs/main
@@ -0,0 +1 @@
+13afe5124825b4f3751f836b40dafda64c1ed062
\ No newline at end of file
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..47d4a5aa69cdef91a53b77f5c5583647a578ca0e
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/config.json
@@ -0,0 +1,35 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.45.0.dev0",
+  "use_cache": true,
+  "vocab_size": 128256
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..2d73a6863086ff9d491c28e49df9fb697cd92c2b
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/generation_config.json
@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "transformers_version": "4.45.0.dev0",
+  "do_sample": true,
+  "temperature": 0.6,
+  "top_p": 0.9
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors.index.json
new file mode 100644
index 0000000000000000000000000000000000000000..d3a1f0f5f401eeadca0c7a6786bd9e877fd42e58
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/artifacts/models--meta-llama--Llama-3.2-3B/snapshots/13afe5124825b4f3751f836b40dafda64c1ed062/model.safetensors.index.json
@@ -0,0 +1,261 @@
+{
+  "metadata": {
+    "total_size": 6425499648
+  },
+  "weight_map": {
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1000/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1000/model-00002-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..b4939b9a52874fa011ce8ddfbfa7d19072beb877
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1000/model-00002-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:710c3570f1f9133ef23165ee96d97f92e2612c6fb0a7e6202317906370a98d68
+size 2247734920
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1800/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1800/model-00002-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..fa5d5fd02fafbb87de2c5ec959226ea850da1cce
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-1800/model-00002-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec13cb22f241240603f88c38935929f0b6be6d77f075956ec35863c430630fdc
+size 2247734920
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-200/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-200/model-00002-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..32671b7a945540a1be13855dad1b72bc00e45af0
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-200/model-00002-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4f491c744960a4266b700c1d08c7527fd4f593356d4e3713624bcf73f0e2ab3
+size 2247734920
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2000/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2000/model-00002-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..053240917338e53808b8ee973c08acc96a34f013
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2000/model-00002-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cfd5dfcd2ccddfd4f2d9d47acf372f87660dec7c11b03194f12a16d4ab45dd55
+size 2247734920
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/generation_config.json
@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": 128001,
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.45.1"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/model-00002-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..a20222ded1be32f1e533cb2231f8f1a0b082f832
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/model-00002-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c0a27080897771123e9438ca60122e385c4380b82362f8e2d6d108dfc923aa79
+size 2247734920
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/model.safetensors.index.json
new file mode 100644
index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/model.safetensors.index.json
@@ -0,0 +1,262 @@
+{
+  "metadata": {
+    "total_size": 7213504512
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00002-of-00002.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/training_args.bin b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..b7adbe5a2dbf27f05befec6cc65fbc93e90eefd0
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2100/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aaa882279512a8c278376c0c601206933f3c5662ddcd3c13c53bf32ae3680d14
+size 6011
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/config.json
@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "meta-llama/Llama-3.2-3B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "torch_dtype": "float16",
+  "transformers_version": "4.45.1",
+  "use_cache": true,
+  "vocab_size": 128256
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/generation_config.json
@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": 128001,
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.45.1"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/latest b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/latest
new file mode 100644
index 0000000000000000000000000000000000000000..adbdf3f70e1e0db88a47bff3edcb0592ed0a2e61
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/latest
@@ -0,0 +1 @@
+global_step2200
\ No newline at end of file
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/model.safetensors.index.json
new file mode 100644
index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/model.safetensors.index.json
@@ -0,0 +1,262 @@
+{
+  "metadata": {
+    "total_size": 7213504512
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00002-of-00002.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/rng_state_0.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..ca3af7294f2219fc064772e6af4abf0481314066
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef233d2806676c682805f85cf51776d27dc06740dc062015cc1e961a14df9e97
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/rng_state_1.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..9a2039c6da7a85a1210a9c30d33ef8b01f90c45a
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24be9519b492ae2ab5da4e7d258ad3f276da346422161ebc6874f26057c028cd
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/rng_state_2.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..c4e626fa4f7fe11f0ad555292fe2005dacbf98ad
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77a120a924d0a14713ef6ebebe2792a637dac07f7fbafa91cf0930019864eb1b
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/scheduler.pt b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..d5d2c93bde9b16b79c2b1c0089b963d8a579156f
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd4ea68e5cb4b15ed93924820290a9a6c1be21d334486b0d094030935b7d9ef2
+size 627
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/tokenizer.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb
+size 17210374
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128256": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|end_of_text|>",
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "[PAD]",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..6abfbda8cd8cb62200c04f19b61f4ad4909a8e68
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/trainer_state.json
@@ -0,0 +1,17193 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.410958904109589,
+  "eval_steps": 10,
+  "global_step": 2200,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001095890410958904,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0636,
+      "step": 1
+    },
+    {
+      "epoch": 0.002191780821917808,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0501,
+      "step": 2
+    },
+    {
+      "epoch": 0.0032876712328767125,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0238,
+      "step": 3
+    },
+    {
+      "epoch": 0.004383561643835616,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0346,
+      "step": 4
+    },
+    {
+      "epoch": 0.005479452054794521,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0108,
+      "step": 5
+    },
+    {
+      "epoch": 0.006575342465753425,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.131,
+      "step": 6
+    },
+    {
+      "epoch": 0.007671232876712329,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0488,
+      "step": 7
+    },
+    {
+      "epoch": 0.008767123287671232,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0905,
+      "step": 8
+    },
+    {
+      "epoch": 0.009863013698630137,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0493,
+      "step": 9
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.1038,
+      "step": 10
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "eval_loss": 3.100078821182251,
+      "eval_runtime": 37.5838,
+      "eval_samples_per_second": 26.607,
+      "eval_steps_per_second": 1.118,
+      "step": 10
+    },
+    {
+      "epoch": 0.012054794520547946,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0804,
+      "step": 11
+    },
+    {
+      "epoch": 0.01315068493150685,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0053,
+      "step": 12
+    },
+    {
+      "epoch": 0.014246575342465753,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0607,
+      "step": 13
+    },
+    {
+      "epoch": 0.015342465753424657,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0539,
+      "step": 14
+    },
+    {
+      "epoch": 0.01643835616438356,
+      "grad_norm": 5.420984745025635,
+      "learning_rate": 1.8248175182481753e-08,
+      "loss": 3.0941,
+      "step": 15
+    },
+    {
+      "epoch": 0.017534246575342465,
+      "grad_norm": 5.76840877532959,
+      "learning_rate": 3.6496350364963505e-08,
+      "loss": 3.036,
+      "step": 16
+    },
+    {
+      "epoch": 0.01863013698630137,
+      "grad_norm": 5.905357360839844,
+      "learning_rate": 5.4744525547445264e-08,
+      "loss": 3.0491,
+      "step": 17
+    },
+    {
+      "epoch": 0.019726027397260273,
+      "grad_norm": 5.981553554534912,
+      "learning_rate": 7.299270072992701e-08,
+      "loss": 3.0775,
+      "step": 18
+    },
+    {
+      "epoch": 0.020821917808219178,
+      "grad_norm": 6.240172863006592,
+      "learning_rate": 9.124087591240876e-08,
+      "loss": 3.0345,
+      "step": 19
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "grad_norm": 5.59375,
+      "learning_rate": 1.0948905109489053e-07,
+      "loss": 3.1063,
+      "step": 20
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "eval_loss": 3.098820686340332,
+      "eval_runtime": 37.2695,
+      "eval_samples_per_second": 26.832,
+      "eval_steps_per_second": 1.127,
+      "step": 20
+    },
+    {
+      "epoch": 0.023013698630136987,
+      "grad_norm": 5.819795608520508,
+      "learning_rate": 1.2773722627737227e-07,
+      "loss": 3.0808,
+      "step": 21
+    },
+    {
+      "epoch": 0.02410958904109589,
+      "grad_norm": 5.7714762687683105,
+      "learning_rate": 1.4598540145985402e-07,
+      "loss": 3.0055,
+      "step": 22
+    },
+    {
+      "epoch": 0.025205479452054796,
+      "grad_norm": 5.51552152633667,
+      "learning_rate": 1.642335766423358e-07,
+      "loss": 3.0396,
+      "step": 23
+    },
+    {
+      "epoch": 0.0263013698630137,
+      "grad_norm": 5.594012260437012,
+      "learning_rate": 1.8248175182481753e-07,
+      "loss": 3.0824,
+      "step": 24
+    },
+    {
+      "epoch": 0.0273972602739726,
+      "grad_norm": 5.1974029541015625,
+      "learning_rate": 2.0072992700729928e-07,
+      "loss": 3.08,
+      "step": 25
+    },
+    {
+      "epoch": 0.028493150684931506,
+      "grad_norm": 4.733968257904053,
+      "learning_rate": 2.1897810218978106e-07,
+      "loss": 3.0597,
+      "step": 26
+    },
+    {
+      "epoch": 0.02958904109589041,
+      "grad_norm": 4.717319011688232,
+      "learning_rate": 2.3722627737226278e-07,
+      "loss": 3.0696,
+      "step": 27
+    },
+    {
+      "epoch": 0.030684931506849315,
+      "grad_norm": 4.6303534507751465,
+      "learning_rate": 2.5547445255474454e-07,
+      "loss": 3.0273,
+      "step": 28
+    },
+    {
+      "epoch": 0.03178082191780822,
+      "grad_norm": 4.413999557495117,
+      "learning_rate": 2.737226277372263e-07,
+      "loss": 3.0202,
+      "step": 29
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "grad_norm": 4.516043186187744,
+      "learning_rate": 2.9197080291970804e-07,
+      "loss": 2.9926,
+      "step": 30
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "eval_loss": 3.0561797618865967,
+      "eval_runtime": 37.0351,
+      "eval_samples_per_second": 27.001,
+      "eval_steps_per_second": 1.134,
+      "step": 30
+    },
+    {
+      "epoch": 0.03397260273972603,
+      "grad_norm": 3.7179207801818848,
+      "learning_rate": 3.102189781021898e-07,
+      "loss": 3.0086,
+      "step": 31
+    },
+    {
+      "epoch": 0.03506849315068493,
+      "grad_norm": 3.7840425968170166,
+      "learning_rate": 3.284671532846716e-07,
+      "loss": 2.9986,
+      "step": 32
+    },
+    {
+      "epoch": 0.03616438356164384,
+      "grad_norm": 3.843116521835327,
+      "learning_rate": 3.467153284671533e-07,
+      "loss": 2.9677,
+      "step": 33
+    },
+    {
+      "epoch": 0.03726027397260274,
+      "grad_norm": 3.8528714179992676,
+      "learning_rate": 3.6496350364963505e-07,
+      "loss": 2.9732,
+      "step": 34
+    },
+    {
+      "epoch": 0.038356164383561646,
+      "grad_norm": 3.76963210105896,
+      "learning_rate": 3.8321167883211686e-07,
+      "loss": 2.9855,
+      "step": 35
+    },
+    {
+      "epoch": 0.03945205479452055,
+      "grad_norm": 3.6240015029907227,
+      "learning_rate": 4.0145985401459856e-07,
+      "loss": 3.0117,
+      "step": 36
+    },
+    {
+      "epoch": 0.040547945205479455,
+      "grad_norm": 3.1380624771118164,
+      "learning_rate": 4.197080291970803e-07,
+      "loss": 2.9481,
+      "step": 37
+    },
+    {
+      "epoch": 0.041643835616438356,
+      "grad_norm": 3.565329074859619,
+      "learning_rate": 4.379562043795621e-07,
+      "loss": 2.9175,
+      "step": 38
+    },
+    {
+      "epoch": 0.042739726027397264,
+      "grad_norm": 3.8062448501586914,
+      "learning_rate": 4.562043795620438e-07,
+      "loss": 3.0074,
+      "step": 39
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "grad_norm": 3.412046432495117,
+      "learning_rate": 4.7445255474452557e-07,
+      "loss": 2.9119,
+      "step": 40
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "eval_loss": 2.9874370098114014,
+      "eval_runtime": 37.0514,
+      "eval_samples_per_second": 26.99,
+      "eval_steps_per_second": 1.134,
+      "step": 40
+    },
+    {
+      "epoch": 0.044931506849315066,
+      "grad_norm": 3.227437973022461,
+      "learning_rate": 4.927007299270073e-07,
+      "loss": 2.9436,
+      "step": 41
+    },
+    {
+      "epoch": 0.046027397260273974,
+      "grad_norm": 2.9633443355560303,
+      "learning_rate": 5.109489051094891e-07,
+      "loss": 2.967,
+      "step": 42
+    },
+    {
+      "epoch": 0.047123287671232875,
+      "grad_norm": 2.942678928375244,
+      "learning_rate": 5.291970802919708e-07,
+      "loss": 2.9375,
+      "step": 43
+    },
+    {
+      "epoch": 0.04821917808219178,
+      "grad_norm": 2.9806888103485107,
+      "learning_rate": 5.474452554744526e-07,
+      "loss": 2.9474,
+      "step": 44
+    },
+    {
+      "epoch": 0.049315068493150684,
+      "grad_norm": 3.000063419342041,
+      "learning_rate": 5.656934306569343e-07,
+      "loss": 2.9289,
+      "step": 45
+    },
+    {
+      "epoch": 0.05041095890410959,
+      "grad_norm": 3.1481194496154785,
+      "learning_rate": 5.839416058394161e-07,
+      "loss": 2.8869,
+      "step": 46
+    },
+    {
+      "epoch": 0.05150684931506849,
+      "grad_norm": 2.949202060699463,
+      "learning_rate": 6.021897810218979e-07,
+      "loss": 2.8783,
+      "step": 47
+    },
+    {
+      "epoch": 0.0526027397260274,
+      "grad_norm": 2.706526279449463,
+      "learning_rate": 6.204379562043796e-07,
+      "loss": 2.9024,
+      "step": 48
+    },
+    {
+      "epoch": 0.0536986301369863,
+      "grad_norm": 2.7317655086517334,
+      "learning_rate": 6.386861313868614e-07,
+      "loss": 2.8325,
+      "step": 49
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "grad_norm": 2.25644588470459,
+      "learning_rate": 6.569343065693432e-07,
+      "loss": 2.9409,
+      "step": 50
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "eval_loss": 2.909320116043091,
+      "eval_runtime": 37.0714,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 50
+    },
+    {
+      "epoch": 0.05589041095890411,
+      "grad_norm": 2.8789613246917725,
+      "learning_rate": 6.751824817518248e-07,
+      "loss": 2.8459,
+      "step": 51
+    },
+    {
+      "epoch": 0.05698630136986301,
+      "grad_norm": 2.686873435974121,
+      "learning_rate": 6.934306569343066e-07,
+      "loss": 2.8653,
+      "step": 52
+    },
+    {
+      "epoch": 0.05808219178082192,
+      "grad_norm": 2.54382586479187,
+      "learning_rate": 7.116788321167883e-07,
+      "loss": 2.7973,
+      "step": 53
+    },
+    {
+      "epoch": 0.05917808219178082,
+      "grad_norm": 2.1871724128723145,
+      "learning_rate": 7.299270072992701e-07,
+      "loss": 2.7982,
+      "step": 54
+    },
+    {
+      "epoch": 0.06027397260273973,
+      "grad_norm": 2.145303249359131,
+      "learning_rate": 7.48175182481752e-07,
+      "loss": 2.8195,
+      "step": 55
+    },
+    {
+      "epoch": 0.06136986301369863,
+      "grad_norm": 2.1635069847106934,
+      "learning_rate": 7.664233576642337e-07,
+      "loss": 2.825,
+      "step": 56
+    },
+    {
+      "epoch": 0.06246575342465754,
+      "grad_norm": 2.3530752658843994,
+      "learning_rate": 7.846715328467154e-07,
+      "loss": 2.8407,
+      "step": 57
+    },
+    {
+      "epoch": 0.06356164383561644,
+      "grad_norm": 2.2455942630767822,
+      "learning_rate": 8.029197080291971e-07,
+      "loss": 2.8182,
+      "step": 58
+    },
+    {
+      "epoch": 0.06465753424657535,
+      "grad_norm": 1.8985673189163208,
+      "learning_rate": 8.211678832116789e-07,
+      "loss": 2.772,
+      "step": 59
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "grad_norm": 1.9699499607086182,
+      "learning_rate": 8.394160583941606e-07,
+      "loss": 2.7367,
+      "step": 60
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "eval_loss": 2.8451859951019287,
+      "eval_runtime": 37.0793,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 60
+    },
+    {
+      "epoch": 0.06684931506849315,
+      "grad_norm": 2.159547805786133,
+      "learning_rate": 8.576642335766425e-07,
+      "loss": 2.7635,
+      "step": 61
+    },
+    {
+      "epoch": 0.06794520547945206,
+      "grad_norm": 2.12705659866333,
+      "learning_rate": 8.759124087591242e-07,
+      "loss": 2.7624,
+      "step": 62
+    },
+    {
+      "epoch": 0.06904109589041096,
+      "grad_norm": 1.8393241167068481,
+      "learning_rate": 8.94160583941606e-07,
+      "loss": 2.7763,
+      "step": 63
+    },
+    {
+      "epoch": 0.07013698630136986,
+      "grad_norm": 1.683522343635559,
+      "learning_rate": 9.124087591240876e-07,
+      "loss": 2.7762,
+      "step": 64
+    },
+    {
+      "epoch": 0.07123287671232877,
+      "grad_norm": 1.9260376691818237,
+      "learning_rate": 9.306569343065694e-07,
+      "loss": 2.7418,
+      "step": 65
+    },
+    {
+      "epoch": 0.07232876712328767,
+      "grad_norm": 2.09322190284729,
+      "learning_rate": 9.489051094890511e-07,
+      "loss": 2.7625,
+      "step": 66
+    },
+    {
+      "epoch": 0.07342465753424658,
+      "grad_norm": 1.9183429479599,
+      "learning_rate": 9.67153284671533e-07,
+      "loss": 2.7453,
+      "step": 67
+    },
+    {
+      "epoch": 0.07452054794520548,
+      "grad_norm": 1.9792778491973877,
+      "learning_rate": 9.854014598540146e-07,
+      "loss": 2.7858,
+      "step": 68
+    },
+    {
+      "epoch": 0.07561643835616438,
+      "grad_norm": 2.1397311687469482,
+      "learning_rate": 1.0036496350364965e-06,
+      "loss": 2.7242,
+      "step": 69
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "grad_norm": 1.8556463718414307,
+      "learning_rate": 1.0218978102189781e-06,
+      "loss": 2.6853,
+      "step": 70
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "eval_loss": 2.7787046432495117,
+      "eval_runtime": 37.1017,
+      "eval_samples_per_second": 26.953,
+      "eval_steps_per_second": 1.132,
+      "step": 70
+    },
+    {
+      "epoch": 0.07780821917808219,
+      "grad_norm": 1.9776800870895386,
+      "learning_rate": 1.04014598540146e-06,
+      "loss": 2.749,
+      "step": 71
+    },
+    {
+      "epoch": 0.0789041095890411,
+      "grad_norm": 2.3972630500793457,
+      "learning_rate": 1.0583941605839416e-06,
+      "loss": 2.6457,
+      "step": 72
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 2.090210199356079,
+      "learning_rate": 1.0766423357664233e-06,
+      "loss": 2.6401,
+      "step": 73
+    },
+    {
+      "epoch": 0.08109589041095891,
+      "grad_norm": 2.4117701053619385,
+      "learning_rate": 1.0948905109489052e-06,
+      "loss": 2.6649,
+      "step": 74
+    },
+    {
+      "epoch": 0.0821917808219178,
+      "grad_norm": 2.4845173358917236,
+      "learning_rate": 1.113138686131387e-06,
+      "loss": 2.6855,
+      "step": 75
+    },
+    {
+      "epoch": 0.08328767123287671,
+      "grad_norm": 1.7779242992401123,
+      "learning_rate": 1.1313868613138687e-06,
+      "loss": 2.6926,
+      "step": 76
+    },
+    {
+      "epoch": 0.08438356164383562,
+      "grad_norm": 2.3163535594940186,
+      "learning_rate": 1.1496350364963505e-06,
+      "loss": 2.6302,
+      "step": 77
+    },
+    {
+      "epoch": 0.08547945205479453,
+      "grad_norm": 2.2296528816223145,
+      "learning_rate": 1.1678832116788322e-06,
+      "loss": 2.6054,
+      "step": 78
+    },
+    {
+      "epoch": 0.08657534246575342,
+      "grad_norm": 2.253431797027588,
+      "learning_rate": 1.1861313868613138e-06,
+      "loss": 2.6104,
+      "step": 79
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "grad_norm": 2.490609645843506,
+      "learning_rate": 1.2043795620437959e-06,
+      "loss": 2.6078,
+      "step": 80
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "eval_loss": 2.706103563308716,
+      "eval_runtime": 37.0892,
+      "eval_samples_per_second": 26.962,
+      "eval_steps_per_second": 1.132,
+      "step": 80
+    },
+    {
+      "epoch": 0.08876712328767124,
+      "grad_norm": 2.3048248291015625,
+      "learning_rate": 1.2226277372262775e-06,
+      "loss": 2.6096,
+      "step": 81
+    },
+    {
+      "epoch": 0.08986301369863013,
+      "grad_norm": 2.440666913986206,
+      "learning_rate": 1.2408759124087592e-06,
+      "loss": 2.6412,
+      "step": 82
+    },
+    {
+      "epoch": 0.09095890410958904,
+      "grad_norm": 1.9456374645233154,
+      "learning_rate": 1.259124087591241e-06,
+      "loss": 2.591,
+      "step": 83
+    },
+    {
+      "epoch": 0.09205479452054795,
+      "grad_norm": 2.598172664642334,
+      "learning_rate": 1.2773722627737229e-06,
+      "loss": 2.6093,
+      "step": 84
+    },
+    {
+      "epoch": 0.09315068493150686,
+      "grad_norm": 2.2587032318115234,
+      "learning_rate": 1.2956204379562045e-06,
+      "loss": 2.5949,
+      "step": 85
+    },
+    {
+      "epoch": 0.09424657534246575,
+      "grad_norm": 2.2927448749542236,
+      "learning_rate": 1.3138686131386864e-06,
+      "loss": 2.5464,
+      "step": 86
+    },
+    {
+      "epoch": 0.09534246575342466,
+      "grad_norm": 2.5940468311309814,
+      "learning_rate": 1.332116788321168e-06,
+      "loss": 2.5613,
+      "step": 87
+    },
+    {
+      "epoch": 0.09643835616438357,
+      "grad_norm": 2.073207378387451,
+      "learning_rate": 1.3503649635036497e-06,
+      "loss": 2.4955,
+      "step": 88
+    },
+    {
+      "epoch": 0.09753424657534246,
+      "grad_norm": 3.211766242980957,
+      "learning_rate": 1.3686131386861315e-06,
+      "loss": 2.5315,
+      "step": 89
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "grad_norm": 2.2590038776397705,
+      "learning_rate": 1.3868613138686132e-06,
+      "loss": 2.4641,
+      "step": 90
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "eval_loss": 2.625087261199951,
+      "eval_runtime": 37.0852,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 90
+    },
+    {
+      "epoch": 0.09972602739726028,
+      "grad_norm": 2.3959836959838867,
+      "learning_rate": 1.405109489051095e-06,
+      "loss": 2.5537,
+      "step": 91
+    },
+    {
+      "epoch": 0.10082191780821918,
+      "grad_norm": 2.596761465072632,
+      "learning_rate": 1.4233576642335767e-06,
+      "loss": 2.5135,
+      "step": 92
+    },
+    {
+      "epoch": 0.10191780821917808,
+      "grad_norm": 2.475759267807007,
+      "learning_rate": 1.4416058394160583e-06,
+      "loss": 2.5436,
+      "step": 93
+    },
+    {
+      "epoch": 0.10301369863013699,
+      "grad_norm": 3.0613298416137695,
+      "learning_rate": 1.4598540145985402e-06,
+      "loss": 2.5066,
+      "step": 94
+    },
+    {
+      "epoch": 0.10410958904109589,
+      "grad_norm": 2.406097650527954,
+      "learning_rate": 1.478102189781022e-06,
+      "loss": 2.4646,
+      "step": 95
+    },
+    {
+      "epoch": 0.1052054794520548,
+      "grad_norm": 2.5872504711151123,
+      "learning_rate": 1.496350364963504e-06,
+      "loss": 2.5351,
+      "step": 96
+    },
+    {
+      "epoch": 0.1063013698630137,
+      "grad_norm": 2.4294471740722656,
+      "learning_rate": 1.5145985401459856e-06,
+      "loss": 2.4953,
+      "step": 97
+    },
+    {
+      "epoch": 0.1073972602739726,
+      "grad_norm": 2.674884080886841,
+      "learning_rate": 1.5328467153284674e-06,
+      "loss": 2.5133,
+      "step": 98
+    },
+    {
+      "epoch": 0.10849315068493151,
+      "grad_norm": 2.5096852779388428,
+      "learning_rate": 1.551094890510949e-06,
+      "loss": 2.4693,
+      "step": 99
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "grad_norm": 2.7390334606170654,
+      "learning_rate": 1.5693430656934307e-06,
+      "loss": 2.4195,
+      "step": 100
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "eval_loss": 2.55997371673584,
+      "eval_runtime": 37.0732,
+      "eval_samples_per_second": 26.974,
+      "eval_steps_per_second": 1.133,
+      "step": 100
+    },
+    {
+      "epoch": 0.11068493150684931,
+      "grad_norm": 2.8887312412261963,
+      "learning_rate": 1.5875912408759126e-06,
+      "loss": 2.4193,
+      "step": 101
+    },
+    {
+      "epoch": 0.11178082191780822,
+      "grad_norm": 3.0721421241760254,
+      "learning_rate": 1.6058394160583942e-06,
+      "loss": 2.4709,
+      "step": 102
+    },
+    {
+      "epoch": 0.11287671232876713,
+      "grad_norm": 2.728156566619873,
+      "learning_rate": 1.624087591240876e-06,
+      "loss": 2.5105,
+      "step": 103
+    },
+    {
+      "epoch": 0.11397260273972602,
+      "grad_norm": 3.1214346885681152,
+      "learning_rate": 1.6423357664233577e-06,
+      "loss": 2.4316,
+      "step": 104
+    },
+    {
+      "epoch": 0.11506849315068493,
+      "grad_norm": 2.9249517917633057,
+      "learning_rate": 1.6605839416058394e-06,
+      "loss": 2.3737,
+      "step": 105
+    },
+    {
+      "epoch": 0.11616438356164384,
+      "grad_norm": 2.6646173000335693,
+      "learning_rate": 1.6788321167883212e-06,
+      "loss": 2.4165,
+      "step": 106
+    },
+    {
+      "epoch": 0.11726027397260275,
+      "grad_norm": 3.108067750930786,
+      "learning_rate": 1.6970802919708029e-06,
+      "loss": 2.4217,
+      "step": 107
+    },
+    {
+      "epoch": 0.11835616438356164,
+      "grad_norm": 3.142338991165161,
+      "learning_rate": 1.715328467153285e-06,
+      "loss": 2.3527,
+      "step": 108
+    },
+    {
+      "epoch": 0.11945205479452055,
+      "grad_norm": 2.9078011512756348,
+      "learning_rate": 1.7335766423357666e-06,
+      "loss": 2.3927,
+      "step": 109
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "grad_norm": 3.40766978263855,
+      "learning_rate": 1.7518248175182485e-06,
+      "loss": 2.4021,
+      "step": 110
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "eval_loss": 2.5031485557556152,
+      "eval_runtime": 36.8147,
+      "eval_samples_per_second": 27.163,
+      "eval_steps_per_second": 1.141,
+      "step": 110
+    },
+    {
+      "epoch": 0.12164383561643835,
+      "grad_norm": 4.396062850952148,
+      "learning_rate": 1.77007299270073e-06,
+      "loss": 2.4373,
+      "step": 111
+    },
+    {
+      "epoch": 0.12273972602739726,
+      "grad_norm": 4.0441365242004395,
+      "learning_rate": 1.788321167883212e-06,
+      "loss": 2.4162,
+      "step": 112
+    },
+    {
+      "epoch": 0.12383561643835617,
+      "grad_norm": 2.8869969844818115,
+      "learning_rate": 1.8065693430656936e-06,
+      "loss": 2.4515,
+      "step": 113
+    },
+    {
+      "epoch": 0.12493150684931507,
+      "grad_norm": 3.4833669662475586,
+      "learning_rate": 1.8248175182481753e-06,
+      "loss": 2.3965,
+      "step": 114
+    },
+    {
+      "epoch": 0.12602739726027398,
+      "grad_norm": 2.9996068477630615,
+      "learning_rate": 1.8430656934306571e-06,
+      "loss": 2.3374,
+      "step": 115
+    },
+    {
+      "epoch": 0.12712328767123288,
+      "grad_norm": 2.701371192932129,
+      "learning_rate": 1.8613138686131388e-06,
+      "loss": 2.3608,
+      "step": 116
+    },
+    {
+      "epoch": 0.12821917808219177,
+      "grad_norm": 2.9491288661956787,
+      "learning_rate": 1.8795620437956206e-06,
+      "loss": 2.355,
+      "step": 117
+    },
+    {
+      "epoch": 0.1293150684931507,
+      "grad_norm": 2.9266560077667236,
+      "learning_rate": 1.8978102189781023e-06,
+      "loss": 2.3475,
+      "step": 118
+    },
+    {
+      "epoch": 0.1304109589041096,
+      "grad_norm": 3.0419797897338867,
+      "learning_rate": 1.916058394160584e-06,
+      "loss": 2.3475,
+      "step": 119
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "grad_norm": 2.4531214237213135,
+      "learning_rate": 1.934306569343066e-06,
+      "loss": 2.3724,
+      "step": 120
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "eval_loss": 2.455423593521118,
+      "eval_runtime": 37.0304,
+      "eval_samples_per_second": 27.005,
+      "eval_steps_per_second": 1.134,
+      "step": 120
+    },
+    {
+      "epoch": 0.1326027397260274,
+      "grad_norm": 3.572502851486206,
+      "learning_rate": 1.952554744525548e-06,
+      "loss": 2.2671,
+      "step": 121
+    },
+    {
+      "epoch": 0.1336986301369863,
+      "grad_norm": 3.5408504009246826,
+      "learning_rate": 1.9708029197080293e-06,
+      "loss": 2.3133,
+      "step": 122
+    },
+    {
+      "epoch": 0.13479452054794522,
+      "grad_norm": 2.488725423812866,
+      "learning_rate": 1.989051094890511e-06,
+      "loss": 2.3141,
+      "step": 123
+    },
+    {
+      "epoch": 0.1358904109589041,
+      "grad_norm": 3.241276264190674,
+      "learning_rate": 2.007299270072993e-06,
+      "loss": 2.318,
+      "step": 124
+    },
+    {
+      "epoch": 0.136986301369863,
+      "grad_norm": 2.853959798812866,
+      "learning_rate": 2.025547445255475e-06,
+      "loss": 2.3712,
+      "step": 125
+    },
+    {
+      "epoch": 0.13808219178082193,
+      "grad_norm": 2.9798169136047363,
+      "learning_rate": 2.0437956204379563e-06,
+      "loss": 2.3072,
+      "step": 126
+    },
+    {
+      "epoch": 0.13917808219178082,
+      "grad_norm": 2.8529601097106934,
+      "learning_rate": 2.062043795620438e-06,
+      "loss": 2.318,
+      "step": 127
+    },
+    {
+      "epoch": 0.14027397260273972,
+      "grad_norm": 3.475337028503418,
+      "learning_rate": 2.08029197080292e-06,
+      "loss": 2.31,
+      "step": 128
+    },
+    {
+      "epoch": 0.14136986301369864,
+      "grad_norm": 4.213955879211426,
+      "learning_rate": 2.0985401459854014e-06,
+      "loss": 2.3734,
+      "step": 129
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "grad_norm": 3.0833330154418945,
+      "learning_rate": 2.1167883211678833e-06,
+      "loss": 2.2602,
+      "step": 130
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "eval_loss": 2.4205682277679443,
+      "eval_runtime": 37.0798,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 130
+    },
+    {
+      "epoch": 0.14356164383561643,
+      "grad_norm": 3.411235809326172,
+      "learning_rate": 2.135036496350365e-06,
+      "loss": 2.3255,
+      "step": 131
+    },
+    {
+      "epoch": 0.14465753424657535,
+      "grad_norm": 2.882053852081299,
+      "learning_rate": 2.1532846715328466e-06,
+      "loss": 2.3483,
+      "step": 132
+    },
+    {
+      "epoch": 0.14575342465753424,
+      "grad_norm": 3.2436420917510986,
+      "learning_rate": 2.171532846715329e-06,
+      "loss": 2.2492,
+      "step": 133
+    },
+    {
+      "epoch": 0.14684931506849316,
+      "grad_norm": 2.929466962814331,
+      "learning_rate": 2.1897810218978103e-06,
+      "loss": 2.2647,
+      "step": 134
+    },
+    {
+      "epoch": 0.14794520547945206,
+      "grad_norm": 2.9344966411590576,
+      "learning_rate": 2.208029197080292e-06,
+      "loss": 2.3061,
+      "step": 135
+    },
+    {
+      "epoch": 0.14904109589041095,
+      "grad_norm": 3.2567203044891357,
+      "learning_rate": 2.226277372262774e-06,
+      "loss": 2.2746,
+      "step": 136
+    },
+    {
+      "epoch": 0.15013698630136987,
+      "grad_norm": 3.1494054794311523,
+      "learning_rate": 2.244525547445256e-06,
+      "loss": 2.2866,
+      "step": 137
+    },
+    {
+      "epoch": 0.15123287671232877,
+      "grad_norm": 2.723390817642212,
+      "learning_rate": 2.2627737226277373e-06,
+      "loss": 2.3109,
+      "step": 138
+    },
+    {
+      "epoch": 0.15232876712328766,
+      "grad_norm": 2.8904967308044434,
+      "learning_rate": 2.281021897810219e-06,
+      "loss": 2.301,
+      "step": 139
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "grad_norm": 2.8899483680725098,
+      "learning_rate": 2.299270072992701e-06,
+      "loss": 2.2637,
+      "step": 140
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "eval_loss": 2.3952724933624268,
+      "eval_runtime": 37.1272,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 140
+    },
+    {
+      "epoch": 0.15452054794520548,
+      "grad_norm": 3.5309066772460938,
+      "learning_rate": 2.3175182481751825e-06,
+      "loss": 2.2572,
+      "step": 141
+    },
+    {
+      "epoch": 0.15561643835616437,
+      "grad_norm": 2.8369667530059814,
+      "learning_rate": 2.3357664233576643e-06,
+      "loss": 2.2712,
+      "step": 142
+    },
+    {
+      "epoch": 0.1567123287671233,
+      "grad_norm": 3.629161834716797,
+      "learning_rate": 2.354014598540146e-06,
+      "loss": 2.2301,
+      "step": 143
+    },
+    {
+      "epoch": 0.1578082191780822,
+      "grad_norm": 3.5313000679016113,
+      "learning_rate": 2.3722627737226276e-06,
+      "loss": 2.2712,
+      "step": 144
+    },
+    {
+      "epoch": 0.1589041095890411,
+      "grad_norm": 3.1328771114349365,
+      "learning_rate": 2.39051094890511e-06,
+      "loss": 2.2706,
+      "step": 145
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 4.6992950439453125,
+      "learning_rate": 2.4087591240875918e-06,
+      "loss": 2.219,
+      "step": 146
+    },
+    {
+      "epoch": 0.1610958904109589,
+      "grad_norm": 5.230221748352051,
+      "learning_rate": 2.427007299270073e-06,
+      "loss": 2.2396,
+      "step": 147
+    },
+    {
+      "epoch": 0.16219178082191782,
+      "grad_norm": 3.5680761337280273,
+      "learning_rate": 2.445255474452555e-06,
+      "loss": 2.2434,
+      "step": 148
+    },
+    {
+      "epoch": 0.1632876712328767,
+      "grad_norm": 3.1642441749572754,
+      "learning_rate": 2.463503649635037e-06,
+      "loss": 2.2856,
+      "step": 149
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "grad_norm": 2.694085121154785,
+      "learning_rate": 2.4817518248175183e-06,
+      "loss": 2.2006,
+      "step": 150
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "eval_loss": 2.37199068069458,
+      "eval_runtime": 37.1342,
+      "eval_samples_per_second": 26.929,
+      "eval_steps_per_second": 1.131,
+      "step": 150
+    },
+    {
+      "epoch": 0.16547945205479453,
+      "grad_norm": 2.9868853092193604,
+      "learning_rate": 2.5e-06,
+      "loss": 2.1694,
+      "step": 151
+    },
+    {
+      "epoch": 0.16657534246575342,
+      "grad_norm": 2.793532371520996,
+      "learning_rate": 2.518248175182482e-06,
+      "loss": 2.2965,
+      "step": 152
+    },
+    {
+      "epoch": 0.16767123287671232,
+      "grad_norm": 2.60152530670166,
+      "learning_rate": 2.5364963503649635e-06,
+      "loss": 2.1954,
+      "step": 153
+    },
+    {
+      "epoch": 0.16876712328767124,
+      "grad_norm": 2.6192219257354736,
+      "learning_rate": 2.5547445255474458e-06,
+      "loss": 2.1909,
+      "step": 154
+    },
+    {
+      "epoch": 0.16986301369863013,
+      "grad_norm": 2.467344045639038,
+      "learning_rate": 2.5729927007299272e-06,
+      "loss": 2.2438,
+      "step": 155
+    },
+    {
+      "epoch": 0.17095890410958905,
+      "grad_norm": 2.772012948989868,
+      "learning_rate": 2.591240875912409e-06,
+      "loss": 2.2213,
+      "step": 156
+    },
+    {
+      "epoch": 0.17205479452054795,
+      "grad_norm": 2.613675355911255,
+      "learning_rate": 2.6094890510948905e-06,
+      "loss": 2.2213,
+      "step": 157
+    },
+    {
+      "epoch": 0.17315068493150684,
+      "grad_norm": 2.849245071411133,
+      "learning_rate": 2.627737226277373e-06,
+      "loss": 2.255,
+      "step": 158
+    },
+    {
+      "epoch": 0.17424657534246576,
+      "grad_norm": 2.8388969898223877,
+      "learning_rate": 2.6459854014598542e-06,
+      "loss": 2.2572,
+      "step": 159
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "grad_norm": 2.5219321250915527,
+      "learning_rate": 2.664233576642336e-06,
+      "loss": 2.2253,
+      "step": 160
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "eval_loss": 2.3509061336517334,
+      "eval_runtime": 37.0953,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 160
+    },
+    {
+      "epoch": 0.17643835616438355,
+      "grad_norm": 2.8460891246795654,
+      "learning_rate": 2.6824817518248175e-06,
+      "loss": 2.2009,
+      "step": 161
+    },
+    {
+      "epoch": 0.17753424657534247,
+      "grad_norm": 2.673377275466919,
+      "learning_rate": 2.7007299270072994e-06,
+      "loss": 2.2211,
+      "step": 162
+    },
+    {
+      "epoch": 0.17863013698630137,
+      "grad_norm": 2.8514182567596436,
+      "learning_rate": 2.718978102189781e-06,
+      "loss": 2.1955,
+      "step": 163
+    },
+    {
+      "epoch": 0.17972602739726026,
+      "grad_norm": 3.530385971069336,
+      "learning_rate": 2.737226277372263e-06,
+      "loss": 2.2536,
+      "step": 164
+    },
+    {
+      "epoch": 0.18082191780821918,
+      "grad_norm": 2.733394145965576,
+      "learning_rate": 2.755474452554745e-06,
+      "loss": 2.2367,
+      "step": 165
+    },
+    {
+      "epoch": 0.18191780821917808,
+      "grad_norm": 3.871150016784668,
+      "learning_rate": 2.7737226277372264e-06,
+      "loss": 2.1949,
+      "step": 166
+    },
+    {
+      "epoch": 0.18301369863013697,
+      "grad_norm": 3.7782530784606934,
+      "learning_rate": 2.7919708029197082e-06,
+      "loss": 2.2979,
+      "step": 167
+    },
+    {
+      "epoch": 0.1841095890410959,
+      "grad_norm": 3.094599485397339,
+      "learning_rate": 2.81021897810219e-06,
+      "loss": 2.2512,
+      "step": 168
+    },
+    {
+      "epoch": 0.1852054794520548,
+      "grad_norm": 3.2166848182678223,
+      "learning_rate": 2.828467153284672e-06,
+      "loss": 2.1784,
+      "step": 169
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "grad_norm": 3.4127073287963867,
+      "learning_rate": 2.8467153284671534e-06,
+      "loss": 2.2241,
+      "step": 170
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "eval_loss": 2.333245038986206,
+      "eval_runtime": 37.0402,
+      "eval_samples_per_second": 26.998,
+      "eval_steps_per_second": 1.134,
+      "step": 170
+    },
+    {
+      "epoch": 0.1873972602739726,
+      "grad_norm": 2.677579879760742,
+      "learning_rate": 2.8649635036496353e-06,
+      "loss": 2.135,
+      "step": 171
+    },
+    {
+      "epoch": 0.1884931506849315,
+      "grad_norm": 2.9602150917053223,
+      "learning_rate": 2.8832116788321167e-06,
+      "loss": 2.1916,
+      "step": 172
+    },
+    {
+      "epoch": 0.18958904109589042,
+      "grad_norm": 2.580122232437134,
+      "learning_rate": 2.901459854014599e-06,
+      "loss": 2.1477,
+      "step": 173
+    },
+    {
+      "epoch": 0.19068493150684931,
+      "grad_norm": 2.652217388153076,
+      "learning_rate": 2.9197080291970804e-06,
+      "loss": 2.2394,
+      "step": 174
+    },
+    {
+      "epoch": 0.1917808219178082,
+      "grad_norm": 2.6887567043304443,
+      "learning_rate": 2.9379562043795623e-06,
+      "loss": 2.2245,
+      "step": 175
+    },
+    {
+      "epoch": 0.19287671232876713,
+      "grad_norm": 2.768165111541748,
+      "learning_rate": 2.956204379562044e-06,
+      "loss": 2.2017,
+      "step": 176
+    },
+    {
+      "epoch": 0.19397260273972602,
+      "grad_norm": 2.9375691413879395,
+      "learning_rate": 2.9744525547445256e-06,
+      "loss": 2.2464,
+      "step": 177
+    },
+    {
+      "epoch": 0.19506849315068492,
+      "grad_norm": 2.8364248275756836,
+      "learning_rate": 2.992700729927008e-06,
+      "loss": 2.2155,
+      "step": 178
+    },
+    {
+      "epoch": 0.19616438356164384,
+      "grad_norm": 2.843487024307251,
+      "learning_rate": 3.0109489051094893e-06,
+      "loss": 2.1577,
+      "step": 179
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "grad_norm": 3.005230665206909,
+      "learning_rate": 3.029197080291971e-06,
+      "loss": 2.1549,
+      "step": 180
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "eval_loss": 2.3209409713745117,
+      "eval_runtime": 37.022,
+      "eval_samples_per_second": 27.011,
+      "eval_steps_per_second": 1.134,
+      "step": 180
+    },
+    {
+      "epoch": 0.19835616438356166,
+      "grad_norm": 2.490260601043701,
+      "learning_rate": 3.0474452554744526e-06,
+      "loss": 2.2313,
+      "step": 181
+    },
+    {
+      "epoch": 0.19945205479452055,
+      "grad_norm": 2.973118543624878,
+      "learning_rate": 3.065693430656935e-06,
+      "loss": 2.2316,
+      "step": 182
+    },
+    {
+      "epoch": 0.20054794520547944,
+      "grad_norm": 3.3308680057525635,
+      "learning_rate": 3.0839416058394163e-06,
+      "loss": 2.2087,
+      "step": 183
+    },
+    {
+      "epoch": 0.20164383561643837,
+      "grad_norm": 3.1440727710723877,
+      "learning_rate": 3.102189781021898e-06,
+      "loss": 2.2256,
+      "step": 184
+    },
+    {
+      "epoch": 0.20273972602739726,
+      "grad_norm": 2.5696325302124023,
+      "learning_rate": 3.1204379562043796e-06,
+      "loss": 2.2093,
+      "step": 185
+    },
+    {
+      "epoch": 0.20383561643835615,
+      "grad_norm": 3.1225037574768066,
+      "learning_rate": 3.1386861313868614e-06,
+      "loss": 2.1425,
+      "step": 186
+    },
+    {
+      "epoch": 0.20493150684931508,
+      "grad_norm": 3.650743246078491,
+      "learning_rate": 3.156934306569343e-06,
+      "loss": 2.1339,
+      "step": 187
+    },
+    {
+      "epoch": 0.20602739726027397,
+      "grad_norm": 2.924072027206421,
+      "learning_rate": 3.175182481751825e-06,
+      "loss": 2.2363,
+      "step": 188
+    },
+    {
+      "epoch": 0.20712328767123286,
+      "grad_norm": 2.9561989307403564,
+      "learning_rate": 3.193430656934307e-06,
+      "loss": 2.1934,
+      "step": 189
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "grad_norm": 2.731393814086914,
+      "learning_rate": 3.2116788321167884e-06,
+      "loss": 2.1718,
+      "step": 190
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "eval_loss": 2.3087782859802246,
+      "eval_runtime": 37.0161,
+      "eval_samples_per_second": 27.015,
+      "eval_steps_per_second": 1.135,
+      "step": 190
+    },
+    {
+      "epoch": 0.20931506849315068,
+      "grad_norm": 2.774442434310913,
+      "learning_rate": 3.2299270072992707e-06,
+      "loss": 2.1841,
+      "step": 191
+    },
+    {
+      "epoch": 0.2104109589041096,
+      "grad_norm": 3.0579135417938232,
+      "learning_rate": 3.248175182481752e-06,
+      "loss": 2.1677,
+      "step": 192
+    },
+    {
+      "epoch": 0.2115068493150685,
+      "grad_norm": 2.732984781265259,
+      "learning_rate": 3.266423357664234e-06,
+      "loss": 2.0975,
+      "step": 193
+    },
+    {
+      "epoch": 0.2126027397260274,
+      "grad_norm": 3.442659854888916,
+      "learning_rate": 3.2846715328467155e-06,
+      "loss": 2.1923,
+      "step": 194
+    },
+    {
+      "epoch": 0.2136986301369863,
+      "grad_norm": 3.064223527908325,
+      "learning_rate": 3.3029197080291973e-06,
+      "loss": 2.1707,
+      "step": 195
+    },
+    {
+      "epoch": 0.2147945205479452,
+      "grad_norm": 2.5138468742370605,
+      "learning_rate": 3.3211678832116788e-06,
+      "loss": 2.1933,
+      "step": 196
+    },
+    {
+      "epoch": 0.2158904109589041,
+      "grad_norm": 3.1076464653015137,
+      "learning_rate": 3.339416058394161e-06,
+      "loss": 2.1725,
+      "step": 197
+    },
+    {
+      "epoch": 0.21698630136986302,
+      "grad_norm": 2.7291388511657715,
+      "learning_rate": 3.3576642335766425e-06,
+      "loss": 2.1722,
+      "step": 198
+    },
+    {
+      "epoch": 0.21808219178082192,
+      "grad_norm": 2.805506467819214,
+      "learning_rate": 3.3759124087591243e-06,
+      "loss": 2.1688,
+      "step": 199
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "grad_norm": 2.6190085411071777,
+      "learning_rate": 3.3941605839416058e-06,
+      "loss": 2.1708,
+      "step": 200
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "eval_loss": 2.295546054840088,
+      "eval_runtime": 37.0133,
+      "eval_samples_per_second": 27.017,
+      "eval_steps_per_second": 1.135,
+      "step": 200
+    },
+    {
+      "epoch": 0.22027397260273973,
+      "grad_norm": 3.1270673274993896,
+      "learning_rate": 3.412408759124088e-06,
+      "loss": 2.1759,
+      "step": 201
+    },
+    {
+      "epoch": 0.22136986301369863,
+      "grad_norm": 3.494178533554077,
+      "learning_rate": 3.43065693430657e-06,
+      "loss": 2.1795,
+      "step": 202
+    },
+    {
+      "epoch": 0.22246575342465755,
+      "grad_norm": 2.7398736476898193,
+      "learning_rate": 3.4489051094890513e-06,
+      "loss": 2.1201,
+      "step": 203
+    },
+    {
+      "epoch": 0.22356164383561644,
+      "grad_norm": 2.981506824493408,
+      "learning_rate": 3.467153284671533e-06,
+      "loss": 2.1693,
+      "step": 204
+    },
+    {
+      "epoch": 0.22465753424657534,
+      "grad_norm": 2.4129178524017334,
+      "learning_rate": 3.4854014598540146e-06,
+      "loss": 2.1125,
+      "step": 205
+    },
+    {
+      "epoch": 0.22575342465753426,
+      "grad_norm": 2.701197385787964,
+      "learning_rate": 3.503649635036497e-06,
+      "loss": 2.1565,
+      "step": 206
+    },
+    {
+      "epoch": 0.22684931506849315,
+      "grad_norm": 2.533306837081909,
+      "learning_rate": 3.5218978102189783e-06,
+      "loss": 2.1487,
+      "step": 207
+    },
+    {
+      "epoch": 0.22794520547945205,
+      "grad_norm": 2.8030788898468018,
+      "learning_rate": 3.54014598540146e-06,
+      "loss": 2.1801,
+      "step": 208
+    },
+    {
+      "epoch": 0.22904109589041097,
+      "grad_norm": 2.895414352416992,
+      "learning_rate": 3.5583941605839416e-06,
+      "loss": 2.189,
+      "step": 209
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "grad_norm": 2.508390426635742,
+      "learning_rate": 3.576642335766424e-06,
+      "loss": 2.2376,
+      "step": 210
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "eval_loss": 2.2900516986846924,
+      "eval_runtime": 36.7645,
+      "eval_samples_per_second": 27.2,
+      "eval_steps_per_second": 1.142,
+      "step": 210
+    },
+    {
+      "epoch": 0.23123287671232876,
+      "grad_norm": 2.6393816471099854,
+      "learning_rate": 3.5948905109489054e-06,
+      "loss": 2.2267,
+      "step": 211
+    },
+    {
+      "epoch": 0.23232876712328768,
+      "grad_norm": 3.0687549114227295,
+      "learning_rate": 3.6131386861313872e-06,
+      "loss": 2.0744,
+      "step": 212
+    },
+    {
+      "epoch": 0.23342465753424657,
+      "grad_norm": 2.6342105865478516,
+      "learning_rate": 3.6313868613138687e-06,
+      "loss": 2.0737,
+      "step": 213
+    },
+    {
+      "epoch": 0.2345205479452055,
+      "grad_norm": 3.5662875175476074,
+      "learning_rate": 3.6496350364963505e-06,
+      "loss": 2.1147,
+      "step": 214
+    },
+    {
+      "epoch": 0.2356164383561644,
+      "grad_norm": 2.5357561111450195,
+      "learning_rate": 3.667883211678833e-06,
+      "loss": 2.1268,
+      "step": 215
+    },
+    {
+      "epoch": 0.23671232876712328,
+      "grad_norm": 3.439711093902588,
+      "learning_rate": 3.6861313868613142e-06,
+      "loss": 2.1118,
+      "step": 216
+    },
+    {
+      "epoch": 0.2378082191780822,
+      "grad_norm": 2.9410204887390137,
+      "learning_rate": 3.704379562043796e-06,
+      "loss": 2.1087,
+      "step": 217
+    },
+    {
+      "epoch": 0.2389041095890411,
+      "grad_norm": 3.3182294368743896,
+      "learning_rate": 3.7226277372262775e-06,
+      "loss": 2.196,
+      "step": 218
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 3.3834011554718018,
+      "learning_rate": 3.7408759124087594e-06,
+      "loss": 2.1633,
+      "step": 219
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "grad_norm": 3.114952325820923,
+      "learning_rate": 3.7591240875912412e-06,
+      "loss": 2.1107,
+      "step": 220
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "eval_loss": 2.277580976486206,
+      "eval_runtime": 36.9846,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 220
+    },
+    {
+      "epoch": 0.2421917808219178,
+      "grad_norm": 3.1151392459869385,
+      "learning_rate": 3.777372262773723e-06,
+      "loss": 2.1471,
+      "step": 221
+    },
+    {
+      "epoch": 0.2432876712328767,
+      "grad_norm": 3.798450469970703,
+      "learning_rate": 3.7956204379562045e-06,
+      "loss": 2.1324,
+      "step": 222
+    },
+    {
+      "epoch": 0.24438356164383562,
+      "grad_norm": 2.6187098026275635,
+      "learning_rate": 3.8138686131386864e-06,
+      "loss": 2.1061,
+      "step": 223
+    },
+    {
+      "epoch": 0.24547945205479452,
+      "grad_norm": 4.014018535614014,
+      "learning_rate": 3.832116788321168e-06,
+      "loss": 2.166,
+      "step": 224
+    },
+    {
+      "epoch": 0.2465753424657534,
+      "grad_norm": 3.0043938159942627,
+      "learning_rate": 3.85036496350365e-06,
+      "loss": 2.1197,
+      "step": 225
+    },
+    {
+      "epoch": 0.24767123287671233,
+      "grad_norm": 3.507066249847412,
+      "learning_rate": 3.868613138686132e-06,
+      "loss": 2.0848,
+      "step": 226
+    },
+    {
+      "epoch": 0.24876712328767123,
+      "grad_norm": 3.5713281631469727,
+      "learning_rate": 3.886861313868613e-06,
+      "loss": 2.1239,
+      "step": 227
+    },
+    {
+      "epoch": 0.24986301369863015,
+      "grad_norm": 3.961082696914673,
+      "learning_rate": 3.905109489051096e-06,
+      "loss": 2.1406,
+      "step": 228
+    },
+    {
+      "epoch": 0.25095890410958904,
+      "grad_norm": 3.1514859199523926,
+      "learning_rate": 3.923357664233577e-06,
+      "loss": 2.1689,
+      "step": 229
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "grad_norm": 3.8486833572387695,
+      "learning_rate": 3.9416058394160585e-06,
+      "loss": 2.1321,
+      "step": 230
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "eval_loss": 2.2620601654052734,
+      "eval_runtime": 37.058,
+      "eval_samples_per_second": 26.985,
+      "eval_steps_per_second": 1.133,
+      "step": 230
+    },
+    {
+      "epoch": 0.25315068493150683,
+      "grad_norm": 2.5626907348632812,
+      "learning_rate": 3.95985401459854e-06,
+      "loss": 2.1777,
+      "step": 231
+    },
+    {
+      "epoch": 0.25424657534246575,
+      "grad_norm": 4.105278015136719,
+      "learning_rate": 3.978102189781022e-06,
+      "loss": 2.2099,
+      "step": 232
+    },
+    {
+      "epoch": 0.2553424657534247,
+      "grad_norm": 2.9818739891052246,
+      "learning_rate": 3.996350364963504e-06,
+      "loss": 2.0705,
+      "step": 233
+    },
+    {
+      "epoch": 0.25643835616438354,
+      "grad_norm": 2.87030291557312,
+      "learning_rate": 4.014598540145986e-06,
+      "loss": 2.1453,
+      "step": 234
+    },
+    {
+      "epoch": 0.25753424657534246,
+      "grad_norm": 2.4847915172576904,
+      "learning_rate": 4.032846715328467e-06,
+      "loss": 2.0461,
+      "step": 235
+    },
+    {
+      "epoch": 0.2586301369863014,
+      "grad_norm": 3.451420783996582,
+      "learning_rate": 4.05109489051095e-06,
+      "loss": 2.1133,
+      "step": 236
+    },
+    {
+      "epoch": 0.25972602739726025,
+      "grad_norm": 2.954723358154297,
+      "learning_rate": 4.069343065693431e-06,
+      "loss": 2.1316,
+      "step": 237
+    },
+    {
+      "epoch": 0.2608219178082192,
+      "grad_norm": 2.5295896530151367,
+      "learning_rate": 4.0875912408759126e-06,
+      "loss": 2.0901,
+      "step": 238
+    },
+    {
+      "epoch": 0.2619178082191781,
+      "grad_norm": 2.6359453201293945,
+      "learning_rate": 4.105839416058395e-06,
+      "loss": 2.117,
+      "step": 239
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "grad_norm": 2.5236105918884277,
+      "learning_rate": 4.124087591240876e-06,
+      "loss": 2.1958,
+      "step": 240
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "eval_loss": 2.250295400619507,
+      "eval_runtime": 37.0811,
+      "eval_samples_per_second": 26.968,
+      "eval_steps_per_second": 1.133,
+      "step": 240
+    },
+    {
+      "epoch": 0.2641095890410959,
+      "grad_norm": 2.581918954849243,
+      "learning_rate": 4.1423357664233586e-06,
+      "loss": 2.1054,
+      "step": 241
+    },
+    {
+      "epoch": 0.2652054794520548,
+      "grad_norm": 2.501332998275757,
+      "learning_rate": 4.16058394160584e-06,
+      "loss": 2.0878,
+      "step": 242
+    },
+    {
+      "epoch": 0.2663013698630137,
+      "grad_norm": 2.6477770805358887,
+      "learning_rate": 4.1788321167883214e-06,
+      "loss": 2.1656,
+      "step": 243
+    },
+    {
+      "epoch": 0.2673972602739726,
+      "grad_norm": 2.7252438068389893,
+      "learning_rate": 4.197080291970803e-06,
+      "loss": 2.0497,
+      "step": 244
+    },
+    {
+      "epoch": 0.2684931506849315,
+      "grad_norm": 2.6056301593780518,
+      "learning_rate": 4.215328467153285e-06,
+      "loss": 2.1411,
+      "step": 245
+    },
+    {
+      "epoch": 0.26958904109589044,
+      "grad_norm": 2.6581931114196777,
+      "learning_rate": 4.233576642335767e-06,
+      "loss": 2.0829,
+      "step": 246
+    },
+    {
+      "epoch": 0.2706849315068493,
+      "grad_norm": 2.6316752433776855,
+      "learning_rate": 4.251824817518249e-06,
+      "loss": 2.0882,
+      "step": 247
+    },
+    {
+      "epoch": 0.2717808219178082,
+      "grad_norm": 2.803823947906494,
+      "learning_rate": 4.27007299270073e-06,
+      "loss": 2.1203,
+      "step": 248
+    },
+    {
+      "epoch": 0.27287671232876715,
+      "grad_norm": 2.9551174640655518,
+      "learning_rate": 4.288321167883212e-06,
+      "loss": 2.0897,
+      "step": 249
+    },
+    {
+      "epoch": 0.273972602739726,
+      "grad_norm": 2.4847161769866943,
+      "learning_rate": 4.306569343065693e-06,
+      "loss": 2.0845,
+      "step": 250
+    },
+    {
+      "epoch": 0.273972602739726,
+      "eval_loss": 2.2491259574890137,
+      "eval_runtime": 37.1024,
+      "eval_samples_per_second": 26.952,
+      "eval_steps_per_second": 1.132,
+      "step": 250
+    },
+    {
+      "epoch": 0.27506849315068493,
+      "grad_norm": 2.6398351192474365,
+      "learning_rate": 4.3248175182481755e-06,
+      "loss": 2.0479,
+      "step": 251
+    },
+    {
+      "epoch": 0.27616438356164386,
+      "grad_norm": 3.0948562622070312,
+      "learning_rate": 4.343065693430658e-06,
+      "loss": 2.1184,
+      "step": 252
+    },
+    {
+      "epoch": 0.2772602739726027,
+      "grad_norm": 2.5378012657165527,
+      "learning_rate": 4.361313868613139e-06,
+      "loss": 2.1099,
+      "step": 253
+    },
+    {
+      "epoch": 0.27835616438356164,
+      "grad_norm": 3.0282225608825684,
+      "learning_rate": 4.379562043795621e-06,
+      "loss": 2.0982,
+      "step": 254
+    },
+    {
+      "epoch": 0.27945205479452057,
+      "grad_norm": 2.6205334663391113,
+      "learning_rate": 4.397810218978102e-06,
+      "loss": 2.1285,
+      "step": 255
+    },
+    {
+      "epoch": 0.28054794520547943,
+      "grad_norm": 3.2056095600128174,
+      "learning_rate": 4.416058394160584e-06,
+      "loss": 2.1221,
+      "step": 256
+    },
+    {
+      "epoch": 0.28164383561643835,
+      "grad_norm": 3.1370127201080322,
+      "learning_rate": 4.434306569343066e-06,
+      "loss": 2.1016,
+      "step": 257
+    },
+    {
+      "epoch": 0.2827397260273973,
+      "grad_norm": 2.494168519973755,
+      "learning_rate": 4.452554744525548e-06,
+      "loss": 2.0754,
+      "step": 258
+    },
+    {
+      "epoch": 0.28383561643835614,
+      "grad_norm": 2.7979183197021484,
+      "learning_rate": 4.4708029197080295e-06,
+      "loss": 2.1218,
+      "step": 259
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "grad_norm": 2.755682945251465,
+      "learning_rate": 4.489051094890512e-06,
+      "loss": 2.1136,
+      "step": 260
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "eval_loss": 2.2433817386627197,
+      "eval_runtime": 37.1121,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 260
+    },
+    {
+      "epoch": 0.286027397260274,
+      "grad_norm": 2.3641891479492188,
+      "learning_rate": 4.507299270072993e-06,
+      "loss": 2.1488,
+      "step": 261
+    },
+    {
+      "epoch": 0.28712328767123285,
+      "grad_norm": 2.741292953491211,
+      "learning_rate": 4.525547445255475e-06,
+      "loss": 2.1048,
+      "step": 262
+    },
+    {
+      "epoch": 0.2882191780821918,
+      "grad_norm": 2.811483383178711,
+      "learning_rate": 4.543795620437956e-06,
+      "loss": 2.0996,
+      "step": 263
+    },
+    {
+      "epoch": 0.2893150684931507,
+      "grad_norm": 2.6705479621887207,
+      "learning_rate": 4.562043795620438e-06,
+      "loss": 2.0652,
+      "step": 264
+    },
+    {
+      "epoch": 0.29041095890410956,
+      "grad_norm": 2.9383747577667236,
+      "learning_rate": 4.580291970802921e-06,
+      "loss": 2.1574,
+      "step": 265
+    },
+    {
+      "epoch": 0.2915068493150685,
+      "grad_norm": 2.591373920440674,
+      "learning_rate": 4.598540145985402e-06,
+      "loss": 2.1655,
+      "step": 266
+    },
+    {
+      "epoch": 0.2926027397260274,
+      "grad_norm": 3.16955828666687,
+      "learning_rate": 4.6167883211678835e-06,
+      "loss": 2.1285,
+      "step": 267
+    },
+    {
+      "epoch": 0.2936986301369863,
+      "grad_norm": 2.66941499710083,
+      "learning_rate": 4.635036496350365e-06,
+      "loss": 2.0777,
+      "step": 268
+    },
+    {
+      "epoch": 0.2947945205479452,
+      "grad_norm": 2.768139600753784,
+      "learning_rate": 4.653284671532847e-06,
+      "loss": 2.0748,
+      "step": 269
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "grad_norm": 2.5951924324035645,
+      "learning_rate": 4.671532846715329e-06,
+      "loss": 2.1289,
+      "step": 270
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "eval_loss": 2.2393550872802734,
+      "eval_runtime": 37.0993,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 270
+    },
+    {
+      "epoch": 0.29698630136986304,
+      "grad_norm": 2.5422160625457764,
+      "learning_rate": 4.689781021897811e-06,
+      "loss": 2.1342,
+      "step": 271
+    },
+    {
+      "epoch": 0.2980821917808219,
+      "grad_norm": 2.540555477142334,
+      "learning_rate": 4.708029197080292e-06,
+      "loss": 2.097,
+      "step": 272
+    },
+    {
+      "epoch": 0.2991780821917808,
+      "grad_norm": 2.995366334915161,
+      "learning_rate": 4.726277372262774e-06,
+      "loss": 2.1452,
+      "step": 273
+    },
+    {
+      "epoch": 0.30027397260273975,
+      "grad_norm": 2.2561020851135254,
+      "learning_rate": 4.744525547445255e-06,
+      "loss": 2.1069,
+      "step": 274
+    },
+    {
+      "epoch": 0.3013698630136986,
+      "grad_norm": 2.53476619720459,
+      "learning_rate": 4.7627737226277375e-06,
+      "loss": 2.0381,
+      "step": 275
+    },
+    {
+      "epoch": 0.30246575342465754,
+      "grad_norm": 2.3415749073028564,
+      "learning_rate": 4.78102189781022e-06,
+      "loss": 2.1027,
+      "step": 276
+    },
+    {
+      "epoch": 0.30356164383561646,
+      "grad_norm": 2.9756431579589844,
+      "learning_rate": 4.799270072992701e-06,
+      "loss": 2.1136,
+      "step": 277
+    },
+    {
+      "epoch": 0.3046575342465753,
+      "grad_norm": 2.606208324432373,
+      "learning_rate": 4.8175182481751835e-06,
+      "loss": 2.0797,
+      "step": 278
+    },
+    {
+      "epoch": 0.30575342465753425,
+      "grad_norm": 2.5089313983917236,
+      "learning_rate": 4.835766423357665e-06,
+      "loss": 2.1026,
+      "step": 279
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "grad_norm": 2.5003838539123535,
+      "learning_rate": 4.854014598540146e-06,
+      "loss": 2.0923,
+      "step": 280
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "eval_loss": 2.2351036071777344,
+      "eval_runtime": 37.111,
+      "eval_samples_per_second": 26.946,
+      "eval_steps_per_second": 1.132,
+      "step": 280
+    },
+    {
+      "epoch": 0.30794520547945203,
+      "grad_norm": 2.403602361679077,
+      "learning_rate": 4.872262773722628e-06,
+      "loss": 2.0388,
+      "step": 281
+    },
+    {
+      "epoch": 0.30904109589041096,
+      "grad_norm": 2.7775959968566895,
+      "learning_rate": 4.89051094890511e-06,
+      "loss": 2.0561,
+      "step": 282
+    },
+    {
+      "epoch": 0.3101369863013699,
+      "grad_norm": 2.7128653526306152,
+      "learning_rate": 4.9087591240875915e-06,
+      "loss": 2.042,
+      "step": 283
+    },
+    {
+      "epoch": 0.31123287671232874,
+      "grad_norm": 2.7029991149902344,
+      "learning_rate": 4.927007299270074e-06,
+      "loss": 2.0814,
+      "step": 284
+    },
+    {
+      "epoch": 0.31232876712328766,
+      "grad_norm": 2.3942933082580566,
+      "learning_rate": 4.945255474452555e-06,
+      "loss": 2.0906,
+      "step": 285
+    },
+    {
+      "epoch": 0.3134246575342466,
+      "grad_norm": 2.5086541175842285,
+      "learning_rate": 4.963503649635037e-06,
+      "loss": 2.0625,
+      "step": 286
+    },
+    {
+      "epoch": 0.31452054794520545,
+      "grad_norm": 2.6668760776519775,
+      "learning_rate": 4.981751824817518e-06,
+      "loss": 2.1538,
+      "step": 287
+    },
+    {
+      "epoch": 0.3156164383561644,
+      "grad_norm": 2.5066750049591064,
+      "learning_rate": 5e-06,
+      "loss": 2.1423,
+      "step": 288
+    },
+    {
+      "epoch": 0.3167123287671233,
+      "grad_norm": 2.84967041015625,
+      "learning_rate": 4.997969130787978e-06,
+      "loss": 2.0943,
+      "step": 289
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "grad_norm": 2.151946544647217,
+      "learning_rate": 4.995938261575955e-06,
+      "loss": 2.1098,
+      "step": 290
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "eval_loss": 2.223247528076172,
+      "eval_runtime": 37.1755,
+      "eval_samples_per_second": 26.899,
+      "eval_steps_per_second": 1.13,
+      "step": 290
+    },
+    {
+      "epoch": 0.3189041095890411,
+      "grad_norm": 2.887166976928711,
+      "learning_rate": 4.993907392363932e-06,
+      "loss": 2.0281,
+      "step": 291
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 2.5930211544036865,
+      "learning_rate": 4.9918765231519095e-06,
+      "loss": 2.1063,
+      "step": 292
+    },
+    {
+      "epoch": 0.32109589041095893,
+      "grad_norm": 2.5568413734436035,
+      "learning_rate": 4.989845653939887e-06,
+      "loss": 2.0313,
+      "step": 293
+    },
+    {
+      "epoch": 0.3221917808219178,
+      "grad_norm": 2.536095380783081,
+      "learning_rate": 4.987814784727864e-06,
+      "loss": 2.0749,
+      "step": 294
+    },
+    {
+      "epoch": 0.3232876712328767,
+      "grad_norm": 2.3385651111602783,
+      "learning_rate": 4.985783915515841e-06,
+      "loss": 2.0666,
+      "step": 295
+    },
+    {
+      "epoch": 0.32438356164383564,
+      "grad_norm": 2.504826068878174,
+      "learning_rate": 4.983753046303819e-06,
+      "loss": 2.0677,
+      "step": 296
+    },
+    {
+      "epoch": 0.3254794520547945,
+      "grad_norm": 2.5289409160614014,
+      "learning_rate": 4.981722177091796e-06,
+      "loss": 2.0828,
+      "step": 297
+    },
+    {
+      "epoch": 0.3265753424657534,
+      "grad_norm": 2.803795099258423,
+      "learning_rate": 4.979691307879773e-06,
+      "loss": 2.0733,
+      "step": 298
+    },
+    {
+      "epoch": 0.32767123287671235,
+      "grad_norm": 2.651273012161255,
+      "learning_rate": 4.9776604386677504e-06,
+      "loss": 2.0987,
+      "step": 299
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "grad_norm": 2.4715864658355713,
+      "learning_rate": 4.975629569455728e-06,
+      "loss": 2.0733,
+      "step": 300
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "eval_loss": 2.214858293533325,
+      "eval_runtime": 37.0559,
+      "eval_samples_per_second": 26.986,
+      "eval_steps_per_second": 1.133,
+      "step": 300
+    },
+    {
+      "epoch": 0.32986301369863014,
+      "grad_norm": 2.2573442459106445,
+      "learning_rate": 4.973598700243705e-06,
+      "loss": 2.1291,
+      "step": 301
+    },
+    {
+      "epoch": 0.33095890410958906,
+      "grad_norm": 2.645966053009033,
+      "learning_rate": 4.971567831031681e-06,
+      "loss": 2.046,
+      "step": 302
+    },
+    {
+      "epoch": 0.3320547945205479,
+      "grad_norm": 2.361392021179199,
+      "learning_rate": 4.9695369618196595e-06,
+      "loss": 2.0447,
+      "step": 303
+    },
+    {
+      "epoch": 0.33315068493150685,
+      "grad_norm": 2.272714376449585,
+      "learning_rate": 4.967506092607636e-06,
+      "loss": 1.9272,
+      "step": 304
+    },
+    {
+      "epoch": 0.33424657534246577,
+      "grad_norm": 2.5846943855285645,
+      "learning_rate": 4.965475223395614e-06,
+      "loss": 2.0199,
+      "step": 305
+    },
+    {
+      "epoch": 0.33534246575342463,
+      "grad_norm": 2.38985013961792,
+      "learning_rate": 4.9634443541835905e-06,
+      "loss": 2.0394,
+      "step": 306
+    },
+    {
+      "epoch": 0.33643835616438356,
+      "grad_norm": 2.3526251316070557,
+      "learning_rate": 4.961413484971569e-06,
+      "loss": 2.0403,
+      "step": 307
+    },
+    {
+      "epoch": 0.3375342465753425,
+      "grad_norm": 2.399531602859497,
+      "learning_rate": 4.959382615759545e-06,
+      "loss": 2.0607,
+      "step": 308
+    },
+    {
+      "epoch": 0.33863013698630134,
+      "grad_norm": 2.50803804397583,
+      "learning_rate": 4.957351746547523e-06,
+      "loss": 2.0688,
+      "step": 309
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "grad_norm": 2.278088092803955,
+      "learning_rate": 4.9553208773355e-06,
+      "loss": 2.0453,
+      "step": 310
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "eval_loss": 2.2185235023498535,
+      "eval_runtime": 36.9712,
+      "eval_samples_per_second": 27.048,
+      "eval_steps_per_second": 1.136,
+      "step": 310
+    },
+    {
+      "epoch": 0.3408219178082192,
+      "grad_norm": 2.555300712585449,
+      "learning_rate": 4.953290008123478e-06,
+      "loss": 2.1726,
+      "step": 311
+    },
+    {
+      "epoch": 0.3419178082191781,
+      "grad_norm": 2.614870548248291,
+      "learning_rate": 4.951259138911454e-06,
+      "loss": 2.0659,
+      "step": 312
+    },
+    {
+      "epoch": 0.343013698630137,
+      "grad_norm": 2.1932661533355713,
+      "learning_rate": 4.9492282696994315e-06,
+      "loss": 1.9981,
+      "step": 313
+    },
+    {
+      "epoch": 0.3441095890410959,
+      "grad_norm": 2.707451581954956,
+      "learning_rate": 4.947197400487409e-06,
+      "loss": 2.1321,
+      "step": 314
+    },
+    {
+      "epoch": 0.3452054794520548,
+      "grad_norm": 2.4317946434020996,
+      "learning_rate": 4.945166531275386e-06,
+      "loss": 2.0548,
+      "step": 315
+    },
+    {
+      "epoch": 0.3463013698630137,
+      "grad_norm": 2.639883279800415,
+      "learning_rate": 4.943135662063363e-06,
+      "loss": 2.0512,
+      "step": 316
+    },
+    {
+      "epoch": 0.3473972602739726,
+      "grad_norm": 3.1968822479248047,
+      "learning_rate": 4.9411047928513406e-06,
+      "loss": 2.0611,
+      "step": 317
+    },
+    {
+      "epoch": 0.34849315068493153,
+      "grad_norm": 2.4361510276794434,
+      "learning_rate": 4.939073923639318e-06,
+      "loss": 2.0974,
+      "step": 318
+    },
+    {
+      "epoch": 0.3495890410958904,
+      "grad_norm": 2.8278262615203857,
+      "learning_rate": 4.937043054427295e-06,
+      "loss": 2.0992,
+      "step": 319
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "grad_norm": 3.01247239112854,
+      "learning_rate": 4.935012185215272e-06,
+      "loss": 2.0372,
+      "step": 320
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "eval_loss": 2.206442356109619,
+      "eval_runtime": 36.8797,
+      "eval_samples_per_second": 27.115,
+      "eval_steps_per_second": 1.139,
+      "step": 320
+    },
+    {
+      "epoch": 0.35178082191780824,
+      "grad_norm": 2.3372156620025635,
+      "learning_rate": 4.93298131600325e-06,
+      "loss": 2.0211,
+      "step": 321
+    },
+    {
+      "epoch": 0.3528767123287671,
+      "grad_norm": 2.91377329826355,
+      "learning_rate": 4.930950446791227e-06,
+      "loss": 2.0879,
+      "step": 322
+    },
+    {
+      "epoch": 0.353972602739726,
+      "grad_norm": 2.3207666873931885,
+      "learning_rate": 4.928919577579204e-06,
+      "loss": 2.056,
+      "step": 323
+    },
+    {
+      "epoch": 0.35506849315068495,
+      "grad_norm": 2.3260514736175537,
+      "learning_rate": 4.9268887083671815e-06,
+      "loss": 2.0684,
+      "step": 324
+    },
+    {
+      "epoch": 0.3561643835616438,
+      "grad_norm": 2.3070554733276367,
+      "learning_rate": 4.924857839155159e-06,
+      "loss": 2.0494,
+      "step": 325
+    },
+    {
+      "epoch": 0.35726027397260274,
+      "grad_norm": 2.189871072769165,
+      "learning_rate": 4.922826969943136e-06,
+      "loss": 2.0531,
+      "step": 326
+    },
+    {
+      "epoch": 0.35835616438356166,
+      "grad_norm": 2.553520917892456,
+      "learning_rate": 4.920796100731113e-06,
+      "loss": 2.0749,
+      "step": 327
+    },
+    {
+      "epoch": 0.3594520547945205,
+      "grad_norm": 2.62296199798584,
+      "learning_rate": 4.918765231519091e-06,
+      "loss": 1.9983,
+      "step": 328
+    },
+    {
+      "epoch": 0.36054794520547945,
+      "grad_norm": 2.4612369537353516,
+      "learning_rate": 4.916734362307068e-06,
+      "loss": 2.0641,
+      "step": 329
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "grad_norm": 2.656157970428467,
+      "learning_rate": 4.914703493095045e-06,
+      "loss": 2.0148,
+      "step": 330
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "eval_loss": 2.2054214477539062,
+      "eval_runtime": 36.886,
+      "eval_samples_per_second": 27.111,
+      "eval_steps_per_second": 1.139,
+      "step": 330
+    },
+    {
+      "epoch": 0.36273972602739724,
+      "grad_norm": 2.2823736667633057,
+      "learning_rate": 4.912672623883022e-06,
+      "loss": 2.0221,
+      "step": 331
+    },
+    {
+      "epoch": 0.36383561643835616,
+      "grad_norm": 2.8403289318084717,
+      "learning_rate": 4.910641754671e-06,
+      "loss": 2.0149,
+      "step": 332
+    },
+    {
+      "epoch": 0.3649315068493151,
+      "grad_norm": 2.453887462615967,
+      "learning_rate": 4.908610885458977e-06,
+      "loss": 2.0411,
+      "step": 333
+    },
+    {
+      "epoch": 0.36602739726027395,
+      "grad_norm": 2.6094300746917725,
+      "learning_rate": 4.906580016246954e-06,
+      "loss": 2.0293,
+      "step": 334
+    },
+    {
+      "epoch": 0.36712328767123287,
+      "grad_norm": 2.4412384033203125,
+      "learning_rate": 4.9045491470349315e-06,
+      "loss": 2.0778,
+      "step": 335
+    },
+    {
+      "epoch": 0.3682191780821918,
+      "grad_norm": 2.4435057640075684,
+      "learning_rate": 4.902518277822909e-06,
+      "loss": 2.0597,
+      "step": 336
+    },
+    {
+      "epoch": 0.3693150684931507,
+      "grad_norm": 2.2390244007110596,
+      "learning_rate": 4.900487408610886e-06,
+      "loss": 2.0654,
+      "step": 337
+    },
+    {
+      "epoch": 0.3704109589041096,
+      "grad_norm": 2.4243953227996826,
+      "learning_rate": 4.898456539398863e-06,
+      "loss": 2.0696,
+      "step": 338
+    },
+    {
+      "epoch": 0.3715068493150685,
+      "grad_norm": 2.375720262527466,
+      "learning_rate": 4.89642567018684e-06,
+      "loss": 2.0866,
+      "step": 339
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "grad_norm": 2.6379668712615967,
+      "learning_rate": 4.894394800974818e-06,
+      "loss": 2.1552,
+      "step": 340
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "eval_loss": 2.196237802505493,
+      "eval_runtime": 36.8953,
+      "eval_samples_per_second": 27.104,
+      "eval_steps_per_second": 1.138,
+      "step": 340
+    },
+    {
+      "epoch": 0.3736986301369863,
+      "grad_norm": 2.417145252227783,
+      "learning_rate": 4.892363931762794e-06,
+      "loss": 2.0024,
+      "step": 341
+    },
+    {
+      "epoch": 0.3747945205479452,
+      "grad_norm": 2.234590768814087,
+      "learning_rate": 4.8903330625507725e-06,
+      "loss": 2.0837,
+      "step": 342
+    },
+    {
+      "epoch": 0.37589041095890413,
+      "grad_norm": 2.404266595840454,
+      "learning_rate": 4.888302193338749e-06,
+      "loss": 2.018,
+      "step": 343
+    },
+    {
+      "epoch": 0.376986301369863,
+      "grad_norm": 2.364802837371826,
+      "learning_rate": 4.886271324126727e-06,
+      "loss": 1.994,
+      "step": 344
+    },
+    {
+      "epoch": 0.3780821917808219,
+      "grad_norm": 2.2262251377105713,
+      "learning_rate": 4.8842404549147034e-06,
+      "loss": 2.0929,
+      "step": 345
+    },
+    {
+      "epoch": 0.37917808219178084,
+      "grad_norm": 2.4854862689971924,
+      "learning_rate": 4.8822095857026816e-06,
+      "loss": 2.0801,
+      "step": 346
+    },
+    {
+      "epoch": 0.3802739726027397,
+      "grad_norm": 2.3861470222473145,
+      "learning_rate": 4.880178716490658e-06,
+      "loss": 2.0496,
+      "step": 347
+    },
+    {
+      "epoch": 0.38136986301369863,
+      "grad_norm": 2.372326612472534,
+      "learning_rate": 4.878147847278636e-06,
+      "loss": 1.9939,
+      "step": 348
+    },
+    {
+      "epoch": 0.38246575342465755,
+      "grad_norm": 2.3593149185180664,
+      "learning_rate": 4.8761169780666125e-06,
+      "loss": 2.0108,
+      "step": 349
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "grad_norm": 2.1802141666412354,
+      "learning_rate": 4.874086108854591e-06,
+      "loss": 1.983,
+      "step": 350
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "eval_loss": 2.1947567462921143,
+      "eval_runtime": 36.8798,
+      "eval_samples_per_second": 27.115,
+      "eval_steps_per_second": 1.139,
+      "step": 350
+    },
+    {
+      "epoch": 0.38465753424657534,
+      "grad_norm": 2.47067928314209,
+      "learning_rate": 4.872055239642567e-06,
+      "loss": 2.0002,
+      "step": 351
+    },
+    {
+      "epoch": 0.38575342465753426,
+      "grad_norm": 2.359609365463257,
+      "learning_rate": 4.870024370430544e-06,
+      "loss": 1.9799,
+      "step": 352
+    },
+    {
+      "epoch": 0.3868493150684931,
+      "grad_norm": 2.3711702823638916,
+      "learning_rate": 4.867993501218522e-06,
+      "loss": 2.0341,
+      "step": 353
+    },
+    {
+      "epoch": 0.38794520547945205,
+      "grad_norm": 2.246647834777832,
+      "learning_rate": 4.865962632006499e-06,
+      "loss": 2.0687,
+      "step": 354
+    },
+    {
+      "epoch": 0.38904109589041097,
+      "grad_norm": 2.3031938076019287,
+      "learning_rate": 4.863931762794476e-06,
+      "loss": 1.9994,
+      "step": 355
+    },
+    {
+      "epoch": 0.39013698630136984,
+      "grad_norm": 2.492159605026245,
+      "learning_rate": 4.8619008935824535e-06,
+      "loss": 2.0195,
+      "step": 356
+    },
+    {
+      "epoch": 0.39123287671232876,
+      "grad_norm": 2.2878570556640625,
+      "learning_rate": 4.859870024370431e-06,
+      "loss": 2.0589,
+      "step": 357
+    },
+    {
+      "epoch": 0.3923287671232877,
+      "grad_norm": 2.444096088409424,
+      "learning_rate": 4.857839155158408e-06,
+      "loss": 2.0796,
+      "step": 358
+    },
+    {
+      "epoch": 0.3934246575342466,
+      "grad_norm": 2.463857650756836,
+      "learning_rate": 4.855808285946385e-06,
+      "loss": 2.0824,
+      "step": 359
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "grad_norm": 2.0893168449401855,
+      "learning_rate": 4.8537774167343626e-06,
+      "loss": 2.0651,
+      "step": 360
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "eval_loss": 2.188523054122925,
+      "eval_runtime": 36.8679,
+      "eval_samples_per_second": 27.124,
+      "eval_steps_per_second": 1.139,
+      "step": 360
+    },
+    {
+      "epoch": 0.3956164383561644,
+      "grad_norm": 2.602850914001465,
+      "learning_rate": 4.85174654752234e-06,
+      "loss": 2.0507,
+      "step": 361
+    },
+    {
+      "epoch": 0.3967123287671233,
+      "grad_norm": 2.4185140132904053,
+      "learning_rate": 4.849715678310317e-06,
+      "loss": 2.0386,
+      "step": 362
+    },
+    {
+      "epoch": 0.3978082191780822,
+      "grad_norm": 2.1492245197296143,
+      "learning_rate": 4.847684809098294e-06,
+      "loss": 2.0376,
+      "step": 363
+    },
+    {
+      "epoch": 0.3989041095890411,
+      "grad_norm": 2.5710153579711914,
+      "learning_rate": 4.845653939886272e-06,
+      "loss": 2.0383,
+      "step": 364
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 2.4986608028411865,
+      "learning_rate": 4.843623070674249e-06,
+      "loss": 2.0464,
+      "step": 365
+    },
+    {
+      "epoch": 0.4010958904109589,
+      "grad_norm": 2.260388135910034,
+      "learning_rate": 4.841592201462226e-06,
+      "loss": 2.0217,
+      "step": 366
+    },
+    {
+      "epoch": 0.4021917808219178,
+      "grad_norm": 2.6500394344329834,
+      "learning_rate": 4.8395613322502035e-06,
+      "loss": 2.0777,
+      "step": 367
+    },
+    {
+      "epoch": 0.40328767123287673,
+      "grad_norm": 2.159274101257324,
+      "learning_rate": 4.837530463038181e-06,
+      "loss": 1.9747,
+      "step": 368
+    },
+    {
+      "epoch": 0.4043835616438356,
+      "grad_norm": 2.4623043537139893,
+      "learning_rate": 4.835499593826158e-06,
+      "loss": 2.039,
+      "step": 369
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "grad_norm": 2.2265007495880127,
+      "learning_rate": 4.833468724614135e-06,
+      "loss": 2.0618,
+      "step": 370
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "eval_loss": 2.184051275253296,
+      "eval_runtime": 36.9284,
+      "eval_samples_per_second": 27.079,
+      "eval_steps_per_second": 1.137,
+      "step": 370
+    },
+    {
+      "epoch": 0.40657534246575344,
+      "grad_norm": 2.300276517868042,
+      "learning_rate": 4.831437855402113e-06,
+      "loss": 2.0382,
+      "step": 371
+    },
+    {
+      "epoch": 0.4076712328767123,
+      "grad_norm": 2.341156482696533,
+      "learning_rate": 4.82940698619009e-06,
+      "loss": 1.9361,
+      "step": 372
+    },
+    {
+      "epoch": 0.40876712328767123,
+      "grad_norm": 2.4196999073028564,
+      "learning_rate": 4.827376116978067e-06,
+      "loss": 2.0087,
+      "step": 373
+    },
+    {
+      "epoch": 0.40986301369863015,
+      "grad_norm": 2.4650514125823975,
+      "learning_rate": 4.8253452477660444e-06,
+      "loss": 2.0687,
+      "step": 374
+    },
+    {
+      "epoch": 0.410958904109589,
+      "grad_norm": 2.393739700317383,
+      "learning_rate": 4.823314378554022e-06,
+      "loss": 2.0421,
+      "step": 375
+    },
+    {
+      "epoch": 0.41205479452054794,
+      "grad_norm": 2.2436327934265137,
+      "learning_rate": 4.821283509341998e-06,
+      "loss": 2.0358,
+      "step": 376
+    },
+    {
+      "epoch": 0.41315068493150686,
+      "grad_norm": 2.1801650524139404,
+      "learning_rate": 4.819252640129976e-06,
+      "loss": 2.0081,
+      "step": 377
+    },
+    {
+      "epoch": 0.41424657534246573,
+      "grad_norm": 2.300882339477539,
+      "learning_rate": 4.817221770917953e-06,
+      "loss": 2.0321,
+      "step": 378
+    },
+    {
+      "epoch": 0.41534246575342465,
+      "grad_norm": 2.2136058807373047,
+      "learning_rate": 4.815190901705931e-06,
+      "loss": 1.9999,
+      "step": 379
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "grad_norm": 2.168034076690674,
+      "learning_rate": 4.813160032493907e-06,
+      "loss": 2.0336,
+      "step": 380
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "eval_loss": 2.1770179271698,
+      "eval_runtime": 37.0015,
+      "eval_samples_per_second": 27.026,
+      "eval_steps_per_second": 1.135,
+      "step": 380
+    },
+    {
+      "epoch": 0.41753424657534244,
+      "grad_norm": 2.1604163646698,
+      "learning_rate": 4.811129163281885e-06,
+      "loss": 2.0212,
+      "step": 381
+    },
+    {
+      "epoch": 0.41863013698630136,
+      "grad_norm": 2.23923659324646,
+      "learning_rate": 4.809098294069862e-06,
+      "loss": 2.0495,
+      "step": 382
+    },
+    {
+      "epoch": 0.4197260273972603,
+      "grad_norm": 2.149949312210083,
+      "learning_rate": 4.80706742485784e-06,
+      "loss": 2.041,
+      "step": 383
+    },
+    {
+      "epoch": 0.4208219178082192,
+      "grad_norm": 2.49120831489563,
+      "learning_rate": 4.805036555645816e-06,
+      "loss": 2.0931,
+      "step": 384
+    },
+    {
+      "epoch": 0.42191780821917807,
+      "grad_norm": 2.2123563289642334,
+      "learning_rate": 4.8030056864337945e-06,
+      "loss": 2.0087,
+      "step": 385
+    },
+    {
+      "epoch": 0.423013698630137,
+      "grad_norm": 2.4235594272613525,
+      "learning_rate": 4.800974817221771e-06,
+      "loss": 2.0656,
+      "step": 386
+    },
+    {
+      "epoch": 0.4241095890410959,
+      "grad_norm": 2.158252239227295,
+      "learning_rate": 4.798943948009749e-06,
+      "loss": 2.0181,
+      "step": 387
+    },
+    {
+      "epoch": 0.4252054794520548,
+      "grad_norm": 2.2412970066070557,
+      "learning_rate": 4.7969130787977254e-06,
+      "loss": 2.0323,
+      "step": 388
+    },
+    {
+      "epoch": 0.4263013698630137,
+      "grad_norm": 2.2520956993103027,
+      "learning_rate": 4.794882209585703e-06,
+      "loss": 2.026,
+      "step": 389
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "grad_norm": 2.168954849243164,
+      "learning_rate": 4.79285134037368e-06,
+      "loss": 2.0459,
+      "step": 390
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "eval_loss": 2.1705472469329834,
+      "eval_runtime": 37.0913,
+      "eval_samples_per_second": 26.96,
+      "eval_steps_per_second": 1.132,
+      "step": 390
+    },
+    {
+      "epoch": 0.4284931506849315,
+      "grad_norm": 2.1397407054901123,
+      "learning_rate": 4.790820471161657e-06,
+      "loss": 1.9995,
+      "step": 391
+    },
+    {
+      "epoch": 0.4295890410958904,
+      "grad_norm": 2.1747751235961914,
+      "learning_rate": 4.7887896019496345e-06,
+      "loss": 2.0088,
+      "step": 392
+    },
+    {
+      "epoch": 0.43068493150684933,
+      "grad_norm": 2.2860991954803467,
+      "learning_rate": 4.786758732737612e-06,
+      "loss": 2.0134,
+      "step": 393
+    },
+    {
+      "epoch": 0.4317808219178082,
+      "grad_norm": 2.2133519649505615,
+      "learning_rate": 4.784727863525589e-06,
+      "loss": 2.0586,
+      "step": 394
+    },
+    {
+      "epoch": 0.4328767123287671,
+      "grad_norm": 2.2147040367126465,
+      "learning_rate": 4.782696994313566e-06,
+      "loss": 1.9748,
+      "step": 395
+    },
+    {
+      "epoch": 0.43397260273972604,
+      "grad_norm": 2.343418836593628,
+      "learning_rate": 4.780666125101544e-06,
+      "loss": 2.0223,
+      "step": 396
+    },
+    {
+      "epoch": 0.4350684931506849,
+      "grad_norm": 2.198935031890869,
+      "learning_rate": 4.778635255889521e-06,
+      "loss": 1.9627,
+      "step": 397
+    },
+    {
+      "epoch": 0.43616438356164383,
+      "grad_norm": 2.1314704418182373,
+      "learning_rate": 4.776604386677498e-06,
+      "loss": 1.9938,
+      "step": 398
+    },
+    {
+      "epoch": 0.43726027397260275,
+      "grad_norm": 2.3005123138427734,
+      "learning_rate": 4.7745735174654755e-06,
+      "loss": 2.0029,
+      "step": 399
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "grad_norm": 2.2826664447784424,
+      "learning_rate": 4.772542648253453e-06,
+      "loss": 2.0585,
+      "step": 400
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "eval_loss": 2.1691415309906006,
+      "eval_runtime": 37.0852,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 400
+    },
+    {
+      "epoch": 0.43945205479452054,
+      "grad_norm": 2.0547878742218018,
+      "learning_rate": 4.77051177904143e-06,
+      "loss": 2.0582,
+      "step": 401
+    },
+    {
+      "epoch": 0.44054794520547946,
+      "grad_norm": 2.200042486190796,
+      "learning_rate": 4.768480909829407e-06,
+      "loss": 1.9936,
+      "step": 402
+    },
+    {
+      "epoch": 0.44164383561643833,
+      "grad_norm": 2.354637861251831,
+      "learning_rate": 4.766450040617385e-06,
+      "loss": 2.0327,
+      "step": 403
+    },
+    {
+      "epoch": 0.44273972602739725,
+      "grad_norm": 2.2332348823547363,
+      "learning_rate": 4.764419171405362e-06,
+      "loss": 2.0052,
+      "step": 404
+    },
+    {
+      "epoch": 0.4438356164383562,
+      "grad_norm": 2.288060426712036,
+      "learning_rate": 4.762388302193339e-06,
+      "loss": 2.0472,
+      "step": 405
+    },
+    {
+      "epoch": 0.4449315068493151,
+      "grad_norm": 2.1368887424468994,
+      "learning_rate": 4.760357432981316e-06,
+      "loss": 2.0305,
+      "step": 406
+    },
+    {
+      "epoch": 0.44602739726027396,
+      "grad_norm": 2.362630844116211,
+      "learning_rate": 4.758326563769294e-06,
+      "loss": 2.002,
+      "step": 407
+    },
+    {
+      "epoch": 0.4471232876712329,
+      "grad_norm": 2.283419370651245,
+      "learning_rate": 4.756295694557271e-06,
+      "loss": 2.0271,
+      "step": 408
+    },
+    {
+      "epoch": 0.4482191780821918,
+      "grad_norm": 2.0400819778442383,
+      "learning_rate": 4.754264825345248e-06,
+      "loss": 2.0256,
+      "step": 409
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "grad_norm": 2.588416814804077,
+      "learning_rate": 4.7522339561332255e-06,
+      "loss": 1.9756,
+      "step": 410
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "eval_loss": 2.17041015625,
+      "eval_runtime": 36.958,
+      "eval_samples_per_second": 27.058,
+      "eval_steps_per_second": 1.136,
+      "step": 410
+    },
+    {
+      "epoch": 0.4504109589041096,
+      "grad_norm": 2.2397494316101074,
+      "learning_rate": 4.750203086921203e-06,
+      "loss": 1.9876,
+      "step": 411
+    },
+    {
+      "epoch": 0.4515068493150685,
+      "grad_norm": 2.2085509300231934,
+      "learning_rate": 4.74817221770918e-06,
+      "loss": 1.9464,
+      "step": 412
+    },
+    {
+      "epoch": 0.4526027397260274,
+      "grad_norm": 2.140974760055542,
+      "learning_rate": 4.7461413484971565e-06,
+      "loss": 2.0179,
+      "step": 413
+    },
+    {
+      "epoch": 0.4536986301369863,
+      "grad_norm": 2.124366044998169,
+      "learning_rate": 4.744110479285135e-06,
+      "loss": 2.0732,
+      "step": 414
+    },
+    {
+      "epoch": 0.4547945205479452,
+      "grad_norm": 2.2463996410369873,
+      "learning_rate": 4.742079610073111e-06,
+      "loss": 1.9616,
+      "step": 415
+    },
+    {
+      "epoch": 0.4558904109589041,
+      "grad_norm": 2.1249961853027344,
+      "learning_rate": 4.740048740861089e-06,
+      "loss": 1.9889,
+      "step": 416
+    },
+    {
+      "epoch": 0.456986301369863,
+      "grad_norm": 2.25842022895813,
+      "learning_rate": 4.738017871649066e-06,
+      "loss": 2.0033,
+      "step": 417
+    },
+    {
+      "epoch": 0.45808219178082193,
+      "grad_norm": 2.271935224533081,
+      "learning_rate": 4.735987002437044e-06,
+      "loss": 2.0148,
+      "step": 418
+    },
+    {
+      "epoch": 0.4591780821917808,
+      "grad_norm": 2.4797163009643555,
+      "learning_rate": 4.73395613322502e-06,
+      "loss": 2.0736,
+      "step": 419
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "grad_norm": 2.1901395320892334,
+      "learning_rate": 4.731925264012998e-06,
+      "loss": 2.0763,
+      "step": 420
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "eval_loss": 2.161660671234131,
+      "eval_runtime": 36.9099,
+      "eval_samples_per_second": 27.093,
+      "eval_steps_per_second": 1.138,
+      "step": 420
+    },
+    {
+      "epoch": 0.46136986301369864,
+      "grad_norm": 2.083536386489868,
+      "learning_rate": 4.729894394800975e-06,
+      "loss": 2.0299,
+      "step": 421
+    },
+    {
+      "epoch": 0.4624657534246575,
+      "grad_norm": 2.1118173599243164,
+      "learning_rate": 4.727863525588953e-06,
+      "loss": 1.9742,
+      "step": 422
+    },
+    {
+      "epoch": 0.46356164383561643,
+      "grad_norm": 2.195802927017212,
+      "learning_rate": 4.725832656376929e-06,
+      "loss": 1.9318,
+      "step": 423
+    },
+    {
+      "epoch": 0.46465753424657535,
+      "grad_norm": 2.318098545074463,
+      "learning_rate": 4.723801787164907e-06,
+      "loss": 1.9875,
+      "step": 424
+    },
+    {
+      "epoch": 0.4657534246575342,
+      "grad_norm": 2.1698496341705322,
+      "learning_rate": 4.721770917952885e-06,
+      "loss": 2.031,
+      "step": 425
+    },
+    {
+      "epoch": 0.46684931506849314,
+      "grad_norm": 2.283935785293579,
+      "learning_rate": 4.719740048740861e-06,
+      "loss": 1.9766,
+      "step": 426
+    },
+    {
+      "epoch": 0.46794520547945206,
+      "grad_norm": 2.4180757999420166,
+      "learning_rate": 4.717709179528839e-06,
+      "loss": 1.9976,
+      "step": 427
+    },
+    {
+      "epoch": 0.469041095890411,
+      "grad_norm": 2.2689030170440674,
+      "learning_rate": 4.715678310316816e-06,
+      "loss": 2.03,
+      "step": 428
+    },
+    {
+      "epoch": 0.47013698630136985,
+      "grad_norm": 2.423708200454712,
+      "learning_rate": 4.713647441104794e-06,
+      "loss": 2.0202,
+      "step": 429
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "grad_norm": 2.184705972671509,
+      "learning_rate": 4.71161657189277e-06,
+      "loss": 1.9655,
+      "step": 430
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "eval_loss": 2.1548638343811035,
+      "eval_runtime": 36.9426,
+      "eval_samples_per_second": 27.069,
+      "eval_steps_per_second": 1.137,
+      "step": 430
+    },
+    {
+      "epoch": 0.4723287671232877,
+      "grad_norm": 2.295851945877075,
+      "learning_rate": 4.709585702680748e-06,
+      "loss": 1.9561,
+      "step": 431
+    },
+    {
+      "epoch": 0.47342465753424656,
+      "grad_norm": 2.1762046813964844,
+      "learning_rate": 4.707554833468725e-06,
+      "loss": 2.0341,
+      "step": 432
+    },
+    {
+      "epoch": 0.4745205479452055,
+      "grad_norm": 2.4684343338012695,
+      "learning_rate": 4.705523964256703e-06,
+      "loss": 2.0898,
+      "step": 433
+    },
+    {
+      "epoch": 0.4756164383561644,
+      "grad_norm": 2.293318271636963,
+      "learning_rate": 4.703493095044679e-06,
+      "loss": 1.9794,
+      "step": 434
+    },
+    {
+      "epoch": 0.4767123287671233,
+      "grad_norm": 2.1884210109710693,
+      "learning_rate": 4.701462225832657e-06,
+      "loss": 1.9979,
+      "step": 435
+    },
+    {
+      "epoch": 0.4778082191780822,
+      "grad_norm": 2.412504196166992,
+      "learning_rate": 4.699431356620634e-06,
+      "loss": 1.9237,
+      "step": 436
+    },
+    {
+      "epoch": 0.4789041095890411,
+      "grad_norm": 2.192824602127075,
+      "learning_rate": 4.697400487408612e-06,
+      "loss": 2.0756,
+      "step": 437
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 2.368100643157959,
+      "learning_rate": 4.695369618196588e-06,
+      "loss": 2.028,
+      "step": 438
+    },
+    {
+      "epoch": 0.4810958904109589,
+      "grad_norm": 2.218454122543335,
+      "learning_rate": 4.693338748984566e-06,
+      "loss": 2.0049,
+      "step": 439
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "grad_norm": 2.356168031692505,
+      "learning_rate": 4.691307879772543e-06,
+      "loss": 1.9958,
+      "step": 440
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "eval_loss": 2.1540727615356445,
+      "eval_runtime": 36.9557,
+      "eval_samples_per_second": 27.059,
+      "eval_steps_per_second": 1.136,
+      "step": 440
+    },
+    {
+      "epoch": 0.4832876712328767,
+      "grad_norm": 2.09956693649292,
+      "learning_rate": 4.68927701056052e-06,
+      "loss": 1.9388,
+      "step": 441
+    },
+    {
+      "epoch": 0.4843835616438356,
+      "grad_norm": 2.5746614933013916,
+      "learning_rate": 4.6872461413484975e-06,
+      "loss": 2.0425,
+      "step": 442
+    },
+    {
+      "epoch": 0.48547945205479454,
+      "grad_norm": 2.2064783573150635,
+      "learning_rate": 4.685215272136475e-06,
+      "loss": 2.0082,
+      "step": 443
+    },
+    {
+      "epoch": 0.4865753424657534,
+      "grad_norm": 2.159029006958008,
+      "learning_rate": 4.683184402924452e-06,
+      "loss": 2.0055,
+      "step": 444
+    },
+    {
+      "epoch": 0.4876712328767123,
+      "grad_norm": 2.5222008228302,
+      "learning_rate": 4.681153533712429e-06,
+      "loss": 2.0007,
+      "step": 445
+    },
+    {
+      "epoch": 0.48876712328767125,
+      "grad_norm": 2.163487672805786,
+      "learning_rate": 4.679122664500407e-06,
+      "loss": 2.0719,
+      "step": 446
+    },
+    {
+      "epoch": 0.4898630136986301,
+      "grad_norm": 2.3383803367614746,
+      "learning_rate": 4.677091795288384e-06,
+      "loss": 1.9623,
+      "step": 447
+    },
+    {
+      "epoch": 0.49095890410958903,
+      "grad_norm": 2.1056675910949707,
+      "learning_rate": 4.675060926076361e-06,
+      "loss": 1.9443,
+      "step": 448
+    },
+    {
+      "epoch": 0.49205479452054796,
+      "grad_norm": 2.145270824432373,
+      "learning_rate": 4.673030056864338e-06,
+      "loss": 2.0469,
+      "step": 449
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "grad_norm": 2.578706741333008,
+      "learning_rate": 4.670999187652316e-06,
+      "loss": 1.9569,
+      "step": 450
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "eval_loss": 2.1527743339538574,
+      "eval_runtime": 36.8961,
+      "eval_samples_per_second": 27.103,
+      "eval_steps_per_second": 1.138,
+      "step": 450
+    },
+    {
+      "epoch": 0.49424657534246574,
+      "grad_norm": 2.1328542232513428,
+      "learning_rate": 4.668968318440293e-06,
+      "loss": 1.9559,
+      "step": 451
+    },
+    {
+      "epoch": 0.49534246575342467,
+      "grad_norm": 2.148820638656616,
+      "learning_rate": 4.66693744922827e-06,
+      "loss": 1.9545,
+      "step": 452
+    },
+    {
+      "epoch": 0.4964383561643836,
+      "grad_norm": 2.250340461730957,
+      "learning_rate": 4.6649065800162475e-06,
+      "loss": 2.0358,
+      "step": 453
+    },
+    {
+      "epoch": 0.49753424657534245,
+      "grad_norm": 2.452397584915161,
+      "learning_rate": 4.662875710804225e-06,
+      "loss": 1.9294,
+      "step": 454
+    },
+    {
+      "epoch": 0.4986301369863014,
+      "grad_norm": 2.140615701675415,
+      "learning_rate": 4.660844841592202e-06,
+      "loss": 2.0084,
+      "step": 455
+    },
+    {
+      "epoch": 0.4997260273972603,
+      "grad_norm": 2.176806926727295,
+      "learning_rate": 4.658813972380179e-06,
+      "loss": 1.9174,
+      "step": 456
+    },
+    {
+      "epoch": 0.5008219178082192,
+      "grad_norm": 2.349031925201416,
+      "learning_rate": 4.656783103168157e-06,
+      "loss": 2.0377,
+      "step": 457
+    },
+    {
+      "epoch": 0.5019178082191781,
+      "grad_norm": 2.436889171600342,
+      "learning_rate": 4.654752233956134e-06,
+      "loss": 1.9133,
+      "step": 458
+    },
+    {
+      "epoch": 0.503013698630137,
+      "grad_norm": 2.116610288619995,
+      "learning_rate": 4.652721364744111e-06,
+      "loss": 1.9993,
+      "step": 459
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "grad_norm": 2.186915636062622,
+      "learning_rate": 4.6506904955320885e-06,
+      "loss": 1.9744,
+      "step": 460
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "eval_loss": 2.1580278873443604,
+      "eval_runtime": 36.9097,
+      "eval_samples_per_second": 27.093,
+      "eval_steps_per_second": 1.138,
+      "step": 460
+    },
+    {
+      "epoch": 0.5052054794520547,
+      "grad_norm": 2.2009212970733643,
+      "learning_rate": 4.648659626320066e-06,
+      "loss": 2.0173,
+      "step": 461
+    },
+    {
+      "epoch": 0.5063013698630137,
+      "grad_norm": 2.117286443710327,
+      "learning_rate": 4.646628757108043e-06,
+      "loss": 2.0238,
+      "step": 462
+    },
+    {
+      "epoch": 0.5073972602739726,
+      "grad_norm": 2.238790273666382,
+      "learning_rate": 4.6445978878960194e-06,
+      "loss": 1.9885,
+      "step": 463
+    },
+    {
+      "epoch": 0.5084931506849315,
+      "grad_norm": 2.3493096828460693,
+      "learning_rate": 4.6425670186839976e-06,
+      "loss": 1.9593,
+      "step": 464
+    },
+    {
+      "epoch": 0.5095890410958904,
+      "grad_norm": 2.3276610374450684,
+      "learning_rate": 4.640536149471974e-06,
+      "loss": 1.9993,
+      "step": 465
+    },
+    {
+      "epoch": 0.5106849315068493,
+      "grad_norm": 2.3671715259552,
+      "learning_rate": 4.638505280259952e-06,
+      "loss": 2.0256,
+      "step": 466
+    },
+    {
+      "epoch": 0.5117808219178083,
+      "grad_norm": 2.251803159713745,
+      "learning_rate": 4.6364744110479285e-06,
+      "loss": 2.0436,
+      "step": 467
+    },
+    {
+      "epoch": 0.5128767123287671,
+      "grad_norm": 2.3023312091827393,
+      "learning_rate": 4.634443541835907e-06,
+      "loss": 1.9968,
+      "step": 468
+    },
+    {
+      "epoch": 0.513972602739726,
+      "grad_norm": 2.3660333156585693,
+      "learning_rate": 4.632412672623883e-06,
+      "loss": 2.0492,
+      "step": 469
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "grad_norm": 2.262179136276245,
+      "learning_rate": 4.630381803411861e-06,
+      "loss": 1.9698,
+      "step": 470
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "eval_loss": 2.1516528129577637,
+      "eval_runtime": 36.7956,
+      "eval_samples_per_second": 27.177,
+      "eval_steps_per_second": 1.141,
+      "step": 470
+    },
+    {
+      "epoch": 0.5161643835616438,
+      "grad_norm": 2.450770616531372,
+      "learning_rate": 4.628350934199838e-06,
+      "loss": 1.8799,
+      "step": 471
+    },
+    {
+      "epoch": 0.5172602739726028,
+      "grad_norm": 2.187458038330078,
+      "learning_rate": 4.626320064987816e-06,
+      "loss": 1.9829,
+      "step": 472
+    },
+    {
+      "epoch": 0.5183561643835617,
+      "grad_norm": 2.1726722717285156,
+      "learning_rate": 4.624289195775792e-06,
+      "loss": 2.012,
+      "step": 473
+    },
+    {
+      "epoch": 0.5194520547945205,
+      "grad_norm": 2.4177136421203613,
+      "learning_rate": 4.62225832656377e-06,
+      "loss": 1.9796,
+      "step": 474
+    },
+    {
+      "epoch": 0.5205479452054794,
+      "grad_norm": 2.2353951930999756,
+      "learning_rate": 4.620227457351747e-06,
+      "loss": 1.99,
+      "step": 475
+    },
+    {
+      "epoch": 0.5216438356164383,
+      "grad_norm": 2.369591236114502,
+      "learning_rate": 4.618196588139724e-06,
+      "loss": 1.9847,
+      "step": 476
+    },
+    {
+      "epoch": 0.5227397260273973,
+      "grad_norm": 2.546703577041626,
+      "learning_rate": 4.616165718927701e-06,
+      "loss": 2.0047,
+      "step": 477
+    },
+    {
+      "epoch": 0.5238356164383562,
+      "grad_norm": 2.3509185314178467,
+      "learning_rate": 4.6141348497156786e-06,
+      "loss": 1.9589,
+      "step": 478
+    },
+    {
+      "epoch": 0.5249315068493151,
+      "grad_norm": 2.3063488006591797,
+      "learning_rate": 4.612103980503656e-06,
+      "loss": 2.045,
+      "step": 479
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "grad_norm": 2.285572052001953,
+      "learning_rate": 4.610073111291633e-06,
+      "loss": 2.0218,
+      "step": 480
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "eval_loss": 2.1472301483154297,
+      "eval_runtime": 36.7689,
+      "eval_samples_per_second": 27.197,
+      "eval_steps_per_second": 1.142,
+      "step": 480
+    },
+    {
+      "epoch": 0.5271232876712328,
+      "grad_norm": 2.22987961769104,
+      "learning_rate": 4.60804224207961e-06,
+      "loss": 1.9955,
+      "step": 481
+    },
+    {
+      "epoch": 0.5282191780821918,
+      "grad_norm": 2.200918197631836,
+      "learning_rate": 4.606011372867588e-06,
+      "loss": 1.9768,
+      "step": 482
+    },
+    {
+      "epoch": 0.5293150684931507,
+      "grad_norm": 2.2441930770874023,
+      "learning_rate": 4.603980503655565e-06,
+      "loss": 2.0088,
+      "step": 483
+    },
+    {
+      "epoch": 0.5304109589041096,
+      "grad_norm": 2.380383014678955,
+      "learning_rate": 4.601949634443542e-06,
+      "loss": 1.9853,
+      "step": 484
+    },
+    {
+      "epoch": 0.5315068493150685,
+      "grad_norm": 2.317995071411133,
+      "learning_rate": 4.5999187652315195e-06,
+      "loss": 1.9183,
+      "step": 485
+    },
+    {
+      "epoch": 0.5326027397260275,
+      "grad_norm": 2.2462668418884277,
+      "learning_rate": 4.597887896019497e-06,
+      "loss": 1.9201,
+      "step": 486
+    },
+    {
+      "epoch": 0.5336986301369863,
+      "grad_norm": 2.1482582092285156,
+      "learning_rate": 4.595857026807474e-06,
+      "loss": 1.9722,
+      "step": 487
+    },
+    {
+      "epoch": 0.5347945205479452,
+      "grad_norm": 2.079831123352051,
+      "learning_rate": 4.593826157595451e-06,
+      "loss": 1.977,
+      "step": 488
+    },
+    {
+      "epoch": 0.5358904109589041,
+      "grad_norm": 2.2169578075408936,
+      "learning_rate": 4.591795288383429e-06,
+      "loss": 1.9615,
+      "step": 489
+    },
+    {
+      "epoch": 0.536986301369863,
+      "grad_norm": 2.087846279144287,
+      "learning_rate": 4.589764419171406e-06,
+      "loss": 2.0402,
+      "step": 490
+    },
+    {
+      "epoch": 0.536986301369863,
+      "eval_loss": 2.1416702270507812,
+      "eval_runtime": 37.1191,
+      "eval_samples_per_second": 26.94,
+      "eval_steps_per_second": 1.131,
+      "step": 490
+    },
+    {
+      "epoch": 0.538082191780822,
+      "grad_norm": 2.2161977291107178,
+      "learning_rate": 4.587733549959383e-06,
+      "loss": 1.9959,
+      "step": 491
+    },
+    {
+      "epoch": 0.5391780821917809,
+      "grad_norm": 2.19530987739563,
+      "learning_rate": 4.5857026807473604e-06,
+      "loss": 1.961,
+      "step": 492
+    },
+    {
+      "epoch": 0.5402739726027397,
+      "grad_norm": 2.1150615215301514,
+      "learning_rate": 4.583671811535338e-06,
+      "loss": 1.946,
+      "step": 493
+    },
+    {
+      "epoch": 0.5413698630136986,
+      "grad_norm": 2.186352491378784,
+      "learning_rate": 4.581640942323315e-06,
+      "loss": 1.9137,
+      "step": 494
+    },
+    {
+      "epoch": 0.5424657534246575,
+      "grad_norm": 2.233370780944824,
+      "learning_rate": 4.579610073111292e-06,
+      "loss": 2.021,
+      "step": 495
+    },
+    {
+      "epoch": 0.5435616438356164,
+      "grad_norm": 2.1326382160186768,
+      "learning_rate": 4.5775792038992695e-06,
+      "loss": 2.0242,
+      "step": 496
+    },
+    {
+      "epoch": 0.5446575342465754,
+      "grad_norm": 2.536837577819824,
+      "learning_rate": 4.575548334687247e-06,
+      "loss": 2.0185,
+      "step": 497
+    },
+    {
+      "epoch": 0.5457534246575343,
+      "grad_norm": 2.193598747253418,
+      "learning_rate": 4.573517465475224e-06,
+      "loss": 2.0533,
+      "step": 498
+    },
+    {
+      "epoch": 0.5468493150684931,
+      "grad_norm": 2.1346311569213867,
+      "learning_rate": 4.571486596263201e-06,
+      "loss": 1.9564,
+      "step": 499
+    },
+    {
+      "epoch": 0.547945205479452,
+      "grad_norm": 2.4294722080230713,
+      "learning_rate": 4.569455727051178e-06,
+      "loss": 2.0003,
+      "step": 500
+    },
+    {
+      "epoch": 0.547945205479452,
+      "eval_loss": 2.1430764198303223,
+      "eval_runtime": 37.0591,
+      "eval_samples_per_second": 26.984,
+      "eval_steps_per_second": 1.133,
+      "step": 500
+    },
+    {
+      "epoch": 0.549041095890411,
+      "grad_norm": 2.2394182682037354,
+      "learning_rate": 4.567424857839156e-06,
+      "loss": 2.0055,
+      "step": 501
+    },
+    {
+      "epoch": 0.5501369863013699,
+      "grad_norm": 2.2543067932128906,
+      "learning_rate": 4.565393988627132e-06,
+      "loss": 1.9837,
+      "step": 502
+    },
+    {
+      "epoch": 0.5512328767123288,
+      "grad_norm": 2.2805464267730713,
+      "learning_rate": 4.5633631194151105e-06,
+      "loss": 2.0495,
+      "step": 503
+    },
+    {
+      "epoch": 0.5523287671232877,
+      "grad_norm": 2.2838146686553955,
+      "learning_rate": 4.561332250203087e-06,
+      "loss": 1.9866,
+      "step": 504
+    },
+    {
+      "epoch": 0.5534246575342465,
+      "grad_norm": 2.286147356033325,
+      "learning_rate": 4.559301380991065e-06,
+      "loss": 1.9805,
+      "step": 505
+    },
+    {
+      "epoch": 0.5545205479452054,
+      "grad_norm": 2.064378023147583,
+      "learning_rate": 4.5572705117790414e-06,
+      "loss": 1.9259,
+      "step": 506
+    },
+    {
+      "epoch": 0.5556164383561644,
+      "grad_norm": 2.4020493030548096,
+      "learning_rate": 4.5552396425670196e-06,
+      "loss": 1.9731,
+      "step": 507
+    },
+    {
+      "epoch": 0.5567123287671233,
+      "grad_norm": 2.11594557762146,
+      "learning_rate": 4.553208773354996e-06,
+      "loss": 2.0126,
+      "step": 508
+    },
+    {
+      "epoch": 0.5578082191780822,
+      "grad_norm": 2.4520673751831055,
+      "learning_rate": 4.551177904142974e-06,
+      "loss": 1.9396,
+      "step": 509
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "grad_norm": 2.302771806716919,
+      "learning_rate": 4.5491470349309505e-06,
+      "loss": 2.0361,
+      "step": 510
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "eval_loss": 2.1408755779266357,
+      "eval_runtime": 36.9881,
+      "eval_samples_per_second": 27.036,
+      "eval_steps_per_second": 1.136,
+      "step": 510
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 2.250927686691284,
+      "learning_rate": 4.547116165718929e-06,
+      "loss": 1.9616,
+      "step": 511
+    },
+    {
+      "epoch": 0.5610958904109589,
+      "grad_norm": 2.4257938861846924,
+      "learning_rate": 4.545085296506905e-06,
+      "loss": 1.9836,
+      "step": 512
+    },
+    {
+      "epoch": 0.5621917808219178,
+      "grad_norm": 2.08693265914917,
+      "learning_rate": 4.543054427294882e-06,
+      "loss": 1.9541,
+      "step": 513
+    },
+    {
+      "epoch": 0.5632876712328767,
+      "grad_norm": 2.133875846862793,
+      "learning_rate": 4.54102355808286e-06,
+      "loss": 1.9377,
+      "step": 514
+    },
+    {
+      "epoch": 0.5643835616438356,
+      "grad_norm": 2.1217594146728516,
+      "learning_rate": 4.538992688870837e-06,
+      "loss": 1.9369,
+      "step": 515
+    },
+    {
+      "epoch": 0.5654794520547946,
+      "grad_norm": 2.117994546890259,
+      "learning_rate": 4.536961819658814e-06,
+      "loss": 1.9666,
+      "step": 516
+    },
+    {
+      "epoch": 0.5665753424657535,
+      "grad_norm": 2.1514523029327393,
+      "learning_rate": 4.5349309504467915e-06,
+      "loss": 1.9704,
+      "step": 517
+    },
+    {
+      "epoch": 0.5676712328767123,
+      "grad_norm": 2.2309377193450928,
+      "learning_rate": 4.532900081234769e-06,
+      "loss": 1.9858,
+      "step": 518
+    },
+    {
+      "epoch": 0.5687671232876712,
+      "grad_norm": 2.100398540496826,
+      "learning_rate": 4.530869212022746e-06,
+      "loss": 1.9573,
+      "step": 519
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "grad_norm": 2.122840166091919,
+      "learning_rate": 4.528838342810723e-06,
+      "loss": 1.9813,
+      "step": 520
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "eval_loss": 2.135481357574463,
+      "eval_runtime": 36.893,
+      "eval_samples_per_second": 27.105,
+      "eval_steps_per_second": 1.138,
+      "step": 520
+    },
+    {
+      "epoch": 0.570958904109589,
+      "grad_norm": 2.1103999614715576,
+      "learning_rate": 4.526807473598701e-06,
+      "loss": 1.9691,
+      "step": 521
+    },
+    {
+      "epoch": 0.572054794520548,
+      "grad_norm": 2.059232473373413,
+      "learning_rate": 4.524776604386678e-06,
+      "loss": 1.9687,
+      "step": 522
+    },
+    {
+      "epoch": 0.5731506849315069,
+      "grad_norm": 2.207000732421875,
+      "learning_rate": 4.522745735174655e-06,
+      "loss": 1.949,
+      "step": 523
+    },
+    {
+      "epoch": 0.5742465753424657,
+      "grad_norm": 2.264000177383423,
+      "learning_rate": 4.520714865962632e-06,
+      "loss": 2.0066,
+      "step": 524
+    },
+    {
+      "epoch": 0.5753424657534246,
+      "grad_norm": 2.121701717376709,
+      "learning_rate": 4.51868399675061e-06,
+      "loss": 1.9879,
+      "step": 525
+    },
+    {
+      "epoch": 0.5764383561643835,
+      "grad_norm": 2.0745644569396973,
+      "learning_rate": 4.516653127538587e-06,
+      "loss": 1.9924,
+      "step": 526
+    },
+    {
+      "epoch": 0.5775342465753425,
+      "grad_norm": 2.2257349491119385,
+      "learning_rate": 4.514622258326564e-06,
+      "loss": 1.9525,
+      "step": 527
+    },
+    {
+      "epoch": 0.5786301369863014,
+      "grad_norm": 2.1678507328033447,
+      "learning_rate": 4.5125913891145415e-06,
+      "loss": 1.9378,
+      "step": 528
+    },
+    {
+      "epoch": 0.5797260273972603,
+      "grad_norm": 2.2800841331481934,
+      "learning_rate": 4.510560519902519e-06,
+      "loss": 1.9685,
+      "step": 529
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "grad_norm": 2.1471869945526123,
+      "learning_rate": 4.508529650690496e-06,
+      "loss": 2.0027,
+      "step": 530
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "eval_loss": 2.1354892253875732,
+      "eval_runtime": 36.8434,
+      "eval_samples_per_second": 27.142,
+      "eval_steps_per_second": 1.14,
+      "step": 530
+    },
+    {
+      "epoch": 0.581917808219178,
+      "grad_norm": 2.2238621711730957,
+      "learning_rate": 4.506498781478473e-06,
+      "loss": 1.9258,
+      "step": 531
+    },
+    {
+      "epoch": 0.583013698630137,
+      "grad_norm": 2.2224624156951904,
+      "learning_rate": 4.504467912266451e-06,
+      "loss": 1.9168,
+      "step": 532
+    },
+    {
+      "epoch": 0.5841095890410959,
+      "grad_norm": 1.9850294589996338,
+      "learning_rate": 4.502437043054428e-06,
+      "loss": 1.9656,
+      "step": 533
+    },
+    {
+      "epoch": 0.5852054794520548,
+      "grad_norm": 2.1303517818450928,
+      "learning_rate": 4.500406173842405e-06,
+      "loss": 2.0349,
+      "step": 534
+    },
+    {
+      "epoch": 0.5863013698630137,
+      "grad_norm": 2.2114832401275635,
+      "learning_rate": 4.4983753046303824e-06,
+      "loss": 1.9878,
+      "step": 535
+    },
+    {
+      "epoch": 0.5873972602739727,
+      "grad_norm": 1.9375886917114258,
+      "learning_rate": 4.49634443541836e-06,
+      "loss": 1.9724,
+      "step": 536
+    },
+    {
+      "epoch": 0.5884931506849315,
+      "grad_norm": 2.185518741607666,
+      "learning_rate": 4.494313566206336e-06,
+      "loss": 1.9584,
+      "step": 537
+    },
+    {
+      "epoch": 0.5895890410958904,
+      "grad_norm": 2.125638484954834,
+      "learning_rate": 4.492282696994314e-06,
+      "loss": 2.0203,
+      "step": 538
+    },
+    {
+      "epoch": 0.5906849315068493,
+      "grad_norm": 1.9456262588500977,
+      "learning_rate": 4.490251827782291e-06,
+      "loss": 1.9523,
+      "step": 539
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "grad_norm": 2.088125705718994,
+      "learning_rate": 4.488220958570269e-06,
+      "loss": 1.9271,
+      "step": 540
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "eval_loss": 2.134958505630493,
+      "eval_runtime": 36.8342,
+      "eval_samples_per_second": 27.149,
+      "eval_steps_per_second": 1.14,
+      "step": 540
+    },
+    {
+      "epoch": 0.5928767123287672,
+      "grad_norm": 2.1423559188842773,
+      "learning_rate": 4.486190089358245e-06,
+      "loss": 1.9467,
+      "step": 541
+    },
+    {
+      "epoch": 0.5939726027397261,
+      "grad_norm": 2.2169737815856934,
+      "learning_rate": 4.484159220146223e-06,
+      "loss": 1.8901,
+      "step": 542
+    },
+    {
+      "epoch": 0.5950684931506849,
+      "grad_norm": 2.419947385787964,
+      "learning_rate": 4.4821283509342e-06,
+      "loss": 2.0064,
+      "step": 543
+    },
+    {
+      "epoch": 0.5961643835616438,
+      "grad_norm": 2.0396435260772705,
+      "learning_rate": 4.480097481722178e-06,
+      "loss": 1.9819,
+      "step": 544
+    },
+    {
+      "epoch": 0.5972602739726027,
+      "grad_norm": 2.210700511932373,
+      "learning_rate": 4.478066612510154e-06,
+      "loss": 1.9578,
+      "step": 545
+    },
+    {
+      "epoch": 0.5983561643835617,
+      "grad_norm": 2.159698009490967,
+      "learning_rate": 4.4760357432981325e-06,
+      "loss": 1.9313,
+      "step": 546
+    },
+    {
+      "epoch": 0.5994520547945206,
+      "grad_norm": 2.125058650970459,
+      "learning_rate": 4.474004874086109e-06,
+      "loss": 1.9549,
+      "step": 547
+    },
+    {
+      "epoch": 0.6005479452054795,
+      "grad_norm": 2.477320432662964,
+      "learning_rate": 4.471974004874087e-06,
+      "loss": 2.0883,
+      "step": 548
+    },
+    {
+      "epoch": 0.6016438356164383,
+      "grad_norm": 2.214901924133301,
+      "learning_rate": 4.4699431356620635e-06,
+      "loss": 1.9184,
+      "step": 549
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "grad_norm": 2.1001312732696533,
+      "learning_rate": 4.467912266450041e-06,
+      "loss": 2.0231,
+      "step": 550
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "eval_loss": 2.1315157413482666,
+      "eval_runtime": 36.8131,
+      "eval_samples_per_second": 27.164,
+      "eval_steps_per_second": 1.141,
+      "step": 550
+    },
+    {
+      "epoch": 0.6038356164383561,
+      "grad_norm": 2.1030538082122803,
+      "learning_rate": 4.465881397238018e-06,
+      "loss": 1.967,
+      "step": 551
+    },
+    {
+      "epoch": 0.6049315068493151,
+      "grad_norm": 2.0390465259552,
+      "learning_rate": 4.463850528025995e-06,
+      "loss": 1.9615,
+      "step": 552
+    },
+    {
+      "epoch": 0.606027397260274,
+      "grad_norm": 2.1331627368927,
+      "learning_rate": 4.4618196588139726e-06,
+      "loss": 1.9667,
+      "step": 553
+    },
+    {
+      "epoch": 0.6071232876712329,
+      "grad_norm": 2.1987671852111816,
+      "learning_rate": 4.45978878960195e-06,
+      "loss": 1.9922,
+      "step": 554
+    },
+    {
+      "epoch": 0.6082191780821918,
+      "grad_norm": 2.127434253692627,
+      "learning_rate": 4.457757920389927e-06,
+      "loss": 1.9913,
+      "step": 555
+    },
+    {
+      "epoch": 0.6093150684931506,
+      "grad_norm": 2.2486324310302734,
+      "learning_rate": 4.455727051177904e-06,
+      "loss": 1.9239,
+      "step": 556
+    },
+    {
+      "epoch": 0.6104109589041096,
+      "grad_norm": 2.127575397491455,
+      "learning_rate": 4.453696181965882e-06,
+      "loss": 1.9521,
+      "step": 557
+    },
+    {
+      "epoch": 0.6115068493150685,
+      "grad_norm": 2.0798254013061523,
+      "learning_rate": 4.451665312753859e-06,
+      "loss": 2.0085,
+      "step": 558
+    },
+    {
+      "epoch": 0.6126027397260274,
+      "grad_norm": 2.287574291229248,
+      "learning_rate": 4.449634443541836e-06,
+      "loss": 1.98,
+      "step": 559
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "grad_norm": 2.2056541442871094,
+      "learning_rate": 4.4476035743298135e-06,
+      "loss": 2.0282,
+      "step": 560
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "eval_loss": 2.1287026405334473,
+      "eval_runtime": 36.7788,
+      "eval_samples_per_second": 27.19,
+      "eval_steps_per_second": 1.142,
+      "step": 560
+    },
+    {
+      "epoch": 0.6147945205479453,
+      "grad_norm": 2.169157028198242,
+      "learning_rate": 4.445572705117791e-06,
+      "loss": 1.9308,
+      "step": 561
+    },
+    {
+      "epoch": 0.6158904109589041,
+      "grad_norm": 2.1035149097442627,
+      "learning_rate": 4.443541835905768e-06,
+      "loss": 1.9507,
+      "step": 562
+    },
+    {
+      "epoch": 0.616986301369863,
+      "grad_norm": 2.1174328327178955,
+      "learning_rate": 4.441510966693745e-06,
+      "loss": 2.0162,
+      "step": 563
+    },
+    {
+      "epoch": 0.6180821917808219,
+      "grad_norm": 2.201641321182251,
+      "learning_rate": 4.439480097481723e-06,
+      "loss": 1.9573,
+      "step": 564
+    },
+    {
+      "epoch": 0.6191780821917808,
+      "grad_norm": 2.1763603687286377,
+      "learning_rate": 4.4374492282697e-06,
+      "loss": 2.0149,
+      "step": 565
+    },
+    {
+      "epoch": 0.6202739726027398,
+      "grad_norm": 2.2866439819335938,
+      "learning_rate": 4.435418359057677e-06,
+      "loss": 1.8836,
+      "step": 566
+    },
+    {
+      "epoch": 0.6213698630136987,
+      "grad_norm": 2.2060160636901855,
+      "learning_rate": 4.433387489845654e-06,
+      "loss": 1.9806,
+      "step": 567
+    },
+    {
+      "epoch": 0.6224657534246575,
+      "grad_norm": 2.0614662170410156,
+      "learning_rate": 4.431356620633632e-06,
+      "loss": 1.9821,
+      "step": 568
+    },
+    {
+      "epoch": 0.6235616438356164,
+      "grad_norm": 2.1055402755737305,
+      "learning_rate": 4.429325751421609e-06,
+      "loss": 1.9365,
+      "step": 569
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "grad_norm": 1.9907679557800293,
+      "learning_rate": 4.427294882209586e-06,
+      "loss": 1.9429,
+      "step": 570
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "eval_loss": 2.1286206245422363,
+      "eval_runtime": 36.7748,
+      "eval_samples_per_second": 27.193,
+      "eval_steps_per_second": 1.142,
+      "step": 570
+    },
+    {
+      "epoch": 0.6257534246575343,
+      "grad_norm": 2.0169243812561035,
+      "learning_rate": 4.4252640129975635e-06,
+      "loss": 1.9643,
+      "step": 571
+    },
+    {
+      "epoch": 0.6268493150684932,
+      "grad_norm": 2.1082401275634766,
+      "learning_rate": 4.423233143785541e-06,
+      "loss": 1.9009,
+      "step": 572
+    },
+    {
+      "epoch": 0.6279452054794521,
+      "grad_norm": 2.202542543411255,
+      "learning_rate": 4.421202274573518e-06,
+      "loss": 1.9694,
+      "step": 573
+    },
+    {
+      "epoch": 0.6290410958904109,
+      "grad_norm": 2.028489828109741,
+      "learning_rate": 4.4191714053614945e-06,
+      "loss": 1.9697,
+      "step": 574
+    },
+    {
+      "epoch": 0.6301369863013698,
+      "grad_norm": 2.134274482727051,
+      "learning_rate": 4.417140536149473e-06,
+      "loss": 1.8672,
+      "step": 575
+    },
+    {
+      "epoch": 0.6312328767123287,
+      "grad_norm": 2.063542366027832,
+      "learning_rate": 4.415109666937449e-06,
+      "loss": 1.909,
+      "step": 576
+    },
+    {
+      "epoch": 0.6323287671232877,
+      "grad_norm": 1.9747039079666138,
+      "learning_rate": 4.413078797725427e-06,
+      "loss": 1.9795,
+      "step": 577
+    },
+    {
+      "epoch": 0.6334246575342466,
+      "grad_norm": 2.063673496246338,
+      "learning_rate": 4.411047928513404e-06,
+      "loss": 1.9283,
+      "step": 578
+    },
+    {
+      "epoch": 0.6345205479452055,
+      "grad_norm": 2.024611711502075,
+      "learning_rate": 4.409017059301382e-06,
+      "loss": 1.9682,
+      "step": 579
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "grad_norm": 2.0471599102020264,
+      "learning_rate": 4.406986190089358e-06,
+      "loss": 1.969,
+      "step": 580
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "eval_loss": 2.1233692169189453,
+      "eval_runtime": 36.8185,
+      "eval_samples_per_second": 27.16,
+      "eval_steps_per_second": 1.141,
+      "step": 580
+    },
+    {
+      "epoch": 0.6367123287671232,
+      "grad_norm": 2.1038553714752197,
+      "learning_rate": 4.404955320877336e-06,
+      "loss": 1.9915,
+      "step": 581
+    },
+    {
+      "epoch": 0.6378082191780822,
+      "grad_norm": 2.1222152709960938,
+      "learning_rate": 4.402924451665313e-06,
+      "loss": 1.9892,
+      "step": 582
+    },
+    {
+      "epoch": 0.6389041095890411,
+      "grad_norm": 2.09137225151062,
+      "learning_rate": 4.400893582453291e-06,
+      "loss": 1.934,
+      "step": 583
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 2.2158281803131104,
+      "learning_rate": 4.398862713241267e-06,
+      "loss": 1.9188,
+      "step": 584
+    },
+    {
+      "epoch": 0.6410958904109589,
+      "grad_norm": 2.34619140625,
+      "learning_rate": 4.396831844029245e-06,
+      "loss": 1.9495,
+      "step": 585
+    },
+    {
+      "epoch": 0.6421917808219179,
+      "grad_norm": 2.2732315063476562,
+      "learning_rate": 4.394800974817222e-06,
+      "loss": 1.9632,
+      "step": 586
+    },
+    {
+      "epoch": 0.6432876712328767,
+      "grad_norm": 2.153338670730591,
+      "learning_rate": 4.392770105605199e-06,
+      "loss": 1.9551,
+      "step": 587
+    },
+    {
+      "epoch": 0.6443835616438356,
+      "grad_norm": 2.2201521396636963,
+      "learning_rate": 4.390739236393176e-06,
+      "loss": 1.9677,
+      "step": 588
+    },
+    {
+      "epoch": 0.6454794520547945,
+      "grad_norm": 1.9894570112228394,
+      "learning_rate": 4.388708367181154e-06,
+      "loss": 1.9826,
+      "step": 589
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "grad_norm": 2.145061492919922,
+      "learning_rate": 4.386677497969131e-06,
+      "loss": 1.9826,
+      "step": 590
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "eval_loss": 2.122312545776367,
+      "eval_runtime": 37.1153,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 590
+    },
+    {
+      "epoch": 0.6476712328767124,
+      "grad_norm": 2.257237672805786,
+      "learning_rate": 4.384646628757108e-06,
+      "loss": 1.9722,
+      "step": 591
+    },
+    {
+      "epoch": 0.6487671232876713,
+      "grad_norm": 2.2406375408172607,
+      "learning_rate": 4.3826157595450855e-06,
+      "loss": 2.0141,
+      "step": 592
+    },
+    {
+      "epoch": 0.6498630136986301,
+      "grad_norm": 1.9656853675842285,
+      "learning_rate": 4.380584890333063e-06,
+      "loss": 1.9435,
+      "step": 593
+    },
+    {
+      "epoch": 0.650958904109589,
+      "grad_norm": 2.1063120365142822,
+      "learning_rate": 4.37855402112104e-06,
+      "loss": 1.9001,
+      "step": 594
+    },
+    {
+      "epoch": 0.6520547945205479,
+      "grad_norm": 2.1594269275665283,
+      "learning_rate": 4.376523151909017e-06,
+      "loss": 1.8596,
+      "step": 595
+    },
+    {
+      "epoch": 0.6531506849315069,
+      "grad_norm": 2.2261908054351807,
+      "learning_rate": 4.3744922826969946e-06,
+      "loss": 2.016,
+      "step": 596
+    },
+    {
+      "epoch": 0.6542465753424658,
+      "grad_norm": 1.9940093755722046,
+      "learning_rate": 4.372461413484972e-06,
+      "loss": 1.9293,
+      "step": 597
+    },
+    {
+      "epoch": 0.6553424657534247,
+      "grad_norm": 2.62186861038208,
+      "learning_rate": 4.370430544272949e-06,
+      "loss": 1.9656,
+      "step": 598
+    },
+    {
+      "epoch": 0.6564383561643835,
+      "grad_norm": 2.3531494140625,
+      "learning_rate": 4.368399675060926e-06,
+      "loss": 1.9984,
+      "step": 599
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "grad_norm": 2.362769365310669,
+      "learning_rate": 4.366368805848904e-06,
+      "loss": 1.9971,
+      "step": 600
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "eval_loss": 2.1172940731048584,
+      "eval_runtime": 37.0498,
+      "eval_samples_per_second": 26.991,
+      "eval_steps_per_second": 1.134,
+      "step": 600
+    },
+    {
+      "epoch": 0.6586301369863014,
+      "grad_norm": 2.3806607723236084,
+      "learning_rate": 4.364337936636881e-06,
+      "loss": 1.9566,
+      "step": 601
+    },
+    {
+      "epoch": 0.6597260273972603,
+      "grad_norm": 2.5531814098358154,
+      "learning_rate": 4.362307067424858e-06,
+      "loss": 1.9002,
+      "step": 602
+    },
+    {
+      "epoch": 0.6608219178082192,
+      "grad_norm": 2.283364772796631,
+      "learning_rate": 4.3602761982128355e-06,
+      "loss": 1.9633,
+      "step": 603
+    },
+    {
+      "epoch": 0.6619178082191781,
+      "grad_norm": 2.180171489715576,
+      "learning_rate": 4.358245329000813e-06,
+      "loss": 1.9277,
+      "step": 604
+    },
+    {
+      "epoch": 0.663013698630137,
+      "grad_norm": 2.0771350860595703,
+      "learning_rate": 4.35621445978879e-06,
+      "loss": 1.9484,
+      "step": 605
+    },
+    {
+      "epoch": 0.6641095890410958,
+      "grad_norm": 2.3772811889648438,
+      "learning_rate": 4.354183590576767e-06,
+      "loss": 1.9564,
+      "step": 606
+    },
+    {
+      "epoch": 0.6652054794520548,
+      "grad_norm": 2.205486297607422,
+      "learning_rate": 4.352152721364745e-06,
+      "loss": 1.9389,
+      "step": 607
+    },
+    {
+      "epoch": 0.6663013698630137,
+      "grad_norm": 2.085190773010254,
+      "learning_rate": 4.350121852152722e-06,
+      "loss": 1.8948,
+      "step": 608
+    },
+    {
+      "epoch": 0.6673972602739726,
+      "grad_norm": 2.4398982524871826,
+      "learning_rate": 4.348090982940699e-06,
+      "loss": 1.9736,
+      "step": 609
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "grad_norm": 2.194087028503418,
+      "learning_rate": 4.3460601137286764e-06,
+      "loss": 1.9607,
+      "step": 610
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "eval_loss": 2.115041971206665,
+      "eval_runtime": 36.9738,
+      "eval_samples_per_second": 27.046,
+      "eval_steps_per_second": 1.136,
+      "step": 610
+    },
+    {
+      "epoch": 0.6695890410958905,
+      "grad_norm": 2.143054485321045,
+      "learning_rate": 4.344029244516654e-06,
+      "loss": 1.9733,
+      "step": 611
+    },
+    {
+      "epoch": 0.6706849315068493,
+      "grad_norm": 2.2484617233276367,
+      "learning_rate": 4.341998375304631e-06,
+      "loss": 1.9417,
+      "step": 612
+    },
+    {
+      "epoch": 0.6717808219178082,
+      "grad_norm": 2.0430805683135986,
+      "learning_rate": 4.339967506092607e-06,
+      "loss": 2.0192,
+      "step": 613
+    },
+    {
+      "epoch": 0.6728767123287671,
+      "grad_norm": 2.061089277267456,
+      "learning_rate": 4.3379366368805855e-06,
+      "loss": 1.9543,
+      "step": 614
+    },
+    {
+      "epoch": 0.673972602739726,
+      "grad_norm": 2.298022747039795,
+      "learning_rate": 4.335905767668562e-06,
+      "loss": 1.9579,
+      "step": 615
+    },
+    {
+      "epoch": 0.675068493150685,
+      "grad_norm": 2.1418962478637695,
+      "learning_rate": 4.33387489845654e-06,
+      "loss": 1.938,
+      "step": 616
+    },
+    {
+      "epoch": 0.6761643835616439,
+      "grad_norm": 2.1005501747131348,
+      "learning_rate": 4.3318440292445165e-06,
+      "loss": 1.972,
+      "step": 617
+    },
+    {
+      "epoch": 0.6772602739726027,
+      "grad_norm": 2.1665523052215576,
+      "learning_rate": 4.329813160032495e-06,
+      "loss": 1.9535,
+      "step": 618
+    },
+    {
+      "epoch": 0.6783561643835616,
+      "grad_norm": 2.144580841064453,
+      "learning_rate": 4.327782290820471e-06,
+      "loss": 1.9188,
+      "step": 619
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "grad_norm": 1.9855908155441284,
+      "learning_rate": 4.325751421608449e-06,
+      "loss": 1.912,
+      "step": 620
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "eval_loss": 2.118319272994995,
+      "eval_runtime": 36.9677,
+      "eval_samples_per_second": 27.051,
+      "eval_steps_per_second": 1.136,
+      "step": 620
+    },
+    {
+      "epoch": 0.6805479452054795,
+      "grad_norm": 2.395538091659546,
+      "learning_rate": 4.323720552396426e-06,
+      "loss": 1.9886,
+      "step": 621
+    },
+    {
+      "epoch": 0.6816438356164384,
+      "grad_norm": 2.190574884414673,
+      "learning_rate": 4.321689683184404e-06,
+      "loss": 1.9326,
+      "step": 622
+    },
+    {
+      "epoch": 0.6827397260273973,
+      "grad_norm": 2.1826560497283936,
+      "learning_rate": 4.31965881397238e-06,
+      "loss": 1.9278,
+      "step": 623
+    },
+    {
+      "epoch": 0.6838356164383562,
+      "grad_norm": 2.151097536087036,
+      "learning_rate": 4.3176279447603574e-06,
+      "loss": 1.9359,
+      "step": 624
+    },
+    {
+      "epoch": 0.684931506849315,
+      "grad_norm": 2.1830122470855713,
+      "learning_rate": 4.315597075548335e-06,
+      "loss": 1.9728,
+      "step": 625
+    },
+    {
+      "epoch": 0.686027397260274,
+      "grad_norm": 2.203325033187866,
+      "learning_rate": 4.313566206336312e-06,
+      "loss": 1.9638,
+      "step": 626
+    },
+    {
+      "epoch": 0.6871232876712329,
+      "grad_norm": 2.7240309715270996,
+      "learning_rate": 4.311535337124289e-06,
+      "loss": 1.9907,
+      "step": 627
+    },
+    {
+      "epoch": 0.6882191780821918,
+      "grad_norm": 2.133599281311035,
+      "learning_rate": 4.3095044679122665e-06,
+      "loss": 1.9352,
+      "step": 628
+    },
+    {
+      "epoch": 0.6893150684931507,
+      "grad_norm": 2.0318551063537598,
+      "learning_rate": 4.307473598700244e-06,
+      "loss": 1.9432,
+      "step": 629
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "grad_norm": 2.502101182937622,
+      "learning_rate": 4.305442729488221e-06,
+      "loss": 1.9641,
+      "step": 630
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "eval_loss": 2.1158926486968994,
+      "eval_runtime": 36.8818,
+      "eval_samples_per_second": 27.114,
+      "eval_steps_per_second": 1.139,
+      "step": 630
+    },
+    {
+      "epoch": 0.6915068493150685,
+      "grad_norm": 2.3620944023132324,
+      "learning_rate": 4.303411860276198e-06,
+      "loss": 1.8841,
+      "step": 631
+    },
+    {
+      "epoch": 0.6926027397260274,
+      "grad_norm": 2.3139240741729736,
+      "learning_rate": 4.301380991064176e-06,
+      "loss": 1.9896,
+      "step": 632
+    },
+    {
+      "epoch": 0.6936986301369863,
+      "grad_norm": 2.314877510070801,
+      "learning_rate": 4.299350121852153e-06,
+      "loss": 1.9655,
+      "step": 633
+    },
+    {
+      "epoch": 0.6947945205479452,
+      "grad_norm": 2.036371946334839,
+      "learning_rate": 4.29731925264013e-06,
+      "loss": 1.9917,
+      "step": 634
+    },
+    {
+      "epoch": 0.6958904109589041,
+      "grad_norm": 2.1859662532806396,
+      "learning_rate": 4.2952883834281075e-06,
+      "loss": 1.9805,
+      "step": 635
+    },
+    {
+      "epoch": 0.6969863013698631,
+      "grad_norm": 2.2317097187042236,
+      "learning_rate": 4.293257514216085e-06,
+      "loss": 1.9402,
+      "step": 636
+    },
+    {
+      "epoch": 0.6980821917808219,
+      "grad_norm": 2.3249905109405518,
+      "learning_rate": 4.291226645004062e-06,
+      "loss": 1.9013,
+      "step": 637
+    },
+    {
+      "epoch": 0.6991780821917808,
+      "grad_norm": 2.199758768081665,
+      "learning_rate": 4.289195775792039e-06,
+      "loss": 1.9783,
+      "step": 638
+    },
+    {
+      "epoch": 0.7002739726027397,
+      "grad_norm": 2.3598549365997314,
+      "learning_rate": 4.287164906580017e-06,
+      "loss": 1.9936,
+      "step": 639
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "grad_norm": 2.1705522537231445,
+      "learning_rate": 4.285134037367994e-06,
+      "loss": 1.9304,
+      "step": 640
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "eval_loss": 2.1055190563201904,
+      "eval_runtime": 36.8673,
+      "eval_samples_per_second": 27.124,
+      "eval_steps_per_second": 1.139,
+      "step": 640
+    },
+    {
+      "epoch": 0.7024657534246576,
+      "grad_norm": 2.3253612518310547,
+      "learning_rate": 4.283103168155971e-06,
+      "loss": 1.9196,
+      "step": 641
+    },
+    {
+      "epoch": 0.7035616438356165,
+      "grad_norm": 2.3054087162017822,
+      "learning_rate": 4.281072298943948e-06,
+      "loss": 1.9053,
+      "step": 642
+    },
+    {
+      "epoch": 0.7046575342465753,
+      "grad_norm": 2.262057304382324,
+      "learning_rate": 4.279041429731926e-06,
+      "loss": 1.8999,
+      "step": 643
+    },
+    {
+      "epoch": 0.7057534246575342,
+      "grad_norm": 2.421391248703003,
+      "learning_rate": 4.277010560519903e-06,
+      "loss": 1.9438,
+      "step": 644
+    },
+    {
+      "epoch": 0.7068493150684931,
+      "grad_norm": 2.2841498851776123,
+      "learning_rate": 4.27497969130788e-06,
+      "loss": 1.9187,
+      "step": 645
+    },
+    {
+      "epoch": 0.707945205479452,
+      "grad_norm": 2.3226938247680664,
+      "learning_rate": 4.2729488220958575e-06,
+      "loss": 1.9275,
+      "step": 646
+    },
+    {
+      "epoch": 0.709041095890411,
+      "grad_norm": 2.2306301593780518,
+      "learning_rate": 4.270917952883835e-06,
+      "loss": 1.9684,
+      "step": 647
+    },
+    {
+      "epoch": 0.7101369863013699,
+      "grad_norm": 2.0551652908325195,
+      "learning_rate": 4.268887083671812e-06,
+      "loss": 1.9849,
+      "step": 648
+    },
+    {
+      "epoch": 0.7112328767123288,
+      "grad_norm": 2.2952210903167725,
+      "learning_rate": 4.266856214459789e-06,
+      "loss": 1.9026,
+      "step": 649
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "grad_norm": 2.1630797386169434,
+      "learning_rate": 4.264825345247766e-06,
+      "loss": 2.0184,
+      "step": 650
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "eval_loss": 2.1037871837615967,
+      "eval_runtime": 36.8221,
+      "eval_samples_per_second": 27.158,
+      "eval_steps_per_second": 1.141,
+      "step": 650
+    },
+    {
+      "epoch": 0.7134246575342466,
+      "grad_norm": 1.964150309562683,
+      "learning_rate": 4.262794476035744e-06,
+      "loss": 1.8932,
+      "step": 651
+    },
+    {
+      "epoch": 0.7145205479452055,
+      "grad_norm": 2.1474685668945312,
+      "learning_rate": 4.26076360682372e-06,
+      "loss": 1.9383,
+      "step": 652
+    },
+    {
+      "epoch": 0.7156164383561644,
+      "grad_norm": 2.199824094772339,
+      "learning_rate": 4.2587327376116984e-06,
+      "loss": 1.8599,
+      "step": 653
+    },
+    {
+      "epoch": 0.7167123287671233,
+      "grad_norm": 2.12156343460083,
+      "learning_rate": 4.256701868399675e-06,
+      "loss": 1.9162,
+      "step": 654
+    },
+    {
+      "epoch": 0.7178082191780822,
+      "grad_norm": 2.3832154273986816,
+      "learning_rate": 4.254670999187653e-06,
+      "loss": 1.9613,
+      "step": 655
+    },
+    {
+      "epoch": 0.718904109589041,
+      "grad_norm": 2.063906192779541,
+      "learning_rate": 4.252640129975629e-06,
+      "loss": 1.9421,
+      "step": 656
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 2.184842586517334,
+      "learning_rate": 4.2506092607636075e-06,
+      "loss": 1.9471,
+      "step": 657
+    },
+    {
+      "epoch": 0.7210958904109589,
+      "grad_norm": 2.1340036392211914,
+      "learning_rate": 4.248578391551584e-06,
+      "loss": 1.931,
+      "step": 658
+    },
+    {
+      "epoch": 0.7221917808219178,
+      "grad_norm": 2.117544651031494,
+      "learning_rate": 4.246547522339562e-06,
+      "loss": 1.8698,
+      "step": 659
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "grad_norm": 2.308344841003418,
+      "learning_rate": 4.2445166531275385e-06,
+      "loss": 1.9217,
+      "step": 660
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "eval_loss": 2.1086323261260986,
+      "eval_runtime": 36.8082,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 660
+    },
+    {
+      "epoch": 0.7243835616438357,
+      "grad_norm": 2.2239017486572266,
+      "learning_rate": 4.242485783915517e-06,
+      "loss": 1.9286,
+      "step": 661
+    },
+    {
+      "epoch": 0.7254794520547945,
+      "grad_norm": 2.263697624206543,
+      "learning_rate": 4.240454914703493e-06,
+      "loss": 1.8821,
+      "step": 662
+    },
+    {
+      "epoch": 0.7265753424657534,
+      "grad_norm": 2.4305059909820557,
+      "learning_rate": 4.23842404549147e-06,
+      "loss": 1.8996,
+      "step": 663
+    },
+    {
+      "epoch": 0.7276712328767123,
+      "grad_norm": 2.2333695888519287,
+      "learning_rate": 4.236393176279448e-06,
+      "loss": 1.8906,
+      "step": 664
+    },
+    {
+      "epoch": 0.7287671232876712,
+      "grad_norm": 2.160982608795166,
+      "learning_rate": 4.234362307067425e-06,
+      "loss": 1.879,
+      "step": 665
+    },
+    {
+      "epoch": 0.7298630136986302,
+      "grad_norm": 2.3707094192504883,
+      "learning_rate": 4.232331437855402e-06,
+      "loss": 1.9577,
+      "step": 666
+    },
+    {
+      "epoch": 0.7309589041095891,
+      "grad_norm": 2.0502827167510986,
+      "learning_rate": 4.2303005686433795e-06,
+      "loss": 1.9801,
+      "step": 667
+    },
+    {
+      "epoch": 0.7320547945205479,
+      "grad_norm": 2.0648176670074463,
+      "learning_rate": 4.228269699431357e-06,
+      "loss": 1.978,
+      "step": 668
+    },
+    {
+      "epoch": 0.7331506849315068,
+      "grad_norm": 2.346316337585449,
+      "learning_rate": 4.226238830219334e-06,
+      "loss": 1.9387,
+      "step": 669
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "grad_norm": 1.9734236001968384,
+      "learning_rate": 4.224207961007311e-06,
+      "loss": 1.9396,
+      "step": 670
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "eval_loss": 2.1104705333709717,
+      "eval_runtime": 36.8144,
+      "eval_samples_per_second": 27.163,
+      "eval_steps_per_second": 1.141,
+      "step": 670
+    },
+    {
+      "epoch": 0.7353424657534247,
+      "grad_norm": 2.3170549869537354,
+      "learning_rate": 4.2221770917952886e-06,
+      "loss": 1.9398,
+      "step": 671
+    },
+    {
+      "epoch": 0.7364383561643836,
+      "grad_norm": 2.0437264442443848,
+      "learning_rate": 4.220146222583266e-06,
+      "loss": 1.9284,
+      "step": 672
+    },
+    {
+      "epoch": 0.7375342465753425,
+      "grad_norm": 2.0808510780334473,
+      "learning_rate": 4.218115353371243e-06,
+      "loss": 1.9425,
+      "step": 673
+    },
+    {
+      "epoch": 0.7386301369863014,
+      "grad_norm": 2.0854742527008057,
+      "learning_rate": 4.21608448415922e-06,
+      "loss": 1.9358,
+      "step": 674
+    },
+    {
+      "epoch": 0.7397260273972602,
+      "grad_norm": 2.131622552871704,
+      "learning_rate": 4.214053614947198e-06,
+      "loss": 1.9048,
+      "step": 675
+    },
+    {
+      "epoch": 0.7408219178082192,
+      "grad_norm": 2.276359796524048,
+      "learning_rate": 4.212022745735175e-06,
+      "loss": 1.9871,
+      "step": 676
+    },
+    {
+      "epoch": 0.7419178082191781,
+      "grad_norm": 2.2194812297821045,
+      "learning_rate": 4.209991876523152e-06,
+      "loss": 1.882,
+      "step": 677
+    },
+    {
+      "epoch": 0.743013698630137,
+      "grad_norm": 2.181891679763794,
+      "learning_rate": 4.2079610073111295e-06,
+      "loss": 1.9546,
+      "step": 678
+    },
+    {
+      "epoch": 0.7441095890410959,
+      "grad_norm": 2.0956673622131348,
+      "learning_rate": 4.205930138099107e-06,
+      "loss": 1.8836,
+      "step": 679
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "grad_norm": 2.057894229888916,
+      "learning_rate": 4.203899268887084e-06,
+      "loss": 1.9842,
+      "step": 680
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "eval_loss": 2.104182720184326,
+      "eval_runtime": 36.8189,
+      "eval_samples_per_second": 27.16,
+      "eval_steps_per_second": 1.141,
+      "step": 680
+    },
+    {
+      "epoch": 0.7463013698630137,
+      "grad_norm": 2.0965662002563477,
+      "learning_rate": 4.201868399675061e-06,
+      "loss": 1.9378,
+      "step": 681
+    },
+    {
+      "epoch": 0.7473972602739726,
+      "grad_norm": 2.1843457221984863,
+      "learning_rate": 4.199837530463039e-06,
+      "loss": 1.9784,
+      "step": 682
+    },
+    {
+      "epoch": 0.7484931506849315,
+      "grad_norm": 2.2044246196746826,
+      "learning_rate": 4.197806661251016e-06,
+      "loss": 1.8642,
+      "step": 683
+    },
+    {
+      "epoch": 0.7495890410958904,
+      "grad_norm": 2.202115058898926,
+      "learning_rate": 4.195775792038993e-06,
+      "loss": 1.929,
+      "step": 684
+    },
+    {
+      "epoch": 0.7506849315068493,
+      "grad_norm": 2.3671462535858154,
+      "learning_rate": 4.19374492282697e-06,
+      "loss": 1.9172,
+      "step": 685
+    },
+    {
+      "epoch": 0.7517808219178083,
+      "grad_norm": 2.0602803230285645,
+      "learning_rate": 4.191714053614948e-06,
+      "loss": 1.8503,
+      "step": 686
+    },
+    {
+      "epoch": 0.7528767123287671,
+      "grad_norm": 2.12145733833313,
+      "learning_rate": 4.189683184402924e-06,
+      "loss": 1.9511,
+      "step": 687
+    },
+    {
+      "epoch": 0.753972602739726,
+      "grad_norm": 2.1049914360046387,
+      "learning_rate": 4.187652315190902e-06,
+      "loss": 1.9043,
+      "step": 688
+    },
+    {
+      "epoch": 0.7550684931506849,
+      "grad_norm": 2.235497236251831,
+      "learning_rate": 4.185621445978879e-06,
+      "loss": 1.9784,
+      "step": 689
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "grad_norm": 2.173222303390503,
+      "learning_rate": 4.183590576766857e-06,
+      "loss": 1.9052,
+      "step": 690
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "eval_loss": 2.0982251167297363,
+      "eval_runtime": 37.1446,
+      "eval_samples_per_second": 26.922,
+      "eval_steps_per_second": 1.131,
+      "step": 690
+    },
+    {
+      "epoch": 0.7572602739726028,
+      "grad_norm": 2.0693562030792236,
+      "learning_rate": 4.181559707554833e-06,
+      "loss": 1.9223,
+      "step": 691
+    },
+    {
+      "epoch": 0.7583561643835617,
+      "grad_norm": 2.209419012069702,
+      "learning_rate": 4.179528838342811e-06,
+      "loss": 1.9709,
+      "step": 692
+    },
+    {
+      "epoch": 0.7594520547945205,
+      "grad_norm": 2.066041946411133,
+      "learning_rate": 4.177497969130788e-06,
+      "loss": 1.889,
+      "step": 693
+    },
+    {
+      "epoch": 0.7605479452054794,
+      "grad_norm": 2.4114325046539307,
+      "learning_rate": 4.175467099918766e-06,
+      "loss": 1.9059,
+      "step": 694
+    },
+    {
+      "epoch": 0.7616438356164383,
+      "grad_norm": 2.0185043811798096,
+      "learning_rate": 4.173436230706742e-06,
+      "loss": 1.9877,
+      "step": 695
+    },
+    {
+      "epoch": 0.7627397260273973,
+      "grad_norm": 2.23392915725708,
+      "learning_rate": 4.1714053614947205e-06,
+      "loss": 1.9199,
+      "step": 696
+    },
+    {
+      "epoch": 0.7638356164383562,
+      "grad_norm": 2.2158737182617188,
+      "learning_rate": 4.169374492282697e-06,
+      "loss": 1.9001,
+      "step": 697
+    },
+    {
+      "epoch": 0.7649315068493151,
+      "grad_norm": 2.1511690616607666,
+      "learning_rate": 4.167343623070675e-06,
+      "loss": 1.9166,
+      "step": 698
+    },
+    {
+      "epoch": 0.766027397260274,
+      "grad_norm": 2.374741554260254,
+      "learning_rate": 4.165312753858652e-06,
+      "loss": 1.9684,
+      "step": 699
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "grad_norm": 2.0409998893737793,
+      "learning_rate": 4.163281884646629e-06,
+      "loss": 1.9177,
+      "step": 700
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "eval_loss": 2.096895933151245,
+      "eval_runtime": 37.066,
+      "eval_samples_per_second": 26.979,
+      "eval_steps_per_second": 1.133,
+      "step": 700
+    },
+    {
+      "epoch": 0.7682191780821918,
+      "grad_norm": 2.3173422813415527,
+      "learning_rate": 4.161251015434607e-06,
+      "loss": 1.9071,
+      "step": 701
+    },
+    {
+      "epoch": 0.7693150684931507,
+      "grad_norm": 2.077842950820923,
+      "learning_rate": 4.159220146222583e-06,
+      "loss": 1.935,
+      "step": 702
+    },
+    {
+      "epoch": 0.7704109589041096,
+      "grad_norm": 2.0518617630004883,
+      "learning_rate": 4.157189277010561e-06,
+      "loss": 1.8942,
+      "step": 703
+    },
+    {
+      "epoch": 0.7715068493150685,
+      "grad_norm": 1.9950882196426392,
+      "learning_rate": 4.155158407798538e-06,
+      "loss": 1.917,
+      "step": 704
+    },
+    {
+      "epoch": 0.7726027397260274,
+      "grad_norm": 2.3222219944000244,
+      "learning_rate": 4.153127538586516e-06,
+      "loss": 1.9551,
+      "step": 705
+    },
+    {
+      "epoch": 0.7736986301369863,
+      "grad_norm": 2.129333972930908,
+      "learning_rate": 4.151096669374492e-06,
+      "loss": 1.9204,
+      "step": 706
+    },
+    {
+      "epoch": 0.7747945205479452,
+      "grad_norm": 2.0357553958892822,
+      "learning_rate": 4.1490658001624705e-06,
+      "loss": 1.9399,
+      "step": 707
+    },
+    {
+      "epoch": 0.7758904109589041,
+      "grad_norm": 2.233694553375244,
+      "learning_rate": 4.147034930950447e-06,
+      "loss": 1.9182,
+      "step": 708
+    },
+    {
+      "epoch": 0.776986301369863,
+      "grad_norm": 2.1391239166259766,
+      "learning_rate": 4.145004061738425e-06,
+      "loss": 1.9157,
+      "step": 709
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "grad_norm": 2.162320613861084,
+      "learning_rate": 4.1429731925264015e-06,
+      "loss": 1.9115,
+      "step": 710
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "eval_loss": 2.0954039096832275,
+      "eval_runtime": 36.9862,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 710
+    },
+    {
+      "epoch": 0.7791780821917809,
+      "grad_norm": 2.2382194995880127,
+      "learning_rate": 4.140942323314379e-06,
+      "loss": 1.8428,
+      "step": 711
+    },
+    {
+      "epoch": 0.7802739726027397,
+      "grad_norm": 2.3286185264587402,
+      "learning_rate": 4.138911454102356e-06,
+      "loss": 1.8906,
+      "step": 712
+    },
+    {
+      "epoch": 0.7813698630136986,
+      "grad_norm": 2.182708501815796,
+      "learning_rate": 4.136880584890333e-06,
+      "loss": 1.9132,
+      "step": 713
+    },
+    {
+      "epoch": 0.7824657534246575,
+      "grad_norm": 1.9978337287902832,
+      "learning_rate": 4.1348497156783106e-06,
+      "loss": 1.9433,
+      "step": 714
+    },
+    {
+      "epoch": 0.7835616438356164,
+      "grad_norm": 2.3092565536499023,
+      "learning_rate": 4.132818846466288e-06,
+      "loss": 1.9374,
+      "step": 715
+    },
+    {
+      "epoch": 0.7846575342465754,
+      "grad_norm": 2.0351054668426514,
+      "learning_rate": 4.130787977254265e-06,
+      "loss": 1.9361,
+      "step": 716
+    },
+    {
+      "epoch": 0.7857534246575343,
+      "grad_norm": 2.14770770072937,
+      "learning_rate": 4.128757108042242e-06,
+      "loss": 1.9553,
+      "step": 717
+    },
+    {
+      "epoch": 0.7868493150684932,
+      "grad_norm": 2.0613555908203125,
+      "learning_rate": 4.12672623883022e-06,
+      "loss": 1.9506,
+      "step": 718
+    },
+    {
+      "epoch": 0.787945205479452,
+      "grad_norm": 2.1312081813812256,
+      "learning_rate": 4.124695369618197e-06,
+      "loss": 2.0079,
+      "step": 719
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "grad_norm": 2.088552474975586,
+      "learning_rate": 4.122664500406174e-06,
+      "loss": 1.9166,
+      "step": 720
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "eval_loss": 2.093032121658325,
+      "eval_runtime": 36.8821,
+      "eval_samples_per_second": 27.113,
+      "eval_steps_per_second": 1.139,
+      "step": 720
+    },
+    {
+      "epoch": 0.7901369863013699,
+      "grad_norm": 2.1748929023742676,
+      "learning_rate": 4.1206336311941515e-06,
+      "loss": 1.9212,
+      "step": 721
+    },
+    {
+      "epoch": 0.7912328767123288,
+      "grad_norm": 2.1492016315460205,
+      "learning_rate": 4.118602761982129e-06,
+      "loss": 1.9452,
+      "step": 722
+    },
+    {
+      "epoch": 0.7923287671232877,
+      "grad_norm": 2.3364970684051514,
+      "learning_rate": 4.116571892770106e-06,
+      "loss": 1.8821,
+      "step": 723
+    },
+    {
+      "epoch": 0.7934246575342466,
+      "grad_norm": 2.318488597869873,
+      "learning_rate": 4.114541023558083e-06,
+      "loss": 1.8597,
+      "step": 724
+    },
+    {
+      "epoch": 0.7945205479452054,
+      "grad_norm": 2.153841257095337,
+      "learning_rate": 4.112510154346061e-06,
+      "loss": 1.9843,
+      "step": 725
+    },
+    {
+      "epoch": 0.7956164383561644,
+      "grad_norm": 2.3464651107788086,
+      "learning_rate": 4.110479285134038e-06,
+      "loss": 1.9105,
+      "step": 726
+    },
+    {
+      "epoch": 0.7967123287671233,
+      "grad_norm": 2.34413743019104,
+      "learning_rate": 4.108448415922015e-06,
+      "loss": 1.9285,
+      "step": 727
+    },
+    {
+      "epoch": 0.7978082191780822,
+      "grad_norm": 2.0822196006774902,
+      "learning_rate": 4.1064175467099924e-06,
+      "loss": 1.9536,
+      "step": 728
+    },
+    {
+      "epoch": 0.7989041095890411,
+      "grad_norm": 2.1978065967559814,
+      "learning_rate": 4.10438667749797e-06,
+      "loss": 1.9487,
+      "step": 729
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 2.4218783378601074,
+      "learning_rate": 4.102355808285947e-06,
+      "loss": 1.8113,
+      "step": 730
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 2.0963101387023926,
+      "eval_runtime": 36.8853,
+      "eval_samples_per_second": 27.111,
+      "eval_steps_per_second": 1.139,
+      "step": 730
+    },
+    {
+      "epoch": 0.8010958904109589,
+      "grad_norm": 2.3887739181518555,
+      "learning_rate": 4.100324939073924e-06,
+      "loss": 2.0024,
+      "step": 731
+    },
+    {
+      "epoch": 0.8021917808219178,
+      "grad_norm": 2.296579360961914,
+      "learning_rate": 4.0982940698619015e-06,
+      "loss": 1.9092,
+      "step": 732
+    },
+    {
+      "epoch": 0.8032876712328767,
+      "grad_norm": 2.3881773948669434,
+      "learning_rate": 4.096263200649879e-06,
+      "loss": 1.9112,
+      "step": 733
+    },
+    {
+      "epoch": 0.8043835616438356,
+      "grad_norm": 2.40120530128479,
+      "learning_rate": 4.094232331437856e-06,
+      "loss": 1.9191,
+      "step": 734
+    },
+    {
+      "epoch": 0.8054794520547945,
+      "grad_norm": 2.184706211090088,
+      "learning_rate": 4.092201462225833e-06,
+      "loss": 1.8857,
+      "step": 735
+    },
+    {
+      "epoch": 0.8065753424657535,
+      "grad_norm": 2.3338732719421387,
+      "learning_rate": 4.090170593013811e-06,
+      "loss": 1.8889,
+      "step": 736
+    },
+    {
+      "epoch": 0.8076712328767123,
+      "grad_norm": 2.4065468311309814,
+      "learning_rate": 4.088139723801787e-06,
+      "loss": 1.9309,
+      "step": 737
+    },
+    {
+      "epoch": 0.8087671232876712,
+      "grad_norm": 2.044992446899414,
+      "learning_rate": 4.086108854589765e-06,
+      "loss": 1.8727,
+      "step": 738
+    },
+    {
+      "epoch": 0.8098630136986301,
+      "grad_norm": 2.23860239982605,
+      "learning_rate": 4.084077985377742e-06,
+      "loss": 1.8936,
+      "step": 739
+    },
+    {
+      "epoch": 0.810958904109589,
+      "grad_norm": 2.1324410438537598,
+      "learning_rate": 4.08204711616572e-06,
+      "loss": 1.9035,
+      "step": 740
+    },
+    {
+      "epoch": 0.810958904109589,
+      "eval_loss": 2.092885732650757,
+      "eval_runtime": 36.7991,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 740
+    },
+    {
+      "epoch": 0.812054794520548,
+      "grad_norm": 2.337340831756592,
+      "learning_rate": 4.080016246953696e-06,
+      "loss": 1.9747,
+      "step": 741
+    },
+    {
+      "epoch": 0.8131506849315069,
+      "grad_norm": 2.262801170349121,
+      "learning_rate": 4.077985377741674e-06,
+      "loss": 1.8817,
+      "step": 742
+    },
+    {
+      "epoch": 0.8142465753424658,
+      "grad_norm": 2.0494258403778076,
+      "learning_rate": 4.075954508529651e-06,
+      "loss": 1.9493,
+      "step": 743
+    },
+    {
+      "epoch": 0.8153424657534246,
+      "grad_norm": 2.085829019546509,
+      "learning_rate": 4.073923639317629e-06,
+      "loss": 1.9611,
+      "step": 744
+    },
+    {
+      "epoch": 0.8164383561643835,
+      "grad_norm": 2.105806827545166,
+      "learning_rate": 4.071892770105605e-06,
+      "loss": 1.9779,
+      "step": 745
+    },
+    {
+      "epoch": 0.8175342465753425,
+      "grad_norm": 2.2664732933044434,
+      "learning_rate": 4.069861900893583e-06,
+      "loss": 1.9009,
+      "step": 746
+    },
+    {
+      "epoch": 0.8186301369863014,
+      "grad_norm": 1.9899768829345703,
+      "learning_rate": 4.06783103168156e-06,
+      "loss": 1.8585,
+      "step": 747
+    },
+    {
+      "epoch": 0.8197260273972603,
+      "grad_norm": 2.242236614227295,
+      "learning_rate": 4.065800162469538e-06,
+      "loss": 1.8904,
+      "step": 748
+    },
+    {
+      "epoch": 0.8208219178082192,
+      "grad_norm": 2.4481139183044434,
+      "learning_rate": 4.063769293257514e-06,
+      "loss": 1.8922,
+      "step": 749
+    },
+    {
+      "epoch": 0.821917808219178,
+      "grad_norm": 2.231722354888916,
+      "learning_rate": 4.061738424045492e-06,
+      "loss": 1.9292,
+      "step": 750
+    },
+    {
+      "epoch": 0.821917808219178,
+      "eval_loss": 2.0909852981567383,
+      "eval_runtime": 36.8255,
+      "eval_samples_per_second": 27.155,
+      "eval_steps_per_second": 1.141,
+      "step": 750
+    },
+    {
+      "epoch": 0.823013698630137,
+      "grad_norm": 2.323070764541626,
+      "learning_rate": 4.059707554833469e-06,
+      "loss": 1.9053,
+      "step": 751
+    },
+    {
+      "epoch": 0.8241095890410959,
+      "grad_norm": 2.137535572052002,
+      "learning_rate": 4.057676685621446e-06,
+      "loss": 1.9272,
+      "step": 752
+    },
+    {
+      "epoch": 0.8252054794520548,
+      "grad_norm": 2.2087643146514893,
+      "learning_rate": 4.0556458164094235e-06,
+      "loss": 1.9541,
+      "step": 753
+    },
+    {
+      "epoch": 0.8263013698630137,
+      "grad_norm": 2.04811692237854,
+      "learning_rate": 4.053614947197401e-06,
+      "loss": 1.922,
+      "step": 754
+    },
+    {
+      "epoch": 0.8273972602739726,
+      "grad_norm": 2.062622308731079,
+      "learning_rate": 4.051584077985378e-06,
+      "loss": 1.8388,
+      "step": 755
+    },
+    {
+      "epoch": 0.8284931506849315,
+      "grad_norm": 2.14631986618042,
+      "learning_rate": 4.049553208773355e-06,
+      "loss": 1.9153,
+      "step": 756
+    },
+    {
+      "epoch": 0.8295890410958904,
+      "grad_norm": 2.353529214859009,
+      "learning_rate": 4.047522339561333e-06,
+      "loss": 1.9623,
+      "step": 757
+    },
+    {
+      "epoch": 0.8306849315068493,
+      "grad_norm": 2.122342348098755,
+      "learning_rate": 4.04549147034931e-06,
+      "loss": 1.9095,
+      "step": 758
+    },
+    {
+      "epoch": 0.8317808219178082,
+      "grad_norm": 2.2540230751037598,
+      "learning_rate": 4.043460601137287e-06,
+      "loss": 1.9068,
+      "step": 759
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "grad_norm": 2.164384365081787,
+      "learning_rate": 4.041429731925264e-06,
+      "loss": 2.0239,
+      "step": 760
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "eval_loss": 2.0904648303985596,
+      "eval_runtime": 36.7617,
+      "eval_samples_per_second": 27.202,
+      "eval_steps_per_second": 1.142,
+      "step": 760
+    },
+    {
+      "epoch": 0.8339726027397261,
+      "grad_norm": 2.2487905025482178,
+      "learning_rate": 4.039398862713242e-06,
+      "loss": 1.9109,
+      "step": 761
+    },
+    {
+      "epoch": 0.8350684931506849,
+      "grad_norm": 2.0947930812835693,
+      "learning_rate": 4.037367993501219e-06,
+      "loss": 1.9133,
+      "step": 762
+    },
+    {
+      "epoch": 0.8361643835616438,
+      "grad_norm": 2.2259862422943115,
+      "learning_rate": 4.035337124289196e-06,
+      "loss": 1.9821,
+      "step": 763
+    },
+    {
+      "epoch": 0.8372602739726027,
+      "grad_norm": 2.0616455078125,
+      "learning_rate": 4.0333062550771735e-06,
+      "loss": 1.9037,
+      "step": 764
+    },
+    {
+      "epoch": 0.8383561643835616,
+      "grad_norm": 2.2704474925994873,
+      "learning_rate": 4.031275385865151e-06,
+      "loss": 1.8664,
+      "step": 765
+    },
+    {
+      "epoch": 0.8394520547945206,
+      "grad_norm": 2.0059332847595215,
+      "learning_rate": 4.029244516653128e-06,
+      "loss": 1.908,
+      "step": 766
+    },
+    {
+      "epoch": 0.8405479452054795,
+      "grad_norm": 2.0462095737457275,
+      "learning_rate": 4.027213647441105e-06,
+      "loss": 1.9397,
+      "step": 767
+    },
+    {
+      "epoch": 0.8416438356164384,
+      "grad_norm": 2.150207996368408,
+      "learning_rate": 4.025182778229083e-06,
+      "loss": 1.8505,
+      "step": 768
+    },
+    {
+      "epoch": 0.8427397260273972,
+      "grad_norm": 2.174863815307617,
+      "learning_rate": 4.02315190901706e-06,
+      "loss": 1.8895,
+      "step": 769
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "grad_norm": 2.1907355785369873,
+      "learning_rate": 4.021121039805037e-06,
+      "loss": 1.9156,
+      "step": 770
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "eval_loss": 2.0900955200195312,
+      "eval_runtime": 36.8004,
+      "eval_samples_per_second": 27.174,
+      "eval_steps_per_second": 1.141,
+      "step": 770
+    },
+    {
+      "epoch": 0.8449315068493151,
+      "grad_norm": 2.2831835746765137,
+      "learning_rate": 4.0190901705930144e-06,
+      "loss": 1.9115,
+      "step": 771
+    },
+    {
+      "epoch": 0.846027397260274,
+      "grad_norm": 2.039520263671875,
+      "learning_rate": 4.017059301380992e-06,
+      "loss": 2.0047,
+      "step": 772
+    },
+    {
+      "epoch": 0.8471232876712329,
+      "grad_norm": 2.147209405899048,
+      "learning_rate": 4.015028432168969e-06,
+      "loss": 1.8845,
+      "step": 773
+    },
+    {
+      "epoch": 0.8482191780821918,
+      "grad_norm": 2.168997287750244,
+      "learning_rate": 4.012997562956945e-06,
+      "loss": 1.9606,
+      "step": 774
+    },
+    {
+      "epoch": 0.8493150684931506,
+      "grad_norm": 2.038966417312622,
+      "learning_rate": 4.0109666937449235e-06,
+      "loss": 1.8734,
+      "step": 775
+    },
+    {
+      "epoch": 0.8504109589041096,
+      "grad_norm": 2.0531768798828125,
+      "learning_rate": 4.0089358245329e-06,
+      "loss": 1.889,
+      "step": 776
+    },
+    {
+      "epoch": 0.8515068493150685,
+      "grad_norm": 2.348008155822754,
+      "learning_rate": 4.006904955320878e-06,
+      "loss": 1.8972,
+      "step": 777
+    },
+    {
+      "epoch": 0.8526027397260274,
+      "grad_norm": 2.1647253036499023,
+      "learning_rate": 4.0048740861088545e-06,
+      "loss": 1.8652,
+      "step": 778
+    },
+    {
+      "epoch": 0.8536986301369863,
+      "grad_norm": 2.1971640586853027,
+      "learning_rate": 4.002843216896833e-06,
+      "loss": 1.9495,
+      "step": 779
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "grad_norm": 2.169632911682129,
+      "learning_rate": 4.000812347684809e-06,
+      "loss": 1.9307,
+      "step": 780
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "eval_loss": 2.087860107421875,
+      "eval_runtime": 36.7982,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 780
+    },
+    {
+      "epoch": 0.8558904109589041,
+      "grad_norm": 2.4424400329589844,
+      "learning_rate": 3.998781478472787e-06,
+      "loss": 1.8849,
+      "step": 781
+    },
+    {
+      "epoch": 0.856986301369863,
+      "grad_norm": 2.2226672172546387,
+      "learning_rate": 3.996750609260764e-06,
+      "loss": 1.8975,
+      "step": 782
+    },
+    {
+      "epoch": 0.8580821917808219,
+      "grad_norm": 2.1898491382598877,
+      "learning_rate": 3.994719740048742e-06,
+      "loss": 1.8751,
+      "step": 783
+    },
+    {
+      "epoch": 0.8591780821917808,
+      "grad_norm": 2.466963052749634,
+      "learning_rate": 3.992688870836718e-06,
+      "loss": 2.0155,
+      "step": 784
+    },
+    {
+      "epoch": 0.8602739726027397,
+      "grad_norm": 2.184039831161499,
+      "learning_rate": 3.990658001624696e-06,
+      "loss": 1.9133,
+      "step": 785
+    },
+    {
+      "epoch": 0.8613698630136987,
+      "grad_norm": 2.089228868484497,
+      "learning_rate": 3.988627132412673e-06,
+      "loss": 1.8707,
+      "step": 786
+    },
+    {
+      "epoch": 0.8624657534246576,
+      "grad_norm": 2.0700666904449463,
+      "learning_rate": 3.98659626320065e-06,
+      "loss": 1.9487,
+      "step": 787
+    },
+    {
+      "epoch": 0.8635616438356164,
+      "grad_norm": 2.1495554447174072,
+      "learning_rate": 3.984565393988627e-06,
+      "loss": 1.9037,
+      "step": 788
+    },
+    {
+      "epoch": 0.8646575342465753,
+      "grad_norm": 2.0916988849639893,
+      "learning_rate": 3.9825345247766046e-06,
+      "loss": 1.9096,
+      "step": 789
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "grad_norm": 2.072387933731079,
+      "learning_rate": 3.980503655564582e-06,
+      "loss": 1.9139,
+      "step": 790
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "eval_loss": 2.0876142978668213,
+      "eval_runtime": 37.1068,
+      "eval_samples_per_second": 26.949,
+      "eval_steps_per_second": 1.132,
+      "step": 790
+    },
+    {
+      "epoch": 0.8668493150684932,
+      "grad_norm": 2.1206343173980713,
+      "learning_rate": 3.978472786352559e-06,
+      "loss": 1.9331,
+      "step": 791
+    },
+    {
+      "epoch": 0.8679452054794521,
+      "grad_norm": 2.20263934135437,
+      "learning_rate": 3.976441917140536e-06,
+      "loss": 1.9297,
+      "step": 792
+    },
+    {
+      "epoch": 0.869041095890411,
+      "grad_norm": 2.027554750442505,
+      "learning_rate": 3.974411047928514e-06,
+      "loss": 1.919,
+      "step": 793
+    },
+    {
+      "epoch": 0.8701369863013698,
+      "grad_norm": 2.0291616916656494,
+      "learning_rate": 3.972380178716491e-06,
+      "loss": 1.9682,
+      "step": 794
+    },
+    {
+      "epoch": 0.8712328767123287,
+      "grad_norm": 2.2719762325286865,
+      "learning_rate": 3.970349309504468e-06,
+      "loss": 1.9407,
+      "step": 795
+    },
+    {
+      "epoch": 0.8723287671232877,
+      "grad_norm": 2.0452375411987305,
+      "learning_rate": 3.9683184402924455e-06,
+      "loss": 1.9487,
+      "step": 796
+    },
+    {
+      "epoch": 0.8734246575342466,
+      "grad_norm": 2.170441150665283,
+      "learning_rate": 3.966287571080423e-06,
+      "loss": 1.8651,
+      "step": 797
+    },
+    {
+      "epoch": 0.8745205479452055,
+      "grad_norm": 2.3370516300201416,
+      "learning_rate": 3.9642567018684e-06,
+      "loss": 1.8599,
+      "step": 798
+    },
+    {
+      "epoch": 0.8756164383561644,
+      "grad_norm": 2.336737632751465,
+      "learning_rate": 3.962225832656377e-06,
+      "loss": 1.9304,
+      "step": 799
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "grad_norm": 1.9708369970321655,
+      "learning_rate": 3.960194963444355e-06,
+      "loss": 1.9125,
+      "step": 800
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "eval_loss": 2.0812273025512695,
+      "eval_runtime": 37.0752,
+      "eval_samples_per_second": 26.972,
+      "eval_steps_per_second": 1.133,
+      "step": 800
+    },
+    {
+      "epoch": 0.8778082191780822,
+      "grad_norm": 2.154219627380371,
+      "learning_rate": 3.958164094232332e-06,
+      "loss": 1.8304,
+      "step": 801
+    },
+    {
+      "epoch": 0.8789041095890411,
+      "grad_norm": 2.237786054611206,
+      "learning_rate": 3.956133225020309e-06,
+      "loss": 1.9147,
+      "step": 802
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 2.07143235206604,
+      "learning_rate": 3.954102355808286e-06,
+      "loss": 1.9128,
+      "step": 803
+    },
+    {
+      "epoch": 0.8810958904109589,
+      "grad_norm": 2.2780392169952393,
+      "learning_rate": 3.952071486596264e-06,
+      "loss": 1.9145,
+      "step": 804
+    },
+    {
+      "epoch": 0.8821917808219178,
+      "grad_norm": 2.2396292686462402,
+      "learning_rate": 3.950040617384241e-06,
+      "loss": 1.9049,
+      "step": 805
+    },
+    {
+      "epoch": 0.8832876712328767,
+      "grad_norm": 2.426511287689209,
+      "learning_rate": 3.948009748172218e-06,
+      "loss": 1.8984,
+      "step": 806
+    },
+    {
+      "epoch": 0.8843835616438356,
+      "grad_norm": 2.1874325275421143,
+      "learning_rate": 3.9459788789601955e-06,
+      "loss": 1.8888,
+      "step": 807
+    },
+    {
+      "epoch": 0.8854794520547945,
+      "grad_norm": 2.038098096847534,
+      "learning_rate": 3.943948009748173e-06,
+      "loss": 1.9582,
+      "step": 808
+    },
+    {
+      "epoch": 0.8865753424657534,
+      "grad_norm": 2.3748130798339844,
+      "learning_rate": 3.94191714053615e-06,
+      "loss": 1.9091,
+      "step": 809
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "grad_norm": 2.2457494735717773,
+      "learning_rate": 3.939886271324127e-06,
+      "loss": 1.9067,
+      "step": 810
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "eval_loss": 2.0766079425811768,
+      "eval_runtime": 36.9896,
+      "eval_samples_per_second": 27.035,
+      "eval_steps_per_second": 1.135,
+      "step": 810
+    },
+    {
+      "epoch": 0.8887671232876713,
+      "grad_norm": 2.156644821166992,
+      "learning_rate": 3.937855402112104e-06,
+      "loss": 1.9347,
+      "step": 811
+    },
+    {
+      "epoch": 0.8898630136986302,
+      "grad_norm": 2.3991472721099854,
+      "learning_rate": 3.935824532900082e-06,
+      "loss": 1.8645,
+      "step": 812
+    },
+    {
+      "epoch": 0.890958904109589,
+      "grad_norm": 2.252729654312134,
+      "learning_rate": 3.933793663688058e-06,
+      "loss": 1.8537,
+      "step": 813
+    },
+    {
+      "epoch": 0.8920547945205479,
+      "grad_norm": 2.2917964458465576,
+      "learning_rate": 3.9317627944760365e-06,
+      "loss": 1.9271,
+      "step": 814
+    },
+    {
+      "epoch": 0.8931506849315068,
+      "grad_norm": 2.217310667037964,
+      "learning_rate": 3.929731925264013e-06,
+      "loss": 1.9406,
+      "step": 815
+    },
+    {
+      "epoch": 0.8942465753424658,
+      "grad_norm": 2.167931079864502,
+      "learning_rate": 3.927701056051991e-06,
+      "loss": 1.8554,
+      "step": 816
+    },
+    {
+      "epoch": 0.8953424657534247,
+      "grad_norm": 2.18337082862854,
+      "learning_rate": 3.9256701868399674e-06,
+      "loss": 1.8795,
+      "step": 817
+    },
+    {
+      "epoch": 0.8964383561643836,
+      "grad_norm": 2.2348973751068115,
+      "learning_rate": 3.9236393176279456e-06,
+      "loss": 1.8831,
+      "step": 818
+    },
+    {
+      "epoch": 0.8975342465753424,
+      "grad_norm": 2.1822657585144043,
+      "learning_rate": 3.921608448415922e-06,
+      "loss": 1.8728,
+      "step": 819
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "grad_norm": 2.5253243446350098,
+      "learning_rate": 3.9195775792039e-06,
+      "loss": 1.8633,
+      "step": 820
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "eval_loss": 2.0769834518432617,
+      "eval_runtime": 36.838,
+      "eval_samples_per_second": 27.146,
+      "eval_steps_per_second": 1.14,
+      "step": 820
+    },
+    {
+      "epoch": 0.8997260273972603,
+      "grad_norm": 2.2109973430633545,
+      "learning_rate": 3.9175467099918765e-06,
+      "loss": 1.912,
+      "step": 821
+    },
+    {
+      "epoch": 0.9008219178082192,
+      "grad_norm": 2.053692102432251,
+      "learning_rate": 3.915515840779855e-06,
+      "loss": 1.8928,
+      "step": 822
+    },
+    {
+      "epoch": 0.9019178082191781,
+      "grad_norm": 2.1418213844299316,
+      "learning_rate": 3.913484971567831e-06,
+      "loss": 1.9399,
+      "step": 823
+    },
+    {
+      "epoch": 0.903013698630137,
+      "grad_norm": 2.3615458011627197,
+      "learning_rate": 3.911454102355808e-06,
+      "loss": 1.9293,
+      "step": 824
+    },
+    {
+      "epoch": 0.9041095890410958,
+      "grad_norm": 2.098254680633545,
+      "learning_rate": 3.909423233143786e-06,
+      "loss": 1.8301,
+      "step": 825
+    },
+    {
+      "epoch": 0.9052054794520548,
+      "grad_norm": 2.4473679065704346,
+      "learning_rate": 3.907392363931763e-06,
+      "loss": 1.9254,
+      "step": 826
+    },
+    {
+      "epoch": 0.9063013698630137,
+      "grad_norm": 2.252843141555786,
+      "learning_rate": 3.90536149471974e-06,
+      "loss": 1.9155,
+      "step": 827
+    },
+    {
+      "epoch": 0.9073972602739726,
+      "grad_norm": 2.1574337482452393,
+      "learning_rate": 3.9033306255077175e-06,
+      "loss": 1.9809,
+      "step": 828
+    },
+    {
+      "epoch": 0.9084931506849315,
+      "grad_norm": 2.2899720668792725,
+      "learning_rate": 3.901299756295695e-06,
+      "loss": 2.0045,
+      "step": 829
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "grad_norm": 2.227285146713257,
+      "learning_rate": 3.899268887083672e-06,
+      "loss": 2.026,
+      "step": 830
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "eval_loss": 2.0764760971069336,
+      "eval_runtime": 36.9195,
+      "eval_samples_per_second": 27.086,
+      "eval_steps_per_second": 1.138,
+      "step": 830
+    },
+    {
+      "epoch": 0.9106849315068493,
+      "grad_norm": 2.1935312747955322,
+      "learning_rate": 3.897238017871649e-06,
+      "loss": 1.8922,
+      "step": 831
+    },
+    {
+      "epoch": 0.9117808219178082,
+      "grad_norm": 2.129692316055298,
+      "learning_rate": 3.8952071486596266e-06,
+      "loss": 1.8981,
+      "step": 832
+    },
+    {
+      "epoch": 0.9128767123287671,
+      "grad_norm": 2.4435646533966064,
+      "learning_rate": 3.893176279447604e-06,
+      "loss": 1.91,
+      "step": 833
+    },
+    {
+      "epoch": 0.913972602739726,
+      "grad_norm": 2.1556732654571533,
+      "learning_rate": 3.891145410235581e-06,
+      "loss": 1.8688,
+      "step": 834
+    },
+    {
+      "epoch": 0.915068493150685,
+      "grad_norm": 2.175891160964966,
+      "learning_rate": 3.889114541023558e-06,
+      "loss": 1.9777,
+      "step": 835
+    },
+    {
+      "epoch": 0.9161643835616439,
+      "grad_norm": 2.220466375350952,
+      "learning_rate": 3.887083671811536e-06,
+      "loss": 1.9152,
+      "step": 836
+    },
+    {
+      "epoch": 0.9172602739726028,
+      "grad_norm": 2.331397294998169,
+      "learning_rate": 3.885052802599513e-06,
+      "loss": 1.892,
+      "step": 837
+    },
+    {
+      "epoch": 0.9183561643835616,
+      "grad_norm": 2.2434773445129395,
+      "learning_rate": 3.88302193338749e-06,
+      "loss": 1.841,
+      "step": 838
+    },
+    {
+      "epoch": 0.9194520547945205,
+      "grad_norm": 2.3248002529144287,
+      "learning_rate": 3.8809910641754675e-06,
+      "loss": 1.9657,
+      "step": 839
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "grad_norm": 2.090252637863159,
+      "learning_rate": 3.878960194963445e-06,
+      "loss": 1.9121,
+      "step": 840
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "eval_loss": 2.0798065662384033,
+      "eval_runtime": 36.8557,
+      "eval_samples_per_second": 27.133,
+      "eval_steps_per_second": 1.14,
+      "step": 840
+    },
+    {
+      "epoch": 0.9216438356164384,
+      "grad_norm": 2.6968672275543213,
+      "learning_rate": 3.876929325751422e-06,
+      "loss": 1.9003,
+      "step": 841
+    },
+    {
+      "epoch": 0.9227397260273973,
+      "grad_norm": 2.356321096420288,
+      "learning_rate": 3.874898456539399e-06,
+      "loss": 1.9471,
+      "step": 842
+    },
+    {
+      "epoch": 0.9238356164383562,
+      "grad_norm": 2.2678606510162354,
+      "learning_rate": 3.872867587327377e-06,
+      "loss": 1.8488,
+      "step": 843
+    },
+    {
+      "epoch": 0.924931506849315,
+      "grad_norm": 2.1122310161590576,
+      "learning_rate": 3.870836718115354e-06,
+      "loss": 1.9247,
+      "step": 844
+    },
+    {
+      "epoch": 0.9260273972602739,
+      "grad_norm": 2.1219136714935303,
+      "learning_rate": 3.868805848903331e-06,
+      "loss": 1.8829,
+      "step": 845
+    },
+    {
+      "epoch": 0.9271232876712329,
+      "grad_norm": 2.175607204437256,
+      "learning_rate": 3.8667749796913084e-06,
+      "loss": 1.887,
+      "step": 846
+    },
+    {
+      "epoch": 0.9282191780821918,
+      "grad_norm": 2.156214714050293,
+      "learning_rate": 3.864744110479286e-06,
+      "loss": 1.9119,
+      "step": 847
+    },
+    {
+      "epoch": 0.9293150684931507,
+      "grad_norm": 2.158461570739746,
+      "learning_rate": 3.862713241267262e-06,
+      "loss": 1.9092,
+      "step": 848
+    },
+    {
+      "epoch": 0.9304109589041096,
+      "grad_norm": 2.180551767349243,
+      "learning_rate": 3.86068237205524e-06,
+      "loss": 1.946,
+      "step": 849
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "grad_norm": 2.1030542850494385,
+      "learning_rate": 3.858651502843217e-06,
+      "loss": 1.9484,
+      "step": 850
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "eval_loss": 2.0767791271209717,
+      "eval_runtime": 36.8047,
+      "eval_samples_per_second": 27.17,
+      "eval_steps_per_second": 1.141,
+      "step": 850
+    },
+    {
+      "epoch": 0.9326027397260274,
+      "grad_norm": 2.240769386291504,
+      "learning_rate": 3.856620633631195e-06,
+      "loss": 1.9114,
+      "step": 851
+    },
+    {
+      "epoch": 0.9336986301369863,
+      "grad_norm": 2.1964619159698486,
+      "learning_rate": 3.854589764419171e-06,
+      "loss": 1.8339,
+      "step": 852
+    },
+    {
+      "epoch": 0.9347945205479452,
+      "grad_norm": 2.2875845432281494,
+      "learning_rate": 3.852558895207149e-06,
+      "loss": 1.8942,
+      "step": 853
+    },
+    {
+      "epoch": 0.9358904109589041,
+      "grad_norm": 2.1462507247924805,
+      "learning_rate": 3.850528025995126e-06,
+      "loss": 1.9838,
+      "step": 854
+    },
+    {
+      "epoch": 0.936986301369863,
+      "grad_norm": 2.0982913970947266,
+      "learning_rate": 3.848497156783104e-06,
+      "loss": 1.9351,
+      "step": 855
+    },
+    {
+      "epoch": 0.938082191780822,
+      "grad_norm": 2.260263442993164,
+      "learning_rate": 3.84646628757108e-06,
+      "loss": 1.8745,
+      "step": 856
+    },
+    {
+      "epoch": 0.9391780821917808,
+      "grad_norm": 2.091778039932251,
+      "learning_rate": 3.8444354183590585e-06,
+      "loss": 1.8888,
+      "step": 857
+    },
+    {
+      "epoch": 0.9402739726027397,
+      "grad_norm": 2.054974317550659,
+      "learning_rate": 3.842404549147035e-06,
+      "loss": 1.8801,
+      "step": 858
+    },
+    {
+      "epoch": 0.9413698630136986,
+      "grad_norm": 2.079641580581665,
+      "learning_rate": 3.840373679935013e-06,
+      "loss": 1.8879,
+      "step": 859
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "grad_norm": 2.3441810607910156,
+      "learning_rate": 3.8383428107229894e-06,
+      "loss": 1.8571,
+      "step": 860
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "eval_loss": 2.0766854286193848,
+      "eval_runtime": 36.8414,
+      "eval_samples_per_second": 27.143,
+      "eval_steps_per_second": 1.14,
+      "step": 860
+    },
+    {
+      "epoch": 0.9435616438356165,
+      "grad_norm": 2.123061180114746,
+      "learning_rate": 3.836311941510967e-06,
+      "loss": 1.8922,
+      "step": 861
+    },
+    {
+      "epoch": 0.9446575342465754,
+      "grad_norm": 2.1573386192321777,
+      "learning_rate": 3.834281072298944e-06,
+      "loss": 1.8713,
+      "step": 862
+    },
+    {
+      "epoch": 0.9457534246575342,
+      "grad_norm": 2.225186824798584,
+      "learning_rate": 3.832250203086921e-06,
+      "loss": 1.925,
+      "step": 863
+    },
+    {
+      "epoch": 0.9468493150684931,
+      "grad_norm": 2.1539998054504395,
+      "learning_rate": 3.8302193338748985e-06,
+      "loss": 1.8721,
+      "step": 864
+    },
+    {
+      "epoch": 0.947945205479452,
+      "grad_norm": 2.075580358505249,
+      "learning_rate": 3.828188464662876e-06,
+      "loss": 1.9489,
+      "step": 865
+    },
+    {
+      "epoch": 0.949041095890411,
+      "grad_norm": 2.1789908409118652,
+      "learning_rate": 3.826157595450853e-06,
+      "loss": 1.9883,
+      "step": 866
+    },
+    {
+      "epoch": 0.9501369863013699,
+      "grad_norm": 2.0838615894317627,
+      "learning_rate": 3.82412672623883e-06,
+      "loss": 1.9149,
+      "step": 867
+    },
+    {
+      "epoch": 0.9512328767123288,
+      "grad_norm": 2.1321213245391846,
+      "learning_rate": 3.822095857026808e-06,
+      "loss": 1.8505,
+      "step": 868
+    },
+    {
+      "epoch": 0.9523287671232876,
+      "grad_norm": 2.1419479846954346,
+      "learning_rate": 3.820064987814785e-06,
+      "loss": 1.9226,
+      "step": 869
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "grad_norm": 2.290163040161133,
+      "learning_rate": 3.818034118602762e-06,
+      "loss": 1.8737,
+      "step": 870
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "eval_loss": 2.073598623275757,
+      "eval_runtime": 36.7896,
+      "eval_samples_per_second": 27.182,
+      "eval_steps_per_second": 1.142,
+      "step": 870
+    },
+    {
+      "epoch": 0.9545205479452055,
+      "grad_norm": 2.1951074600219727,
+      "learning_rate": 3.8160032493907395e-06,
+      "loss": 1.8699,
+      "step": 871
+    },
+    {
+      "epoch": 0.9556164383561644,
+      "grad_norm": 2.25960111618042,
+      "learning_rate": 3.813972380178717e-06,
+      "loss": 1.8813,
+      "step": 872
+    },
+    {
+      "epoch": 0.9567123287671233,
+      "grad_norm": 2.4376907348632812,
+      "learning_rate": 3.811941510966694e-06,
+      "loss": 1.8782,
+      "step": 873
+    },
+    {
+      "epoch": 0.9578082191780822,
+      "grad_norm": 1.980772614479065,
+      "learning_rate": 3.809910641754671e-06,
+      "loss": 1.9205,
+      "step": 874
+    },
+    {
+      "epoch": 0.958904109589041,
+      "grad_norm": 2.0699028968811035,
+      "learning_rate": 3.8078797725426486e-06,
+      "loss": 1.8674,
+      "step": 875
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 2.2765867710113525,
+      "learning_rate": 3.8058489033306254e-06,
+      "loss": 1.81,
+      "step": 876
+    },
+    {
+      "epoch": 0.9610958904109589,
+      "grad_norm": 2.0998799800872803,
+      "learning_rate": 3.803818034118603e-06,
+      "loss": 1.9061,
+      "step": 877
+    },
+    {
+      "epoch": 0.9621917808219178,
+      "grad_norm": 2.2362914085388184,
+      "learning_rate": 3.80178716490658e-06,
+      "loss": 1.8798,
+      "step": 878
+    },
+    {
+      "epoch": 0.9632876712328767,
+      "grad_norm": 2.0789835453033447,
+      "learning_rate": 3.7997562956945577e-06,
+      "loss": 1.9491,
+      "step": 879
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "grad_norm": 2.0458688735961914,
+      "learning_rate": 3.7977254264825345e-06,
+      "loss": 1.931,
+      "step": 880
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "eval_loss": 2.0672221183776855,
+      "eval_runtime": 36.8218,
+      "eval_samples_per_second": 27.158,
+      "eval_steps_per_second": 1.141,
+      "step": 880
+    },
+    {
+      "epoch": 0.9654794520547946,
+      "grad_norm": 2.361626625061035,
+      "learning_rate": 3.7956945572705122e-06,
+      "loss": 1.8815,
+      "step": 881
+    },
+    {
+      "epoch": 0.9665753424657534,
+      "grad_norm": 2.294160842895508,
+      "learning_rate": 3.793663688058489e-06,
+      "loss": 1.8649,
+      "step": 882
+    },
+    {
+      "epoch": 0.9676712328767123,
+      "grad_norm": 2.2242612838745117,
+      "learning_rate": 3.791632818846467e-06,
+      "loss": 1.9535,
+      "step": 883
+    },
+    {
+      "epoch": 0.9687671232876712,
+      "grad_norm": 2.220269203186035,
+      "learning_rate": 3.7896019496344436e-06,
+      "loss": 1.9506,
+      "step": 884
+    },
+    {
+      "epoch": 0.9698630136986301,
+      "grad_norm": 2.112518072128296,
+      "learning_rate": 3.787571080422421e-06,
+      "loss": 1.9487,
+      "step": 885
+    },
+    {
+      "epoch": 0.9709589041095891,
+      "grad_norm": 2.17695951461792,
+      "learning_rate": 3.785540211210398e-06,
+      "loss": 1.9467,
+      "step": 886
+    },
+    {
+      "epoch": 0.972054794520548,
+      "grad_norm": 2.0990567207336426,
+      "learning_rate": 3.7835093419983755e-06,
+      "loss": 1.9216,
+      "step": 887
+    },
+    {
+      "epoch": 0.9731506849315068,
+      "grad_norm": 2.222566604614258,
+      "learning_rate": 3.7814784727863527e-06,
+      "loss": 1.9043,
+      "step": 888
+    },
+    {
+      "epoch": 0.9742465753424657,
+      "grad_norm": 2.09890079498291,
+      "learning_rate": 3.77944760357433e-06,
+      "loss": 1.884,
+      "step": 889
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "grad_norm": 2.136141061782837,
+      "learning_rate": 3.7774167343623073e-06,
+      "loss": 2.0032,
+      "step": 890
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "eval_loss": 2.0665438175201416,
+      "eval_runtime": 37.1207,
+      "eval_samples_per_second": 26.939,
+      "eval_steps_per_second": 1.131,
+      "step": 890
+    },
+    {
+      "epoch": 0.9764383561643836,
+      "grad_norm": 2.267444372177124,
+      "learning_rate": 3.7753858651502846e-06,
+      "loss": 1.947,
+      "step": 891
+    },
+    {
+      "epoch": 0.9775342465753425,
+      "grad_norm": 2.1141815185546875,
+      "learning_rate": 3.773354995938262e-06,
+      "loss": 1.8705,
+      "step": 892
+    },
+    {
+      "epoch": 0.9786301369863014,
+      "grad_norm": 2.1036782264709473,
+      "learning_rate": 3.771324126726239e-06,
+      "loss": 1.9381,
+      "step": 893
+    },
+    {
+      "epoch": 0.9797260273972602,
+      "grad_norm": 2.081982135772705,
+      "learning_rate": 3.7692932575142164e-06,
+      "loss": 1.9079,
+      "step": 894
+    },
+    {
+      "epoch": 0.9808219178082191,
+      "grad_norm": 2.0815446376800537,
+      "learning_rate": 3.7672623883021937e-06,
+      "loss": 1.9417,
+      "step": 895
+    },
+    {
+      "epoch": 0.9819178082191781,
+      "grad_norm": 2.0531511306762695,
+      "learning_rate": 3.765231519090171e-06,
+      "loss": 1.876,
+      "step": 896
+    },
+    {
+      "epoch": 0.983013698630137,
+      "grad_norm": 2.212940216064453,
+      "learning_rate": 3.7632006498781482e-06,
+      "loss": 1.8008,
+      "step": 897
+    },
+    {
+      "epoch": 0.9841095890410959,
+      "grad_norm": 2.1541531085968018,
+      "learning_rate": 3.761169780666125e-06,
+      "loss": 1.9039,
+      "step": 898
+    },
+    {
+      "epoch": 0.9852054794520548,
+      "grad_norm": 2.113884925842285,
+      "learning_rate": 3.7591389114541028e-06,
+      "loss": 1.8562,
+      "step": 899
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "grad_norm": 2.1503701210021973,
+      "learning_rate": 3.7571080422420796e-06,
+      "loss": 1.9052,
+      "step": 900
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "eval_loss": 2.0664408206939697,
+      "eval_runtime": 37.0876,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 900
+    },
+    {
+      "epoch": 0.9873972602739726,
+      "grad_norm": 2.1584348678588867,
+      "learning_rate": 3.7550771730300573e-06,
+      "loss": 1.9296,
+      "step": 901
+    },
+    {
+      "epoch": 0.9884931506849315,
+      "grad_norm": 2.164318084716797,
+      "learning_rate": 3.753046303818034e-06,
+      "loss": 1.8834,
+      "step": 902
+    },
+    {
+      "epoch": 0.9895890410958904,
+      "grad_norm": 2.0222678184509277,
+      "learning_rate": 3.751015434606012e-06,
+      "loss": 1.9426,
+      "step": 903
+    },
+    {
+      "epoch": 0.9906849315068493,
+      "grad_norm": 2.116835832595825,
+      "learning_rate": 3.7489845653939887e-06,
+      "loss": 1.9195,
+      "step": 904
+    },
+    {
+      "epoch": 0.9917808219178083,
+      "grad_norm": 2.2708873748779297,
+      "learning_rate": 3.7469536961819664e-06,
+      "loss": 1.9631,
+      "step": 905
+    },
+    {
+      "epoch": 0.9928767123287672,
+      "grad_norm": 2.1174001693725586,
+      "learning_rate": 3.7449228269699433e-06,
+      "loss": 1.8602,
+      "step": 906
+    },
+    {
+      "epoch": 0.993972602739726,
+      "grad_norm": 2.091830015182495,
+      "learning_rate": 3.742891957757921e-06,
+      "loss": 1.8906,
+      "step": 907
+    },
+    {
+      "epoch": 0.9950684931506849,
+      "grad_norm": 2.007119655609131,
+      "learning_rate": 3.740861088545898e-06,
+      "loss": 1.9122,
+      "step": 908
+    },
+    {
+      "epoch": 0.9961643835616438,
+      "grad_norm": 2.070396661758423,
+      "learning_rate": 3.7388302193338755e-06,
+      "loss": 1.9106,
+      "step": 909
+    },
+    {
+      "epoch": 0.9972602739726028,
+      "grad_norm": 2.080331325531006,
+      "learning_rate": 3.7367993501218524e-06,
+      "loss": 1.934,
+      "step": 910
+    },
+    {
+      "epoch": 0.9972602739726028,
+      "eval_loss": 2.0694408416748047,
+      "eval_runtime": 36.9953,
+      "eval_samples_per_second": 27.03,
+      "eval_steps_per_second": 1.135,
+      "step": 910
+    },
+    {
+      "epoch": 0.9983561643835617,
+      "grad_norm": 2.0920615196228027,
+      "learning_rate": 3.7347684809098297e-06,
+      "loss": 1.8679,
+      "step": 911
+    },
+    {
+      "epoch": 0.9994520547945206,
+      "grad_norm": 2.112732172012329,
+      "learning_rate": 3.732737611697807e-06,
+      "loss": 1.8723,
+      "step": 912
+    },
+    {
+      "epoch": 1.0005479452054795,
+      "grad_norm": 2.1449220180511475,
+      "learning_rate": 3.7307067424857842e-06,
+      "loss": 1.8509,
+      "step": 913
+    },
+    {
+      "epoch": 1.0016438356164383,
+      "grad_norm": 2.2179951667785645,
+      "learning_rate": 3.7286758732737615e-06,
+      "loss": 1.7911,
+      "step": 914
+    },
+    {
+      "epoch": 1.0027397260273974,
+      "grad_norm": 2.1392273902893066,
+      "learning_rate": 3.7266450040617388e-06,
+      "loss": 1.8475,
+      "step": 915
+    },
+    {
+      "epoch": 1.0038356164383562,
+      "grad_norm": 2.190265417098999,
+      "learning_rate": 3.724614134849716e-06,
+      "loss": 1.8394,
+      "step": 916
+    },
+    {
+      "epoch": 1.004931506849315,
+      "grad_norm": 2.1833760738372803,
+      "learning_rate": 3.7225832656376933e-06,
+      "loss": 1.9696,
+      "step": 917
+    },
+    {
+      "epoch": 1.006027397260274,
+      "grad_norm": 2.2504308223724365,
+      "learning_rate": 3.7205523964256706e-06,
+      "loss": 1.8135,
+      "step": 918
+    },
+    {
+      "epoch": 1.0071232876712328,
+      "grad_norm": 2.3909389972686768,
+      "learning_rate": 3.718521527213648e-06,
+      "loss": 1.9331,
+      "step": 919
+    },
+    {
+      "epoch": 1.0082191780821919,
+      "grad_norm": 2.3367602825164795,
+      "learning_rate": 3.716490658001625e-06,
+      "loss": 1.9026,
+      "step": 920
+    },
+    {
+      "epoch": 1.0082191780821919,
+      "eval_loss": 2.066087484359741,
+      "eval_runtime": 36.8701,
+      "eval_samples_per_second": 27.122,
+      "eval_steps_per_second": 1.139,
+      "step": 920
+    },
+    {
+      "epoch": 1.0093150684931507,
+      "grad_norm": 2.1732144355773926,
+      "learning_rate": 3.7144597887896024e-06,
+      "loss": 1.8888,
+      "step": 921
+    },
+    {
+      "epoch": 1.0104109589041095,
+      "grad_norm": 2.4632201194763184,
+      "learning_rate": 3.7124289195775797e-06,
+      "loss": 1.7725,
+      "step": 922
+    },
+    {
+      "epoch": 1.0115068493150685,
+      "grad_norm": 2.4887173175811768,
+      "learning_rate": 3.710398050365557e-06,
+      "loss": 1.9125,
+      "step": 923
+    },
+    {
+      "epoch": 1.0126027397260273,
+      "grad_norm": 2.4684619903564453,
+      "learning_rate": 3.708367181153534e-06,
+      "loss": 1.8829,
+      "step": 924
+    },
+    {
+      "epoch": 1.0136986301369864,
+      "grad_norm": 2.348788022994995,
+      "learning_rate": 3.7063363119415115e-06,
+      "loss": 1.8542,
+      "step": 925
+    },
+    {
+      "epoch": 1.0147945205479452,
+      "grad_norm": 2.4432766437530518,
+      "learning_rate": 3.7043054427294884e-06,
+      "loss": 1.903,
+      "step": 926
+    },
+    {
+      "epoch": 1.0158904109589042,
+      "grad_norm": 2.425910234451294,
+      "learning_rate": 3.702274573517466e-06,
+      "loss": 1.8532,
+      "step": 927
+    },
+    {
+      "epoch": 1.016986301369863,
+      "grad_norm": 2.447777032852173,
+      "learning_rate": 3.700243704305443e-06,
+      "loss": 1.8277,
+      "step": 928
+    },
+    {
+      "epoch": 1.0180821917808218,
+      "grad_norm": 2.1830408573150635,
+      "learning_rate": 3.6982128350934206e-06,
+      "loss": 1.8546,
+      "step": 929
+    },
+    {
+      "epoch": 1.0191780821917809,
+      "grad_norm": 2.2692818641662598,
+      "learning_rate": 3.6961819658813975e-06,
+      "loss": 1.8356,
+      "step": 930
+    },
+    {
+      "epoch": 1.0191780821917809,
+      "eval_loss": 2.0714614391326904,
+      "eval_runtime": 36.8661,
+      "eval_samples_per_second": 27.125,
+      "eval_steps_per_second": 1.139,
+      "step": 930
+    },
+    {
+      "epoch": 1.0202739726027397,
+      "grad_norm": 2.5044169425964355,
+      "learning_rate": 3.694151096669375e-06,
+      "loss": 1.8173,
+      "step": 931
+    },
+    {
+      "epoch": 1.0213698630136987,
+      "grad_norm": 2.419637680053711,
+      "learning_rate": 3.692120227457352e-06,
+      "loss": 1.9418,
+      "step": 932
+    },
+    {
+      "epoch": 1.0224657534246575,
+      "grad_norm": 2.282288074493408,
+      "learning_rate": 3.6900893582453297e-06,
+      "loss": 1.8698,
+      "step": 933
+    },
+    {
+      "epoch": 1.0235616438356165,
+      "grad_norm": 2.3215551376342773,
+      "learning_rate": 3.6880584890333066e-06,
+      "loss": 1.8935,
+      "step": 934
+    },
+    {
+      "epoch": 1.0246575342465754,
+      "grad_norm": 2.5938258171081543,
+      "learning_rate": 3.6860276198212834e-06,
+      "loss": 1.79,
+      "step": 935
+    },
+    {
+      "epoch": 1.0257534246575342,
+      "grad_norm": 2.215024471282959,
+      "learning_rate": 3.683996750609261e-06,
+      "loss": 1.8768,
+      "step": 936
+    },
+    {
+      "epoch": 1.0268493150684932,
+      "grad_norm": 2.239609956741333,
+      "learning_rate": 3.681965881397238e-06,
+      "loss": 1.8887,
+      "step": 937
+    },
+    {
+      "epoch": 1.027945205479452,
+      "grad_norm": 2.3134827613830566,
+      "learning_rate": 3.6799350121852157e-06,
+      "loss": 1.8897,
+      "step": 938
+    },
+    {
+      "epoch": 1.029041095890411,
+      "grad_norm": 2.4100260734558105,
+      "learning_rate": 3.6779041429731925e-06,
+      "loss": 1.909,
+      "step": 939
+    },
+    {
+      "epoch": 1.0301369863013699,
+      "grad_norm": 2.26204776763916,
+      "learning_rate": 3.6758732737611702e-06,
+      "loss": 1.8514,
+      "step": 940
+    },
+    {
+      "epoch": 1.0301369863013699,
+      "eval_loss": 2.0666027069091797,
+      "eval_runtime": 36.8336,
+      "eval_samples_per_second": 27.149,
+      "eval_steps_per_second": 1.14,
+      "step": 940
+    },
+    {
+      "epoch": 1.0312328767123287,
+      "grad_norm": 2.22210955619812,
+      "learning_rate": 3.673842404549147e-06,
+      "loss": 1.8743,
+      "step": 941
+    },
+    {
+      "epoch": 1.0323287671232877,
+      "grad_norm": 2.2371468544006348,
+      "learning_rate": 3.671811535337125e-06,
+      "loss": 1.8573,
+      "step": 942
+    },
+    {
+      "epoch": 1.0334246575342465,
+      "grad_norm": 2.346745014190674,
+      "learning_rate": 3.6697806661251016e-06,
+      "loss": 1.8614,
+      "step": 943
+    },
+    {
+      "epoch": 1.0345205479452055,
+      "grad_norm": 2.4765074253082275,
+      "learning_rate": 3.6677497969130793e-06,
+      "loss": 1.8654,
+      "step": 944
+    },
+    {
+      "epoch": 1.0356164383561643,
+      "grad_norm": 2.3023204803466797,
+      "learning_rate": 3.665718927701056e-06,
+      "loss": 1.9901,
+      "step": 945
+    },
+    {
+      "epoch": 1.0367123287671234,
+      "grad_norm": 2.322421073913574,
+      "learning_rate": 3.663688058489034e-06,
+      "loss": 1.8775,
+      "step": 946
+    },
+    {
+      "epoch": 1.0378082191780822,
+      "grad_norm": 2.2324275970458984,
+      "learning_rate": 3.6616571892770107e-06,
+      "loss": 1.9303,
+      "step": 947
+    },
+    {
+      "epoch": 1.038904109589041,
+      "grad_norm": 2.372067928314209,
+      "learning_rate": 3.659626320064988e-06,
+      "loss": 1.8673,
+      "step": 948
+    },
+    {
+      "epoch": 1.04,
+      "grad_norm": 2.2404983043670654,
+      "learning_rate": 3.6575954508529653e-06,
+      "loss": 1.8025,
+      "step": 949
+    },
+    {
+      "epoch": 1.0410958904109588,
+      "grad_norm": 2.180673599243164,
+      "learning_rate": 3.6555645816409426e-06,
+      "loss": 1.8563,
+      "step": 950
+    },
+    {
+      "epoch": 1.0410958904109588,
+      "eval_loss": 2.061262845993042,
+      "eval_runtime": 36.7727,
+      "eval_samples_per_second": 27.194,
+      "eval_steps_per_second": 1.142,
+      "step": 950
+    },
+    {
+      "epoch": 1.0421917808219179,
+      "grad_norm": 2.299043893814087,
+      "learning_rate": 3.65353371242892e-06,
+      "loss": 1.8453,
+      "step": 951
+    },
+    {
+      "epoch": 1.0432876712328767,
+      "grad_norm": 2.4878089427948,
+      "learning_rate": 3.651502843216897e-06,
+      "loss": 1.8901,
+      "step": 952
+    },
+    {
+      "epoch": 1.0443835616438357,
+      "grad_norm": 2.150404930114746,
+      "learning_rate": 3.6494719740048744e-06,
+      "loss": 1.7371,
+      "step": 953
+    },
+    {
+      "epoch": 1.0454794520547945,
+      "grad_norm": 2.177750587463379,
+      "learning_rate": 3.6474411047928517e-06,
+      "loss": 1.8814,
+      "step": 954
+    },
+    {
+      "epoch": 1.0465753424657533,
+      "grad_norm": 2.230797052383423,
+      "learning_rate": 3.645410235580829e-06,
+      "loss": 1.8323,
+      "step": 955
+    },
+    {
+      "epoch": 1.0476712328767124,
+      "grad_norm": 2.2484378814697266,
+      "learning_rate": 3.6433793663688062e-06,
+      "loss": 1.8749,
+      "step": 956
+    },
+    {
+      "epoch": 1.0487671232876712,
+      "grad_norm": 2.201279640197754,
+      "learning_rate": 3.6413484971567835e-06,
+      "loss": 1.8518,
+      "step": 957
+    },
+    {
+      "epoch": 1.0498630136986302,
+      "grad_norm": 2.1125736236572266,
+      "learning_rate": 3.6393176279447608e-06,
+      "loss": 1.846,
+      "step": 958
+    },
+    {
+      "epoch": 1.050958904109589,
+      "grad_norm": 2.3078789710998535,
+      "learning_rate": 3.637286758732738e-06,
+      "loss": 1.8791,
+      "step": 959
+    },
+    {
+      "epoch": 1.0520547945205478,
+      "grad_norm": 2.2911148071289062,
+      "learning_rate": 3.6352558895207153e-06,
+      "loss": 1.8901,
+      "step": 960
+    },
+    {
+      "epoch": 1.0520547945205478,
+      "eval_loss": 2.0583066940307617,
+      "eval_runtime": 36.7535,
+      "eval_samples_per_second": 27.208,
+      "eval_steps_per_second": 1.143,
+      "step": 960
+    },
+    {
+      "epoch": 1.0531506849315069,
+      "grad_norm": 2.3990273475646973,
+      "learning_rate": 3.633225020308692e-06,
+      "loss": 1.9071,
+      "step": 961
+    },
+    {
+      "epoch": 1.0542465753424657,
+      "grad_norm": 2.358685255050659,
+      "learning_rate": 3.63119415109667e-06,
+      "loss": 1.8982,
+      "step": 962
+    },
+    {
+      "epoch": 1.0553424657534247,
+      "grad_norm": 2.1145992279052734,
+      "learning_rate": 3.6291632818846467e-06,
+      "loss": 1.8649,
+      "step": 963
+    },
+    {
+      "epoch": 1.0564383561643835,
+      "grad_norm": 2.2302935123443604,
+      "learning_rate": 3.6271324126726244e-06,
+      "loss": 1.8911,
+      "step": 964
+    },
+    {
+      "epoch": 1.0575342465753426,
+      "grad_norm": 2.344797372817993,
+      "learning_rate": 3.6251015434606013e-06,
+      "loss": 1.8,
+      "step": 965
+    },
+    {
+      "epoch": 1.0586301369863014,
+      "grad_norm": 2.2514073848724365,
+      "learning_rate": 3.623070674248579e-06,
+      "loss": 1.8128,
+      "step": 966
+    },
+    {
+      "epoch": 1.0597260273972602,
+      "grad_norm": 2.541221857070923,
+      "learning_rate": 3.621039805036556e-06,
+      "loss": 1.8504,
+      "step": 967
+    },
+    {
+      "epoch": 1.0608219178082192,
+      "grad_norm": 2.3211770057678223,
+      "learning_rate": 3.6190089358245335e-06,
+      "loss": 1.8329,
+      "step": 968
+    },
+    {
+      "epoch": 1.061917808219178,
+      "grad_norm": 2.1008756160736084,
+      "learning_rate": 3.6169780666125104e-06,
+      "loss": 1.8972,
+      "step": 969
+    },
+    {
+      "epoch": 1.063013698630137,
+      "grad_norm": 2.4148898124694824,
+      "learning_rate": 3.614947197400488e-06,
+      "loss": 1.8748,
+      "step": 970
+    },
+    {
+      "epoch": 1.063013698630137,
+      "eval_loss": 2.0592916011810303,
+      "eval_runtime": 36.7663,
+      "eval_samples_per_second": 27.199,
+      "eval_steps_per_second": 1.142,
+      "step": 970
+    },
+    {
+      "epoch": 1.0641095890410959,
+      "grad_norm": 2.4116220474243164,
+      "learning_rate": 3.612916328188465e-06,
+      "loss": 1.8915,
+      "step": 971
+    },
+    {
+      "epoch": 1.0652054794520547,
+      "grad_norm": 2.257411241531372,
+      "learning_rate": 3.610885458976442e-06,
+      "loss": 1.7901,
+      "step": 972
+    },
+    {
+      "epoch": 1.0663013698630137,
+      "grad_norm": 2.166875123977661,
+      "learning_rate": 3.6088545897644195e-06,
+      "loss": 1.8969,
+      "step": 973
+    },
+    {
+      "epoch": 1.0673972602739725,
+      "grad_norm": 2.464716911315918,
+      "learning_rate": 3.6068237205523963e-06,
+      "loss": 1.8844,
+      "step": 974
+    },
+    {
+      "epoch": 1.0684931506849316,
+      "grad_norm": 2.337334156036377,
+      "learning_rate": 3.604792851340374e-06,
+      "loss": 1.8192,
+      "step": 975
+    },
+    {
+      "epoch": 1.0695890410958904,
+      "grad_norm": 2.117663621902466,
+      "learning_rate": 3.602761982128351e-06,
+      "loss": 1.8233,
+      "step": 976
+    },
+    {
+      "epoch": 1.0706849315068494,
+      "grad_norm": 2.3353586196899414,
+      "learning_rate": 3.6007311129163286e-06,
+      "loss": 1.9133,
+      "step": 977
+    },
+    {
+      "epoch": 1.0717808219178082,
+      "grad_norm": 2.444279909133911,
+      "learning_rate": 3.5987002437043054e-06,
+      "loss": 1.8803,
+      "step": 978
+    },
+    {
+      "epoch": 1.072876712328767,
+      "grad_norm": 2.3458383083343506,
+      "learning_rate": 3.596669374492283e-06,
+      "loss": 1.8108,
+      "step": 979
+    },
+    {
+      "epoch": 1.073972602739726,
+      "grad_norm": 2.207977771759033,
+      "learning_rate": 3.59463850528026e-06,
+      "loss": 1.8423,
+      "step": 980
+    },
+    {
+      "epoch": 1.073972602739726,
+      "eval_loss": 2.0578773021698,
+      "eval_runtime": 36.8123,
+      "eval_samples_per_second": 27.165,
+      "eval_steps_per_second": 1.141,
+      "step": 980
+    },
+    {
+      "epoch": 1.0750684931506849,
+      "grad_norm": 2.378127336502075,
+      "learning_rate": 3.5926076360682377e-06,
+      "loss": 1.9268,
+      "step": 981
+    },
+    {
+      "epoch": 1.076164383561644,
+      "grad_norm": 2.248105525970459,
+      "learning_rate": 3.5905767668562145e-06,
+      "loss": 1.8111,
+      "step": 982
+    },
+    {
+      "epoch": 1.0772602739726027,
+      "grad_norm": 2.275592088699341,
+      "learning_rate": 3.5885458976441922e-06,
+      "loss": 1.9099,
+      "step": 983
+    },
+    {
+      "epoch": 1.0783561643835617,
+      "grad_norm": 3.108377456665039,
+      "learning_rate": 3.586515028432169e-06,
+      "loss": 1.8728,
+      "step": 984
+    },
+    {
+      "epoch": 1.0794520547945206,
+      "grad_norm": 2.2702999114990234,
+      "learning_rate": 3.5844841592201464e-06,
+      "loss": 1.9007,
+      "step": 985
+    },
+    {
+      "epoch": 1.0805479452054794,
+      "grad_norm": 3.218647003173828,
+      "learning_rate": 3.5824532900081237e-06,
+      "loss": 1.9157,
+      "step": 986
+    },
+    {
+      "epoch": 1.0816438356164384,
+      "grad_norm": 2.3393027782440186,
+      "learning_rate": 3.580422420796101e-06,
+      "loss": 1.7936,
+      "step": 987
+    },
+    {
+      "epoch": 1.0827397260273972,
+      "grad_norm": 2.420541286468506,
+      "learning_rate": 3.578391551584078e-06,
+      "loss": 1.8608,
+      "step": 988
+    },
+    {
+      "epoch": 1.0838356164383562,
+      "grad_norm": 2.3354556560516357,
+      "learning_rate": 3.5763606823720555e-06,
+      "loss": 1.8684,
+      "step": 989
+    },
+    {
+      "epoch": 1.084931506849315,
+      "grad_norm": 2.200282573699951,
+      "learning_rate": 3.5743298131600328e-06,
+      "loss": 1.7814,
+      "step": 990
+    },
+    {
+      "epoch": 1.084931506849315,
+      "eval_loss": 2.0565249919891357,
+      "eval_runtime": 36.952,
+      "eval_samples_per_second": 27.062,
+      "eval_steps_per_second": 1.137,
+      "step": 990
+    },
+    {
+      "epoch": 1.086027397260274,
+      "grad_norm": 2.149033308029175,
+      "learning_rate": 3.57229894394801e-06,
+      "loss": 1.8708,
+      "step": 991
+    },
+    {
+      "epoch": 1.087123287671233,
+      "grad_norm": 2.4663679599761963,
+      "learning_rate": 3.5702680747359873e-06,
+      "loss": 1.8947,
+      "step": 992
+    },
+    {
+      "epoch": 1.0882191780821917,
+      "grad_norm": 2.3142261505126953,
+      "learning_rate": 3.5682372055239646e-06,
+      "loss": 1.8727,
+      "step": 993
+    },
+    {
+      "epoch": 1.0893150684931507,
+      "grad_norm": 2.2223033905029297,
+      "learning_rate": 3.566206336311942e-06,
+      "loss": 1.8203,
+      "step": 994
+    },
+    {
+      "epoch": 1.0904109589041096,
+      "grad_norm": 2.423855781555176,
+      "learning_rate": 3.564175467099919e-06,
+      "loss": 1.8687,
+      "step": 995
+    },
+    {
+      "epoch": 1.0915068493150686,
+      "grad_norm": 2.4022302627563477,
+      "learning_rate": 3.5621445978878964e-06,
+      "loss": 1.8215,
+      "step": 996
+    },
+    {
+      "epoch": 1.0926027397260274,
+      "grad_norm": 2.378236770629883,
+      "learning_rate": 3.5601137286758737e-06,
+      "loss": 1.7994,
+      "step": 997
+    },
+    {
+      "epoch": 1.0936986301369862,
+      "grad_norm": 2.4499096870422363,
+      "learning_rate": 3.5580828594638505e-06,
+      "loss": 1.8779,
+      "step": 998
+    },
+    {
+      "epoch": 1.0947945205479452,
+      "grad_norm": 2.3181755542755127,
+      "learning_rate": 3.5560519902518282e-06,
+      "loss": 1.8341,
+      "step": 999
+    },
+    {
+      "epoch": 1.095890410958904,
+      "grad_norm": 2.2406623363494873,
+      "learning_rate": 3.554021121039805e-06,
+      "loss": 1.874,
+      "step": 1000
+    },
+    {
+      "epoch": 1.095890410958904,
+      "eval_loss": 2.0523672103881836,
+      "eval_runtime": 37.0454,
+      "eval_samples_per_second": 26.994,
+      "eval_steps_per_second": 1.134,
+      "step": 1000
+    },
+    {
+      "epoch": 1.096986301369863,
+      "grad_norm": 2.254667043685913,
+      "learning_rate": 3.551990251827783e-06,
+      "loss": 1.9017,
+      "step": 1001
+    },
+    {
+      "epoch": 1.098082191780822,
+      "grad_norm": 2.2846896648406982,
+      "learning_rate": 3.5499593826157596e-06,
+      "loss": 1.8505,
+      "step": 1002
+    },
+    {
+      "epoch": 1.099178082191781,
+      "grad_norm": 2.183469772338867,
+      "learning_rate": 3.5479285134037373e-06,
+      "loss": 1.8266,
+      "step": 1003
+    },
+    {
+      "epoch": 1.1002739726027397,
+      "grad_norm": 2.151798963546753,
+      "learning_rate": 3.545897644191714e-06,
+      "loss": 1.8487,
+      "step": 1004
+    },
+    {
+      "epoch": 1.1013698630136985,
+      "grad_norm": 2.2039146423339844,
+      "learning_rate": 3.543866774979692e-06,
+      "loss": 1.8327,
+      "step": 1005
+    },
+    {
+      "epoch": 1.1024657534246576,
+      "grad_norm": 2.2091665267944336,
+      "learning_rate": 3.5418359057676687e-06,
+      "loss": 1.8214,
+      "step": 1006
+    },
+    {
+      "epoch": 1.1035616438356164,
+      "grad_norm": 2.2821404933929443,
+      "learning_rate": 3.5398050365556464e-06,
+      "loss": 1.8591,
+      "step": 1007
+    },
+    {
+      "epoch": 1.1046575342465754,
+      "grad_norm": 2.1726973056793213,
+      "learning_rate": 3.5377741673436233e-06,
+      "loss": 1.8265,
+      "step": 1008
+    },
+    {
+      "epoch": 1.1057534246575342,
+      "grad_norm": 2.2870233058929443,
+      "learning_rate": 3.535743298131601e-06,
+      "loss": 1.8783,
+      "step": 1009
+    },
+    {
+      "epoch": 1.106849315068493,
+      "grad_norm": 2.148080587387085,
+      "learning_rate": 3.533712428919578e-06,
+      "loss": 1.8272,
+      "step": 1010
+    },
+    {
+      "epoch": 1.106849315068493,
+      "eval_loss": 2.049504041671753,
+      "eval_runtime": 36.8717,
+      "eval_samples_per_second": 27.121,
+      "eval_steps_per_second": 1.139,
+      "step": 1010
+    },
+    {
+      "epoch": 1.107945205479452,
+      "grad_norm": 2.182035446166992,
+      "learning_rate": 3.5316815597075547e-06,
+      "loss": 1.8771,
+      "step": 1011
+    },
+    {
+      "epoch": 1.1090410958904109,
+      "grad_norm": 2.135085344314575,
+      "learning_rate": 3.5296506904955324e-06,
+      "loss": 1.9031,
+      "step": 1012
+    },
+    {
+      "epoch": 1.11013698630137,
+      "grad_norm": 2.1017255783081055,
+      "learning_rate": 3.5276198212835093e-06,
+      "loss": 1.8341,
+      "step": 1013
+    },
+    {
+      "epoch": 1.1112328767123287,
+      "grad_norm": 2.2773115634918213,
+      "learning_rate": 3.525588952071487e-06,
+      "loss": 1.8564,
+      "step": 1014
+    },
+    {
+      "epoch": 1.1123287671232878,
+      "grad_norm": 2.2773115634918213,
+      "learning_rate": 3.525588952071487e-06,
+      "loss": 1.8971,
+      "step": 1015
+    },
+    {
+      "epoch": 1.1134246575342466,
+      "grad_norm": 2.273375988006592,
+      "learning_rate": 3.523558082859464e-06,
+      "loss": 1.9074,
+      "step": 1016
+    },
+    {
+      "epoch": 1.1145205479452054,
+      "grad_norm": 2.2055182456970215,
+      "learning_rate": 3.5215272136474415e-06,
+      "loss": 1.8724,
+      "step": 1017
+    },
+    {
+      "epoch": 1.1156164383561644,
+      "grad_norm": 2.246126651763916,
+      "learning_rate": 3.5194963444354184e-06,
+      "loss": 1.8387,
+      "step": 1018
+    },
+    {
+      "epoch": 1.1167123287671232,
+      "grad_norm": 2.3076183795928955,
+      "learning_rate": 3.517465475223396e-06,
+      "loss": 1.8387,
+      "step": 1019
+    },
+    {
+      "epoch": 1.1178082191780823,
+      "grad_norm": 2.3365488052368164,
+      "learning_rate": 3.515434606011373e-06,
+      "loss": 1.8063,
+      "step": 1020
+    },
+    {
+      "epoch": 1.1178082191780823,
+      "eval_loss": 2.0528781414031982,
+      "eval_runtime": 36.8885,
+      "eval_samples_per_second": 27.109,
+      "eval_steps_per_second": 1.139,
+      "step": 1020
+    },
+    {
+      "epoch": 1.118904109589041,
+      "grad_norm": 2.319324493408203,
+      "learning_rate": 3.5134037367993506e-06,
+      "loss": 1.8843,
+      "step": 1021
+    },
+    {
+      "epoch": 1.12,
+      "grad_norm": 2.238720655441284,
+      "learning_rate": 3.5113728675873275e-06,
+      "loss": 1.9375,
+      "step": 1022
+    },
+    {
+      "epoch": 1.121095890410959,
+      "grad_norm": 2.2824833393096924,
+      "learning_rate": 3.5093419983753047e-06,
+      "loss": 1.8749,
+      "step": 1023
+    },
+    {
+      "epoch": 1.1221917808219177,
+      "grad_norm": 2.5120770931243896,
+      "learning_rate": 3.507311129163282e-06,
+      "loss": 1.8774,
+      "step": 1024
+    },
+    {
+      "epoch": 1.1232876712328768,
+      "grad_norm": 2.1762094497680664,
+      "learning_rate": 3.5052802599512593e-06,
+      "loss": 1.8018,
+      "step": 1025
+    },
+    {
+      "epoch": 1.1243835616438356,
+      "grad_norm": 2.2530343532562256,
+      "learning_rate": 3.5032493907392366e-06,
+      "loss": 1.8451,
+      "step": 1026
+    },
+    {
+      "epoch": 1.1254794520547946,
+      "grad_norm": 2.21649169921875,
+      "learning_rate": 3.501218521527214e-06,
+      "loss": 1.8613,
+      "step": 1027
+    },
+    {
+      "epoch": 1.1265753424657534,
+      "grad_norm": 2.306431293487549,
+      "learning_rate": 3.499187652315191e-06,
+      "loss": 1.7947,
+      "step": 1028
+    },
+    {
+      "epoch": 1.1276712328767124,
+      "grad_norm": 2.203005313873291,
+      "learning_rate": 3.4971567831031684e-06,
+      "loss": 1.8483,
+      "step": 1029
+    },
+    {
+      "epoch": 1.1287671232876713,
+      "grad_norm": 2.1645009517669678,
+      "learning_rate": 3.4951259138911457e-06,
+      "loss": 1.8915,
+      "step": 1030
+    },
+    {
+      "epoch": 1.1287671232876713,
+      "eval_loss": 2.053393602371216,
+      "eval_runtime": 36.855,
+      "eval_samples_per_second": 27.133,
+      "eval_steps_per_second": 1.14,
+      "step": 1030
+    },
+    {
+      "epoch": 1.12986301369863,
+      "grad_norm": 2.2863385677337646,
+      "learning_rate": 3.493095044679123e-06,
+      "loss": 1.8073,
+      "step": 1031
+    },
+    {
+      "epoch": 1.130958904109589,
+      "grad_norm": 2.1439812183380127,
+      "learning_rate": 3.4910641754671002e-06,
+      "loss": 1.9191,
+      "step": 1032
+    },
+    {
+      "epoch": 1.132054794520548,
+      "grad_norm": 2.224228858947754,
+      "learning_rate": 3.4890333062550775e-06,
+      "loss": 1.7478,
+      "step": 1033
+    },
+    {
+      "epoch": 1.133150684931507,
+      "grad_norm": 2.319059133529663,
+      "learning_rate": 3.4870024370430548e-06,
+      "loss": 1.8779,
+      "step": 1034
+    },
+    {
+      "epoch": 1.1342465753424658,
+      "grad_norm": 2.5003139972686768,
+      "learning_rate": 3.484971567831032e-06,
+      "loss": 1.776,
+      "step": 1035
+    },
+    {
+      "epoch": 1.1353424657534246,
+      "grad_norm": 2.2032580375671387,
+      "learning_rate": 3.482940698619009e-06,
+      "loss": 1.8221,
+      "step": 1036
+    },
+    {
+      "epoch": 1.1364383561643836,
+      "grad_norm": 2.5253560543060303,
+      "learning_rate": 3.4809098294069866e-06,
+      "loss": 1.8952,
+      "step": 1037
+    },
+    {
+      "epoch": 1.1375342465753424,
+      "grad_norm": 2.3215854167938232,
+      "learning_rate": 3.4788789601949634e-06,
+      "loss": 1.847,
+      "step": 1038
+    },
+    {
+      "epoch": 1.1386301369863014,
+      "grad_norm": 2.2543280124664307,
+      "learning_rate": 3.476848090982941e-06,
+      "loss": 1.8689,
+      "step": 1039
+    },
+    {
+      "epoch": 1.1397260273972603,
+      "grad_norm": 2.1592042446136475,
+      "learning_rate": 3.474817221770918e-06,
+      "loss": 1.8375,
+      "step": 1040
+    },
+    {
+      "epoch": 1.1397260273972603,
+      "eval_loss": 2.0524044036865234,
+      "eval_runtime": 36.8163,
+      "eval_samples_per_second": 27.162,
+      "eval_steps_per_second": 1.141,
+      "step": 1040
+    },
+    {
+      "epoch": 1.140821917808219,
+      "grad_norm": 2.2874810695648193,
+      "learning_rate": 3.4727863525588957e-06,
+      "loss": 1.9442,
+      "step": 1041
+    },
+    {
+      "epoch": 1.141917808219178,
+      "grad_norm": 2.288395404815674,
+      "learning_rate": 3.4707554833468725e-06,
+      "loss": 1.8881,
+      "step": 1042
+    },
+    {
+      "epoch": 1.143013698630137,
+      "grad_norm": 2.2264153957366943,
+      "learning_rate": 3.4687246141348502e-06,
+      "loss": 1.8518,
+      "step": 1043
+    },
+    {
+      "epoch": 1.144109589041096,
+      "grad_norm": 2.4413652420043945,
+      "learning_rate": 3.466693744922827e-06,
+      "loss": 1.8146,
+      "step": 1044
+    },
+    {
+      "epoch": 1.1452054794520548,
+      "grad_norm": 2.3597898483276367,
+      "learning_rate": 3.464662875710805e-06,
+      "loss": 1.8421,
+      "step": 1045
+    },
+    {
+      "epoch": 1.1463013698630138,
+      "grad_norm": 2.2489402294158936,
+      "learning_rate": 3.4626320064987817e-06,
+      "loss": 1.8643,
+      "step": 1046
+    },
+    {
+      "epoch": 1.1473972602739726,
+      "grad_norm": 2.3958940505981445,
+      "learning_rate": 3.4606011372867594e-06,
+      "loss": 1.8522,
+      "step": 1047
+    },
+    {
+      "epoch": 1.1484931506849314,
+      "grad_norm": 2.295349597930908,
+      "learning_rate": 3.458570268074736e-06,
+      "loss": 1.8677,
+      "step": 1048
+    },
+    {
+      "epoch": 1.1495890410958904,
+      "grad_norm": 2.2571280002593994,
+      "learning_rate": 3.4565393988627135e-06,
+      "loss": 1.8393,
+      "step": 1049
+    },
+    {
+      "epoch": 1.1506849315068493,
+      "grad_norm": 2.173983335494995,
+      "learning_rate": 3.4545085296506908e-06,
+      "loss": 1.8267,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1506849315068493,
+      "eval_loss": 2.0489604473114014,
+      "eval_runtime": 36.7678,
+      "eval_samples_per_second": 27.198,
+      "eval_steps_per_second": 1.142,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1517808219178083,
+      "grad_norm": 2.3395209312438965,
+      "learning_rate": 3.452477660438668e-06,
+      "loss": 1.8032,
+      "step": 1051
+    },
+    {
+      "epoch": 1.152876712328767,
+      "grad_norm": 2.1894757747650146,
+      "learning_rate": 3.4504467912266453e-06,
+      "loss": 1.8834,
+      "step": 1052
+    },
+    {
+      "epoch": 1.1539726027397261,
+      "grad_norm": 2.1978437900543213,
+      "learning_rate": 3.4484159220146226e-06,
+      "loss": 1.7897,
+      "step": 1053
+    },
+    {
+      "epoch": 1.155068493150685,
+      "grad_norm": 2.275792360305786,
+      "learning_rate": 3.4463850528026e-06,
+      "loss": 1.8566,
+      "step": 1054
+    },
+    {
+      "epoch": 1.1561643835616437,
+      "grad_norm": 2.2659800052642822,
+      "learning_rate": 3.444354183590577e-06,
+      "loss": 1.8307,
+      "step": 1055
+    },
+    {
+      "epoch": 1.1572602739726028,
+      "grad_norm": 2.232149600982666,
+      "learning_rate": 3.4423233143785544e-06,
+      "loss": 1.8701,
+      "step": 1056
+    },
+    {
+      "epoch": 1.1583561643835616,
+      "grad_norm": 2.3636281490325928,
+      "learning_rate": 3.4402924451665317e-06,
+      "loss": 1.8008,
+      "step": 1057
+    },
+    {
+      "epoch": 1.1594520547945206,
+      "grad_norm": 2.271533489227295,
+      "learning_rate": 3.438261575954509e-06,
+      "loss": 1.7751,
+      "step": 1058
+    },
+    {
+      "epoch": 1.1605479452054794,
+      "grad_norm": 2.3286566734313965,
+      "learning_rate": 3.4362307067424862e-06,
+      "loss": 1.865,
+      "step": 1059
+    },
+    {
+      "epoch": 1.1616438356164385,
+      "grad_norm": 2.46665096282959,
+      "learning_rate": 3.434199837530463e-06,
+      "loss": 1.9001,
+      "step": 1060
+    },
+    {
+      "epoch": 1.1616438356164385,
+      "eval_loss": 2.0477516651153564,
+      "eval_runtime": 36.7818,
+      "eval_samples_per_second": 27.187,
+      "eval_steps_per_second": 1.142,
+      "step": 1060
+    },
+    {
+      "epoch": 1.1627397260273973,
+      "grad_norm": 2.307744264602661,
+      "learning_rate": 3.432168968318441e-06,
+      "loss": 1.8637,
+      "step": 1061
+    },
+    {
+      "epoch": 1.163835616438356,
+      "grad_norm": 2.3935985565185547,
+      "learning_rate": 3.4301380991064176e-06,
+      "loss": 1.7867,
+      "step": 1062
+    },
+    {
+      "epoch": 1.1649315068493151,
+      "grad_norm": 2.290971040725708,
+      "learning_rate": 3.4281072298943953e-06,
+      "loss": 1.8708,
+      "step": 1063
+    },
+    {
+      "epoch": 1.166027397260274,
+      "grad_norm": 2.191629409790039,
+      "learning_rate": 3.426076360682372e-06,
+      "loss": 1.9031,
+      "step": 1064
+    },
+    {
+      "epoch": 1.167123287671233,
+      "grad_norm": 2.284426689147949,
+      "learning_rate": 3.42404549147035e-06,
+      "loss": 1.9026,
+      "step": 1065
+    },
+    {
+      "epoch": 1.1682191780821918,
+      "grad_norm": 2.2567849159240723,
+      "learning_rate": 3.4220146222583267e-06,
+      "loss": 1.8507,
+      "step": 1066
+    },
+    {
+      "epoch": 1.1693150684931506,
+      "grad_norm": 2.2093160152435303,
+      "learning_rate": 3.4199837530463044e-06,
+      "loss": 1.8001,
+      "step": 1067
+    },
+    {
+      "epoch": 1.1704109589041096,
+      "grad_norm": 2.092560291290283,
+      "learning_rate": 3.4179528838342813e-06,
+      "loss": 1.8393,
+      "step": 1068
+    },
+    {
+      "epoch": 1.1715068493150684,
+      "grad_norm": 2.313749074935913,
+      "learning_rate": 3.415922014622259e-06,
+      "loss": 1.9385,
+      "step": 1069
+    },
+    {
+      "epoch": 1.1726027397260275,
+      "grad_norm": 2.488673210144043,
+      "learning_rate": 3.413891145410236e-06,
+      "loss": 1.8924,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1726027397260275,
+      "eval_loss": 2.0482401847839355,
+      "eval_runtime": 36.7919,
+      "eval_samples_per_second": 27.18,
+      "eval_steps_per_second": 1.142,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1736986301369863,
+      "grad_norm": 2.0804412364959717,
+      "learning_rate": 3.4118602761982135e-06,
+      "loss": 1.881,
+      "step": 1071
+    },
+    {
+      "epoch": 1.174794520547945,
+      "grad_norm": 2.165426254272461,
+      "learning_rate": 3.4098294069861904e-06,
+      "loss": 1.9035,
+      "step": 1072
+    },
+    {
+      "epoch": 1.1758904109589041,
+      "grad_norm": 2.3034675121307373,
+      "learning_rate": 3.4077985377741673e-06,
+      "loss": 1.8554,
+      "step": 1073
+    },
+    {
+      "epoch": 1.176986301369863,
+      "grad_norm": 2.303159713745117,
+      "learning_rate": 3.405767668562145e-06,
+      "loss": 1.8255,
+      "step": 1074
+    },
+    {
+      "epoch": 1.178082191780822,
+      "grad_norm": 2.2402753829956055,
+      "learning_rate": 3.403736799350122e-06,
+      "loss": 1.9087,
+      "step": 1075
+    },
+    {
+      "epoch": 1.1791780821917808,
+      "grad_norm": 2.251638650894165,
+      "learning_rate": 3.4017059301380995e-06,
+      "loss": 1.8878,
+      "step": 1076
+    },
+    {
+      "epoch": 1.1802739726027398,
+      "grad_norm": 2.2661657333374023,
+      "learning_rate": 3.3996750609260764e-06,
+      "loss": 1.8244,
+      "step": 1077
+    },
+    {
+      "epoch": 1.1813698630136986,
+      "grad_norm": 2.3280904293060303,
+      "learning_rate": 3.397644191714054e-06,
+      "loss": 1.9328,
+      "step": 1078
+    },
+    {
+      "epoch": 1.1824657534246574,
+      "grad_norm": 2.315779447555542,
+      "learning_rate": 3.395613322502031e-06,
+      "loss": 1.8289,
+      "step": 1079
+    },
+    {
+      "epoch": 1.1835616438356165,
+      "grad_norm": 2.136967182159424,
+      "learning_rate": 3.3935824532900086e-06,
+      "loss": 1.8307,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1835616438356165,
+      "eval_loss": 2.0473902225494385,
+      "eval_runtime": 36.8094,
+      "eval_samples_per_second": 27.167,
+      "eval_steps_per_second": 1.141,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1846575342465753,
+      "grad_norm": 2.3701720237731934,
+      "learning_rate": 3.3915515840779855e-06,
+      "loss": 1.8123,
+      "step": 1081
+    },
+    {
+      "epoch": 1.1857534246575343,
+      "grad_norm": 2.2662320137023926,
+      "learning_rate": 3.389520714865963e-06,
+      "loss": 1.85,
+      "step": 1082
+    },
+    {
+      "epoch": 1.1868493150684931,
+      "grad_norm": 2.2151782512664795,
+      "learning_rate": 3.38748984565394e-06,
+      "loss": 1.845,
+      "step": 1083
+    },
+    {
+      "epoch": 1.1879452054794521,
+      "grad_norm": 2.2209115028381348,
+      "learning_rate": 3.3854589764419177e-06,
+      "loss": 1.8787,
+      "step": 1084
+    },
+    {
+      "epoch": 1.189041095890411,
+      "grad_norm": 2.4760477542877197,
+      "learning_rate": 3.3834281072298946e-06,
+      "loss": 1.818,
+      "step": 1085
+    },
+    {
+      "epoch": 1.1901369863013698,
+      "grad_norm": 2.309034585952759,
+      "learning_rate": 3.381397238017872e-06,
+      "loss": 1.8387,
+      "step": 1086
+    },
+    {
+      "epoch": 1.1912328767123288,
+      "grad_norm": 2.2174596786499023,
+      "learning_rate": 3.379366368805849e-06,
+      "loss": 1.8135,
+      "step": 1087
+    },
+    {
+      "epoch": 1.1923287671232876,
+      "grad_norm": 2.3043556213378906,
+      "learning_rate": 3.3773354995938264e-06,
+      "loss": 1.9156,
+      "step": 1088
+    },
+    {
+      "epoch": 1.1934246575342466,
+      "grad_norm": 2.2240524291992188,
+      "learning_rate": 3.3753046303818037e-06,
+      "loss": 1.8636,
+      "step": 1089
+    },
+    {
+      "epoch": 1.1945205479452055,
+      "grad_norm": 2.1108691692352295,
+      "learning_rate": 3.373273761169781e-06,
+      "loss": 1.8458,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1945205479452055,
+      "eval_loss": 2.04378604888916,
+      "eval_runtime": 36.9058,
+      "eval_samples_per_second": 27.096,
+      "eval_steps_per_second": 1.138,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1956164383561645,
+      "grad_norm": 2.1893398761749268,
+      "learning_rate": 3.3712428919577582e-06,
+      "loss": 1.8446,
+      "step": 1091
+    },
+    {
+      "epoch": 1.1967123287671233,
+      "grad_norm": 2.305298089981079,
+      "learning_rate": 3.3692120227457355e-06,
+      "loss": 1.8283,
+      "step": 1092
+    },
+    {
+      "epoch": 1.1978082191780821,
+      "grad_norm": 2.364433765411377,
+      "learning_rate": 3.3671811535337128e-06,
+      "loss": 1.8708,
+      "step": 1093
+    },
+    {
+      "epoch": 1.1989041095890411,
+      "grad_norm": 2.259892225265503,
+      "learning_rate": 3.36515028432169e-06,
+      "loss": 1.7966,
+      "step": 1094
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 2.2487263679504395,
+      "learning_rate": 3.3631194151096673e-06,
+      "loss": 1.8238,
+      "step": 1095
+    },
+    {
+      "epoch": 1.201095890410959,
+      "grad_norm": 2.1592557430267334,
+      "learning_rate": 3.3610885458976446e-06,
+      "loss": 1.8651,
+      "step": 1096
+    },
+    {
+      "epoch": 1.2021917808219178,
+      "grad_norm": 2.257737159729004,
+      "learning_rate": 3.359057676685622e-06,
+      "loss": 1.8926,
+      "step": 1097
+    },
+    {
+      "epoch": 1.2032876712328768,
+      "grad_norm": 2.0702707767486572,
+      "learning_rate": 3.357026807473599e-06,
+      "loss": 1.8833,
+      "step": 1098
+    },
+    {
+      "epoch": 1.2043835616438356,
+      "grad_norm": 2.1568455696105957,
+      "learning_rate": 3.354995938261576e-06,
+      "loss": 1.7732,
+      "step": 1099
+    },
+    {
+      "epoch": 1.2054794520547945,
+      "grad_norm": 2.294715404510498,
+      "learning_rate": 3.3529650690495537e-06,
+      "loss": 1.8205,
+      "step": 1100
+    },
+    {
+      "epoch": 1.2054794520547945,
+      "eval_loss": 2.044201612472534,
+      "eval_runtime": 36.9861,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 1100
+    },
+    {
+      "epoch": 1.2065753424657535,
+      "grad_norm": 2.3234570026397705,
+      "learning_rate": 3.3509341998375305e-06,
+      "loss": 1.7935,
+      "step": 1101
+    },
+    {
+      "epoch": 1.2076712328767123,
+      "grad_norm": 2.307224750518799,
+      "learning_rate": 3.3489033306255082e-06,
+      "loss": 1.8254,
+      "step": 1102
+    },
+    {
+      "epoch": 1.208767123287671,
+      "grad_norm": 2.3725457191467285,
+      "learning_rate": 3.346872461413485e-06,
+      "loss": 1.8493,
+      "step": 1103
+    },
+    {
+      "epoch": 1.2098630136986301,
+      "grad_norm": 2.2057435512542725,
+      "learning_rate": 3.344841592201463e-06,
+      "loss": 1.8241,
+      "step": 1104
+    },
+    {
+      "epoch": 1.210958904109589,
+      "grad_norm": 2.232947826385498,
+      "learning_rate": 3.3428107229894397e-06,
+      "loss": 1.8057,
+      "step": 1105
+    },
+    {
+      "epoch": 1.212054794520548,
+      "grad_norm": 2.4268925189971924,
+      "learning_rate": 3.3407798537774174e-06,
+      "loss": 1.8334,
+      "step": 1106
+    },
+    {
+      "epoch": 1.2131506849315068,
+      "grad_norm": 2.3134701251983643,
+      "learning_rate": 3.338748984565394e-06,
+      "loss": 1.8516,
+      "step": 1107
+    },
+    {
+      "epoch": 1.2142465753424658,
+      "grad_norm": 2.375873327255249,
+      "learning_rate": 3.336718115353372e-06,
+      "loss": 1.8942,
+      "step": 1108
+    },
+    {
+      "epoch": 1.2153424657534246,
+      "grad_norm": 2.622528076171875,
+      "learning_rate": 3.3346872461413488e-06,
+      "loss": 1.9027,
+      "step": 1109
+    },
+    {
+      "epoch": 1.2164383561643834,
+      "grad_norm": 2.290865898132324,
+      "learning_rate": 3.3326563769293256e-06,
+      "loss": 1.8258,
+      "step": 1110
+    },
+    {
+      "epoch": 1.2164383561643834,
+      "eval_loss": 2.0408973693847656,
+      "eval_runtime": 36.8596,
+      "eval_samples_per_second": 27.13,
+      "eval_steps_per_second": 1.139,
+      "step": 1110
+    },
+    {
+      "epoch": 1.2175342465753425,
+      "grad_norm": 2.1596908569335938,
+      "learning_rate": 3.3306255077173033e-06,
+      "loss": 1.8066,
+      "step": 1111
+    },
+    {
+      "epoch": 1.2186301369863013,
+      "grad_norm": 2.4652187824249268,
+      "learning_rate": 3.32859463850528e-06,
+      "loss": 1.8372,
+      "step": 1112
+    },
+    {
+      "epoch": 1.2197260273972603,
+      "grad_norm": 2.282440185546875,
+      "learning_rate": 3.326563769293258e-06,
+      "loss": 1.8535,
+      "step": 1113
+    },
+    {
+      "epoch": 1.2208219178082191,
+      "grad_norm": 2.2390308380126953,
+      "learning_rate": 3.3245329000812347e-06,
+      "loss": 1.7719,
+      "step": 1114
+    },
+    {
+      "epoch": 1.2219178082191782,
+      "grad_norm": 2.368565320968628,
+      "learning_rate": 3.3225020308692124e-06,
+      "loss": 1.8207,
+      "step": 1115
+    },
+    {
+      "epoch": 1.223013698630137,
+      "grad_norm": 2.463942766189575,
+      "learning_rate": 3.3204711616571893e-06,
+      "loss": 1.8402,
+      "step": 1116
+    },
+    {
+      "epoch": 1.2241095890410958,
+      "grad_norm": 2.266096830368042,
+      "learning_rate": 3.318440292445167e-06,
+      "loss": 1.7854,
+      "step": 1117
+    },
+    {
+      "epoch": 1.2252054794520548,
+      "grad_norm": 2.3463494777679443,
+      "learning_rate": 3.316409423233144e-06,
+      "loss": 1.8449,
+      "step": 1118
+    },
+    {
+      "epoch": 1.2263013698630136,
+      "grad_norm": 2.381727933883667,
+      "learning_rate": 3.3143785540211215e-06,
+      "loss": 1.8085,
+      "step": 1119
+    },
+    {
+      "epoch": 1.2273972602739727,
+      "grad_norm": 2.3251113891601562,
+      "learning_rate": 3.3123476848090984e-06,
+      "loss": 1.8922,
+      "step": 1120
+    },
+    {
+      "epoch": 1.2273972602739727,
+      "eval_loss": 2.0425429344177246,
+      "eval_runtime": 36.851,
+      "eval_samples_per_second": 27.136,
+      "eval_steps_per_second": 1.14,
+      "step": 1120
+    },
+    {
+      "epoch": 1.2284931506849315,
+      "grad_norm": 2.2314298152923584,
+      "learning_rate": 3.310316815597076e-06,
+      "loss": 1.8474,
+      "step": 1121
+    },
+    {
+      "epoch": 1.2295890410958905,
+      "grad_norm": 2.1847472190856934,
+      "learning_rate": 3.308285946385053e-06,
+      "loss": 1.8454,
+      "step": 1122
+    },
+    {
+      "epoch": 1.2306849315068493,
+      "grad_norm": 2.249422788619995,
+      "learning_rate": 3.30625507717303e-06,
+      "loss": 1.9257,
+      "step": 1123
+    },
+    {
+      "epoch": 1.2317808219178081,
+      "grad_norm": 2.223930835723877,
+      "learning_rate": 3.3042242079610075e-06,
+      "loss": 1.7816,
+      "step": 1124
+    },
+    {
+      "epoch": 1.2328767123287672,
+      "grad_norm": 2.3209826946258545,
+      "learning_rate": 3.3021933387489847e-06,
+      "loss": 1.8773,
+      "step": 1125
+    },
+    {
+      "epoch": 1.233972602739726,
+      "grad_norm": 2.2288055419921875,
+      "learning_rate": 3.300162469536962e-06,
+      "loss": 1.8931,
+      "step": 1126
+    },
+    {
+      "epoch": 1.235068493150685,
+      "grad_norm": 2.201449155807495,
+      "learning_rate": 3.2981316003249393e-06,
+      "loss": 1.8149,
+      "step": 1127
+    },
+    {
+      "epoch": 1.2361643835616438,
+      "grad_norm": 2.215824842453003,
+      "learning_rate": 3.2961007311129166e-06,
+      "loss": 1.8949,
+      "step": 1128
+    },
+    {
+      "epoch": 1.2372602739726029,
+      "grad_norm": 2.133671522140503,
+      "learning_rate": 3.294069861900894e-06,
+      "loss": 1.8669,
+      "step": 1129
+    },
+    {
+      "epoch": 1.2383561643835617,
+      "grad_norm": 2.2195792198181152,
+      "learning_rate": 3.292038992688871e-06,
+      "loss": 1.9245,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2383561643835617,
+      "eval_loss": 2.0442771911621094,
+      "eval_runtime": 36.8077,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2394520547945205,
+      "grad_norm": 2.310866117477417,
+      "learning_rate": 3.2900081234768484e-06,
+      "loss": 1.8523,
+      "step": 1131
+    },
+    {
+      "epoch": 1.2405479452054795,
+      "grad_norm": 2.318300247192383,
+      "learning_rate": 3.2879772542648257e-06,
+      "loss": 1.8665,
+      "step": 1132
+    },
+    {
+      "epoch": 1.2416438356164383,
+      "grad_norm": 2.082564115524292,
+      "learning_rate": 3.285946385052803e-06,
+      "loss": 1.8962,
+      "step": 1133
+    },
+    {
+      "epoch": 1.2427397260273974,
+      "grad_norm": 2.1982104778289795,
+      "learning_rate": 3.2839155158407802e-06,
+      "loss": 1.7722,
+      "step": 1134
+    },
+    {
+      "epoch": 1.2438356164383562,
+      "grad_norm": 2.330782890319824,
+      "learning_rate": 3.2818846466287575e-06,
+      "loss": 1.8787,
+      "step": 1135
+    },
+    {
+      "epoch": 1.244931506849315,
+      "grad_norm": 2.136747121810913,
+      "learning_rate": 3.2798537774167344e-06,
+      "loss": 1.7739,
+      "step": 1136
+    },
+    {
+      "epoch": 1.246027397260274,
+      "grad_norm": 2.205862283706665,
+      "learning_rate": 3.277822908204712e-06,
+      "loss": 1.9127,
+      "step": 1137
+    },
+    {
+      "epoch": 1.2471232876712328,
+      "grad_norm": 2.2060093879699707,
+      "learning_rate": 3.275792038992689e-06,
+      "loss": 1.8379,
+      "step": 1138
+    },
+    {
+      "epoch": 1.2482191780821918,
+      "grad_norm": 2.255772829055786,
+      "learning_rate": 3.2737611697806666e-06,
+      "loss": 1.7931,
+      "step": 1139
+    },
+    {
+      "epoch": 1.2493150684931507,
+      "grad_norm": 2.240095615386963,
+      "learning_rate": 3.2717303005686435e-06,
+      "loss": 1.8466,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2493150684931507,
+      "eval_loss": 2.0429375171661377,
+      "eval_runtime": 36.7983,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2504109589041095,
+      "grad_norm": 2.1491949558258057,
+      "learning_rate": 3.269699431356621e-06,
+      "loss": 1.8137,
+      "step": 1141
+    },
+    {
+      "epoch": 1.2515068493150685,
+      "grad_norm": 2.3854610919952393,
+      "learning_rate": 3.267668562144598e-06,
+      "loss": 1.8486,
+      "step": 1142
+    },
+    {
+      "epoch": 1.2526027397260273,
+      "grad_norm": 2.3498709201812744,
+      "learning_rate": 3.2656376929325757e-06,
+      "loss": 1.837,
+      "step": 1143
+    },
+    {
+      "epoch": 1.2536986301369863,
+      "grad_norm": 2.1651649475097656,
+      "learning_rate": 3.2636068237205526e-06,
+      "loss": 1.8,
+      "step": 1144
+    },
+    {
+      "epoch": 1.2547945205479452,
+      "grad_norm": 2.2500200271606445,
+      "learning_rate": 3.2615759545085303e-06,
+      "loss": 1.8691,
+      "step": 1145
+    },
+    {
+      "epoch": 1.2558904109589042,
+      "grad_norm": 2.2178192138671875,
+      "learning_rate": 3.259545085296507e-06,
+      "loss": 1.8345,
+      "step": 1146
+    },
+    {
+      "epoch": 1.256986301369863,
+      "grad_norm": 2.238114595413208,
+      "learning_rate": 3.257514216084484e-06,
+      "loss": 1.8935,
+      "step": 1147
+    },
+    {
+      "epoch": 1.2580821917808218,
+      "grad_norm": 2.37042498588562,
+      "learning_rate": 3.2554833468724617e-06,
+      "loss": 1.7872,
+      "step": 1148
+    },
+    {
+      "epoch": 1.2591780821917808,
+      "grad_norm": 2.1831908226013184,
+      "learning_rate": 3.2534524776604385e-06,
+      "loss": 1.8913,
+      "step": 1149
+    },
+    {
+      "epoch": 1.2602739726027397,
+      "grad_norm": 2.251382827758789,
+      "learning_rate": 3.2514216084484162e-06,
+      "loss": 1.8611,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2602739726027397,
+      "eval_loss": 2.041306257247925,
+      "eval_runtime": 36.8199,
+      "eval_samples_per_second": 27.159,
+      "eval_steps_per_second": 1.141,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2613698630136987,
+      "grad_norm": 2.2690420150756836,
+      "learning_rate": 3.249390739236393e-06,
+      "loss": 1.8283,
+      "step": 1151
+    },
+    {
+      "epoch": 1.2624657534246575,
+      "grad_norm": 2.255833387374878,
+      "learning_rate": 3.2473598700243708e-06,
+      "loss": 1.8563,
+      "step": 1152
+    },
+    {
+      "epoch": 1.2635616438356165,
+      "grad_norm": 2.1362855434417725,
+      "learning_rate": 3.2453290008123476e-06,
+      "loss": 1.809,
+      "step": 1153
+    },
+    {
+      "epoch": 1.2646575342465753,
+      "grad_norm": 2.1227219104766846,
+      "learning_rate": 3.2432981316003253e-06,
+      "loss": 1.848,
+      "step": 1154
+    },
+    {
+      "epoch": 1.2657534246575342,
+      "grad_norm": 2.3639719486236572,
+      "learning_rate": 3.241267262388302e-06,
+      "loss": 1.8472,
+      "step": 1155
+    },
+    {
+      "epoch": 1.2668493150684932,
+      "grad_norm": 2.334404468536377,
+      "learning_rate": 3.23923639317628e-06,
+      "loss": 1.8073,
+      "step": 1156
+    },
+    {
+      "epoch": 1.267945205479452,
+      "grad_norm": 2.231656312942505,
+      "learning_rate": 3.2372055239642567e-06,
+      "loss": 1.8446,
+      "step": 1157
+    },
+    {
+      "epoch": 1.269041095890411,
+      "grad_norm": 2.371375560760498,
+      "learning_rate": 3.2351746547522344e-06,
+      "loss": 1.8002,
+      "step": 1158
+    },
+    {
+      "epoch": 1.2701369863013698,
+      "grad_norm": 2.5212879180908203,
+      "learning_rate": 3.2331437855402113e-06,
+      "loss": 1.863,
+      "step": 1159
+    },
+    {
+      "epoch": 1.2712328767123289,
+      "grad_norm": 2.3285911083221436,
+      "learning_rate": 3.2311129163281885e-06,
+      "loss": 1.9101,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2712328767123289,
+      "eval_loss": 2.038487195968628,
+      "eval_runtime": 36.8002,
+      "eval_samples_per_second": 27.174,
+      "eval_steps_per_second": 1.141,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2723287671232877,
+      "grad_norm": 2.1316046714782715,
+      "learning_rate": 3.229082047116166e-06,
+      "loss": 1.7952,
+      "step": 1161
+    },
+    {
+      "epoch": 1.2734246575342465,
+      "grad_norm": 2.4078125953674316,
+      "learning_rate": 3.227051177904143e-06,
+      "loss": 1.869,
+      "step": 1162
+    },
+    {
+      "epoch": 1.2745205479452055,
+      "grad_norm": 2.363672971725464,
+      "learning_rate": 3.2250203086921204e-06,
+      "loss": 1.8478,
+      "step": 1163
+    },
+    {
+      "epoch": 1.2756164383561643,
+      "grad_norm": 2.240558385848999,
+      "learning_rate": 3.2229894394800977e-06,
+      "loss": 1.8483,
+      "step": 1164
+    },
+    {
+      "epoch": 1.2767123287671232,
+      "grad_norm": 2.9142000675201416,
+      "learning_rate": 3.220958570268075e-06,
+      "loss": 1.8789,
+      "step": 1165
+    },
+    {
+      "epoch": 1.2778082191780822,
+      "grad_norm": 2.1930558681488037,
+      "learning_rate": 3.218927701056052e-06,
+      "loss": 1.7659,
+      "step": 1166
+    },
+    {
+      "epoch": 1.2789041095890412,
+      "grad_norm": 2.312471866607666,
+      "learning_rate": 3.2168968318440295e-06,
+      "loss": 1.8374,
+      "step": 1167
+    },
+    {
+      "epoch": 1.28,
+      "grad_norm": 2.6222236156463623,
+      "learning_rate": 3.2148659626320068e-06,
+      "loss": 1.7539,
+      "step": 1168
+    },
+    {
+      "epoch": 1.2810958904109588,
+      "grad_norm": 2.3013322353363037,
+      "learning_rate": 3.212835093419984e-06,
+      "loss": 1.9011,
+      "step": 1169
+    },
+    {
+      "epoch": 1.2821917808219179,
+      "grad_norm": 2.215240716934204,
+      "learning_rate": 3.2108042242079613e-06,
+      "loss": 1.8032,
+      "step": 1170
+    },
+    {
+      "epoch": 1.2821917808219179,
+      "eval_loss": 2.0365872383117676,
+      "eval_runtime": 36.8265,
+      "eval_samples_per_second": 27.154,
+      "eval_steps_per_second": 1.14,
+      "step": 1170
+    },
+    {
+      "epoch": 1.2832876712328767,
+      "grad_norm": 2.3620481491088867,
+      "learning_rate": 3.2087733549959386e-06,
+      "loss": 1.8258,
+      "step": 1171
+    },
+    {
+      "epoch": 1.2843835616438355,
+      "grad_norm": 2.2650208473205566,
+      "learning_rate": 3.206742485783916e-06,
+      "loss": 1.8888,
+      "step": 1172
+    },
+    {
+      "epoch": 1.2854794520547945,
+      "grad_norm": 2.2564077377319336,
+      "learning_rate": 3.2047116165718927e-06,
+      "loss": 1.7874,
+      "step": 1173
+    },
+    {
+      "epoch": 1.2865753424657536,
+      "grad_norm": 2.2855281829833984,
+      "learning_rate": 3.2026807473598704e-06,
+      "loss": 1.8003,
+      "step": 1174
+    },
+    {
+      "epoch": 1.2876712328767124,
+      "grad_norm": 2.353893756866455,
+      "learning_rate": 3.2006498781478473e-06,
+      "loss": 1.8466,
+      "step": 1175
+    },
+    {
+      "epoch": 1.2887671232876712,
+      "grad_norm": 2.1893045902252197,
+      "learning_rate": 3.198619008935825e-06,
+      "loss": 1.8652,
+      "step": 1176
+    },
+    {
+      "epoch": 1.2898630136986302,
+      "grad_norm": 2.6629793643951416,
+      "learning_rate": 3.196588139723802e-06,
+      "loss": 1.797,
+      "step": 1177
+    },
+    {
+      "epoch": 1.290958904109589,
+      "grad_norm": 2.2427361011505127,
+      "learning_rate": 3.1945572705117795e-06,
+      "loss": 1.8294,
+      "step": 1178
+    },
+    {
+      "epoch": 1.2920547945205478,
+      "grad_norm": 2.321878671646118,
+      "learning_rate": 3.1925264012997564e-06,
+      "loss": 1.8494,
+      "step": 1179
+    },
+    {
+      "epoch": 1.2931506849315069,
+      "grad_norm": 2.6928281784057617,
+      "learning_rate": 3.190495532087734e-06,
+      "loss": 1.8574,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2931506849315069,
+      "eval_loss": 2.0426652431488037,
+      "eval_runtime": 36.8482,
+      "eval_samples_per_second": 27.138,
+      "eval_steps_per_second": 1.14,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2942465753424657,
+      "grad_norm": 2.382164239883423,
+      "learning_rate": 3.188464662875711e-06,
+      "loss": 1.8518,
+      "step": 1181
+    },
+    {
+      "epoch": 1.2953424657534247,
+      "grad_norm": 2.4109256267547607,
+      "learning_rate": 3.1864337936636886e-06,
+      "loss": 1.9083,
+      "step": 1182
+    },
+    {
+      "epoch": 1.2964383561643835,
+      "grad_norm": 2.468062400817871,
+      "learning_rate": 3.1844029244516655e-06,
+      "loss": 1.8247,
+      "step": 1183
+    },
+    {
+      "epoch": 1.2975342465753426,
+      "grad_norm": 2.459378242492676,
+      "learning_rate": 3.182372055239643e-06,
+      "loss": 1.791,
+      "step": 1184
+    },
+    {
+      "epoch": 1.2986301369863014,
+      "grad_norm": 2.3331072330474854,
+      "learning_rate": 3.18034118602762e-06,
+      "loss": 1.7869,
+      "step": 1185
+    },
+    {
+      "epoch": 1.2997260273972602,
+      "grad_norm": 2.339428424835205,
+      "learning_rate": 3.1783103168155973e-06,
+      "loss": 1.8729,
+      "step": 1186
+    },
+    {
+      "epoch": 1.3008219178082192,
+      "grad_norm": 2.3033406734466553,
+      "learning_rate": 3.1762794476035746e-06,
+      "loss": 1.8455,
+      "step": 1187
+    },
+    {
+      "epoch": 1.301917808219178,
+      "grad_norm": 2.2480714321136475,
+      "learning_rate": 3.174248578391552e-06,
+      "loss": 1.8233,
+      "step": 1188
+    },
+    {
+      "epoch": 1.303013698630137,
+      "grad_norm": 2.2953624725341797,
+      "learning_rate": 3.172217709179529e-06,
+      "loss": 1.8261,
+      "step": 1189
+    },
+    {
+      "epoch": 1.3041095890410959,
+      "grad_norm": 2.494307518005371,
+      "learning_rate": 3.1701868399675064e-06,
+      "loss": 1.8153,
+      "step": 1190
+    },
+    {
+      "epoch": 1.3041095890410959,
+      "eval_loss": 2.037719964981079,
+      "eval_runtime": 36.9788,
+      "eval_samples_per_second": 27.043,
+      "eval_steps_per_second": 1.136,
+      "step": 1190
+    },
+    {
+      "epoch": 1.305205479452055,
+      "grad_norm": 2.314159631729126,
+      "learning_rate": 3.1681559707554837e-06,
+      "loss": 1.8406,
+      "step": 1191
+    },
+    {
+      "epoch": 1.3063013698630137,
+      "grad_norm": 2.4262588024139404,
+      "learning_rate": 3.166125101543461e-06,
+      "loss": 1.7803,
+      "step": 1192
+    },
+    {
+      "epoch": 1.3073972602739725,
+      "grad_norm": 2.3800835609436035,
+      "learning_rate": 3.1640942323314382e-06,
+      "loss": 1.8521,
+      "step": 1193
+    },
+    {
+      "epoch": 1.3084931506849315,
+      "grad_norm": 2.234551429748535,
+      "learning_rate": 3.1620633631194155e-06,
+      "loss": 1.845,
+      "step": 1194
+    },
+    {
+      "epoch": 1.3095890410958904,
+      "grad_norm": 2.167635440826416,
+      "learning_rate": 3.1600324939073928e-06,
+      "loss": 1.789,
+      "step": 1195
+    },
+    {
+      "epoch": 1.3106849315068494,
+      "grad_norm": 2.320247173309326,
+      "learning_rate": 3.15800162469537e-06,
+      "loss": 1.8311,
+      "step": 1196
+    },
+    {
+      "epoch": 1.3117808219178082,
+      "grad_norm": 2.2715561389923096,
+      "learning_rate": 3.155970755483347e-06,
+      "loss": 1.8308,
+      "step": 1197
+    },
+    {
+      "epoch": 1.3128767123287672,
+      "grad_norm": 2.2535455226898193,
+      "learning_rate": 3.1539398862713246e-06,
+      "loss": 1.8358,
+      "step": 1198
+    },
+    {
+      "epoch": 1.313972602739726,
+      "grad_norm": 2.2399163246154785,
+      "learning_rate": 3.1519090170593015e-06,
+      "loss": 1.8513,
+      "step": 1199
+    },
+    {
+      "epoch": 1.3150684931506849,
+      "grad_norm": 2.2984132766723633,
+      "learning_rate": 3.149878147847279e-06,
+      "loss": 1.8377,
+      "step": 1200
+    },
+    {
+      "epoch": 1.3150684931506849,
+      "eval_loss": 2.0334064960479736,
+      "eval_runtime": 37.0346,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1200
+    },
+    {
+      "epoch": 1.316164383561644,
+      "grad_norm": 2.1296892166137695,
+      "learning_rate": 3.147847278635256e-06,
+      "loss": 1.9023,
+      "step": 1201
+    },
+    {
+      "epoch": 1.3172602739726027,
+      "grad_norm": 2.1034350395202637,
+      "learning_rate": 3.1458164094232337e-06,
+      "loss": 1.917,
+      "step": 1202
+    },
+    {
+      "epoch": 1.3183561643835615,
+      "grad_norm": 2.258863687515259,
+      "learning_rate": 3.1437855402112106e-06,
+      "loss": 1.8093,
+      "step": 1203
+    },
+    {
+      "epoch": 1.3194520547945205,
+      "grad_norm": 2.20119309425354,
+      "learning_rate": 3.1417546709991883e-06,
+      "loss": 1.7796,
+      "step": 1204
+    },
+    {
+      "epoch": 1.3205479452054796,
+      "grad_norm": 2.3005869388580322,
+      "learning_rate": 3.139723801787165e-06,
+      "loss": 1.8496,
+      "step": 1205
+    },
+    {
+      "epoch": 1.3216438356164384,
+      "grad_norm": 2.2009031772613525,
+      "learning_rate": 3.137692932575143e-06,
+      "loss": 1.8244,
+      "step": 1206
+    },
+    {
+      "epoch": 1.3227397260273972,
+      "grad_norm": 2.3467633724212646,
+      "learning_rate": 3.1356620633631197e-06,
+      "loss": 1.8803,
+      "step": 1207
+    },
+    {
+      "epoch": 1.3238356164383562,
+      "grad_norm": 2.318711519241333,
+      "learning_rate": 3.1336311941510974e-06,
+      "loss": 1.7999,
+      "step": 1208
+    },
+    {
+      "epoch": 1.324931506849315,
+      "grad_norm": 2.182602882385254,
+      "learning_rate": 3.1316003249390742e-06,
+      "loss": 1.836,
+      "step": 1209
+    },
+    {
+      "epoch": 1.3260273972602739,
+      "grad_norm": 2.217298984527588,
+      "learning_rate": 3.129569455727051e-06,
+      "loss": 1.8776,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3260273972602739,
+      "eval_loss": 2.033573627471924,
+      "eval_runtime": 36.8029,
+      "eval_samples_per_second": 27.172,
+      "eval_steps_per_second": 1.141,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3271232876712329,
+      "grad_norm": 2.2814879417419434,
+      "learning_rate": 3.1275385865150288e-06,
+      "loss": 1.8739,
+      "step": 1211
+    },
+    {
+      "epoch": 1.3282191780821917,
+      "grad_norm": 2.1692471504211426,
+      "learning_rate": 3.1255077173030056e-06,
+      "loss": 1.8539,
+      "step": 1212
+    },
+    {
+      "epoch": 1.3293150684931507,
+      "grad_norm": 2.2472000122070312,
+      "learning_rate": 3.1234768480909833e-06,
+      "loss": 1.8527,
+      "step": 1213
+    },
+    {
+      "epoch": 1.3304109589041095,
+      "grad_norm": 2.1969852447509766,
+      "learning_rate": 3.12144597887896e-06,
+      "loss": 1.8429,
+      "step": 1214
+    },
+    {
+      "epoch": 1.3315068493150686,
+      "grad_norm": 2.141629219055176,
+      "learning_rate": 3.119415109666938e-06,
+      "loss": 1.868,
+      "step": 1215
+    },
+    {
+      "epoch": 1.3326027397260274,
+      "grad_norm": 2.282562494277954,
+      "learning_rate": 3.1173842404549147e-06,
+      "loss": 1.8913,
+      "step": 1216
+    },
+    {
+      "epoch": 1.3336986301369862,
+      "grad_norm": 2.2375850677490234,
+      "learning_rate": 3.1153533712428924e-06,
+      "loss": 1.8462,
+      "step": 1217
+    },
+    {
+      "epoch": 1.3347945205479452,
+      "grad_norm": 2.3529837131500244,
+      "learning_rate": 3.1133225020308693e-06,
+      "loss": 1.8445,
+      "step": 1218
+    },
+    {
+      "epoch": 1.335890410958904,
+      "grad_norm": 2.385857343673706,
+      "learning_rate": 3.111291632818847e-06,
+      "loss": 1.89,
+      "step": 1219
+    },
+    {
+      "epoch": 1.336986301369863,
+      "grad_norm": 2.301253080368042,
+      "learning_rate": 3.109260763606824e-06,
+      "loss": 1.8882,
+      "step": 1220
+    },
+    {
+      "epoch": 1.336986301369863,
+      "eval_loss": 2.040025234222412,
+      "eval_runtime": 36.839,
+      "eval_samples_per_second": 27.145,
+      "eval_steps_per_second": 1.14,
+      "step": 1220
+    },
+    {
+      "epoch": 1.3380821917808219,
+      "grad_norm": 2.3772335052490234,
+      "learning_rate": 3.1072298943948015e-06,
+      "loss": 1.8545,
+      "step": 1221
+    },
+    {
+      "epoch": 1.339178082191781,
+      "grad_norm": 2.387830972671509,
+      "learning_rate": 3.1051990251827784e-06,
+      "loss": 1.8653,
+      "step": 1222
+    },
+    {
+      "epoch": 1.3402739726027397,
+      "grad_norm": 2.4285190105438232,
+      "learning_rate": 3.1031681559707557e-06,
+      "loss": 1.8268,
+      "step": 1223
+    },
+    {
+      "epoch": 1.3413698630136985,
+      "grad_norm": 2.1892075538635254,
+      "learning_rate": 3.101137286758733e-06,
+      "loss": 1.8818,
+      "step": 1224
+    },
+    {
+      "epoch": 1.3424657534246576,
+      "grad_norm": 2.2510695457458496,
+      "learning_rate": 3.09910641754671e-06,
+      "loss": 1.879,
+      "step": 1225
+    },
+    {
+      "epoch": 1.3435616438356164,
+      "grad_norm": 2.505671977996826,
+      "learning_rate": 3.0970755483346875e-06,
+      "loss": 1.872,
+      "step": 1226
+    },
+    {
+      "epoch": 1.3446575342465754,
+      "grad_norm": 2.5315067768096924,
+      "learning_rate": 3.0950446791226648e-06,
+      "loss": 1.8651,
+      "step": 1227
+    },
+    {
+      "epoch": 1.3457534246575342,
+      "grad_norm": 2.223600387573242,
+      "learning_rate": 3.093013809910642e-06,
+      "loss": 1.9167,
+      "step": 1228
+    },
+    {
+      "epoch": 1.3468493150684933,
+      "grad_norm": 2.357802391052246,
+      "learning_rate": 3.0909829406986193e-06,
+      "loss": 1.8593,
+      "step": 1229
+    },
+    {
+      "epoch": 1.347945205479452,
+      "grad_norm": 2.3102364540100098,
+      "learning_rate": 3.0889520714865966e-06,
+      "loss": 1.8994,
+      "step": 1230
+    },
+    {
+      "epoch": 1.347945205479452,
+      "eval_loss": 2.0332930088043213,
+      "eval_runtime": 36.989,
+      "eval_samples_per_second": 27.035,
+      "eval_steps_per_second": 1.135,
+      "step": 1230
+    },
+    {
+      "epoch": 1.3490410958904109,
+      "grad_norm": 2.085080623626709,
+      "learning_rate": 3.086921202274574e-06,
+      "loss": 1.8741,
+      "step": 1231
+    },
+    {
+      "epoch": 1.35013698630137,
+      "grad_norm": 2.267305612564087,
+      "learning_rate": 3.084890333062551e-06,
+      "loss": 1.8426,
+      "step": 1232
+    },
+    {
+      "epoch": 1.3512328767123287,
+      "grad_norm": 2.181033134460449,
+      "learning_rate": 3.0828594638505284e-06,
+      "loss": 1.8199,
+      "step": 1233
+    },
+    {
+      "epoch": 1.3523287671232875,
+      "grad_norm": 2.288386821746826,
+      "learning_rate": 3.0808285946385057e-06,
+      "loss": 1.8469,
+      "step": 1234
+    },
+    {
+      "epoch": 1.3534246575342466,
+      "grad_norm": 2.1516127586364746,
+      "learning_rate": 3.078797725426483e-06,
+      "loss": 1.8056,
+      "step": 1235
+    },
+    {
+      "epoch": 1.3545205479452056,
+      "grad_norm": 2.2699198722839355,
+      "learning_rate": 3.07676685621446e-06,
+      "loss": 1.7535,
+      "step": 1236
+    },
+    {
+      "epoch": 1.3556164383561644,
+      "grad_norm": 2.193251132965088,
+      "learning_rate": 3.0747359870024375e-06,
+      "loss": 1.8407,
+      "step": 1237
+    },
+    {
+      "epoch": 1.3567123287671232,
+      "grad_norm": 2.2217774391174316,
+      "learning_rate": 3.0727051177904144e-06,
+      "loss": 1.8062,
+      "step": 1238
+    },
+    {
+      "epoch": 1.3578082191780823,
+      "grad_norm": 2.2376203536987305,
+      "learning_rate": 3.070674248578392e-06,
+      "loss": 1.829,
+      "step": 1239
+    },
+    {
+      "epoch": 1.358904109589041,
+      "grad_norm": 2.387559413909912,
+      "learning_rate": 3.068643379366369e-06,
+      "loss": 1.8459,
+      "step": 1240
+    },
+    {
+      "epoch": 1.358904109589041,
+      "eval_loss": 2.029298782348633,
+      "eval_runtime": 37.0491,
+      "eval_samples_per_second": 26.991,
+      "eval_steps_per_second": 1.134,
+      "step": 1240
+    },
+    {
+      "epoch": 1.3599999999999999,
+      "grad_norm": 2.1940178871154785,
+      "learning_rate": 3.0666125101543466e-06,
+      "loss": 1.831,
+      "step": 1241
+    },
+    {
+      "epoch": 1.361095890410959,
+      "grad_norm": 2.145514488220215,
+      "learning_rate": 3.0645816409423235e-06,
+      "loss": 1.7919,
+      "step": 1242
+    },
+    {
+      "epoch": 1.362191780821918,
+      "grad_norm": 2.319164276123047,
+      "learning_rate": 3.062550771730301e-06,
+      "loss": 1.8472,
+      "step": 1243
+    },
+    {
+      "epoch": 1.3632876712328768,
+      "grad_norm": 2.218986749649048,
+      "learning_rate": 3.060519902518278e-06,
+      "loss": 1.8184,
+      "step": 1244
+    },
+    {
+      "epoch": 1.3643835616438356,
+      "grad_norm": 2.3755338191986084,
+      "learning_rate": 3.0584890333062557e-06,
+      "loss": 1.8166,
+      "step": 1245
+    },
+    {
+      "epoch": 1.3654794520547946,
+      "grad_norm": 2.3186213970184326,
+      "learning_rate": 3.0564581640942326e-06,
+      "loss": 1.7599,
+      "step": 1246
+    },
+    {
+      "epoch": 1.3665753424657534,
+      "grad_norm": 2.193958282470703,
+      "learning_rate": 3.0544272948822094e-06,
+      "loss": 1.8404,
+      "step": 1247
+    },
+    {
+      "epoch": 1.3676712328767122,
+      "grad_norm": 2.40419602394104,
+      "learning_rate": 3.052396425670187e-06,
+      "loss": 1.821,
+      "step": 1248
+    },
+    {
+      "epoch": 1.3687671232876713,
+      "grad_norm": 2.445542573928833,
+      "learning_rate": 3.050365556458164e-06,
+      "loss": 1.8518,
+      "step": 1249
+    },
+    {
+      "epoch": 1.36986301369863,
+      "grad_norm": 2.2776567935943604,
+      "learning_rate": 3.0483346872461417e-06,
+      "loss": 1.8121,
+      "step": 1250
+    },
+    {
+      "epoch": 1.36986301369863,
+      "eval_loss": 2.030043125152588,
+      "eval_runtime": 37.0721,
+      "eval_samples_per_second": 26.974,
+      "eval_steps_per_second": 1.133,
+      "step": 1250
+    },
+    {
+      "epoch": 1.370958904109589,
+      "grad_norm": 2.420448064804077,
+      "learning_rate": 3.0463038180341185e-06,
+      "loss": 1.8168,
+      "step": 1251
+    },
+    {
+      "epoch": 1.372054794520548,
+      "grad_norm": 2.252065658569336,
+      "learning_rate": 3.0442729488220962e-06,
+      "loss": 1.8757,
+      "step": 1252
+    },
+    {
+      "epoch": 1.373150684931507,
+      "grad_norm": 2.4772560596466064,
+      "learning_rate": 3.042242079610073e-06,
+      "loss": 1.8711,
+      "step": 1253
+    },
+    {
+      "epoch": 1.3742465753424657,
+      "grad_norm": 2.3986129760742188,
+      "learning_rate": 3.0402112103980508e-06,
+      "loss": 1.8636,
+      "step": 1254
+    },
+    {
+      "epoch": 1.3753424657534246,
+      "grad_norm": 2.257274866104126,
+      "learning_rate": 3.0381803411860276e-06,
+      "loss": 1.7858,
+      "step": 1255
+    },
+    {
+      "epoch": 1.3764383561643836,
+      "grad_norm": 2.2373499870300293,
+      "learning_rate": 3.0361494719740053e-06,
+      "loss": 1.817,
+      "step": 1256
+    },
+    {
+      "epoch": 1.3775342465753424,
+      "grad_norm": 2.292388916015625,
+      "learning_rate": 3.034118602761982e-06,
+      "loss": 1.827,
+      "step": 1257
+    },
+    {
+      "epoch": 1.3786301369863014,
+      "grad_norm": 2.4618642330169678,
+      "learning_rate": 3.03208773354996e-06,
+      "loss": 1.8218,
+      "step": 1258
+    },
+    {
+      "epoch": 1.3797260273972602,
+      "grad_norm": 2.271958112716675,
+      "learning_rate": 3.0300568643379367e-06,
+      "loss": 1.8092,
+      "step": 1259
+    },
+    {
+      "epoch": 1.3808219178082193,
+      "grad_norm": 2.3990538120269775,
+      "learning_rate": 3.028025995125914e-06,
+      "loss": 1.8633,
+      "step": 1260
+    },
+    {
+      "epoch": 1.3808219178082193,
+      "eval_loss": 2.028458595275879,
+      "eval_runtime": 37.0659,
+      "eval_samples_per_second": 26.979,
+      "eval_steps_per_second": 1.133,
+      "step": 1260
+    },
+    {
+      "epoch": 1.381917808219178,
+      "grad_norm": 2.208803653717041,
+      "learning_rate": 3.0259951259138913e-06,
+      "loss": 1.8463,
+      "step": 1261
+    },
+    {
+      "epoch": 1.383013698630137,
+      "grad_norm": 2.3269028663635254,
+      "learning_rate": 3.0239642567018686e-06,
+      "loss": 1.7784,
+      "step": 1262
+    },
+    {
+      "epoch": 1.384109589041096,
+      "grad_norm": 2.367910623550415,
+      "learning_rate": 3.021933387489846e-06,
+      "loss": 1.8281,
+      "step": 1263
+    },
+    {
+      "epoch": 1.3852054794520547,
+      "grad_norm": 2.273686647415161,
+      "learning_rate": 3.019902518277823e-06,
+      "loss": 1.7832,
+      "step": 1264
+    },
+    {
+      "epoch": 1.3863013698630138,
+      "grad_norm": 2.421964168548584,
+      "learning_rate": 3.0178716490658004e-06,
+      "loss": 1.8407,
+      "step": 1265
+    },
+    {
+      "epoch": 1.3873972602739726,
+      "grad_norm": 2.2627696990966797,
+      "learning_rate": 3.0158407798537777e-06,
+      "loss": 1.8718,
+      "step": 1266
+    },
+    {
+      "epoch": 1.3884931506849316,
+      "grad_norm": 2.1981618404388428,
+      "learning_rate": 3.013809910641755e-06,
+      "loss": 1.7763,
+      "step": 1267
+    },
+    {
+      "epoch": 1.3895890410958904,
+      "grad_norm": 2.245253562927246,
+      "learning_rate": 3.0117790414297322e-06,
+      "loss": 1.7915,
+      "step": 1268
+    },
+    {
+      "epoch": 1.3906849315068492,
+      "grad_norm": 2.2292444705963135,
+      "learning_rate": 3.0097481722177095e-06,
+      "loss": 1.7909,
+      "step": 1269
+    },
+    {
+      "epoch": 1.3917808219178083,
+      "grad_norm": 2.2290430068969727,
+      "learning_rate": 3.0077173030056868e-06,
+      "loss": 1.8364,
+      "step": 1270
+    },
+    {
+      "epoch": 1.3917808219178083,
+      "eval_loss": 2.0290327072143555,
+      "eval_runtime": 37.0666,
+      "eval_samples_per_second": 26.978,
+      "eval_steps_per_second": 1.133,
+      "step": 1270
+    },
+    {
+      "epoch": 1.392876712328767,
+      "grad_norm": 2.220881700515747,
+      "learning_rate": 3.005686433793664e-06,
+      "loss": 1.7968,
+      "step": 1271
+    },
+    {
+      "epoch": 1.393972602739726,
+      "grad_norm": 2.375332832336426,
+      "learning_rate": 3.0036555645816413e-06,
+      "loss": 1.8502,
+      "step": 1272
+    },
+    {
+      "epoch": 1.395068493150685,
+      "grad_norm": 2.3065083026885986,
+      "learning_rate": 3.001624695369618e-06,
+      "loss": 1.8116,
+      "step": 1273
+    },
+    {
+      "epoch": 1.396164383561644,
+      "grad_norm": 2.7262511253356934,
+      "learning_rate": 2.999593826157596e-06,
+      "loss": 1.8641,
+      "step": 1274
+    },
+    {
+      "epoch": 1.3972602739726028,
+      "grad_norm": 2.3381216526031494,
+      "learning_rate": 2.9975629569455727e-06,
+      "loss": 1.8128,
+      "step": 1275
+    },
+    {
+      "epoch": 1.3983561643835616,
+      "grad_norm": 2.1994192600250244,
+      "learning_rate": 2.9955320877335504e-06,
+      "loss": 1.8563,
+      "step": 1276
+    },
+    {
+      "epoch": 1.3994520547945206,
+      "grad_norm": 2.3810958862304688,
+      "learning_rate": 2.9935012185215273e-06,
+      "loss": 1.8196,
+      "step": 1277
+    },
+    {
+      "epoch": 1.4005479452054794,
+      "grad_norm": 2.4343955516815186,
+      "learning_rate": 2.991470349309505e-06,
+      "loss": 1.8564,
+      "step": 1278
+    },
+    {
+      "epoch": 1.4016438356164382,
+      "grad_norm": 2.1929988861083984,
+      "learning_rate": 2.989439480097482e-06,
+      "loss": 1.8119,
+      "step": 1279
+    },
+    {
+      "epoch": 1.4027397260273973,
+      "grad_norm": 2.2305309772491455,
+      "learning_rate": 2.9874086108854595e-06,
+      "loss": 1.8289,
+      "step": 1280
+    },
+    {
+      "epoch": 1.4027397260273973,
+      "eval_loss": 2.032022476196289,
+      "eval_runtime": 37.084,
+      "eval_samples_per_second": 26.966,
+      "eval_steps_per_second": 1.133,
+      "step": 1280
+    },
+    {
+      "epoch": 1.403835616438356,
+      "grad_norm": 2.6255314350128174,
+      "learning_rate": 2.9853777416734364e-06,
+      "loss": 1.8176,
+      "step": 1281
+    },
+    {
+      "epoch": 1.4049315068493151,
+      "grad_norm": 2.238891363143921,
+      "learning_rate": 2.983346872461414e-06,
+      "loss": 1.8679,
+      "step": 1282
+    },
+    {
+      "epoch": 1.406027397260274,
+      "grad_norm": 2.9972007274627686,
+      "learning_rate": 2.981316003249391e-06,
+      "loss": 1.8839,
+      "step": 1283
+    },
+    {
+      "epoch": 1.407123287671233,
+      "grad_norm": 2.3255932331085205,
+      "learning_rate": 2.9792851340373678e-06,
+      "loss": 1.8451,
+      "step": 1284
+    },
+    {
+      "epoch": 1.4082191780821918,
+      "grad_norm": 2.290395498275757,
+      "learning_rate": 2.9772542648253455e-06,
+      "loss": 1.8244,
+      "step": 1285
+    },
+    {
+      "epoch": 1.4093150684931506,
+      "grad_norm": 2.288728713989258,
+      "learning_rate": 2.9752233956133223e-06,
+      "loss": 1.8628,
+      "step": 1286
+    },
+    {
+      "epoch": 1.4104109589041096,
+      "grad_norm": 2.177381992340088,
+      "learning_rate": 2.9731925264013e-06,
+      "loss": 1.8111,
+      "step": 1287
+    },
+    {
+      "epoch": 1.4115068493150684,
+      "grad_norm": 2.214956760406494,
+      "learning_rate": 2.971161657189277e-06,
+      "loss": 1.8377,
+      "step": 1288
+    },
+    {
+      "epoch": 1.4126027397260275,
+      "grad_norm": 2.3859663009643555,
+      "learning_rate": 2.9691307879772546e-06,
+      "loss": 1.8154,
+      "step": 1289
+    },
+    {
+      "epoch": 1.4136986301369863,
+      "grad_norm": 2.0873706340789795,
+      "learning_rate": 2.9670999187652314e-06,
+      "loss": 1.8098,
+      "step": 1290
+    },
+    {
+      "epoch": 1.4136986301369863,
+      "eval_loss": 2.029996633529663,
+      "eval_runtime": 37.0979,
+      "eval_samples_per_second": 26.956,
+      "eval_steps_per_second": 1.132,
+      "step": 1290
+    },
+    {
+      "epoch": 1.4147945205479453,
+      "grad_norm": 2.2632675170898438,
+      "learning_rate": 2.965069049553209e-06,
+      "loss": 1.7974,
+      "step": 1291
+    },
+    {
+      "epoch": 1.415890410958904,
+      "grad_norm": 2.208515167236328,
+      "learning_rate": 2.963038180341186e-06,
+      "loss": 1.8856,
+      "step": 1292
+    },
+    {
+      "epoch": 1.416986301369863,
+      "grad_norm": 2.176525115966797,
+      "learning_rate": 2.9610073111291637e-06,
+      "loss": 1.7783,
+      "step": 1293
+    },
+    {
+      "epoch": 1.418082191780822,
+      "grad_norm": 2.226682186126709,
+      "learning_rate": 2.9589764419171405e-06,
+      "loss": 1.865,
+      "step": 1294
+    },
+    {
+      "epoch": 1.4191780821917808,
+      "grad_norm": 2.3544883728027344,
+      "learning_rate": 2.9569455727051182e-06,
+      "loss": 1.8808,
+      "step": 1295
+    },
+    {
+      "epoch": 1.4202739726027398,
+      "grad_norm": 2.1729962825775146,
+      "learning_rate": 2.954914703493095e-06,
+      "loss": 1.7927,
+      "step": 1296
+    },
+    {
+      "epoch": 1.4213698630136986,
+      "grad_norm": 2.168168067932129,
+      "learning_rate": 2.9528838342810724e-06,
+      "loss": 1.8253,
+      "step": 1297
+    },
+    {
+      "epoch": 1.4224657534246576,
+      "grad_norm": 2.2479910850524902,
+      "learning_rate": 2.9508529650690496e-06,
+      "loss": 1.7976,
+      "step": 1298
+    },
+    {
+      "epoch": 1.4235616438356165,
+      "grad_norm": 2.1733837127685547,
+      "learning_rate": 2.948822095857027e-06,
+      "loss": 1.8015,
+      "step": 1299
+    },
+    {
+      "epoch": 1.4246575342465753,
+      "grad_norm": 2.2553391456604004,
+      "learning_rate": 2.946791226645004e-06,
+      "loss": 1.8545,
+      "step": 1300
+    },
+    {
+      "epoch": 1.4246575342465753,
+      "eval_loss": 2.0245048999786377,
+      "eval_runtime": 37.0943,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 1300
+    },
+    {
+      "epoch": 1.4257534246575343,
+      "grad_norm": 2.2817769050598145,
+      "learning_rate": 2.9447603574329815e-06,
+      "loss": 1.8017,
+      "step": 1301
+    },
+    {
+      "epoch": 1.426849315068493,
+      "grad_norm": 2.170865297317505,
+      "learning_rate": 2.9427294882209587e-06,
+      "loss": 1.7738,
+      "step": 1302
+    },
+    {
+      "epoch": 1.427945205479452,
+      "grad_norm": 2.296919107437134,
+      "learning_rate": 2.940698619008936e-06,
+      "loss": 1.8098,
+      "step": 1303
+    },
+    {
+      "epoch": 1.429041095890411,
+      "grad_norm": 2.476266622543335,
+      "learning_rate": 2.9386677497969133e-06,
+      "loss": 1.8229,
+      "step": 1304
+    },
+    {
+      "epoch": 1.43013698630137,
+      "grad_norm": 2.400918483734131,
+      "learning_rate": 2.9366368805848906e-06,
+      "loss": 1.8057,
+      "step": 1305
+    },
+    {
+      "epoch": 1.4312328767123288,
+      "grad_norm": 2.393892765045166,
+      "learning_rate": 2.934606011372868e-06,
+      "loss": 1.8568,
+      "step": 1306
+    },
+    {
+      "epoch": 1.4323287671232876,
+      "grad_norm": 2.2551307678222656,
+      "learning_rate": 2.932575142160845e-06,
+      "loss": 1.8493,
+      "step": 1307
+    },
+    {
+      "epoch": 1.4334246575342466,
+      "grad_norm": 2.3091611862182617,
+      "learning_rate": 2.9305442729488224e-06,
+      "loss": 1.8297,
+      "step": 1308
+    },
+    {
+      "epoch": 1.4345205479452054,
+      "grad_norm": 2.231614828109741,
+      "learning_rate": 2.9285134037367997e-06,
+      "loss": 1.8345,
+      "step": 1309
+    },
+    {
+      "epoch": 1.4356164383561643,
+      "grad_norm": 2.24942946434021,
+      "learning_rate": 2.9264825345247765e-06,
+      "loss": 1.8284,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4356164383561643,
+      "eval_loss": 2.024843215942383,
+      "eval_runtime": 36.7795,
+      "eval_samples_per_second": 27.189,
+      "eval_steps_per_second": 1.142,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4367123287671233,
+      "grad_norm": 2.2798643112182617,
+      "learning_rate": 2.9244516653127542e-06,
+      "loss": 1.8554,
+      "step": 1311
+    },
+    {
+      "epoch": 1.4378082191780823,
+      "grad_norm": 2.342576265335083,
+      "learning_rate": 2.922420796100731e-06,
+      "loss": 1.8296,
+      "step": 1312
+    },
+    {
+      "epoch": 1.4389041095890411,
+      "grad_norm": 2.268016815185547,
+      "learning_rate": 2.9203899268887088e-06,
+      "loss": 1.7467,
+      "step": 1313
+    },
+    {
+      "epoch": 1.44,
+      "grad_norm": 2.2520217895507812,
+      "learning_rate": 2.9183590576766856e-06,
+      "loss": 1.8748,
+      "step": 1314
+    },
+    {
+      "epoch": 1.441095890410959,
+      "grad_norm": 2.2091598510742188,
+      "learning_rate": 2.9163281884646633e-06,
+      "loss": 1.7715,
+      "step": 1315
+    },
+    {
+      "epoch": 1.4421917808219178,
+      "grad_norm": 2.197575092315674,
+      "learning_rate": 2.91429731925264e-06,
+      "loss": 1.8265,
+      "step": 1316
+    },
+    {
+      "epoch": 1.4432876712328766,
+      "grad_norm": 2.261122465133667,
+      "learning_rate": 2.912266450040618e-06,
+      "loss": 1.8758,
+      "step": 1317
+    },
+    {
+      "epoch": 1.4443835616438356,
+      "grad_norm": 2.3291778564453125,
+      "learning_rate": 2.9102355808285947e-06,
+      "loss": 1.8213,
+      "step": 1318
+    },
+    {
+      "epoch": 1.4454794520547944,
+      "grad_norm": 2.2182250022888184,
+      "learning_rate": 2.9082047116165724e-06,
+      "loss": 1.822,
+      "step": 1319
+    },
+    {
+      "epoch": 1.4465753424657535,
+      "grad_norm": 2.295978546142578,
+      "learning_rate": 2.9061738424045493e-06,
+      "loss": 1.8008,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4465753424657535,
+      "eval_loss": 2.0230674743652344,
+      "eval_runtime": 37.0013,
+      "eval_samples_per_second": 27.026,
+      "eval_steps_per_second": 1.135,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4476712328767123,
+      "grad_norm": 2.3187007904052734,
+      "learning_rate": 2.904142973192527e-06,
+      "loss": 1.7923,
+      "step": 1321
+    },
+    {
+      "epoch": 1.4487671232876713,
+      "grad_norm": 2.242851972579956,
+      "learning_rate": 2.902112103980504e-06,
+      "loss": 1.8874,
+      "step": 1322
+    },
+    {
+      "epoch": 1.4498630136986301,
+      "grad_norm": 2.2560081481933594,
+      "learning_rate": 2.900081234768481e-06,
+      "loss": 1.8284,
+      "step": 1323
+    },
+    {
+      "epoch": 1.450958904109589,
+      "grad_norm": 2.3017799854278564,
+      "learning_rate": 2.8980503655564584e-06,
+      "loss": 1.8263,
+      "step": 1324
+    },
+    {
+      "epoch": 1.452054794520548,
+      "grad_norm": 2.3016278743743896,
+      "learning_rate": 2.8960194963444357e-06,
+      "loss": 1.8174,
+      "step": 1325
+    },
+    {
+      "epoch": 1.4531506849315068,
+      "grad_norm": 2.1965372562408447,
+      "learning_rate": 2.893988627132413e-06,
+      "loss": 1.7812,
+      "step": 1326
+    },
+    {
+      "epoch": 1.4542465753424658,
+      "grad_norm": 2.302969455718994,
+      "learning_rate": 2.8919577579203902e-06,
+      "loss": 1.8115,
+      "step": 1327
+    },
+    {
+      "epoch": 1.4553424657534246,
+      "grad_norm": 2.338066339492798,
+      "learning_rate": 2.8899268887083675e-06,
+      "loss": 1.7535,
+      "step": 1328
+    },
+    {
+      "epoch": 1.4564383561643837,
+      "grad_norm": 2.446503162384033,
+      "learning_rate": 2.8878960194963448e-06,
+      "loss": 1.8052,
+      "step": 1329
+    },
+    {
+      "epoch": 1.4575342465753425,
+      "grad_norm": 2.2319273948669434,
+      "learning_rate": 2.885865150284322e-06,
+      "loss": 1.8226,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4575342465753425,
+      "eval_loss": 2.022364854812622,
+      "eval_runtime": 37.0901,
+      "eval_samples_per_second": 26.961,
+      "eval_steps_per_second": 1.132,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4586301369863013,
+      "grad_norm": 2.391080856323242,
+      "learning_rate": 2.8838342810722993e-06,
+      "loss": 1.8529,
+      "step": 1331
+    },
+    {
+      "epoch": 1.4597260273972603,
+      "grad_norm": 2.274463415145874,
+      "learning_rate": 2.8818034118602766e-06,
+      "loss": 1.8043,
+      "step": 1332
+    },
+    {
+      "epoch": 1.4608219178082191,
+      "grad_norm": 2.4040963649749756,
+      "learning_rate": 2.879772542648254e-06,
+      "loss": 1.7842,
+      "step": 1333
+    },
+    {
+      "epoch": 1.4619178082191782,
+      "grad_norm": 2.268629312515259,
+      "learning_rate": 2.8777416734362307e-06,
+      "loss": 1.859,
+      "step": 1334
+    },
+    {
+      "epoch": 1.463013698630137,
+      "grad_norm": 2.4727842807769775,
+      "learning_rate": 2.8757108042242084e-06,
+      "loss": 1.7615,
+      "step": 1335
+    },
+    {
+      "epoch": 1.464109589041096,
+      "grad_norm": 2.179358959197998,
+      "learning_rate": 2.8736799350121853e-06,
+      "loss": 1.9085,
+      "step": 1336
+    },
+    {
+      "epoch": 1.4652054794520548,
+      "grad_norm": 2.220917224884033,
+      "learning_rate": 2.871649065800163e-06,
+      "loss": 1.8618,
+      "step": 1337
+    },
+    {
+      "epoch": 1.4663013698630136,
+      "grad_norm": 2.3807291984558105,
+      "learning_rate": 2.86961819658814e-06,
+      "loss": 1.88,
+      "step": 1338
+    },
+    {
+      "epoch": 1.4673972602739727,
+      "grad_norm": 2.272507667541504,
+      "learning_rate": 2.8675873273761175e-06,
+      "loss": 1.8045,
+      "step": 1339
+    },
+    {
+      "epoch": 1.4684931506849315,
+      "grad_norm": 2.1832377910614014,
+      "learning_rate": 2.8655564581640944e-06,
+      "loss": 1.8273,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4684931506849315,
+      "eval_loss": 2.0218358039855957,
+      "eval_runtime": 37.1223,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4695890410958903,
+      "grad_norm": 2.2655251026153564,
+      "learning_rate": 2.863525588952072e-06,
+      "loss": 1.8162,
+      "step": 1341
+    },
+    {
+      "epoch": 1.4706849315068493,
+      "grad_norm": 2.191852331161499,
+      "learning_rate": 2.861494719740049e-06,
+      "loss": 1.8778,
+      "step": 1342
+    },
+    {
+      "epoch": 1.4717808219178083,
+      "grad_norm": 2.1869187355041504,
+      "learning_rate": 2.8594638505280266e-06,
+      "loss": 1.7695,
+      "step": 1343
+    },
+    {
+      "epoch": 1.4728767123287672,
+      "grad_norm": 2.248108148574829,
+      "learning_rate": 2.8574329813160035e-06,
+      "loss": 1.7759,
+      "step": 1344
+    },
+    {
+      "epoch": 1.473972602739726,
+      "grad_norm": 2.2599704265594482,
+      "learning_rate": 2.855402112103981e-06,
+      "loss": 1.8214,
+      "step": 1345
+    },
+    {
+      "epoch": 1.475068493150685,
+      "grad_norm": 2.2920355796813965,
+      "learning_rate": 2.853371242891958e-06,
+      "loss": 1.8061,
+      "step": 1346
+    },
+    {
+      "epoch": 1.4761643835616438,
+      "grad_norm": 2.26715087890625,
+      "learning_rate": 2.851340373679935e-06,
+      "loss": 1.8244,
+      "step": 1347
+    },
+    {
+      "epoch": 1.4772602739726026,
+      "grad_norm": 2.349169969558716,
+      "learning_rate": 2.8493095044679126e-06,
+      "loss": 1.8401,
+      "step": 1348
+    },
+    {
+      "epoch": 1.4783561643835617,
+      "grad_norm": 2.3571698665618896,
+      "learning_rate": 2.8472786352558894e-06,
+      "loss": 1.8913,
+      "step": 1349
+    },
+    {
+      "epoch": 1.4794520547945205,
+      "grad_norm": 2.334057331085205,
+      "learning_rate": 2.845247766043867e-06,
+      "loss": 1.8434,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4794520547945205,
+      "eval_loss": 2.0251200199127197,
+      "eval_runtime": 37.1298,
+      "eval_samples_per_second": 26.933,
+      "eval_steps_per_second": 1.131,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4805479452054795,
+      "grad_norm": 2.363516330718994,
+      "learning_rate": 2.843216896831844e-06,
+      "loss": 1.7846,
+      "step": 1351
+    },
+    {
+      "epoch": 1.4816438356164383,
+      "grad_norm": 2.372403860092163,
+      "learning_rate": 2.8411860276198217e-06,
+      "loss": 1.7791,
+      "step": 1352
+    },
+    {
+      "epoch": 1.4827397260273973,
+      "grad_norm": 2.2666590213775635,
+      "learning_rate": 2.8391551584077985e-06,
+      "loss": 1.7969,
+      "step": 1353
+    },
+    {
+      "epoch": 1.4838356164383562,
+      "grad_norm": 2.3651468753814697,
+      "learning_rate": 2.8371242891957762e-06,
+      "loss": 1.8565,
+      "step": 1354
+    },
+    {
+      "epoch": 1.484931506849315,
+      "grad_norm": 2.3559513092041016,
+      "learning_rate": 2.835093419983753e-06,
+      "loss": 1.8765,
+      "step": 1355
+    },
+    {
+      "epoch": 1.486027397260274,
+      "grad_norm": 2.232041358947754,
+      "learning_rate": 2.833062550771731e-06,
+      "loss": 1.805,
+      "step": 1356
+    },
+    {
+      "epoch": 1.4871232876712328,
+      "grad_norm": 2.2423202991485596,
+      "learning_rate": 2.8310316815597076e-06,
+      "loss": 1.8137,
+      "step": 1357
+    },
+    {
+      "epoch": 1.4882191780821918,
+      "grad_norm": 2.3454222679138184,
+      "learning_rate": 2.8290008123476853e-06,
+      "loss": 1.7839,
+      "step": 1358
+    },
+    {
+      "epoch": 1.4893150684931507,
+      "grad_norm": 2.351266384124756,
+      "learning_rate": 2.826969943135662e-06,
+      "loss": 1.8208,
+      "step": 1359
+    },
+    {
+      "epoch": 1.4904109589041097,
+      "grad_norm": 2.3025388717651367,
+      "learning_rate": 2.8249390739236395e-06,
+      "loss": 1.8059,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4904109589041097,
+      "eval_loss": 2.0235962867736816,
+      "eval_runtime": 37.1359,
+      "eval_samples_per_second": 26.928,
+      "eval_steps_per_second": 1.131,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4915068493150685,
+      "grad_norm": 2.2277750968933105,
+      "learning_rate": 2.8229082047116167e-06,
+      "loss": 1.8238,
+      "step": 1361
+    },
+    {
+      "epoch": 1.4926027397260273,
+      "grad_norm": 2.318286180496216,
+      "learning_rate": 2.820877335499594e-06,
+      "loss": 1.7867,
+      "step": 1362
+    },
+    {
+      "epoch": 1.4936986301369863,
+      "grad_norm": 2.296259880065918,
+      "learning_rate": 2.8188464662875713e-06,
+      "loss": 1.8068,
+      "step": 1363
+    },
+    {
+      "epoch": 1.4947945205479451,
+      "grad_norm": 2.2617714405059814,
+      "learning_rate": 2.8168155970755486e-06,
+      "loss": 1.7607,
+      "step": 1364
+    },
+    {
+      "epoch": 1.4958904109589042,
+      "grad_norm": 2.318294048309326,
+      "learning_rate": 2.814784727863526e-06,
+      "loss": 1.7872,
+      "step": 1365
+    },
+    {
+      "epoch": 1.496986301369863,
+      "grad_norm": 2.260286808013916,
+      "learning_rate": 2.812753858651503e-06,
+      "loss": 1.8013,
+      "step": 1366
+    },
+    {
+      "epoch": 1.498082191780822,
+      "grad_norm": 2.222074508666992,
+      "learning_rate": 2.8107229894394804e-06,
+      "loss": 1.8338,
+      "step": 1367
+    },
+    {
+      "epoch": 1.4991780821917808,
+      "grad_norm": 2.6403002738952637,
+      "learning_rate": 2.8086921202274577e-06,
+      "loss": 1.8126,
+      "step": 1368
+    },
+    {
+      "epoch": 1.5002739726027396,
+      "grad_norm": 2.2439680099487305,
+      "learning_rate": 2.806661251015435e-06,
+      "loss": 1.7669,
+      "step": 1369
+    },
+    {
+      "epoch": 1.5013698630136987,
+      "grad_norm": 2.3914687633514404,
+      "learning_rate": 2.8046303818034122e-06,
+      "loss": 1.8421,
+      "step": 1370
+    },
+    {
+      "epoch": 1.5013698630136987,
+      "eval_loss": 2.0222702026367188,
+      "eval_runtime": 37.1598,
+      "eval_samples_per_second": 26.911,
+      "eval_steps_per_second": 1.13,
+      "step": 1370
+    },
+    {
+      "epoch": 1.5024657534246575,
+      "grad_norm": 2.3510093688964844,
+      "learning_rate": 2.802599512591389e-06,
+      "loss": 1.8902,
+      "step": 1371
+    },
+    {
+      "epoch": 1.5035616438356163,
+      "grad_norm": 2.4427363872528076,
+      "learning_rate": 2.8005686433793668e-06,
+      "loss": 1.9376,
+      "step": 1372
+    },
+    {
+      "epoch": 1.5046575342465753,
+      "grad_norm": 2.299356460571289,
+      "learning_rate": 2.7985377741673436e-06,
+      "loss": 1.812,
+      "step": 1373
+    },
+    {
+      "epoch": 1.5057534246575344,
+      "grad_norm": 2.3738906383514404,
+      "learning_rate": 2.7965069049553213e-06,
+      "loss": 1.7666,
+      "step": 1374
+    },
+    {
+      "epoch": 1.5068493150684932,
+      "grad_norm": 2.3620493412017822,
+      "learning_rate": 2.794476035743298e-06,
+      "loss": 1.824,
+      "step": 1375
+    },
+    {
+      "epoch": 1.507945205479452,
+      "grad_norm": 2.539048671722412,
+      "learning_rate": 2.792445166531276e-06,
+      "loss": 1.8712,
+      "step": 1376
+    },
+    {
+      "epoch": 1.509041095890411,
+      "grad_norm": 2.2931456565856934,
+      "learning_rate": 2.7904142973192527e-06,
+      "loss": 1.8414,
+      "step": 1377
+    },
+    {
+      "epoch": 1.5101369863013698,
+      "grad_norm": 2.4368224143981934,
+      "learning_rate": 2.7883834281072304e-06,
+      "loss": 1.8419,
+      "step": 1378
+    },
+    {
+      "epoch": 1.5112328767123286,
+      "grad_norm": 2.298302173614502,
+      "learning_rate": 2.7863525588952073e-06,
+      "loss": 1.7846,
+      "step": 1379
+    },
+    {
+      "epoch": 1.5123287671232877,
+      "grad_norm": 2.3726806640625,
+      "learning_rate": 2.784321689683185e-06,
+      "loss": 1.7733,
+      "step": 1380
+    },
+    {
+      "epoch": 1.5123287671232877,
+      "eval_loss": 2.0201807022094727,
+      "eval_runtime": 37.1012,
+      "eval_samples_per_second": 26.953,
+      "eval_steps_per_second": 1.132,
+      "step": 1380
+    },
+    {
+      "epoch": 1.5134246575342467,
+      "grad_norm": 2.6515002250671387,
+      "learning_rate": 2.782290820471162e-06,
+      "loss": 1.7764,
+      "step": 1381
+    },
+    {
+      "epoch": 1.5145205479452055,
+      "grad_norm": 2.273272752761841,
+      "learning_rate": 2.7802599512591395e-06,
+      "loss": 1.7747,
+      "step": 1382
+    },
+    {
+      "epoch": 1.5156164383561643,
+      "grad_norm": 2.570009708404541,
+      "learning_rate": 2.7782290820471164e-06,
+      "loss": 1.8612,
+      "step": 1383
+    },
+    {
+      "epoch": 1.5167123287671234,
+      "grad_norm": 2.567878007888794,
+      "learning_rate": 2.7761982128350932e-06,
+      "loss": 1.8402,
+      "step": 1384
+    },
+    {
+      "epoch": 1.5178082191780822,
+      "grad_norm": 2.3768470287323,
+      "learning_rate": 2.774167343623071e-06,
+      "loss": 1.8473,
+      "step": 1385
+    },
+    {
+      "epoch": 1.518904109589041,
+      "grad_norm": 2.1383397579193115,
+      "learning_rate": 2.772136474411048e-06,
+      "loss": 1.8593,
+      "step": 1386
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 2.4437034130096436,
+      "learning_rate": 2.7701056051990255e-06,
+      "loss": 1.792,
+      "step": 1387
+    },
+    {
+      "epoch": 1.521095890410959,
+      "grad_norm": 2.433933734893799,
+      "learning_rate": 2.7680747359870023e-06,
+      "loss": 1.7312,
+      "step": 1388
+    },
+    {
+      "epoch": 1.5221917808219179,
+      "grad_norm": 2.340994119644165,
+      "learning_rate": 2.76604386677498e-06,
+      "loss": 1.8326,
+      "step": 1389
+    },
+    {
+      "epoch": 1.5232876712328767,
+      "grad_norm": 2.4730165004730225,
+      "learning_rate": 2.764012997562957e-06,
+      "loss": 1.8129,
+      "step": 1390
+    },
+    {
+      "epoch": 1.5232876712328767,
+      "eval_loss": 2.0158324241638184,
+      "eval_runtime": 37.0997,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 1390
+    },
+    {
+      "epoch": 1.5243835616438357,
+      "grad_norm": 2.2898151874542236,
+      "learning_rate": 2.7619821283509346e-06,
+      "loss": 1.8726,
+      "step": 1391
+    },
+    {
+      "epoch": 1.5254794520547945,
+      "grad_norm": 2.255439519882202,
+      "learning_rate": 2.7599512591389114e-06,
+      "loss": 1.7587,
+      "step": 1392
+    },
+    {
+      "epoch": 1.5265753424657533,
+      "grad_norm": 2.4269862174987793,
+      "learning_rate": 2.757920389926889e-06,
+      "loss": 1.8498,
+      "step": 1393
+    },
+    {
+      "epoch": 1.5276712328767124,
+      "grad_norm": 2.3411412239074707,
+      "learning_rate": 2.755889520714866e-06,
+      "loss": 1.8103,
+      "step": 1394
+    },
+    {
+      "epoch": 1.5287671232876714,
+      "grad_norm": 2.2859277725219727,
+      "learning_rate": 2.7538586515028437e-06,
+      "loss": 1.7909,
+      "step": 1395
+    },
+    {
+      "epoch": 1.52986301369863,
+      "grad_norm": 2.2675201892852783,
+      "learning_rate": 2.7518277822908205e-06,
+      "loss": 1.7672,
+      "step": 1396
+    },
+    {
+      "epoch": 1.530958904109589,
+      "grad_norm": 2.240610361099243,
+      "learning_rate": 2.749796913078798e-06,
+      "loss": 1.8162,
+      "step": 1397
+    },
+    {
+      "epoch": 1.532054794520548,
+      "grad_norm": 2.272045373916626,
+      "learning_rate": 2.747766043866775e-06,
+      "loss": 1.8593,
+      "step": 1398
+    },
+    {
+      "epoch": 1.5331506849315069,
+      "grad_norm": 2.1736183166503906,
+      "learning_rate": 2.7457351746547524e-06,
+      "loss": 1.8154,
+      "step": 1399
+    },
+    {
+      "epoch": 1.5342465753424657,
+      "grad_norm": 2.197085380554199,
+      "learning_rate": 2.7437043054427297e-06,
+      "loss": 1.8758,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5342465753424657,
+      "eval_loss": 2.017298698425293,
+      "eval_runtime": 37.1666,
+      "eval_samples_per_second": 26.906,
+      "eval_steps_per_second": 1.13,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5353424657534247,
+      "grad_norm": 2.2274084091186523,
+      "learning_rate": 2.741673436230707e-06,
+      "loss": 1.9241,
+      "step": 1401
+    },
+    {
+      "epoch": 1.5364383561643835,
+      "grad_norm": 2.3320305347442627,
+      "learning_rate": 2.739642567018684e-06,
+      "loss": 1.8209,
+      "step": 1402
+    },
+    {
+      "epoch": 1.5375342465753423,
+      "grad_norm": 2.370651960372925,
+      "learning_rate": 2.7376116978066615e-06,
+      "loss": 1.7897,
+      "step": 1403
+    },
+    {
+      "epoch": 1.5386301369863014,
+      "grad_norm": 2.387906312942505,
+      "learning_rate": 2.7355808285946388e-06,
+      "loss": 1.8385,
+      "step": 1404
+    },
+    {
+      "epoch": 1.5397260273972604,
+      "grad_norm": 2.340228796005249,
+      "learning_rate": 2.733549959382616e-06,
+      "loss": 1.8279,
+      "step": 1405
+    },
+    {
+      "epoch": 1.5408219178082192,
+      "grad_norm": 2.338785409927368,
+      "learning_rate": 2.7315190901705933e-06,
+      "loss": 1.7887,
+      "step": 1406
+    },
+    {
+      "epoch": 1.541917808219178,
+      "grad_norm": 2.350057601928711,
+      "learning_rate": 2.7294882209585706e-06,
+      "loss": 1.8772,
+      "step": 1407
+    },
+    {
+      "epoch": 1.543013698630137,
+      "grad_norm": 2.1129322052001953,
+      "learning_rate": 2.727457351746548e-06,
+      "loss": 1.8393,
+      "step": 1408
+    },
+    {
+      "epoch": 1.5441095890410959,
+      "grad_norm": 2.282553195953369,
+      "learning_rate": 2.725426482534525e-06,
+      "loss": 1.7893,
+      "step": 1409
+    },
+    {
+      "epoch": 1.5452054794520547,
+      "grad_norm": 2.3291871547698975,
+      "learning_rate": 2.723395613322502e-06,
+      "loss": 1.7921,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5452054794520547,
+      "eval_loss": 2.0166046619415283,
+      "eval_runtime": 36.7906,
+      "eval_samples_per_second": 27.181,
+      "eval_steps_per_second": 1.142,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5463013698630137,
+      "grad_norm": 2.3050696849823,
+      "learning_rate": 2.7213647441104797e-06,
+      "loss": 1.8292,
+      "step": 1411
+    },
+    {
+      "epoch": 1.5473972602739727,
+      "grad_norm": 2.2762458324432373,
+      "learning_rate": 2.7193338748984565e-06,
+      "loss": 1.811,
+      "step": 1412
+    },
+    {
+      "epoch": 1.5484931506849315,
+      "grad_norm": 2.5544228553771973,
+      "learning_rate": 2.7173030056864342e-06,
+      "loss": 1.8203,
+      "step": 1413
+    },
+    {
+      "epoch": 1.5495890410958904,
+      "grad_norm": 2.6392276287078857,
+      "learning_rate": 2.715272136474411e-06,
+      "loss": 1.9724,
+      "step": 1414
+    },
+    {
+      "epoch": 1.5506849315068494,
+      "grad_norm": 2.445366382598877,
+      "learning_rate": 2.713241267262389e-06,
+      "loss": 1.7589,
+      "step": 1415
+    },
+    {
+      "epoch": 1.5517808219178082,
+      "grad_norm": 2.5305662155151367,
+      "learning_rate": 2.7112103980503656e-06,
+      "loss": 1.8311,
+      "step": 1416
+    },
+    {
+      "epoch": 1.552876712328767,
+      "grad_norm": 2.405482769012451,
+      "learning_rate": 2.7091795288383433e-06,
+      "loss": 1.7893,
+      "step": 1417
+    },
+    {
+      "epoch": 1.553972602739726,
+      "grad_norm": 2.258361577987671,
+      "learning_rate": 2.70714865962632e-06,
+      "loss": 1.8489,
+      "step": 1418
+    },
+    {
+      "epoch": 1.555068493150685,
+      "grad_norm": 2.1798577308654785,
+      "learning_rate": 2.705117790414298e-06,
+      "loss": 1.7983,
+      "step": 1419
+    },
+    {
+      "epoch": 1.5561643835616439,
+      "grad_norm": 2.1985363960266113,
+      "learning_rate": 2.7030869212022747e-06,
+      "loss": 1.7934,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5561643835616439,
+      "eval_loss": 2.0219907760620117,
+      "eval_runtime": 36.9416,
+      "eval_samples_per_second": 27.07,
+      "eval_steps_per_second": 1.137,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5572602739726027,
+      "grad_norm": 2.273252248764038,
+      "learning_rate": 2.7010560519902516e-06,
+      "loss": 1.8571,
+      "step": 1421
+    },
+    {
+      "epoch": 1.5583561643835617,
+      "grad_norm": 2.22920298576355,
+      "learning_rate": 2.6990251827782293e-06,
+      "loss": 1.7785,
+      "step": 1422
+    },
+    {
+      "epoch": 1.5594520547945205,
+      "grad_norm": 2.1784098148345947,
+      "learning_rate": 2.696994313566206e-06,
+      "loss": 1.8053,
+      "step": 1423
+    },
+    {
+      "epoch": 1.5605479452054793,
+      "grad_norm": 2.223404884338379,
+      "learning_rate": 2.694963444354184e-06,
+      "loss": 1.7965,
+      "step": 1424
+    },
+    {
+      "epoch": 1.5616438356164384,
+      "grad_norm": 2.2004787921905518,
+      "learning_rate": 2.6929325751421607e-06,
+      "loss": 1.8851,
+      "step": 1425
+    },
+    {
+      "epoch": 1.5627397260273974,
+      "grad_norm": 2.202467441558838,
+      "learning_rate": 2.6909017059301384e-06,
+      "loss": 1.7884,
+      "step": 1426
+    },
+    {
+      "epoch": 1.563835616438356,
+      "grad_norm": 2.337430238723755,
+      "learning_rate": 2.6888708367181153e-06,
+      "loss": 1.8543,
+      "step": 1427
+    },
+    {
+      "epoch": 1.564931506849315,
+      "grad_norm": 2.322848081588745,
+      "learning_rate": 2.686839967506093e-06,
+      "loss": 1.7607,
+      "step": 1428
+    },
+    {
+      "epoch": 1.566027397260274,
+      "grad_norm": 2.277765989303589,
+      "learning_rate": 2.68480909829407e-06,
+      "loss": 1.8354,
+      "step": 1429
+    },
+    {
+      "epoch": 1.5671232876712329,
+      "grad_norm": 2.3229448795318604,
+      "learning_rate": 2.6827782290820475e-06,
+      "loss": 1.7277,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5671232876712329,
+      "eval_loss": 2.0220985412597656,
+      "eval_runtime": 37.0054,
+      "eval_samples_per_second": 27.023,
+      "eval_steps_per_second": 1.135,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5682191780821917,
+      "grad_norm": 2.49545955657959,
+      "learning_rate": 2.6807473598700244e-06,
+      "loss": 1.7351,
+      "step": 1431
+    },
+    {
+      "epoch": 1.5693150684931507,
+      "grad_norm": 2.530553102493286,
+      "learning_rate": 2.678716490658002e-06,
+      "loss": 1.8614,
+      "step": 1432
+    },
+    {
+      "epoch": 1.5704109589041098,
+      "grad_norm": 2.3352842330932617,
+      "learning_rate": 2.676685621445979e-06,
+      "loss": 1.8218,
+      "step": 1433
+    },
+    {
+      "epoch": 1.5715068493150683,
+      "grad_norm": 2.3533058166503906,
+      "learning_rate": 2.674654752233956e-06,
+      "loss": 1.8282,
+      "step": 1434
+    },
+    {
+      "epoch": 1.5726027397260274,
+      "grad_norm": 2.5509746074676514,
+      "learning_rate": 2.6726238830219335e-06,
+      "loss": 1.8333,
+      "step": 1435
+    },
+    {
+      "epoch": 1.5736986301369864,
+      "grad_norm": 2.185175895690918,
+      "learning_rate": 2.6705930138099107e-06,
+      "loss": 1.8299,
+      "step": 1436
+    },
+    {
+      "epoch": 1.5747945205479452,
+      "grad_norm": 2.2338390350341797,
+      "learning_rate": 2.668562144597888e-06,
+      "loss": 1.857,
+      "step": 1437
+    },
+    {
+      "epoch": 1.575890410958904,
+      "grad_norm": 2.4094536304473877,
+      "learning_rate": 2.6665312753858653e-06,
+      "loss": 1.7333,
+      "step": 1438
+    },
+    {
+      "epoch": 1.576986301369863,
+      "grad_norm": 2.4650957584381104,
+      "learning_rate": 2.6645004061738426e-06,
+      "loss": 1.8237,
+      "step": 1439
+    },
+    {
+      "epoch": 1.5780821917808219,
+      "grad_norm": 2.427415132522583,
+      "learning_rate": 2.66246953696182e-06,
+      "loss": 1.8206,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5780821917808219,
+      "eval_loss": 2.0222699642181396,
+      "eval_runtime": 37.0252,
+      "eval_samples_per_second": 27.009,
+      "eval_steps_per_second": 1.134,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5791780821917807,
+      "grad_norm": 2.3878679275512695,
+      "learning_rate": 2.660438667749797e-06,
+      "loss": 1.775,
+      "step": 1441
+    },
+    {
+      "epoch": 1.5802739726027397,
+      "grad_norm": 2.317819595336914,
+      "learning_rate": 2.6584077985377744e-06,
+      "loss": 1.7479,
+      "step": 1442
+    },
+    {
+      "epoch": 1.5813698630136988,
+      "grad_norm": 2.34951114654541,
+      "learning_rate": 2.6563769293257517e-06,
+      "loss": 1.8025,
+      "step": 1443
+    },
+    {
+      "epoch": 1.5824657534246576,
+      "grad_norm": 2.2849795818328857,
+      "learning_rate": 2.654346060113729e-06,
+      "loss": 1.8152,
+      "step": 1444
+    },
+    {
+      "epoch": 1.5835616438356164,
+      "grad_norm": 2.211489677429199,
+      "learning_rate": 2.6523151909017062e-06,
+      "loss": 1.7666,
+      "step": 1445
+    },
+    {
+      "epoch": 1.5846575342465754,
+      "grad_norm": 2.1876230239868164,
+      "learning_rate": 2.6502843216896835e-06,
+      "loss": 1.7933,
+      "step": 1446
+    },
+    {
+      "epoch": 1.5857534246575342,
+      "grad_norm": 2.367628335952759,
+      "learning_rate": 2.6482534524776603e-06,
+      "loss": 1.7756,
+      "step": 1447
+    },
+    {
+      "epoch": 1.586849315068493,
+      "grad_norm": 2.1198348999023438,
+      "learning_rate": 2.646222583265638e-06,
+      "loss": 1.8908,
+      "step": 1448
+    },
+    {
+      "epoch": 1.587945205479452,
+      "grad_norm": 2.192788600921631,
+      "learning_rate": 2.644191714053615e-06,
+      "loss": 1.8432,
+      "step": 1449
+    },
+    {
+      "epoch": 1.589041095890411,
+      "grad_norm": 2.4480655193328857,
+      "learning_rate": 2.6421608448415926e-06,
+      "loss": 1.8223,
+      "step": 1450
+    },
+    {
+      "epoch": 1.589041095890411,
+      "eval_loss": 2.019976854324341,
+      "eval_runtime": 37.0117,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1450
+    },
+    {
+      "epoch": 1.59013698630137,
+      "grad_norm": 2.573078155517578,
+      "learning_rate": 2.6401299756295694e-06,
+      "loss": 1.7881,
+      "step": 1451
+    },
+    {
+      "epoch": 1.5912328767123287,
+      "grad_norm": 2.2820355892181396,
+      "learning_rate": 2.638099106417547e-06,
+      "loss": 1.7703,
+      "step": 1452
+    },
+    {
+      "epoch": 1.5923287671232877,
+      "grad_norm": 2.526477575302124,
+      "learning_rate": 2.636068237205524e-06,
+      "loss": 1.8281,
+      "step": 1453
+    },
+    {
+      "epoch": 1.5934246575342466,
+      "grad_norm": 2.873647689819336,
+      "learning_rate": 2.6340373679935017e-06,
+      "loss": 1.7771,
+      "step": 1454
+    },
+    {
+      "epoch": 1.5945205479452054,
+      "grad_norm": 2.3782413005828857,
+      "learning_rate": 2.6320064987814785e-06,
+      "loss": 1.8864,
+      "step": 1455
+    },
+    {
+      "epoch": 1.5956164383561644,
+      "grad_norm": 2.3153343200683594,
+      "learning_rate": 2.6299756295694562e-06,
+      "loss": 1.734,
+      "step": 1456
+    },
+    {
+      "epoch": 1.5967123287671234,
+      "grad_norm": 2.631880283355713,
+      "learning_rate": 2.627944760357433e-06,
+      "loss": 1.8295,
+      "step": 1457
+    },
+    {
+      "epoch": 1.5978082191780822,
+      "grad_norm": 2.5827910900115967,
+      "learning_rate": 2.6259138911454104e-06,
+      "loss": 1.8738,
+      "step": 1458
+    },
+    {
+      "epoch": 1.598904109589041,
+      "grad_norm": 2.3519935607910156,
+      "learning_rate": 2.6238830219333877e-06,
+      "loss": 1.8832,
+      "step": 1459
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 2.5236215591430664,
+      "learning_rate": 2.621852152721365e-06,
+      "loss": 1.8595,
+      "step": 1460
+    },
+    {
+      "epoch": 1.6,
+      "eval_loss": 2.0170512199401855,
+      "eval_runtime": 37.0302,
+      "eval_samples_per_second": 27.005,
+      "eval_steps_per_second": 1.134,
+      "step": 1460
+    },
+    {
+      "epoch": 1.601095890410959,
+      "grad_norm": 2.725083112716675,
+      "learning_rate": 2.619821283509342e-06,
+      "loss": 1.7476,
+      "step": 1461
+    },
+    {
+      "epoch": 1.6021917808219177,
+      "grad_norm": 2.212644100189209,
+      "learning_rate": 2.6177904142973195e-06,
+      "loss": 1.8475,
+      "step": 1462
+    },
+    {
+      "epoch": 1.6032876712328767,
+      "grad_norm": 2.346808910369873,
+      "learning_rate": 2.6157595450852968e-06,
+      "loss": 1.7762,
+      "step": 1463
+    },
+    {
+      "epoch": 1.6043835616438358,
+      "grad_norm": 2.702331304550171,
+      "learning_rate": 2.613728675873274e-06,
+      "loss": 1.8296,
+      "step": 1464
+    },
+    {
+      "epoch": 1.6054794520547944,
+      "grad_norm": 2.4175899028778076,
+      "learning_rate": 2.6116978066612513e-06,
+      "loss": 1.8476,
+      "step": 1465
+    },
+    {
+      "epoch": 1.6065753424657534,
+      "grad_norm": 2.325457811355591,
+      "learning_rate": 2.6096669374492286e-06,
+      "loss": 1.8706,
+      "step": 1466
+    },
+    {
+      "epoch": 1.6076712328767124,
+      "grad_norm": 2.383842945098877,
+      "learning_rate": 2.607636068237206e-06,
+      "loss": 1.8246,
+      "step": 1467
+    },
+    {
+      "epoch": 1.6087671232876712,
+      "grad_norm": 2.289515256881714,
+      "learning_rate": 2.605605199025183e-06,
+      "loss": 1.7728,
+      "step": 1468
+    },
+    {
+      "epoch": 1.60986301369863,
+      "grad_norm": 2.5116689205169678,
+      "learning_rate": 2.6035743298131604e-06,
+      "loss": 1.8359,
+      "step": 1469
+    },
+    {
+      "epoch": 1.610958904109589,
+      "grad_norm": 2.279695749282837,
+      "learning_rate": 2.6015434606011377e-06,
+      "loss": 1.8155,
+      "step": 1470
+    },
+    {
+      "epoch": 1.610958904109589,
+      "eval_loss": 2.0142691135406494,
+      "eval_runtime": 37.0119,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1470
+    },
+    {
+      "epoch": 1.612054794520548,
+      "grad_norm": 2.4397690296173096,
+      "learning_rate": 2.5995125913891145e-06,
+      "loss": 1.8424,
+      "step": 1471
+    },
+    {
+      "epoch": 1.6131506849315067,
+      "grad_norm": 2.433577299118042,
+      "learning_rate": 2.5974817221770922e-06,
+      "loss": 1.8196,
+      "step": 1472
+    },
+    {
+      "epoch": 1.6142465753424657,
+      "grad_norm": 2.365459442138672,
+      "learning_rate": 2.595450852965069e-06,
+      "loss": 1.8124,
+      "step": 1473
+    },
+    {
+      "epoch": 1.6153424657534248,
+      "grad_norm": 2.253133535385132,
+      "learning_rate": 2.593419983753047e-06,
+      "loss": 1.8443,
+      "step": 1474
+    },
+    {
+      "epoch": 1.6164383561643836,
+      "grad_norm": 2.227625608444214,
+      "learning_rate": 2.5913891145410236e-06,
+      "loss": 1.8524,
+      "step": 1475
+    },
+    {
+      "epoch": 1.6175342465753424,
+      "grad_norm": 2.403702974319458,
+      "learning_rate": 2.5893582453290013e-06,
+      "loss": 1.7983,
+      "step": 1476
+    },
+    {
+      "epoch": 1.6186301369863014,
+      "grad_norm": 2.24507999420166,
+      "learning_rate": 2.587327376116978e-06,
+      "loss": 1.8563,
+      "step": 1477
+    },
+    {
+      "epoch": 1.6197260273972602,
+      "grad_norm": 2.284700393676758,
+      "learning_rate": 2.585296506904956e-06,
+      "loss": 1.7958,
+      "step": 1478
+    },
+    {
+      "epoch": 1.620821917808219,
+      "grad_norm": 2.288215398788452,
+      "learning_rate": 2.5832656376929327e-06,
+      "loss": 1.8021,
+      "step": 1479
+    },
+    {
+      "epoch": 1.621917808219178,
+      "grad_norm": 2.2312939167022705,
+      "learning_rate": 2.5812347684809104e-06,
+      "loss": 1.7856,
+      "step": 1480
+    },
+    {
+      "epoch": 1.621917808219178,
+      "eval_loss": 2.01674222946167,
+      "eval_runtime": 36.9943,
+      "eval_samples_per_second": 27.031,
+      "eval_steps_per_second": 1.135,
+      "step": 1480
+    },
+    {
+      "epoch": 1.6230136986301371,
+      "grad_norm": 2.319363832473755,
+      "learning_rate": 2.5792038992688873e-06,
+      "loss": 1.7263,
+      "step": 1481
+    },
+    {
+      "epoch": 1.624109589041096,
+      "grad_norm": 2.4342033863067627,
+      "learning_rate": 2.577173030056865e-06,
+      "loss": 1.897,
+      "step": 1482
+    },
+    {
+      "epoch": 1.6252054794520547,
+      "grad_norm": 2.3052282333374023,
+      "learning_rate": 2.575142160844842e-06,
+      "loss": 1.7165,
+      "step": 1483
+    },
+    {
+      "epoch": 1.6263013698630138,
+      "grad_norm": 2.2476744651794434,
+      "learning_rate": 2.5731112916328187e-06,
+      "loss": 1.8347,
+      "step": 1484
+    },
+    {
+      "epoch": 1.6273972602739726,
+      "grad_norm": 2.3128387928009033,
+      "learning_rate": 2.5710804224207964e-06,
+      "loss": 1.8403,
+      "step": 1485
+    },
+    {
+      "epoch": 1.6284931506849314,
+      "grad_norm": 2.3483974933624268,
+      "learning_rate": 2.5690495532087733e-06,
+      "loss": 1.781,
+      "step": 1486
+    },
+    {
+      "epoch": 1.6295890410958904,
+      "grad_norm": 2.328298568725586,
+      "learning_rate": 2.567018683996751e-06,
+      "loss": 1.7356,
+      "step": 1487
+    },
+    {
+      "epoch": 1.6306849315068495,
+      "grad_norm": 2.2756922245025635,
+      "learning_rate": 2.564987814784728e-06,
+      "loss": 1.8449,
+      "step": 1488
+    },
+    {
+      "epoch": 1.6317808219178083,
+      "grad_norm": 2.363874912261963,
+      "learning_rate": 2.5629569455727055e-06,
+      "loss": 1.7767,
+      "step": 1489
+    },
+    {
+      "epoch": 1.632876712328767,
+      "grad_norm": 2.2872767448425293,
+      "learning_rate": 2.5609260763606824e-06,
+      "loss": 1.8052,
+      "step": 1490
+    },
+    {
+      "epoch": 1.632876712328767,
+      "eval_loss": 2.0134243965148926,
+      "eval_runtime": 37.029,
+      "eval_samples_per_second": 27.006,
+      "eval_steps_per_second": 1.134,
+      "step": 1490
+    },
+    {
+      "epoch": 1.633972602739726,
+      "grad_norm": 2.354851007461548,
+      "learning_rate": 2.55889520714866e-06,
+      "loss": 1.7953,
+      "step": 1491
+    },
+    {
+      "epoch": 1.635068493150685,
+      "grad_norm": 2.4440531730651855,
+      "learning_rate": 2.556864337936637e-06,
+      "loss": 1.9063,
+      "step": 1492
+    },
+    {
+      "epoch": 1.6361643835616437,
+      "grad_norm": 2.296835422515869,
+      "learning_rate": 2.5548334687246146e-06,
+      "loss": 1.7961,
+      "step": 1493
+    },
+    {
+      "epoch": 1.6372602739726028,
+      "grad_norm": 2.7421159744262695,
+      "learning_rate": 2.5528025995125915e-06,
+      "loss": 1.809,
+      "step": 1494
+    },
+    {
+      "epoch": 1.6383561643835618,
+      "grad_norm": 2.819125175476074,
+      "learning_rate": 2.550771730300569e-06,
+      "loss": 1.853,
+      "step": 1495
+    },
+    {
+      "epoch": 1.6394520547945204,
+      "grad_norm": 2.2536680698394775,
+      "learning_rate": 2.548740861088546e-06,
+      "loss": 1.8297,
+      "step": 1496
+    },
+    {
+      "epoch": 1.6405479452054794,
+      "grad_norm": 2.7130825519561768,
+      "learning_rate": 2.5467099918765233e-06,
+      "loss": 1.843,
+      "step": 1497
+    },
+    {
+      "epoch": 1.6416438356164385,
+      "grad_norm": 3.0625038146972656,
+      "learning_rate": 2.5446791226645006e-06,
+      "loss": 1.7462,
+      "step": 1498
+    },
+    {
+      "epoch": 1.6427397260273973,
+      "grad_norm": 2.3095271587371826,
+      "learning_rate": 2.542648253452478e-06,
+      "loss": 1.7363,
+      "step": 1499
+    },
+    {
+      "epoch": 1.643835616438356,
+      "grad_norm": 2.536547899246216,
+      "learning_rate": 2.540617384240455e-06,
+      "loss": 1.8422,
+      "step": 1500
+    },
+    {
+      "epoch": 1.643835616438356,
+      "eval_loss": 2.0126097202301025,
+      "eval_runtime": 37.0469,
+      "eval_samples_per_second": 26.993,
+      "eval_steps_per_second": 1.134,
+      "step": 1500
+    },
+    {
+      "epoch": 1.644931506849315,
+      "grad_norm": 2.7454309463500977,
+      "learning_rate": 2.5385865150284324e-06,
+      "loss": 1.8183,
+      "step": 1501
+    },
+    {
+      "epoch": 1.6460273972602741,
+      "grad_norm": 2.3443901538848877,
+      "learning_rate": 2.5365556458164097e-06,
+      "loss": 1.7827,
+      "step": 1502
+    },
+    {
+      "epoch": 1.6471232876712327,
+      "grad_norm": 2.2412657737731934,
+      "learning_rate": 2.534524776604387e-06,
+      "loss": 1.797,
+      "step": 1503
+    },
+    {
+      "epoch": 1.6482191780821918,
+      "grad_norm": 2.6193699836730957,
+      "learning_rate": 2.5324939073923642e-06,
+      "loss": 1.8103,
+      "step": 1504
+    },
+    {
+      "epoch": 1.6493150684931508,
+      "grad_norm": 2.363919258117676,
+      "learning_rate": 2.5304630381803415e-06,
+      "loss": 1.6987,
+      "step": 1505
+    },
+    {
+      "epoch": 1.6504109589041096,
+      "grad_norm": 2.303164482116699,
+      "learning_rate": 2.5284321689683188e-06,
+      "loss": 1.8005,
+      "step": 1506
+    },
+    {
+      "epoch": 1.6515068493150684,
+      "grad_norm": 2.444174289703369,
+      "learning_rate": 2.526401299756296e-06,
+      "loss": 1.8009,
+      "step": 1507
+    },
+    {
+      "epoch": 1.6526027397260274,
+      "grad_norm": 2.418454647064209,
+      "learning_rate": 2.524370430544273e-06,
+      "loss": 1.782,
+      "step": 1508
+    },
+    {
+      "epoch": 1.6536986301369863,
+      "grad_norm": 2.48840069770813,
+      "learning_rate": 2.5223395613322506e-06,
+      "loss": 1.8143,
+      "step": 1509
+    },
+    {
+      "epoch": 1.654794520547945,
+      "grad_norm": 2.2814347743988037,
+      "learning_rate": 2.5203086921202274e-06,
+      "loss": 1.7687,
+      "step": 1510
+    },
+    {
+      "epoch": 1.654794520547945,
+      "eval_loss": 2.0148253440856934,
+      "eval_runtime": 36.7833,
+      "eval_samples_per_second": 27.186,
+      "eval_steps_per_second": 1.142,
+      "step": 1510
+    },
+    {
+      "epoch": 1.655890410958904,
+      "grad_norm": 2.228933572769165,
+      "learning_rate": 2.518277822908205e-06,
+      "loss": 1.8416,
+      "step": 1511
+    },
+    {
+      "epoch": 1.6569863013698631,
+      "grad_norm": 2.3696742057800293,
+      "learning_rate": 2.516246953696182e-06,
+      "loss": 1.7868,
+      "step": 1512
+    },
+    {
+      "epoch": 1.658082191780822,
+      "grad_norm": 2.216360092163086,
+      "learning_rate": 2.5142160844841597e-06,
+      "loss": 1.8312,
+      "step": 1513
+    },
+    {
+      "epoch": 1.6591780821917808,
+      "grad_norm": 2.454345226287842,
+      "learning_rate": 2.5121852152721365e-06,
+      "loss": 1.8309,
+      "step": 1514
+    },
+    {
+      "epoch": 1.6602739726027398,
+      "grad_norm": 2.2106776237487793,
+      "learning_rate": 2.5101543460601142e-06,
+      "loss": 1.8155,
+      "step": 1515
+    },
+    {
+      "epoch": 1.6613698630136986,
+      "grad_norm": 2.289006233215332,
+      "learning_rate": 2.508123476848091e-06,
+      "loss": 1.8162,
+      "step": 1516
+    },
+    {
+      "epoch": 1.6624657534246574,
+      "grad_norm": 2.297246217727661,
+      "learning_rate": 2.506092607636069e-06,
+      "loss": 1.7559,
+      "step": 1517
+    },
+    {
+      "epoch": 1.6635616438356164,
+      "grad_norm": 2.3909482955932617,
+      "learning_rate": 2.5040617384240457e-06,
+      "loss": 1.8389,
+      "step": 1518
+    },
+    {
+      "epoch": 1.6646575342465755,
+      "grad_norm": 2.398171901702881,
+      "learning_rate": 2.5020308692120234e-06,
+      "loss": 1.7785,
+      "step": 1519
+    },
+    {
+      "epoch": 1.6657534246575343,
+      "grad_norm": 2.377707004547119,
+      "learning_rate": 2.5e-06,
+      "loss": 1.768,
+      "step": 1520
+    },
+    {
+      "epoch": 1.6657534246575343,
+      "eval_loss": 2.012815475463867,
+      "eval_runtime": 36.9603,
+      "eval_samples_per_second": 27.056,
+      "eval_steps_per_second": 1.136,
+      "step": 1520
+    },
+    {
+      "epoch": 1.666849315068493,
+      "grad_norm": 2.336160898208618,
+      "learning_rate": 2.4979691307879775e-06,
+      "loss": 1.8653,
+      "step": 1521
+    },
+    {
+      "epoch": 1.6679452054794521,
+      "grad_norm": 2.288424253463745,
+      "learning_rate": 2.4959382615759548e-06,
+      "loss": 1.7798,
+      "step": 1522
+    },
+    {
+      "epoch": 1.669041095890411,
+      "grad_norm": 2.2599472999572754,
+      "learning_rate": 2.493907392363932e-06,
+      "loss": 1.7736,
+      "step": 1523
+    },
+    {
+      "epoch": 1.6701369863013698,
+      "grad_norm": 2.242276430130005,
+      "learning_rate": 2.4918765231519093e-06,
+      "loss": 1.8753,
+      "step": 1524
+    },
+    {
+      "epoch": 1.6712328767123288,
+      "grad_norm": 2.1897733211517334,
+      "learning_rate": 2.4898456539398866e-06,
+      "loss": 1.8353,
+      "step": 1525
+    },
+    {
+      "epoch": 1.6723287671232878,
+      "grad_norm": 2.254009246826172,
+      "learning_rate": 2.487814784727864e-06,
+      "loss": 1.7825,
+      "step": 1526
+    },
+    {
+      "epoch": 1.6734246575342466,
+      "grad_norm": 2.3704216480255127,
+      "learning_rate": 2.4857839155158407e-06,
+      "loss": 1.8742,
+      "step": 1527
+    },
+    {
+      "epoch": 1.6745205479452054,
+      "grad_norm": 2.1590118408203125,
+      "learning_rate": 2.483753046303818e-06,
+      "loss": 1.7974,
+      "step": 1528
+    },
+    {
+      "epoch": 1.6756164383561645,
+      "grad_norm": 2.2389416694641113,
+      "learning_rate": 2.4817221770917953e-06,
+      "loss": 1.8007,
+      "step": 1529
+    },
+    {
+      "epoch": 1.6767123287671233,
+      "grad_norm": 2.324131488800049,
+      "learning_rate": 2.4796913078797725e-06,
+      "loss": 1.8079,
+      "step": 1530
+    },
+    {
+      "epoch": 1.6767123287671233,
+      "eval_loss": 2.013923168182373,
+      "eval_runtime": 37.0828,
+      "eval_samples_per_second": 26.967,
+      "eval_steps_per_second": 1.133,
+      "step": 1530
+    },
+    {
+      "epoch": 1.677808219178082,
+      "grad_norm": 2.406456708908081,
+      "learning_rate": 2.47766043866775e-06,
+      "loss": 1.8356,
+      "step": 1531
+    },
+    {
+      "epoch": 1.6789041095890411,
+      "grad_norm": 2.289919137954712,
+      "learning_rate": 2.475629569455727e-06,
+      "loss": 1.8239,
+      "step": 1532
+    },
+    {
+      "epoch": 1.6800000000000002,
+      "grad_norm": 2.2617733478546143,
+      "learning_rate": 2.4735987002437044e-06,
+      "loss": 1.8025,
+      "step": 1533
+    },
+    {
+      "epoch": 1.6810958904109587,
+      "grad_norm": 2.133652925491333,
+      "learning_rate": 2.4715678310316816e-06,
+      "loss": 1.7743,
+      "step": 1534
+    },
+    {
+      "epoch": 1.6821917808219178,
+      "grad_norm": 2.215322256088257,
+      "learning_rate": 2.469536961819659e-06,
+      "loss": 1.7979,
+      "step": 1535
+    },
+    {
+      "epoch": 1.6832876712328768,
+      "grad_norm": 2.336021900177002,
+      "learning_rate": 2.467506092607636e-06,
+      "loss": 1.8027,
+      "step": 1536
+    },
+    {
+      "epoch": 1.6843835616438356,
+      "grad_norm": 2.3567142486572266,
+      "learning_rate": 2.4654752233956135e-06,
+      "loss": 1.7877,
+      "step": 1537
+    },
+    {
+      "epoch": 1.6854794520547944,
+      "grad_norm": 2.2642927169799805,
+      "learning_rate": 2.4634443541835907e-06,
+      "loss": 1.7695,
+      "step": 1538
+    },
+    {
+      "epoch": 1.6865753424657535,
+      "grad_norm": 2.353303909301758,
+      "learning_rate": 2.461413484971568e-06,
+      "loss": 1.867,
+      "step": 1539
+    },
+    {
+      "epoch": 1.6876712328767123,
+      "grad_norm": 2.252074956893921,
+      "learning_rate": 2.4593826157595453e-06,
+      "loss": 1.8009,
+      "step": 1540
+    },
+    {
+      "epoch": 1.6876712328767123,
+      "eval_loss": 2.0113608837127686,
+      "eval_runtime": 37.0875,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 1540
+    },
+    {
+      "epoch": 1.688767123287671,
+      "grad_norm": 2.2303407192230225,
+      "learning_rate": 2.4573517465475226e-06,
+      "loss": 1.8881,
+      "step": 1541
+    },
+    {
+      "epoch": 1.6898630136986301,
+      "grad_norm": 2.3359081745147705,
+      "learning_rate": 2.4553208773355e-06,
+      "loss": 1.7388,
+      "step": 1542
+    },
+    {
+      "epoch": 1.6909589041095892,
+      "grad_norm": 2.3783583641052246,
+      "learning_rate": 2.453290008123477e-06,
+      "loss": 1.7993,
+      "step": 1543
+    },
+    {
+      "epoch": 1.692054794520548,
+      "grad_norm": 2.159264326095581,
+      "learning_rate": 2.4512591389114544e-06,
+      "loss": 1.81,
+      "step": 1544
+    },
+    {
+      "epoch": 1.6931506849315068,
+      "grad_norm": 2.2232203483581543,
+      "learning_rate": 2.4492282696994317e-06,
+      "loss": 1.8095,
+      "step": 1545
+    },
+    {
+      "epoch": 1.6942465753424658,
+      "grad_norm": 2.1755056381225586,
+      "learning_rate": 2.447197400487409e-06,
+      "loss": 1.8149,
+      "step": 1546
+    },
+    {
+      "epoch": 1.6953424657534246,
+      "grad_norm": 2.325871467590332,
+      "learning_rate": 2.4451665312753862e-06,
+      "loss": 1.8154,
+      "step": 1547
+    },
+    {
+      "epoch": 1.6964383561643834,
+      "grad_norm": 2.286848545074463,
+      "learning_rate": 2.4431356620633635e-06,
+      "loss": 1.8409,
+      "step": 1548
+    },
+    {
+      "epoch": 1.6975342465753425,
+      "grad_norm": 2.359510898590088,
+      "learning_rate": 2.4411047928513408e-06,
+      "loss": 1.843,
+      "step": 1549
+    },
+    {
+      "epoch": 1.6986301369863015,
+      "grad_norm": 2.474428176879883,
+      "learning_rate": 2.439073923639318e-06,
+      "loss": 1.8089,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6986301369863015,
+      "eval_loss": 2.0077731609344482,
+      "eval_runtime": 37.1092,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6997260273972603,
+      "grad_norm": 2.3662912845611572,
+      "learning_rate": 2.4370430544272953e-06,
+      "loss": 1.8176,
+      "step": 1551
+    },
+    {
+      "epoch": 1.7008219178082191,
+      "grad_norm": 2.448948860168457,
+      "learning_rate": 2.435012185215272e-06,
+      "loss": 1.7776,
+      "step": 1552
+    },
+    {
+      "epoch": 1.7019178082191782,
+      "grad_norm": 2.333739757537842,
+      "learning_rate": 2.4329813160032495e-06,
+      "loss": 1.7552,
+      "step": 1553
+    },
+    {
+      "epoch": 1.703013698630137,
+      "grad_norm": 2.170365571975708,
+      "learning_rate": 2.4309504467912267e-06,
+      "loss": 1.7759,
+      "step": 1554
+    },
+    {
+      "epoch": 1.7041095890410958,
+      "grad_norm": 2.2343010902404785,
+      "learning_rate": 2.428919577579204e-06,
+      "loss": 1.8073,
+      "step": 1555
+    },
+    {
+      "epoch": 1.7052054794520548,
+      "grad_norm": 2.2245287895202637,
+      "learning_rate": 2.4268887083671813e-06,
+      "loss": 1.7826,
+      "step": 1556
+    },
+    {
+      "epoch": 1.7063013698630138,
+      "grad_norm": 2.339942216873169,
+      "learning_rate": 2.4248578391551586e-06,
+      "loss": 1.8766,
+      "step": 1557
+    },
+    {
+      "epoch": 1.7073972602739727,
+      "grad_norm": 2.1585116386413574,
+      "learning_rate": 2.422826969943136e-06,
+      "loss": 1.7605,
+      "step": 1558
+    },
+    {
+      "epoch": 1.7084931506849315,
+      "grad_norm": 2.2108142375946045,
+      "learning_rate": 2.420796100731113e-06,
+      "loss": 1.7855,
+      "step": 1559
+    },
+    {
+      "epoch": 1.7095890410958905,
+      "grad_norm": 2.192457675933838,
+      "learning_rate": 2.4187652315190904e-06,
+      "loss": 1.7489,
+      "step": 1560
+    },
+    {
+      "epoch": 1.7095890410958905,
+      "eval_loss": 2.0085604190826416,
+      "eval_runtime": 37.1105,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1560
+    },
+    {
+      "epoch": 1.7106849315068493,
+      "grad_norm": 2.291106939315796,
+      "learning_rate": 2.4167343623070677e-06,
+      "loss": 1.7981,
+      "step": 1561
+    },
+    {
+      "epoch": 1.7117808219178081,
+      "grad_norm": 2.509237766265869,
+      "learning_rate": 2.414703493095045e-06,
+      "loss": 1.8146,
+      "step": 1562
+    },
+    {
+      "epoch": 1.7128767123287671,
+      "grad_norm": 2.497162342071533,
+      "learning_rate": 2.4126726238830222e-06,
+      "loss": 1.807,
+      "step": 1563
+    },
+    {
+      "epoch": 1.7139726027397262,
+      "grad_norm": 2.256833076477051,
+      "learning_rate": 2.410641754670999e-06,
+      "loss": 1.8304,
+      "step": 1564
+    },
+    {
+      "epoch": 1.7150684931506848,
+      "grad_norm": 2.2523772716522217,
+      "learning_rate": 2.4086108854589763e-06,
+      "loss": 1.7794,
+      "step": 1565
+    },
+    {
+      "epoch": 1.7161643835616438,
+      "grad_norm": 2.289182424545288,
+      "learning_rate": 2.4065800162469536e-06,
+      "loss": 1.8425,
+      "step": 1566
+    },
+    {
+      "epoch": 1.7172602739726028,
+      "grad_norm": 2.734487771987915,
+      "learning_rate": 2.404549147034931e-06,
+      "loss": 1.7791,
+      "step": 1567
+    },
+    {
+      "epoch": 1.7183561643835616,
+      "grad_norm": 2.3307888507843018,
+      "learning_rate": 2.402518277822908e-06,
+      "loss": 1.8064,
+      "step": 1568
+    },
+    {
+      "epoch": 1.7194520547945205,
+      "grad_norm": 2.5016298294067383,
+      "learning_rate": 2.4004874086108854e-06,
+      "loss": 1.8544,
+      "step": 1569
+    },
+    {
+      "epoch": 1.7205479452054795,
+      "grad_norm": 2.142246961593628,
+      "learning_rate": 2.3984565393988627e-06,
+      "loss": 1.7653,
+      "step": 1570
+    },
+    {
+      "epoch": 1.7205479452054795,
+      "eval_loss": 2.0068864822387695,
+      "eval_runtime": 37.1078,
+      "eval_samples_per_second": 26.949,
+      "eval_steps_per_second": 1.132,
+      "step": 1570
+    },
+    {
+      "epoch": 1.7216438356164385,
+      "grad_norm": 2.5692505836486816,
+      "learning_rate": 2.39642567018684e-06,
+      "loss": 1.8444,
+      "step": 1571
+    },
+    {
+      "epoch": 1.7227397260273971,
+      "grad_norm": 2.2596213817596436,
+      "learning_rate": 2.3943948009748173e-06,
+      "loss": 1.7694,
+      "step": 1572
+    },
+    {
+      "epoch": 1.7238356164383561,
+      "grad_norm": 2.230984926223755,
+      "learning_rate": 2.3923639317627945e-06,
+      "loss": 1.786,
+      "step": 1573
+    },
+    {
+      "epoch": 1.7249315068493152,
+      "grad_norm": 2.366178274154663,
+      "learning_rate": 2.390333062550772e-06,
+      "loss": 1.8138,
+      "step": 1574
+    },
+    {
+      "epoch": 1.726027397260274,
+      "grad_norm": 2.23783016204834,
+      "learning_rate": 2.388302193338749e-06,
+      "loss": 1.7615,
+      "step": 1575
+    },
+    {
+      "epoch": 1.7271232876712328,
+      "grad_norm": 2.3766677379608154,
+      "learning_rate": 2.3862713241267264e-06,
+      "loss": 1.6811,
+      "step": 1576
+    },
+    {
+      "epoch": 1.7282191780821918,
+      "grad_norm": 2.3434040546417236,
+      "learning_rate": 2.3842404549147037e-06,
+      "loss": 1.8806,
+      "step": 1577
+    },
+    {
+      "epoch": 1.7293150684931506,
+      "grad_norm": 2.2444276809692383,
+      "learning_rate": 2.382209585702681e-06,
+      "loss": 1.7968,
+      "step": 1578
+    },
+    {
+      "epoch": 1.7304109589041095,
+      "grad_norm": 2.2524919509887695,
+      "learning_rate": 2.380178716490658e-06,
+      "loss": 1.7602,
+      "step": 1579
+    },
+    {
+      "epoch": 1.7315068493150685,
+      "grad_norm": 2.3310587406158447,
+      "learning_rate": 2.3781478472786355e-06,
+      "loss": 1.8414,
+      "step": 1580
+    },
+    {
+      "epoch": 1.7315068493150685,
+      "eval_loss": 1.9981482028961182,
+      "eval_runtime": 37.0827,
+      "eval_samples_per_second": 26.967,
+      "eval_steps_per_second": 1.133,
+      "step": 1580
+    },
+    {
+      "epoch": 1.7326027397260275,
+      "grad_norm": 2.4081802368164062,
+      "learning_rate": 2.3761169780666128e-06,
+      "loss": 1.8002,
+      "step": 1581
+    },
+    {
+      "epoch": 1.7336986301369863,
+      "grad_norm": 2.293506622314453,
+      "learning_rate": 2.37408610885459e-06,
+      "loss": 1.8365,
+      "step": 1582
+    },
+    {
+      "epoch": 1.7347945205479451,
+      "grad_norm": 2.2015790939331055,
+      "learning_rate": 2.3720552396425673e-06,
+      "loss": 1.8571,
+      "step": 1583
+    },
+    {
+      "epoch": 1.7358904109589042,
+      "grad_norm": 2.308809518814087,
+      "learning_rate": 2.3700243704305446e-06,
+      "loss": 1.7821,
+      "step": 1584
+    },
+    {
+      "epoch": 1.736986301369863,
+      "grad_norm": 2.376842975616455,
+      "learning_rate": 2.367993501218522e-06,
+      "loss": 1.8701,
+      "step": 1585
+    },
+    {
+      "epoch": 1.7380821917808218,
+      "grad_norm": 2.4102022647857666,
+      "learning_rate": 2.365962632006499e-06,
+      "loss": 1.8055,
+      "step": 1586
+    },
+    {
+      "epoch": 1.7391780821917808,
+      "grad_norm": 2.2734079360961914,
+      "learning_rate": 2.3639317627944764e-06,
+      "loss": 1.8117,
+      "step": 1587
+    },
+    {
+      "epoch": 1.7402739726027399,
+      "grad_norm": 2.3426599502563477,
+      "learning_rate": 2.3619008935824537e-06,
+      "loss": 1.8337,
+      "step": 1588
+    },
+    {
+      "epoch": 1.7413698630136987,
+      "grad_norm": 2.340383768081665,
+      "learning_rate": 2.3598700243704305e-06,
+      "loss": 1.8247,
+      "step": 1589
+    },
+    {
+      "epoch": 1.7424657534246575,
+      "grad_norm": 2.2561380863189697,
+      "learning_rate": 2.357839155158408e-06,
+      "loss": 1.8249,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7424657534246575,
+      "eval_loss": 1.9979513883590698,
+      "eval_runtime": 37.0948,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7435616438356165,
+      "grad_norm": 2.2459654808044434,
+      "learning_rate": 2.355808285946385e-06,
+      "loss": 1.8273,
+      "step": 1591
+    },
+    {
+      "epoch": 1.7446575342465753,
+      "grad_norm": 2.2017974853515625,
+      "learning_rate": 2.3537774167343624e-06,
+      "loss": 1.828,
+      "step": 1592
+    },
+    {
+      "epoch": 1.7457534246575341,
+      "grad_norm": 2.286304235458374,
+      "learning_rate": 2.3517465475223396e-06,
+      "loss": 1.8497,
+      "step": 1593
+    },
+    {
+      "epoch": 1.7468493150684932,
+      "grad_norm": 2.296494722366333,
+      "learning_rate": 2.349715678310317e-06,
+      "loss": 1.8155,
+      "step": 1594
+    },
+    {
+      "epoch": 1.7479452054794522,
+      "grad_norm": 2.2989203929901123,
+      "learning_rate": 2.347684809098294e-06,
+      "loss": 1.8214,
+      "step": 1595
+    },
+    {
+      "epoch": 1.749041095890411,
+      "grad_norm": 2.3034965991973877,
+      "learning_rate": 2.3456539398862715e-06,
+      "loss": 1.7792,
+      "step": 1596
+    },
+    {
+      "epoch": 1.7501369863013698,
+      "grad_norm": 2.3823719024658203,
+      "learning_rate": 2.3436230706742487e-06,
+      "loss": 1.8326,
+      "step": 1597
+    },
+    {
+      "epoch": 1.7512328767123289,
+      "grad_norm": 2.231095552444458,
+      "learning_rate": 2.341592201462226e-06,
+      "loss": 1.7713,
+      "step": 1598
+    },
+    {
+      "epoch": 1.7523287671232877,
+      "grad_norm": 2.287524938583374,
+      "learning_rate": 2.3395613322502033e-06,
+      "loss": 1.8143,
+      "step": 1599
+    },
+    {
+      "epoch": 1.7534246575342465,
+      "grad_norm": 2.267552614212036,
+      "learning_rate": 2.3375304630381806e-06,
+      "loss": 1.7883,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7534246575342465,
+      "eval_loss": 1.998821496963501,
+      "eval_runtime": 37.1222,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7545205479452055,
+      "grad_norm": 2.2986714839935303,
+      "learning_rate": 2.335499593826158e-06,
+      "loss": 1.7615,
+      "step": 1601
+    },
+    {
+      "epoch": 1.7556164383561645,
+      "grad_norm": 2.2913246154785156,
+      "learning_rate": 2.333468724614135e-06,
+      "loss": 1.7584,
+      "step": 1602
+    },
+    {
+      "epoch": 1.7567123287671231,
+      "grad_norm": 2.2486486434936523,
+      "learning_rate": 2.3314378554021124e-06,
+      "loss": 1.8089,
+      "step": 1603
+    },
+    {
+      "epoch": 1.7578082191780822,
+      "grad_norm": 2.319581985473633,
+      "learning_rate": 2.3294069861900897e-06,
+      "loss": 1.8364,
+      "step": 1604
+    },
+    {
+      "epoch": 1.7589041095890412,
+      "grad_norm": 2.2775609493255615,
+      "learning_rate": 2.327376116978067e-06,
+      "loss": 1.7661,
+      "step": 1605
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 2.3445472717285156,
+      "learning_rate": 2.3253452477660442e-06,
+      "loss": 1.7583,
+      "step": 1606
+    },
+    {
+      "epoch": 1.7610958904109588,
+      "grad_norm": 2.420865058898926,
+      "learning_rate": 2.3233143785540215e-06,
+      "loss": 1.8411,
+      "step": 1607
+    },
+    {
+      "epoch": 1.7621917808219179,
+      "grad_norm": 2.294821262359619,
+      "learning_rate": 2.3212835093419988e-06,
+      "loss": 1.7812,
+      "step": 1608
+    },
+    {
+      "epoch": 1.7632876712328767,
+      "grad_norm": 2.332784414291382,
+      "learning_rate": 2.319252640129976e-06,
+      "loss": 1.8252,
+      "step": 1609
+    },
+    {
+      "epoch": 1.7643835616438355,
+      "grad_norm": 2.433619499206543,
+      "learning_rate": 2.3172217709179533e-06,
+      "loss": 1.7709,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7643835616438355,
+      "eval_loss": 2.0010287761688232,
+      "eval_runtime": 36.7858,
+      "eval_samples_per_second": 27.184,
+      "eval_steps_per_second": 1.142,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7654794520547945,
+      "grad_norm": 2.532618522644043,
+      "learning_rate": 2.3151909017059306e-06,
+      "loss": 1.8677,
+      "step": 1611
+    },
+    {
+      "epoch": 1.7665753424657535,
+      "grad_norm": 2.354578733444214,
+      "learning_rate": 2.313160032493908e-06,
+      "loss": 1.8412,
+      "step": 1612
+    },
+    {
+      "epoch": 1.7676712328767124,
+      "grad_norm": 2.382774591445923,
+      "learning_rate": 2.311129163281885e-06,
+      "loss": 1.7543,
+      "step": 1613
+    },
+    {
+      "epoch": 1.7687671232876712,
+      "grad_norm": 2.3443398475646973,
+      "learning_rate": 2.309098294069862e-06,
+      "loss": 1.8267,
+      "step": 1614
+    },
+    {
+      "epoch": 1.7698630136986302,
+      "grad_norm": 2.3804502487182617,
+      "learning_rate": 2.3070674248578393e-06,
+      "loss": 1.7582,
+      "step": 1615
+    },
+    {
+      "epoch": 1.770958904109589,
+      "grad_norm": 2.3163580894470215,
+      "learning_rate": 2.3050365556458166e-06,
+      "loss": 1.7539,
+      "step": 1616
+    },
+    {
+      "epoch": 1.7720547945205478,
+      "grad_norm": 2.605738639831543,
+      "learning_rate": 2.303005686433794e-06,
+      "loss": 1.8124,
+      "step": 1617
+    },
+    {
+      "epoch": 1.7731506849315068,
+      "grad_norm": 2.3859915733337402,
+      "learning_rate": 2.300974817221771e-06,
+      "loss": 1.8271,
+      "step": 1618
+    },
+    {
+      "epoch": 1.7742465753424659,
+      "grad_norm": 2.3715479373931885,
+      "learning_rate": 2.2989439480097484e-06,
+      "loss": 1.7823,
+      "step": 1619
+    },
+    {
+      "epoch": 1.7753424657534247,
+      "grad_norm": 2.441100597381592,
+      "learning_rate": 2.2969130787977257e-06,
+      "loss": 1.7461,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7753424657534247,
+      "eval_loss": 2.002152442932129,
+      "eval_runtime": 37.0246,
+      "eval_samples_per_second": 27.009,
+      "eval_steps_per_second": 1.134,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7764383561643835,
+      "grad_norm": 2.4154021739959717,
+      "learning_rate": 2.294882209585703e-06,
+      "loss": 1.8012,
+      "step": 1621
+    },
+    {
+      "epoch": 1.7775342465753425,
+      "grad_norm": 2.333400249481201,
+      "learning_rate": 2.2928513403736802e-06,
+      "loss": 1.8222,
+      "step": 1622
+    },
+    {
+      "epoch": 1.7786301369863013,
+      "grad_norm": 2.3183937072753906,
+      "learning_rate": 2.2908204711616575e-06,
+      "loss": 1.7561,
+      "step": 1623
+    },
+    {
+      "epoch": 1.7797260273972602,
+      "grad_norm": 2.302823066711426,
+      "learning_rate": 2.2887896019496348e-06,
+      "loss": 1.7975,
+      "step": 1624
+    },
+    {
+      "epoch": 1.7808219178082192,
+      "grad_norm": 2.3702917098999023,
+      "learning_rate": 2.286758732737612e-06,
+      "loss": 1.8478,
+      "step": 1625
+    },
+    {
+      "epoch": 1.7819178082191782,
+      "grad_norm": 2.4997549057006836,
+      "learning_rate": 2.284727863525589e-06,
+      "loss": 1.7855,
+      "step": 1626
+    },
+    {
+      "epoch": 1.783013698630137,
+      "grad_norm": 2.296191930770874,
+      "learning_rate": 2.282696994313566e-06,
+      "loss": 1.7922,
+      "step": 1627
+    },
+    {
+      "epoch": 1.7841095890410958,
+      "grad_norm": 2.2567873001098633,
+      "learning_rate": 2.2806661251015434e-06,
+      "loss": 1.8455,
+      "step": 1628
+    },
+    {
+      "epoch": 1.7852054794520549,
+      "grad_norm": 2.393336534500122,
+      "learning_rate": 2.2786352558895207e-06,
+      "loss": 1.8141,
+      "step": 1629
+    },
+    {
+      "epoch": 1.7863013698630137,
+      "grad_norm": 2.610466480255127,
+      "learning_rate": 2.276604386677498e-06,
+      "loss": 1.7875,
+      "step": 1630
+    },
+    {
+      "epoch": 1.7863013698630137,
+      "eval_loss": 1.9985637664794922,
+      "eval_runtime": 37.1001,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 1630
+    },
+    {
+      "epoch": 1.7873972602739725,
+      "grad_norm": 2.2163283824920654,
+      "learning_rate": 2.2745735174654753e-06,
+      "loss": 1.7479,
+      "step": 1631
+    },
+    {
+      "epoch": 1.7884931506849315,
+      "grad_norm": 2.2641165256500244,
+      "learning_rate": 2.2725426482534526e-06,
+      "loss": 1.7749,
+      "step": 1632
+    },
+    {
+      "epoch": 1.7895890410958906,
+      "grad_norm": 2.2786946296691895,
+      "learning_rate": 2.27051177904143e-06,
+      "loss": 1.8403,
+      "step": 1633
+    },
+    {
+      "epoch": 1.7906849315068492,
+      "grad_norm": 2.660356044769287,
+      "learning_rate": 2.268480909829407e-06,
+      "loss": 1.7994,
+      "step": 1634
+    },
+    {
+      "epoch": 1.7917808219178082,
+      "grad_norm": 2.322181224822998,
+      "learning_rate": 2.2664500406173844e-06,
+      "loss": 1.7917,
+      "step": 1635
+    },
+    {
+      "epoch": 1.7928767123287672,
+      "grad_norm": 2.3020682334899902,
+      "learning_rate": 2.2644191714053617e-06,
+      "loss": 1.7983,
+      "step": 1636
+    },
+    {
+      "epoch": 1.793972602739726,
+      "grad_norm": 2.5280368328094482,
+      "learning_rate": 2.262388302193339e-06,
+      "loss": 1.8378,
+      "step": 1637
+    },
+    {
+      "epoch": 1.7950684931506848,
+      "grad_norm": 2.3975653648376465,
+      "learning_rate": 2.260357432981316e-06,
+      "loss": 1.8081,
+      "step": 1638
+    },
+    {
+      "epoch": 1.7961643835616439,
+      "grad_norm": 2.419806957244873,
+      "learning_rate": 2.2583265637692935e-06,
+      "loss": 1.793,
+      "step": 1639
+    },
+    {
+      "epoch": 1.7972602739726027,
+      "grad_norm": 2.4601471424102783,
+      "learning_rate": 2.2562956945572708e-06,
+      "loss": 1.8116,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7972602739726027,
+      "eval_loss": 1.9951634407043457,
+      "eval_runtime": 37.1095,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7983561643835615,
+      "grad_norm": 2.17830753326416,
+      "learning_rate": 2.254264825345248e-06,
+      "loss": 1.7845,
+      "step": 1641
+    },
+    {
+      "epoch": 1.7994520547945205,
+      "grad_norm": 2.3332247734069824,
+      "learning_rate": 2.2522339561332253e-06,
+      "loss": 1.7945,
+      "step": 1642
+    },
+    {
+      "epoch": 1.8005479452054796,
+      "grad_norm": 2.3841960430145264,
+      "learning_rate": 2.2502030869212026e-06,
+      "loss": 1.7994,
+      "step": 1643
+    },
+    {
+      "epoch": 1.8016438356164384,
+      "grad_norm": 2.3669073581695557,
+      "learning_rate": 2.24817221770918e-06,
+      "loss": 1.8053,
+      "step": 1644
+    },
+    {
+      "epoch": 1.8027397260273972,
+      "grad_norm": 2.406858205795288,
+      "learning_rate": 2.246141348497157e-06,
+      "loss": 1.8647,
+      "step": 1645
+    },
+    {
+      "epoch": 1.8038356164383562,
+      "grad_norm": 2.2854161262512207,
+      "learning_rate": 2.2441104792851344e-06,
+      "loss": 1.8312,
+      "step": 1646
+    },
+    {
+      "epoch": 1.804931506849315,
+      "grad_norm": 2.252976655960083,
+      "learning_rate": 2.2420796100731117e-06,
+      "loss": 1.7834,
+      "step": 1647
+    },
+    {
+      "epoch": 1.8060273972602738,
+      "grad_norm": 2.4131624698638916,
+      "learning_rate": 2.240048740861089e-06,
+      "loss": 1.8204,
+      "step": 1648
+    },
+    {
+      "epoch": 1.8071232876712329,
+      "grad_norm": 2.4727940559387207,
+      "learning_rate": 2.2380178716490662e-06,
+      "loss": 1.8326,
+      "step": 1649
+    },
+    {
+      "epoch": 1.808219178082192,
+      "grad_norm": 2.4432709217071533,
+      "learning_rate": 2.2359870024370435e-06,
+      "loss": 1.6911,
+      "step": 1650
+    },
+    {
+      "epoch": 1.808219178082192,
+      "eval_loss": 1.996634840965271,
+      "eval_runtime": 37.072,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 1650
+    },
+    {
+      "epoch": 1.8093150684931507,
+      "grad_norm": 2.3810362815856934,
+      "learning_rate": 2.2339561332250204e-06,
+      "loss": 1.8256,
+      "step": 1651
+    },
+    {
+      "epoch": 1.8104109589041095,
+      "grad_norm": 2.4489142894744873,
+      "learning_rate": 2.2319252640129976e-06,
+      "loss": 1.8368,
+      "step": 1652
+    },
+    {
+      "epoch": 1.8115068493150686,
+      "grad_norm": 2.458686113357544,
+      "learning_rate": 2.229894394800975e-06,
+      "loss": 1.8176,
+      "step": 1653
+    },
+    {
+      "epoch": 1.8126027397260274,
+      "grad_norm": 2.3977842330932617,
+      "learning_rate": 2.227863525588952e-06,
+      "loss": 1.7516,
+      "step": 1654
+    },
+    {
+      "epoch": 1.8136986301369862,
+      "grad_norm": 2.3777663707733154,
+      "learning_rate": 2.2258326563769295e-06,
+      "loss": 1.732,
+      "step": 1655
+    },
+    {
+      "epoch": 1.8147945205479452,
+      "grad_norm": 2.2874817848205566,
+      "learning_rate": 2.2238017871649067e-06,
+      "loss": 1.749,
+      "step": 1656
+    },
+    {
+      "epoch": 1.8158904109589042,
+      "grad_norm": 2.288114547729492,
+      "learning_rate": 2.221770917952884e-06,
+      "loss": 1.7357,
+      "step": 1657
+    },
+    {
+      "epoch": 1.816986301369863,
+      "grad_norm": 2.3532187938690186,
+      "learning_rate": 2.2197400487408613e-06,
+      "loss": 1.7953,
+      "step": 1658
+    },
+    {
+      "epoch": 1.8180821917808219,
+      "grad_norm": 2.2829396724700928,
+      "learning_rate": 2.2177091795288386e-06,
+      "loss": 1.8358,
+      "step": 1659
+    },
+    {
+      "epoch": 1.819178082191781,
+      "grad_norm": 2.3282570838928223,
+      "learning_rate": 2.215678310316816e-06,
+      "loss": 1.7448,
+      "step": 1660
+    },
+    {
+      "epoch": 1.819178082191781,
+      "eval_loss": 1.999374270439148,
+      "eval_runtime": 37.0739,
+      "eval_samples_per_second": 26.973,
+      "eval_steps_per_second": 1.133,
+      "step": 1660
+    },
+    {
+      "epoch": 1.8202739726027397,
+      "grad_norm": 2.474106788635254,
+      "learning_rate": 2.213647441104793e-06,
+      "loss": 1.7979,
+      "step": 1661
+    },
+    {
+      "epoch": 1.8213698630136985,
+      "grad_norm": 2.159498453140259,
+      "learning_rate": 2.2116165718927704e-06,
+      "loss": 1.8075,
+      "step": 1662
+    },
+    {
+      "epoch": 1.8224657534246576,
+      "grad_norm": 2.2503573894500732,
+      "learning_rate": 2.2095857026807473e-06,
+      "loss": 1.7742,
+      "step": 1663
+    },
+    {
+      "epoch": 1.8235616438356166,
+      "grad_norm": 2.4465060234069824,
+      "learning_rate": 2.2075548334687245e-06,
+      "loss": 1.8067,
+      "step": 1664
+    },
+    {
+      "epoch": 1.8246575342465754,
+      "grad_norm": 2.462158679962158,
+      "learning_rate": 2.205523964256702e-06,
+      "loss": 1.7456,
+      "step": 1665
+    },
+    {
+      "epoch": 1.8257534246575342,
+      "grad_norm": 2.354051113128662,
+      "learning_rate": 2.203493095044679e-06,
+      "loss": 1.7688,
+      "step": 1666
+    },
+    {
+      "epoch": 1.8268493150684932,
+      "grad_norm": 2.3165128231048584,
+      "learning_rate": 2.2014622258326564e-06,
+      "loss": 1.7986,
+      "step": 1667
+    },
+    {
+      "epoch": 1.827945205479452,
+      "grad_norm": 2.3131113052368164,
+      "learning_rate": 2.1994313566206336e-06,
+      "loss": 1.8348,
+      "step": 1668
+    },
+    {
+      "epoch": 1.8290410958904109,
+      "grad_norm": 2.303086996078491,
+      "learning_rate": 2.197400487408611e-06,
+      "loss": 1.8413,
+      "step": 1669
+    },
+    {
+      "epoch": 1.83013698630137,
+      "grad_norm": 2.392995834350586,
+      "learning_rate": 2.195369618196588e-06,
+      "loss": 1.7663,
+      "step": 1670
+    },
+    {
+      "epoch": 1.83013698630137,
+      "eval_loss": 1.9936721324920654,
+      "eval_runtime": 37.0717,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 1670
+    },
+    {
+      "epoch": 1.831232876712329,
+      "grad_norm": 2.323596477508545,
+      "learning_rate": 2.1933387489845655e-06,
+      "loss": 1.7463,
+      "step": 1671
+    },
+    {
+      "epoch": 1.8323287671232875,
+      "grad_norm": 2.3553171157836914,
+      "learning_rate": 2.1913078797725427e-06,
+      "loss": 1.7844,
+      "step": 1672
+    },
+    {
+      "epoch": 1.8334246575342465,
+      "grad_norm": 2.3724591732025146,
+      "learning_rate": 2.18927701056052e-06,
+      "loss": 1.7955,
+      "step": 1673
+    },
+    {
+      "epoch": 1.8345205479452056,
+      "grad_norm": 2.4300284385681152,
+      "learning_rate": 2.1872461413484973e-06,
+      "loss": 1.8697,
+      "step": 1674
+    },
+    {
+      "epoch": 1.8356164383561644,
+      "grad_norm": 2.4300284385681152,
+      "learning_rate": 2.1872461413484973e-06,
+      "loss": 1.8065,
+      "step": 1675
+    },
+    {
+      "epoch": 1.8367123287671232,
+      "grad_norm": 2.523627519607544,
+      "learning_rate": 2.1852152721364746e-06,
+      "loss": 1.801,
+      "step": 1676
+    },
+    {
+      "epoch": 1.8378082191780822,
+      "grad_norm": 2.2884891033172607,
+      "learning_rate": 2.183184402924452e-06,
+      "loss": 1.7512,
+      "step": 1677
+    },
+    {
+      "epoch": 1.838904109589041,
+      "grad_norm": 2.3114986419677734,
+      "learning_rate": 2.181153533712429e-06,
+      "loss": 1.8245,
+      "step": 1678
+    },
+    {
+      "epoch": 1.8399999999999999,
+      "grad_norm": 2.326855421066284,
+      "learning_rate": 2.1791226645004064e-06,
+      "loss": 1.7921,
+      "step": 1679
+    },
+    {
+      "epoch": 1.841095890410959,
+      "grad_norm": 2.3448753356933594,
+      "learning_rate": 2.1770917952883837e-06,
+      "loss": 1.7709,
+      "step": 1680
+    },
+    {
+      "epoch": 1.841095890410959,
+      "eval_loss": 1.996195673942566,
+      "eval_runtime": 37.0481,
+      "eval_samples_per_second": 26.992,
+      "eval_steps_per_second": 1.134,
+      "step": 1680
+    },
+    {
+      "epoch": 1.842191780821918,
+      "grad_norm": 2.223470687866211,
+      "learning_rate": 2.175060926076361e-06,
+      "loss": 1.8468,
+      "step": 1681
+    },
+    {
+      "epoch": 1.8432876712328767,
+      "grad_norm": 2.29726505279541,
+      "learning_rate": 2.1730300568643382e-06,
+      "loss": 1.8753,
+      "step": 1682
+    },
+    {
+      "epoch": 1.8443835616438355,
+      "grad_norm": 2.3869571685791016,
+      "learning_rate": 2.1709991876523155e-06,
+      "loss": 1.7835,
+      "step": 1683
+    },
+    {
+      "epoch": 1.8454794520547946,
+      "grad_norm": 2.192608594894409,
+      "learning_rate": 2.1689683184402928e-06,
+      "loss": 1.7556,
+      "step": 1684
+    },
+    {
+      "epoch": 1.8465753424657534,
+      "grad_norm": 2.355975866317749,
+      "learning_rate": 2.16693744922827e-06,
+      "loss": 1.8323,
+      "step": 1685
+    },
+    {
+      "epoch": 1.8476712328767122,
+      "grad_norm": 2.243938446044922,
+      "learning_rate": 2.1649065800162473e-06,
+      "loss": 1.7942,
+      "step": 1686
+    },
+    {
+      "epoch": 1.8487671232876712,
+      "grad_norm": 2.337069511413574,
+      "learning_rate": 2.1628757108042246e-06,
+      "loss": 1.7709,
+      "step": 1687
+    },
+    {
+      "epoch": 1.8498630136986303,
+      "grad_norm": 2.257504463195801,
+      "learning_rate": 2.160844841592202e-06,
+      "loss": 1.7496,
+      "step": 1688
+    },
+    {
+      "epoch": 1.850958904109589,
+      "grad_norm": 2.287336826324463,
+      "learning_rate": 2.1588139723801787e-06,
+      "loss": 1.7454,
+      "step": 1689
+    },
+    {
+      "epoch": 1.8520547945205479,
+      "grad_norm": 2.2864668369293213,
+      "learning_rate": 2.156783103168156e-06,
+      "loss": 1.7807,
+      "step": 1690
+    },
+    {
+      "epoch": 1.8520547945205479,
+      "eval_loss": 1.9940879344940186,
+      "eval_runtime": 37.0261,
+      "eval_samples_per_second": 27.008,
+      "eval_steps_per_second": 1.134,
+      "step": 1690
+    },
+    {
+      "epoch": 1.853150684931507,
+      "grad_norm": 2.1782145500183105,
+      "learning_rate": 2.1547522339561333e-06,
+      "loss": 1.7567,
+      "step": 1691
+    },
+    {
+      "epoch": 1.8542465753424657,
+      "grad_norm": 2.252023458480835,
+      "learning_rate": 2.1527213647441106e-06,
+      "loss": 1.811,
+      "step": 1692
+    },
+    {
+      "epoch": 1.8553424657534245,
+      "grad_norm": 2.2532849311828613,
+      "learning_rate": 2.150690495532088e-06,
+      "loss": 1.8491,
+      "step": 1693
+    },
+    {
+      "epoch": 1.8564383561643836,
+      "grad_norm": 2.35648250579834,
+      "learning_rate": 2.148659626320065e-06,
+      "loss": 1.8125,
+      "step": 1694
+    },
+    {
+      "epoch": 1.8575342465753426,
+      "grad_norm": 2.3490521907806396,
+      "learning_rate": 2.1466287571080424e-06,
+      "loss": 1.8035,
+      "step": 1695
+    },
+    {
+      "epoch": 1.8586301369863014,
+      "grad_norm": 2.2851879596710205,
+      "learning_rate": 2.1445978878960197e-06,
+      "loss": 1.7755,
+      "step": 1696
+    },
+    {
+      "epoch": 1.8597260273972602,
+      "grad_norm": 2.166754961013794,
+      "learning_rate": 2.142567018683997e-06,
+      "loss": 1.7861,
+      "step": 1697
+    },
+    {
+      "epoch": 1.8608219178082193,
+      "grad_norm": 2.2356410026550293,
+      "learning_rate": 2.140536149471974e-06,
+      "loss": 1.8412,
+      "step": 1698
+    },
+    {
+      "epoch": 1.861917808219178,
+      "grad_norm": 2.215618133544922,
+      "learning_rate": 2.1385052802599515e-06,
+      "loss": 1.7623,
+      "step": 1699
+    },
+    {
+      "epoch": 1.8630136986301369,
+      "grad_norm": 2.283298969268799,
+      "learning_rate": 2.1364744110479288e-06,
+      "loss": 1.7692,
+      "step": 1700
+    },
+    {
+      "epoch": 1.8630136986301369,
+      "eval_loss": 1.9961501359939575,
+      "eval_runtime": 37.0338,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1700
+    },
+    {
+      "epoch": 1.864109589041096,
+      "grad_norm": 2.292818546295166,
+      "learning_rate": 2.134443541835906e-06,
+      "loss": 1.8251,
+      "step": 1701
+    },
+    {
+      "epoch": 1.865205479452055,
+      "grad_norm": 2.235895872116089,
+      "learning_rate": 2.132412672623883e-06,
+      "loss": 1.7726,
+      "step": 1702
+    },
+    {
+      "epoch": 1.8663013698630135,
+      "grad_norm": 2.239413022994995,
+      "learning_rate": 2.13038180341186e-06,
+      "loss": 1.7509,
+      "step": 1703
+    },
+    {
+      "epoch": 1.8673972602739726,
+      "grad_norm": 2.4125585556030273,
+      "learning_rate": 2.1283509341998374e-06,
+      "loss": 1.7973,
+      "step": 1704
+    },
+    {
+      "epoch": 1.8684931506849316,
+      "grad_norm": 2.3774733543395996,
+      "learning_rate": 2.1263200649878147e-06,
+      "loss": 1.8526,
+      "step": 1705
+    },
+    {
+      "epoch": 1.8695890410958904,
+      "grad_norm": 2.344918966293335,
+      "learning_rate": 2.124289195775792e-06,
+      "loss": 1.7708,
+      "step": 1706
+    },
+    {
+      "epoch": 1.8706849315068492,
+      "grad_norm": 2.1933865547180176,
+      "learning_rate": 2.1222583265637693e-06,
+      "loss": 1.7935,
+      "step": 1707
+    },
+    {
+      "epoch": 1.8717808219178083,
+      "grad_norm": 2.4939582347869873,
+      "learning_rate": 2.1202274573517465e-06,
+      "loss": 1.706,
+      "step": 1708
+    },
+    {
+      "epoch": 1.872876712328767,
+      "grad_norm": 2.407532215118408,
+      "learning_rate": 2.118196588139724e-06,
+      "loss": 1.7522,
+      "step": 1709
+    },
+    {
+      "epoch": 1.8739726027397259,
+      "grad_norm": 2.297900915145874,
+      "learning_rate": 2.116165718927701e-06,
+      "loss": 1.7558,
+      "step": 1710
+    },
+    {
+      "epoch": 1.8739726027397259,
+      "eval_loss": 1.9946404695510864,
+      "eval_runtime": 36.7569,
+      "eval_samples_per_second": 27.206,
+      "eval_steps_per_second": 1.143,
+      "step": 1710
+    },
+    {
+      "epoch": 1.875068493150685,
+      "grad_norm": 2.2298660278320312,
+      "learning_rate": 2.1141348497156784e-06,
+      "loss": 1.7836,
+      "step": 1711
+    },
+    {
+      "epoch": 1.876164383561644,
+      "grad_norm": 2.308060884475708,
+      "learning_rate": 2.1121039805036556e-06,
+      "loss": 1.7616,
+      "step": 1712
+    },
+    {
+      "epoch": 1.8772602739726028,
+      "grad_norm": 2.288381338119507,
+      "learning_rate": 2.110073111291633e-06,
+      "loss": 1.7502,
+      "step": 1713
+    },
+    {
+      "epoch": 1.8783561643835616,
+      "grad_norm": 2.2191097736358643,
+      "learning_rate": 2.10804224207961e-06,
+      "loss": 1.7946,
+      "step": 1714
+    },
+    {
+      "epoch": 1.8794520547945206,
+      "grad_norm": 2.211777925491333,
+      "learning_rate": 2.1060113728675875e-06,
+      "loss": 1.8238,
+      "step": 1715
+    },
+    {
+      "epoch": 1.8805479452054794,
+      "grad_norm": 2.3946406841278076,
+      "learning_rate": 2.1039805036555647e-06,
+      "loss": 1.8146,
+      "step": 1716
+    },
+    {
+      "epoch": 1.8816438356164382,
+      "grad_norm": 2.2949914932250977,
+      "learning_rate": 2.101949634443542e-06,
+      "loss": 1.6877,
+      "step": 1717
+    },
+    {
+      "epoch": 1.8827397260273973,
+      "grad_norm": 2.2905256748199463,
+      "learning_rate": 2.0999187652315193e-06,
+      "loss": 1.7868,
+      "step": 1718
+    },
+    {
+      "epoch": 1.8838356164383563,
+      "grad_norm": 2.289280891418457,
+      "learning_rate": 2.0978878960194966e-06,
+      "loss": 1.7387,
+      "step": 1719
+    },
+    {
+      "epoch": 1.884931506849315,
+      "grad_norm": 2.35497784614563,
+      "learning_rate": 2.095857026807474e-06,
+      "loss": 1.8002,
+      "step": 1720
+    },
+    {
+      "epoch": 1.884931506849315,
+      "eval_loss": 1.9928913116455078,
+      "eval_runtime": 36.9289,
+      "eval_samples_per_second": 27.079,
+      "eval_steps_per_second": 1.137,
+      "step": 1720
+    },
+    {
+      "epoch": 1.886027397260274,
+      "grad_norm": 2.4404468536376953,
+      "learning_rate": 2.093826157595451e-06,
+      "loss": 1.7981,
+      "step": 1721
+    },
+    {
+      "epoch": 1.887123287671233,
+      "grad_norm": 2.315718173980713,
+      "learning_rate": 2.0917952883834284e-06,
+      "loss": 1.7581,
+      "step": 1722
+    },
+    {
+      "epoch": 1.8882191780821918,
+      "grad_norm": 2.2984843254089355,
+      "learning_rate": 2.0897644191714057e-06,
+      "loss": 1.8153,
+      "step": 1723
+    },
+    {
+      "epoch": 1.8893150684931506,
+      "grad_norm": 2.3332557678222656,
+      "learning_rate": 2.087733549959383e-06,
+      "loss": 1.831,
+      "step": 1724
+    },
+    {
+      "epoch": 1.8904109589041096,
+      "grad_norm": 2.3394649028778076,
+      "learning_rate": 2.0857026807473602e-06,
+      "loss": 1.7486,
+      "step": 1725
+    },
+    {
+      "epoch": 1.8915068493150686,
+      "grad_norm": 2.2985246181488037,
+      "learning_rate": 2.0836718115353375e-06,
+      "loss": 1.721,
+      "step": 1726
+    },
+    {
+      "epoch": 1.8926027397260274,
+      "grad_norm": 2.318504810333252,
+      "learning_rate": 2.0816409423233144e-06,
+      "loss": 1.8469,
+      "step": 1727
+    },
+    {
+      "epoch": 1.8936986301369862,
+      "grad_norm": 2.4938831329345703,
+      "learning_rate": 2.0796100731112916e-06,
+      "loss": 1.7525,
+      "step": 1728
+    },
+    {
+      "epoch": 1.8947945205479453,
+      "grad_norm": 2.3391780853271484,
+      "learning_rate": 2.077579203899269e-06,
+      "loss": 1.7733,
+      "step": 1729
+    },
+    {
+      "epoch": 1.895890410958904,
+      "grad_norm": 2.437540054321289,
+      "learning_rate": 2.075548334687246e-06,
+      "loss": 1.7344,
+      "step": 1730
+    },
+    {
+      "epoch": 1.895890410958904,
+      "eval_loss": 1.990493655204773,
+      "eval_runtime": 37.0347,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1730
+    },
+    {
+      "epoch": 1.896986301369863,
+      "grad_norm": 2.413628578186035,
+      "learning_rate": 2.0735174654752235e-06,
+      "loss": 1.7978,
+      "step": 1731
+    },
+    {
+      "epoch": 1.898082191780822,
+      "grad_norm": 2.3673973083496094,
+      "learning_rate": 2.0714865962632007e-06,
+      "loss": 1.7414,
+      "step": 1732
+    },
+    {
+      "epoch": 1.899178082191781,
+      "grad_norm": 2.3961968421936035,
+      "learning_rate": 2.069455727051178e-06,
+      "loss": 1.7854,
+      "step": 1733
+    },
+    {
+      "epoch": 1.9002739726027398,
+      "grad_norm": 2.5583736896514893,
+      "learning_rate": 2.0674248578391553e-06,
+      "loss": 1.7629,
+      "step": 1734
+    },
+    {
+      "epoch": 1.9013698630136986,
+      "grad_norm": 2.256418228149414,
+      "learning_rate": 2.0653939886271326e-06,
+      "loss": 1.7845,
+      "step": 1735
+    },
+    {
+      "epoch": 1.9024657534246576,
+      "grad_norm": 2.4235012531280518,
+      "learning_rate": 2.06336311941511e-06,
+      "loss": 1.8151,
+      "step": 1736
+    },
+    {
+      "epoch": 1.9035616438356164,
+      "grad_norm": 2.2945358753204346,
+      "learning_rate": 2.061332250203087e-06,
+      "loss": 1.7665,
+      "step": 1737
+    },
+    {
+      "epoch": 1.9046575342465752,
+      "grad_norm": 2.371809959411621,
+      "learning_rate": 2.0593013809910644e-06,
+      "loss": 1.7355,
+      "step": 1738
+    },
+    {
+      "epoch": 1.9057534246575343,
+      "grad_norm": 2.26947021484375,
+      "learning_rate": 2.0572705117790417e-06,
+      "loss": 1.7702,
+      "step": 1739
+    },
+    {
+      "epoch": 1.9068493150684933,
+      "grad_norm": 2.314237356185913,
+      "learning_rate": 2.055239642567019e-06,
+      "loss": 1.8084,
+      "step": 1740
+    },
+    {
+      "epoch": 1.9068493150684933,
+      "eval_loss": 1.9961000680923462,
+      "eval_runtime": 37.0775,
+      "eval_samples_per_second": 26.971,
+      "eval_steps_per_second": 1.133,
+      "step": 1740
+    },
+    {
+      "epoch": 1.907945205479452,
+      "grad_norm": 2.248680591583252,
+      "learning_rate": 2.0532087733549962e-06,
+      "loss": 1.7898,
+      "step": 1741
+    },
+    {
+      "epoch": 1.909041095890411,
+      "grad_norm": 2.336378812789917,
+      "learning_rate": 2.0511779041429735e-06,
+      "loss": 1.795,
+      "step": 1742
+    },
+    {
+      "epoch": 1.91013698630137,
+      "grad_norm": 2.2420554161071777,
+      "learning_rate": 2.0491470349309508e-06,
+      "loss": 1.76,
+      "step": 1743
+    },
+    {
+      "epoch": 1.9112328767123288,
+      "grad_norm": 2.3305511474609375,
+      "learning_rate": 2.047116165718928e-06,
+      "loss": 1.7959,
+      "step": 1744
+    },
+    {
+      "epoch": 1.9123287671232876,
+      "grad_norm": 2.2677786350250244,
+      "learning_rate": 2.0450852965069053e-06,
+      "loss": 1.7922,
+      "step": 1745
+    },
+    {
+      "epoch": 1.9134246575342466,
+      "grad_norm": 2.224148988723755,
+      "learning_rate": 2.0430544272948826e-06,
+      "loss": 1.841,
+      "step": 1746
+    },
+    {
+      "epoch": 1.9145205479452054,
+      "grad_norm": 2.172091245651245,
+      "learning_rate": 2.04102355808286e-06,
+      "loss": 1.784,
+      "step": 1747
+    },
+    {
+      "epoch": 1.9156164383561642,
+      "grad_norm": 2.319962501525879,
+      "learning_rate": 2.038992688870837e-06,
+      "loss": 1.7238,
+      "step": 1748
+    },
+    {
+      "epoch": 1.9167123287671233,
+      "grad_norm": 2.299760103225708,
+      "learning_rate": 2.0369618196588144e-06,
+      "loss": 1.7501,
+      "step": 1749
+    },
+    {
+      "epoch": 1.9178082191780823,
+      "grad_norm": 2.218407154083252,
+      "learning_rate": 2.0349309504467917e-06,
+      "loss": 1.7628,
+      "step": 1750
+    },
+    {
+      "epoch": 1.9178082191780823,
+      "eval_loss": 1.9934331178665161,
+      "eval_runtime": 37.0986,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 1750
+    },
+    {
+      "epoch": 1.9189041095890411,
+      "grad_norm": 2.4121744632720947,
+      "learning_rate": 2.032900081234769e-06,
+      "loss": 1.8128,
+      "step": 1751
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 2.253225326538086,
+      "learning_rate": 2.030869212022746e-06,
+      "loss": 1.8161,
+      "step": 1752
+    },
+    {
+      "epoch": 1.921095890410959,
+      "grad_norm": 2.28702449798584,
+      "learning_rate": 2.028838342810723e-06,
+      "loss": 1.7769,
+      "step": 1753
+    },
+    {
+      "epoch": 1.9221917808219178,
+      "grad_norm": 2.3004465103149414,
+      "learning_rate": 2.0268074735987004e-06,
+      "loss": 1.7631,
+      "step": 1754
+    },
+    {
+      "epoch": 1.9232876712328766,
+      "grad_norm": 2.2733960151672363,
+      "learning_rate": 2.0247766043866777e-06,
+      "loss": 1.75,
+      "step": 1755
+    },
+    {
+      "epoch": 1.9243835616438356,
+      "grad_norm": 2.274019241333008,
+      "learning_rate": 2.022745735174655e-06,
+      "loss": 1.8317,
+      "step": 1756
+    },
+    {
+      "epoch": 1.9254794520547946,
+      "grad_norm": 2.3094100952148438,
+      "learning_rate": 2.020714865962632e-06,
+      "loss": 1.7841,
+      "step": 1757
+    },
+    {
+      "epoch": 1.9265753424657535,
+      "grad_norm": 2.269321918487549,
+      "learning_rate": 2.0186839967506095e-06,
+      "loss": 1.8052,
+      "step": 1758
+    },
+    {
+      "epoch": 1.9276712328767123,
+      "grad_norm": 2.5324459075927734,
+      "learning_rate": 2.0166531275385868e-06,
+      "loss": 1.7916,
+      "step": 1759
+    },
+    {
+      "epoch": 1.9287671232876713,
+      "grad_norm": 2.302569627761841,
+      "learning_rate": 2.014622258326564e-06,
+      "loss": 1.7499,
+      "step": 1760
+    },
+    {
+      "epoch": 1.9287671232876713,
+      "eval_loss": 1.9879337549209595,
+      "eval_runtime": 37.1186,
+      "eval_samples_per_second": 26.941,
+      "eval_steps_per_second": 1.132,
+      "step": 1760
+    },
+    {
+      "epoch": 1.9298630136986301,
+      "grad_norm": 2.29470157623291,
+      "learning_rate": 2.0125913891145413e-06,
+      "loss": 1.7922,
+      "step": 1761
+    },
+    {
+      "epoch": 1.930958904109589,
+      "grad_norm": 2.4307754039764404,
+      "learning_rate": 2.0105605199025186e-06,
+      "loss": 1.7802,
+      "step": 1762
+    },
+    {
+      "epoch": 1.932054794520548,
+      "grad_norm": 2.179588794708252,
+      "learning_rate": 2.008529650690496e-06,
+      "loss": 1.7391,
+      "step": 1763
+    },
+    {
+      "epoch": 1.933150684931507,
+      "grad_norm": 2.5824451446533203,
+      "learning_rate": 2.0064987814784727e-06,
+      "loss": 1.8106,
+      "step": 1764
+    },
+    {
+      "epoch": 1.9342465753424658,
+      "grad_norm": 2.32308292388916,
+      "learning_rate": 2.00446791226645e-06,
+      "loss": 1.7689,
+      "step": 1765
+    },
+    {
+      "epoch": 1.9353424657534246,
+      "grad_norm": 2.262533664703369,
+      "learning_rate": 2.0024370430544273e-06,
+      "loss": 1.7604,
+      "step": 1766
+    },
+    {
+      "epoch": 1.9364383561643836,
+      "grad_norm": 2.325253963470459,
+      "learning_rate": 2.0004061738424045e-06,
+      "loss": 1.7695,
+      "step": 1767
+    },
+    {
+      "epoch": 1.9375342465753425,
+      "grad_norm": 2.2855470180511475,
+      "learning_rate": 1.998375304630382e-06,
+      "loss": 1.8233,
+      "step": 1768
+    },
+    {
+      "epoch": 1.9386301369863013,
+      "grad_norm": 2.3688673973083496,
+      "learning_rate": 1.996344435418359e-06,
+      "loss": 1.7799,
+      "step": 1769
+    },
+    {
+      "epoch": 1.9397260273972603,
+      "grad_norm": 2.4682323932647705,
+      "learning_rate": 1.9943135662063364e-06,
+      "loss": 1.8036,
+      "step": 1770
+    },
+    {
+      "epoch": 1.9397260273972603,
+      "eval_loss": 1.9906389713287354,
+      "eval_runtime": 37.0959,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 1770
+    },
+    {
+      "epoch": 1.9408219178082193,
+      "grad_norm": 2.369166374206543,
+      "learning_rate": 1.9922826969943136e-06,
+      "loss": 1.7878,
+      "step": 1771
+    },
+    {
+      "epoch": 1.941917808219178,
+      "grad_norm": 2.4669666290283203,
+      "learning_rate": 1.990251827782291e-06,
+      "loss": 1.8228,
+      "step": 1772
+    },
+    {
+      "epoch": 1.943013698630137,
+      "grad_norm": 2.3083250522613525,
+      "learning_rate": 1.988220958570268e-06,
+      "loss": 1.7997,
+      "step": 1773
+    },
+    {
+      "epoch": 1.944109589041096,
+      "grad_norm": 2.4884281158447266,
+      "learning_rate": 1.9861900893582455e-06,
+      "loss": 1.8105,
+      "step": 1774
+    },
+    {
+      "epoch": 1.9452054794520548,
+      "grad_norm": 2.3504858016967773,
+      "learning_rate": 1.9841592201462227e-06,
+      "loss": 1.8159,
+      "step": 1775
+    },
+    {
+      "epoch": 1.9463013698630136,
+      "grad_norm": 2.618431329727173,
+      "learning_rate": 1.9821283509342e-06,
+      "loss": 1.7352,
+      "step": 1776
+    },
+    {
+      "epoch": 1.9473972602739726,
+      "grad_norm": 2.3452653884887695,
+      "learning_rate": 1.9800974817221773e-06,
+      "loss": 1.7451,
+      "step": 1777
+    },
+    {
+      "epoch": 1.9484931506849315,
+      "grad_norm": 2.5966849327087402,
+      "learning_rate": 1.9780666125101546e-06,
+      "loss": 1.7965,
+      "step": 1778
+    },
+    {
+      "epoch": 1.9495890410958903,
+      "grad_norm": 2.3017661571502686,
+      "learning_rate": 1.976035743298132e-06,
+      "loss": 1.7629,
+      "step": 1779
+    },
+    {
+      "epoch": 1.9506849315068493,
+      "grad_norm": 2.3766098022460938,
+      "learning_rate": 1.974004874086109e-06,
+      "loss": 1.8403,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9506849315068493,
+      "eval_loss": 1.9936374425888062,
+      "eval_runtime": 37.1158,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9517808219178083,
+      "grad_norm": 2.2327704429626465,
+      "learning_rate": 1.9719740048740864e-06,
+      "loss": 1.7096,
+      "step": 1781
+    },
+    {
+      "epoch": 1.9528767123287671,
+      "grad_norm": 2.458707809448242,
+      "learning_rate": 1.9699431356620637e-06,
+      "loss": 1.7636,
+      "step": 1782
+    },
+    {
+      "epoch": 1.953972602739726,
+      "grad_norm": 2.4143803119659424,
+      "learning_rate": 1.967912266450041e-06,
+      "loss": 1.757,
+      "step": 1783
+    },
+    {
+      "epoch": 1.955068493150685,
+      "grad_norm": 2.465148448944092,
+      "learning_rate": 1.9658813972380182e-06,
+      "loss": 1.7526,
+      "step": 1784
+    },
+    {
+      "epoch": 1.9561643835616438,
+      "grad_norm": 2.356210231781006,
+      "learning_rate": 1.9638505280259955e-06,
+      "loss": 1.8705,
+      "step": 1785
+    },
+    {
+      "epoch": 1.9572602739726026,
+      "grad_norm": 2.7456002235412598,
+      "learning_rate": 1.9618196588139728e-06,
+      "loss": 1.7574,
+      "step": 1786
+    },
+    {
+      "epoch": 1.9583561643835616,
+      "grad_norm": 2.3724262714385986,
+      "learning_rate": 1.95978878960195e-06,
+      "loss": 1.7854,
+      "step": 1787
+    },
+    {
+      "epoch": 1.9594520547945207,
+      "grad_norm": 2.275827646255493,
+      "learning_rate": 1.9577579203899273e-06,
+      "loss": 1.8447,
+      "step": 1788
+    },
+    {
+      "epoch": 1.9605479452054795,
+      "grad_norm": 2.369992256164551,
+      "learning_rate": 1.955727051177904e-06,
+      "loss": 1.8299,
+      "step": 1789
+    },
+    {
+      "epoch": 1.9616438356164383,
+      "grad_norm": 2.462372064590454,
+      "learning_rate": 1.9536961819658815e-06,
+      "loss": 1.7596,
+      "step": 1790
+    },
+    {
+      "epoch": 1.9616438356164383,
+      "eval_loss": 1.9947243928909302,
+      "eval_runtime": 37.111,
+      "eval_samples_per_second": 26.946,
+      "eval_steps_per_second": 1.132,
+      "step": 1790
+    },
+    {
+      "epoch": 1.9627397260273973,
+      "grad_norm": 2.3040034770965576,
+      "learning_rate": 1.9516653127538587e-06,
+      "loss": 1.7391,
+      "step": 1791
+    },
+    {
+      "epoch": 1.9638356164383561,
+      "grad_norm": 2.449328899383545,
+      "learning_rate": 1.949634443541836e-06,
+      "loss": 1.7428,
+      "step": 1792
+    },
+    {
+      "epoch": 1.964931506849315,
+      "grad_norm": 2.2679781913757324,
+      "learning_rate": 1.9476035743298133e-06,
+      "loss": 1.738,
+      "step": 1793
+    },
+    {
+      "epoch": 1.966027397260274,
+      "grad_norm": 2.3444032669067383,
+      "learning_rate": 1.9455727051177906e-06,
+      "loss": 1.8089,
+      "step": 1794
+    },
+    {
+      "epoch": 1.967123287671233,
+      "grad_norm": 2.3005306720733643,
+      "learning_rate": 1.943541835905768e-06,
+      "loss": 1.7718,
+      "step": 1795
+    },
+    {
+      "epoch": 1.9682191780821918,
+      "grad_norm": 2.6022629737854004,
+      "learning_rate": 1.941510966693745e-06,
+      "loss": 1.8325,
+      "step": 1796
+    },
+    {
+      "epoch": 1.9693150684931506,
+      "grad_norm": 2.407585382461548,
+      "learning_rate": 1.9394800974817224e-06,
+      "loss": 1.737,
+      "step": 1797
+    },
+    {
+      "epoch": 1.9704109589041097,
+      "grad_norm": 2.349339723587036,
+      "learning_rate": 1.9374492282696997e-06,
+      "loss": 1.81,
+      "step": 1798
+    },
+    {
+      "epoch": 1.9715068493150685,
+      "grad_norm": 2.296192169189453,
+      "learning_rate": 1.935418359057677e-06,
+      "loss": 1.8424,
+      "step": 1799
+    },
+    {
+      "epoch": 1.9726027397260273,
+      "grad_norm": 2.175875663757324,
+      "learning_rate": 1.9333874898456542e-06,
+      "loss": 1.7977,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9726027397260273,
+      "eval_loss": 1.9929841756820679,
+      "eval_runtime": 37.1159,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9736986301369863,
+      "grad_norm": 2.37428617477417,
+      "learning_rate": 1.931356620633631e-06,
+      "loss": 1.8248,
+      "step": 1801
+    },
+    {
+      "epoch": 1.9747945205479454,
+      "grad_norm": 2.297621965408325,
+      "learning_rate": 1.9293257514216083e-06,
+      "loss": 1.7215,
+      "step": 1802
+    },
+    {
+      "epoch": 1.9758904109589042,
+      "grad_norm": 2.217275857925415,
+      "learning_rate": 1.9272948822095856e-06,
+      "loss": 1.7828,
+      "step": 1803
+    },
+    {
+      "epoch": 1.976986301369863,
+      "grad_norm": 2.250854015350342,
+      "learning_rate": 1.925264012997563e-06,
+      "loss": 1.7245,
+      "step": 1804
+    },
+    {
+      "epoch": 1.978082191780822,
+      "grad_norm": 2.2556822299957275,
+      "learning_rate": 1.92323314378554e-06,
+      "loss": 1.7715,
+      "step": 1805
+    },
+    {
+      "epoch": 1.9791780821917808,
+      "grad_norm": 2.316389322280884,
+      "learning_rate": 1.9212022745735174e-06,
+      "loss": 1.8063,
+      "step": 1806
+    },
+    {
+      "epoch": 1.9802739726027396,
+      "grad_norm": 2.4358153343200684,
+      "learning_rate": 1.9191714053614947e-06,
+      "loss": 1.747,
+      "step": 1807
+    },
+    {
+      "epoch": 1.9813698630136987,
+      "grad_norm": 2.437631368637085,
+      "learning_rate": 1.917140536149472e-06,
+      "loss": 1.7485,
+      "step": 1808
+    },
+    {
+      "epoch": 1.9824657534246577,
+      "grad_norm": 2.3554601669311523,
+      "learning_rate": 1.9151096669374493e-06,
+      "loss": 1.7195,
+      "step": 1809
+    },
+    {
+      "epoch": 1.9835616438356163,
+      "grad_norm": 2.575559616088867,
+      "learning_rate": 1.9130787977254266e-06,
+      "loss": 1.8368,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9835616438356163,
+      "eval_loss": 1.9903514385223389,
+      "eval_runtime": 36.8082,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9846575342465753,
+      "grad_norm": 2.3074066638946533,
+      "learning_rate": 1.911047928513404e-06,
+      "loss": 1.7952,
+      "step": 1811
+    },
+    {
+      "epoch": 1.9857534246575343,
+      "grad_norm": 2.3162307739257812,
+      "learning_rate": 1.909017059301381e-06,
+      "loss": 1.7455,
+      "step": 1812
+    },
+    {
+      "epoch": 1.9868493150684932,
+      "grad_norm": 2.3085129261016846,
+      "learning_rate": 1.9069861900893586e-06,
+      "loss": 1.7513,
+      "step": 1813
+    },
+    {
+      "epoch": 1.987945205479452,
+      "grad_norm": 2.3454785346984863,
+      "learning_rate": 1.9049553208773354e-06,
+      "loss": 1.803,
+      "step": 1814
+    },
+    {
+      "epoch": 1.989041095890411,
+      "grad_norm": 2.4522876739501953,
+      "learning_rate": 1.9029244516653127e-06,
+      "loss": 1.7567,
+      "step": 1815
+    },
+    {
+      "epoch": 1.9901369863013698,
+      "grad_norm": 2.3876044750213623,
+      "learning_rate": 1.90089358245329e-06,
+      "loss": 1.7126,
+      "step": 1816
+    },
+    {
+      "epoch": 1.9912328767123286,
+      "grad_norm": 2.4061570167541504,
+      "learning_rate": 1.8988627132412673e-06,
+      "loss": 1.7618,
+      "step": 1817
+    },
+    {
+      "epoch": 1.9923287671232877,
+      "grad_norm": 2.307445764541626,
+      "learning_rate": 1.8968318440292445e-06,
+      "loss": 1.8014,
+      "step": 1818
+    },
+    {
+      "epoch": 1.9934246575342467,
+      "grad_norm": 2.4414806365966797,
+      "learning_rate": 1.8948009748172218e-06,
+      "loss": 1.7202,
+      "step": 1819
+    },
+    {
+      "epoch": 1.9945205479452055,
+      "grad_norm": 2.3933329582214355,
+      "learning_rate": 1.892770105605199e-06,
+      "loss": 1.8133,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9945205479452055,
+      "eval_loss": 1.991702675819397,
+      "eval_runtime": 37.0127,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9956164383561643,
+      "grad_norm": 2.237699508666992,
+      "learning_rate": 1.8907392363931764e-06,
+      "loss": 1.8026,
+      "step": 1821
+    },
+    {
+      "epoch": 1.9967123287671233,
+      "grad_norm": 2.290004014968872,
+      "learning_rate": 1.8887083671811536e-06,
+      "loss": 1.7689,
+      "step": 1822
+    },
+    {
+      "epoch": 1.9978082191780822,
+      "grad_norm": 2.2264676094055176,
+      "learning_rate": 1.886677497969131e-06,
+      "loss": 1.7862,
+      "step": 1823
+    },
+    {
+      "epoch": 1.998904109589041,
+      "grad_norm": 2.2764949798583984,
+      "learning_rate": 1.8846466287571082e-06,
+      "loss": 1.6763,
+      "step": 1824
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 2.3024826049804688,
+      "learning_rate": 1.8826157595450855e-06,
+      "loss": 1.8076,
+      "step": 1825
+    },
+    {
+      "epoch": 2.001095890410959,
+      "grad_norm": 2.362882614135742,
+      "learning_rate": 1.8805848903330625e-06,
+      "loss": 1.7009,
+      "step": 1826
+    },
+    {
+      "epoch": 2.0021917808219176,
+      "grad_norm": 2.432370185852051,
+      "learning_rate": 1.8785540211210398e-06,
+      "loss": 1.8712,
+      "step": 1827
+    },
+    {
+      "epoch": 2.0032876712328767,
+      "grad_norm": 2.3584744930267334,
+      "learning_rate": 1.876523151909017e-06,
+      "loss": 1.7869,
+      "step": 1828
+    },
+    {
+      "epoch": 2.0043835616438357,
+      "grad_norm": 2.534205198287964,
+      "learning_rate": 1.8744922826969944e-06,
+      "loss": 1.7376,
+      "step": 1829
+    },
+    {
+      "epoch": 2.0054794520547947,
+      "grad_norm": 2.58241605758667,
+      "learning_rate": 1.8724614134849716e-06,
+      "loss": 1.7204,
+      "step": 1830
+    },
+    {
+      "epoch": 2.0054794520547947,
+      "eval_loss": 1.9912636280059814,
+      "eval_runtime": 37.0969,
+      "eval_samples_per_second": 26.956,
+      "eval_steps_per_second": 1.132,
+      "step": 1830
+    },
+    {
+      "epoch": 2.0065753424657533,
+      "grad_norm": 2.3742504119873047,
+      "learning_rate": 1.870430544272949e-06,
+      "loss": 1.7537,
+      "step": 1831
+    },
+    {
+      "epoch": 2.0076712328767123,
+      "grad_norm": 2.6560916900634766,
+      "learning_rate": 1.8683996750609262e-06,
+      "loss": 1.8524,
+      "step": 1832
+    },
+    {
+      "epoch": 2.0087671232876714,
+      "grad_norm": 2.523041248321533,
+      "learning_rate": 1.8663688058489035e-06,
+      "loss": 1.7405,
+      "step": 1833
+    },
+    {
+      "epoch": 2.00986301369863,
+      "grad_norm": 2.5444304943084717,
+      "learning_rate": 1.8643379366368807e-06,
+      "loss": 1.7588,
+      "step": 1834
+    },
+    {
+      "epoch": 2.010958904109589,
+      "grad_norm": 2.708984613418579,
+      "learning_rate": 1.862307067424858e-06,
+      "loss": 1.7686,
+      "step": 1835
+    },
+    {
+      "epoch": 2.012054794520548,
+      "grad_norm": 2.726691961288452,
+      "learning_rate": 1.8602761982128353e-06,
+      "loss": 1.7733,
+      "step": 1836
+    },
+    {
+      "epoch": 2.013150684931507,
+      "grad_norm": 2.5085904598236084,
+      "learning_rate": 1.8582453290008126e-06,
+      "loss": 1.81,
+      "step": 1837
+    },
+    {
+      "epoch": 2.0142465753424657,
+      "grad_norm": 2.4699466228485107,
+      "learning_rate": 1.8562144597887898e-06,
+      "loss": 1.7369,
+      "step": 1838
+    },
+    {
+      "epoch": 2.0153424657534247,
+      "grad_norm": 2.531027317047119,
+      "learning_rate": 1.854183590576767e-06,
+      "loss": 1.7611,
+      "step": 1839
+    },
+    {
+      "epoch": 2.0164383561643837,
+      "grad_norm": 2.5061161518096924,
+      "learning_rate": 1.8521527213647442e-06,
+      "loss": 1.811,
+      "step": 1840
+    },
+    {
+      "epoch": 2.0164383561643837,
+      "eval_loss": 1.9893144369125366,
+      "eval_runtime": 37.1262,
+      "eval_samples_per_second": 26.935,
+      "eval_steps_per_second": 1.131,
+      "step": 1840
+    },
+    {
+      "epoch": 2.0175342465753423,
+      "grad_norm": 2.5739994049072266,
+      "learning_rate": 1.8501218521527215e-06,
+      "loss": 1.7153,
+      "step": 1841
+    },
+    {
+      "epoch": 2.0186301369863013,
+      "grad_norm": 2.5673015117645264,
+      "learning_rate": 1.8480909829406987e-06,
+      "loss": 1.6879,
+      "step": 1842
+    },
+    {
+      "epoch": 2.0197260273972604,
+      "grad_norm": 2.717430353164673,
+      "learning_rate": 1.846060113728676e-06,
+      "loss": 1.723,
+      "step": 1843
+    },
+    {
+      "epoch": 2.020821917808219,
+      "grad_norm": 2.4905285835266113,
+      "learning_rate": 1.8440292445166533e-06,
+      "loss": 1.7651,
+      "step": 1844
+    },
+    {
+      "epoch": 2.021917808219178,
+      "grad_norm": 2.6798791885375977,
+      "learning_rate": 1.8419983753046306e-06,
+      "loss": 1.7308,
+      "step": 1845
+    },
+    {
+      "epoch": 2.023013698630137,
+      "grad_norm": 2.591614007949829,
+      "learning_rate": 1.8399675060926078e-06,
+      "loss": 1.786,
+      "step": 1846
+    },
+    {
+      "epoch": 2.024109589041096,
+      "grad_norm": 2.5405824184417725,
+      "learning_rate": 1.8379366368805851e-06,
+      "loss": 1.7699,
+      "step": 1847
+    },
+    {
+      "epoch": 2.0252054794520546,
+      "grad_norm": 2.4189934730529785,
+      "learning_rate": 1.8359057676685624e-06,
+      "loss": 1.732,
+      "step": 1848
+    },
+    {
+      "epoch": 2.0263013698630137,
+      "grad_norm": 2.5857982635498047,
+      "learning_rate": 1.8338748984565397e-06,
+      "loss": 1.7123,
+      "step": 1849
+    },
+    {
+      "epoch": 2.0273972602739727,
+      "grad_norm": 2.4710893630981445,
+      "learning_rate": 1.831844029244517e-06,
+      "loss": 1.7419,
+      "step": 1850
+    },
+    {
+      "epoch": 2.0273972602739727,
+      "eval_loss": 1.9874999523162842,
+      "eval_runtime": 37.1588,
+      "eval_samples_per_second": 26.912,
+      "eval_steps_per_second": 1.13,
+      "step": 1850
+    },
+    {
+      "epoch": 2.0284931506849313,
+      "grad_norm": 2.3414645195007324,
+      "learning_rate": 1.829813160032494e-06,
+      "loss": 1.7102,
+      "step": 1851
+    },
+    {
+      "epoch": 2.0295890410958903,
+      "grad_norm": 2.4148900508880615,
+      "learning_rate": 1.8277822908204713e-06,
+      "loss": 1.7287,
+      "step": 1852
+    },
+    {
+      "epoch": 2.0306849315068494,
+      "grad_norm": 2.5501346588134766,
+      "learning_rate": 1.8257514216084486e-06,
+      "loss": 1.7633,
+      "step": 1853
+    },
+    {
+      "epoch": 2.0317808219178084,
+      "grad_norm": 2.52215576171875,
+      "learning_rate": 1.8237205523964258e-06,
+      "loss": 1.7385,
+      "step": 1854
+    },
+    {
+      "epoch": 2.032876712328767,
+      "grad_norm": 2.433335065841675,
+      "learning_rate": 1.8216896831844031e-06,
+      "loss": 1.7566,
+      "step": 1855
+    },
+    {
+      "epoch": 2.033972602739726,
+      "grad_norm": 2.441232442855835,
+      "learning_rate": 1.8196588139723804e-06,
+      "loss": 1.8027,
+      "step": 1856
+    },
+    {
+      "epoch": 2.035068493150685,
+      "grad_norm": 2.4601082801818848,
+      "learning_rate": 1.8176279447603577e-06,
+      "loss": 1.6859,
+      "step": 1857
+    },
+    {
+      "epoch": 2.0361643835616436,
+      "grad_norm": 2.5008108615875244,
+      "learning_rate": 1.815597075548335e-06,
+      "loss": 1.7294,
+      "step": 1858
+    },
+    {
+      "epoch": 2.0372602739726027,
+      "grad_norm": 2.433915615081787,
+      "learning_rate": 1.8135662063363122e-06,
+      "loss": 1.755,
+      "step": 1859
+    },
+    {
+      "epoch": 2.0383561643835617,
+      "grad_norm": 2.4938697814941406,
+      "learning_rate": 1.8115353371242895e-06,
+      "loss": 1.7904,
+      "step": 1860
+    },
+    {
+      "epoch": 2.0383561643835617,
+      "eval_loss": 1.9939266443252563,
+      "eval_runtime": 37.1539,
+      "eval_samples_per_second": 26.915,
+      "eval_steps_per_second": 1.13,
+      "step": 1860
+    },
+    {
+      "epoch": 2.0394520547945207,
+      "grad_norm": 2.382324695587158,
+      "learning_rate": 1.8095044679122668e-06,
+      "loss": 1.6856,
+      "step": 1861
+    },
+    {
+      "epoch": 2.0405479452054793,
+      "grad_norm": 2.5250437259674072,
+      "learning_rate": 1.807473598700244e-06,
+      "loss": 1.7434,
+      "step": 1862
+    },
+    {
+      "epoch": 2.0416438356164384,
+      "grad_norm": 2.63969087600708,
+      "learning_rate": 1.805442729488221e-06,
+      "loss": 1.7343,
+      "step": 1863
+    },
+    {
+      "epoch": 2.0427397260273974,
+      "grad_norm": 2.4852068424224854,
+      "learning_rate": 1.8034118602761982e-06,
+      "loss": 1.7494,
+      "step": 1864
+    },
+    {
+      "epoch": 2.043835616438356,
+      "grad_norm": 2.393967390060425,
+      "learning_rate": 1.8013809910641754e-06,
+      "loss": 1.7503,
+      "step": 1865
+    },
+    {
+      "epoch": 2.044931506849315,
+      "grad_norm": 2.4447648525238037,
+      "learning_rate": 1.7993501218521527e-06,
+      "loss": 1.7539,
+      "step": 1866
+    },
+    {
+      "epoch": 2.046027397260274,
+      "grad_norm": 2.483530044555664,
+      "learning_rate": 1.79731925264013e-06,
+      "loss": 1.7001,
+      "step": 1867
+    },
+    {
+      "epoch": 2.047123287671233,
+      "grad_norm": 2.56669282913208,
+      "learning_rate": 1.7952883834281073e-06,
+      "loss": 1.7828,
+      "step": 1868
+    },
+    {
+      "epoch": 2.0482191780821917,
+      "grad_norm": 2.4894964694976807,
+      "learning_rate": 1.7932575142160846e-06,
+      "loss": 1.7208,
+      "step": 1869
+    },
+    {
+      "epoch": 2.0493150684931507,
+      "grad_norm": 2.4713854789733887,
+      "learning_rate": 1.7912266450040618e-06,
+      "loss": 1.7687,
+      "step": 1870
+    },
+    {
+      "epoch": 2.0493150684931507,
+      "eval_loss": 1.9938212633132935,
+      "eval_runtime": 37.1298,
+      "eval_samples_per_second": 26.933,
+      "eval_steps_per_second": 1.131,
+      "step": 1870
+    },
+    {
+      "epoch": 2.0504109589041097,
+      "grad_norm": 2.6204898357391357,
+      "learning_rate": 1.789195775792039e-06,
+      "loss": 1.7718,
+      "step": 1871
+    },
+    {
+      "epoch": 2.0515068493150683,
+      "grad_norm": 2.5194199085235596,
+      "learning_rate": 1.7871649065800164e-06,
+      "loss": 1.6653,
+      "step": 1872
+    },
+    {
+      "epoch": 2.0526027397260274,
+      "grad_norm": 2.5077965259552,
+      "learning_rate": 1.7851340373679937e-06,
+      "loss": 1.8098,
+      "step": 1873
+    },
+    {
+      "epoch": 2.0536986301369864,
+      "grad_norm": 2.5841965675354004,
+      "learning_rate": 1.783103168155971e-06,
+      "loss": 1.7126,
+      "step": 1874
+    },
+    {
+      "epoch": 2.0547945205479454,
+      "grad_norm": 2.617910623550415,
+      "learning_rate": 1.7810722989439482e-06,
+      "loss": 1.7294,
+      "step": 1875
+    },
+    {
+      "epoch": 2.055890410958904,
+      "grad_norm": 2.519253969192505,
+      "learning_rate": 1.7790414297319253e-06,
+      "loss": 1.7497,
+      "step": 1876
+    },
+    {
+      "epoch": 2.056986301369863,
+      "grad_norm": 2.5518412590026855,
+      "learning_rate": 1.7770105605199025e-06,
+      "loss": 1.7188,
+      "step": 1877
+    },
+    {
+      "epoch": 2.058082191780822,
+      "grad_norm": 2.5372824668884277,
+      "learning_rate": 1.7749796913078798e-06,
+      "loss": 1.6998,
+      "step": 1878
+    },
+    {
+      "epoch": 2.0591780821917807,
+      "grad_norm": 2.6649036407470703,
+      "learning_rate": 1.772948822095857e-06,
+      "loss": 1.7621,
+      "step": 1879
+    },
+    {
+      "epoch": 2.0602739726027397,
+      "grad_norm": 2.5216212272644043,
+      "learning_rate": 1.7709179528838344e-06,
+      "loss": 1.7963,
+      "step": 1880
+    },
+    {
+      "epoch": 2.0602739726027397,
+      "eval_loss": 1.993401050567627,
+      "eval_runtime": 37.0954,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 1880
+    },
+    {
+      "epoch": 2.0613698630136987,
+      "grad_norm": 2.5226478576660156,
+      "learning_rate": 1.7688870836718116e-06,
+      "loss": 1.7841,
+      "step": 1881
+    },
+    {
+      "epoch": 2.0624657534246573,
+      "grad_norm": 2.584319829940796,
+      "learning_rate": 1.766856214459789e-06,
+      "loss": 1.7449,
+      "step": 1882
+    },
+    {
+      "epoch": 2.0635616438356164,
+      "grad_norm": 2.5305304527282715,
+      "learning_rate": 1.7648253452477662e-06,
+      "loss": 1.7373,
+      "step": 1883
+    },
+    {
+      "epoch": 2.0646575342465754,
+      "grad_norm": 2.48797607421875,
+      "learning_rate": 1.7627944760357435e-06,
+      "loss": 1.722,
+      "step": 1884
+    },
+    {
+      "epoch": 2.0657534246575344,
+      "grad_norm": 2.4461028575897217,
+      "learning_rate": 1.7607636068237208e-06,
+      "loss": 1.7983,
+      "step": 1885
+    },
+    {
+      "epoch": 2.066849315068493,
+      "grad_norm": 2.624058961868286,
+      "learning_rate": 1.758732737611698e-06,
+      "loss": 1.7709,
+      "step": 1886
+    },
+    {
+      "epoch": 2.067945205479452,
+      "grad_norm": 2.848062515258789,
+      "learning_rate": 1.7567018683996753e-06,
+      "loss": 1.7513,
+      "step": 1887
+    },
+    {
+      "epoch": 2.069041095890411,
+      "grad_norm": 2.5219593048095703,
+      "learning_rate": 1.7546709991876524e-06,
+      "loss": 1.7387,
+      "step": 1888
+    },
+    {
+      "epoch": 2.0701369863013697,
+      "grad_norm": 2.5296876430511475,
+      "learning_rate": 1.7526401299756296e-06,
+      "loss": 1.7378,
+      "step": 1889
+    },
+    {
+      "epoch": 2.0712328767123287,
+      "grad_norm": 2.5452613830566406,
+      "learning_rate": 1.750609260763607e-06,
+      "loss": 1.7402,
+      "step": 1890
+    },
+    {
+      "epoch": 2.0712328767123287,
+      "eval_loss": 1.9914054870605469,
+      "eval_runtime": 37.0548,
+      "eval_samples_per_second": 26.987,
+      "eval_steps_per_second": 1.133,
+      "step": 1890
+    },
+    {
+      "epoch": 2.0723287671232877,
+      "grad_norm": 2.653596878051758,
+      "learning_rate": 1.7485783915515842e-06,
+      "loss": 1.7369,
+      "step": 1891
+    },
+    {
+      "epoch": 2.0734246575342468,
+      "grad_norm": 2.615581750869751,
+      "learning_rate": 1.7465475223395615e-06,
+      "loss": 1.7317,
+      "step": 1892
+    },
+    {
+      "epoch": 2.0745205479452054,
+      "grad_norm": 2.610304832458496,
+      "learning_rate": 1.7445166531275387e-06,
+      "loss": 1.7121,
+      "step": 1893
+    },
+    {
+      "epoch": 2.0756164383561644,
+      "grad_norm": 2.6463494300842285,
+      "learning_rate": 1.742485783915516e-06,
+      "loss": 1.8523,
+      "step": 1894
+    },
+    {
+      "epoch": 2.0767123287671234,
+      "grad_norm": 2.4814131259918213,
+      "learning_rate": 1.7404549147034933e-06,
+      "loss": 1.8239,
+      "step": 1895
+    },
+    {
+      "epoch": 2.077808219178082,
+      "grad_norm": 2.7416090965270996,
+      "learning_rate": 1.7384240454914706e-06,
+      "loss": 1.669,
+      "step": 1896
+    },
+    {
+      "epoch": 2.078904109589041,
+      "grad_norm": 2.9458305835723877,
+      "learning_rate": 1.7363931762794478e-06,
+      "loss": 1.7504,
+      "step": 1897
+    },
+    {
+      "epoch": 2.08,
+      "grad_norm": 2.5118932723999023,
+      "learning_rate": 1.7343623070674251e-06,
+      "loss": 1.7854,
+      "step": 1898
+    },
+    {
+      "epoch": 2.081095890410959,
+      "grad_norm": 2.479421615600586,
+      "learning_rate": 1.7323314378554024e-06,
+      "loss": 1.7325,
+      "step": 1899
+    },
+    {
+      "epoch": 2.0821917808219177,
+      "grad_norm": 2.7400424480438232,
+      "learning_rate": 1.7303005686433797e-06,
+      "loss": 1.7434,
+      "step": 1900
+    },
+    {
+      "epoch": 2.0821917808219177,
+      "eval_loss": 1.9887213706970215,
+      "eval_runtime": 37.0484,
+      "eval_samples_per_second": 26.992,
+      "eval_steps_per_second": 1.134,
+      "step": 1900
+    },
+    {
+      "epoch": 2.0832876712328767,
+      "grad_norm": 2.587714433670044,
+      "learning_rate": 1.7282696994313567e-06,
+      "loss": 1.7335,
+      "step": 1901
+    },
+    {
+      "epoch": 2.0843835616438358,
+      "grad_norm": 2.526818037033081,
+      "learning_rate": 1.726238830219334e-06,
+      "loss": 1.7538,
+      "step": 1902
+    },
+    {
+      "epoch": 2.0854794520547943,
+      "grad_norm": 2.595221519470215,
+      "learning_rate": 1.7242079610073113e-06,
+      "loss": 1.7331,
+      "step": 1903
+    },
+    {
+      "epoch": 2.0865753424657534,
+      "grad_norm": 2.5018110275268555,
+      "learning_rate": 1.7221770917952886e-06,
+      "loss": 1.8069,
+      "step": 1904
+    },
+    {
+      "epoch": 2.0876712328767124,
+      "grad_norm": 2.4378256797790527,
+      "learning_rate": 1.7201462225832658e-06,
+      "loss": 1.757,
+      "step": 1905
+    },
+    {
+      "epoch": 2.0887671232876714,
+      "grad_norm": 2.651167392730713,
+      "learning_rate": 1.7181153533712431e-06,
+      "loss": 1.7696,
+      "step": 1906
+    },
+    {
+      "epoch": 2.08986301369863,
+      "grad_norm": 2.550239086151123,
+      "learning_rate": 1.7160844841592204e-06,
+      "loss": 1.6551,
+      "step": 1907
+    },
+    {
+      "epoch": 2.090958904109589,
+      "grad_norm": 2.4192144870758057,
+      "learning_rate": 1.7140536149471977e-06,
+      "loss": 1.7031,
+      "step": 1908
+    },
+    {
+      "epoch": 2.092054794520548,
+      "grad_norm": 2.5327398777008057,
+      "learning_rate": 1.712022745735175e-06,
+      "loss": 1.734,
+      "step": 1909
+    },
+    {
+      "epoch": 2.0931506849315067,
+      "grad_norm": 2.5962865352630615,
+      "learning_rate": 1.7099918765231522e-06,
+      "loss": 1.732,
+      "step": 1910
+    },
+    {
+      "epoch": 2.0931506849315067,
+      "eval_loss": 1.9915482997894287,
+      "eval_runtime": 36.7856,
+      "eval_samples_per_second": 27.185,
+      "eval_steps_per_second": 1.142,
+      "step": 1910
+    },
+    {
+      "epoch": 2.0942465753424657,
+      "grad_norm": 2.561943531036377,
+      "learning_rate": 1.7079610073111295e-06,
+      "loss": 1.7556,
+      "step": 1911
+    },
+    {
+      "epoch": 2.0953424657534248,
+      "grad_norm": 2.529829263687134,
+      "learning_rate": 1.7059301380991068e-06,
+      "loss": 1.6839,
+      "step": 1912
+    },
+    {
+      "epoch": 2.0964383561643833,
+      "grad_norm": 2.636559009552002,
+      "learning_rate": 1.7038992688870836e-06,
+      "loss": 1.7393,
+      "step": 1913
+    },
+    {
+      "epoch": 2.0975342465753424,
+      "grad_norm": 2.558166742324829,
+      "learning_rate": 1.701868399675061e-06,
+      "loss": 1.7346,
+      "step": 1914
+    },
+    {
+      "epoch": 2.0986301369863014,
+      "grad_norm": 2.5193750858306885,
+      "learning_rate": 1.6998375304630382e-06,
+      "loss": 1.7666,
+      "step": 1915
+    },
+    {
+      "epoch": 2.0997260273972604,
+      "grad_norm": 2.498244047164917,
+      "learning_rate": 1.6978066612510155e-06,
+      "loss": 1.827,
+      "step": 1916
+    },
+    {
+      "epoch": 2.100821917808219,
+      "grad_norm": 2.799603223800659,
+      "learning_rate": 1.6957757920389927e-06,
+      "loss": 1.788,
+      "step": 1917
+    },
+    {
+      "epoch": 2.101917808219178,
+      "grad_norm": 2.733391761779785,
+      "learning_rate": 1.69374492282697e-06,
+      "loss": 1.7487,
+      "step": 1918
+    },
+    {
+      "epoch": 2.103013698630137,
+      "grad_norm": 2.5135183334350586,
+      "learning_rate": 1.6917140536149473e-06,
+      "loss": 1.7114,
+      "step": 1919
+    },
+    {
+      "epoch": 2.1041095890410957,
+      "grad_norm": 2.5275800228118896,
+      "learning_rate": 1.6896831844029246e-06,
+      "loss": 1.797,
+      "step": 1920
+    },
+    {
+      "epoch": 2.1041095890410957,
+      "eval_loss": 1.9926362037658691,
+      "eval_runtime": 36.9326,
+      "eval_samples_per_second": 27.076,
+      "eval_steps_per_second": 1.137,
+      "step": 1920
+    },
+    {
+      "epoch": 2.1052054794520547,
+      "grad_norm": 2.682774782180786,
+      "learning_rate": 1.6876523151909018e-06,
+      "loss": 1.7581,
+      "step": 1921
+    },
+    {
+      "epoch": 2.1063013698630138,
+      "grad_norm": 2.625620126724243,
+      "learning_rate": 1.6856214459788791e-06,
+      "loss": 1.7606,
+      "step": 1922
+    },
+    {
+      "epoch": 2.107397260273973,
+      "grad_norm": 2.5005173683166504,
+      "learning_rate": 1.6835905767668564e-06,
+      "loss": 1.7139,
+      "step": 1923
+    },
+    {
+      "epoch": 2.1084931506849314,
+      "grad_norm": 2.6731526851654053,
+      "learning_rate": 1.6815597075548337e-06,
+      "loss": 1.7174,
+      "step": 1924
+    },
+    {
+      "epoch": 2.1095890410958904,
+      "grad_norm": 2.5407724380493164,
+      "learning_rate": 1.679528838342811e-06,
+      "loss": 1.8049,
+      "step": 1925
+    },
+    {
+      "epoch": 2.1106849315068494,
+      "grad_norm": 2.502427101135254,
+      "learning_rate": 1.677497969130788e-06,
+      "loss": 1.7068,
+      "step": 1926
+    },
+    {
+      "epoch": 2.111780821917808,
+      "grad_norm": 2.5063414573669434,
+      "learning_rate": 1.6754670999187653e-06,
+      "loss": 1.7174,
+      "step": 1927
+    },
+    {
+      "epoch": 2.112876712328767,
+      "grad_norm": 2.524588108062744,
+      "learning_rate": 1.6734362307067426e-06,
+      "loss": 1.7442,
+      "step": 1928
+    },
+    {
+      "epoch": 2.113972602739726,
+      "grad_norm": 2.531909465789795,
+      "learning_rate": 1.6714053614947198e-06,
+      "loss": 1.692,
+      "step": 1929
+    },
+    {
+      "epoch": 2.115068493150685,
+      "grad_norm": 2.5104269981384277,
+      "learning_rate": 1.669374492282697e-06,
+      "loss": 1.7039,
+      "step": 1930
+    },
+    {
+      "epoch": 2.115068493150685,
+      "eval_loss": 1.9927221536636353,
+      "eval_runtime": 37.004,
+      "eval_samples_per_second": 27.024,
+      "eval_steps_per_second": 1.135,
+      "step": 1930
+    },
+    {
+      "epoch": 2.1161643835616437,
+      "grad_norm": 2.5443670749664307,
+      "learning_rate": 1.6673436230706744e-06,
+      "loss": 1.7896,
+      "step": 1931
+    },
+    {
+      "epoch": 2.1172602739726027,
+      "grad_norm": 2.6495745182037354,
+      "learning_rate": 1.6653127538586517e-06,
+      "loss": 1.7735,
+      "step": 1932
+    },
+    {
+      "epoch": 2.118356164383562,
+      "grad_norm": 2.7233121395111084,
+      "learning_rate": 1.663281884646629e-06,
+      "loss": 1.7186,
+      "step": 1933
+    },
+    {
+      "epoch": 2.1194520547945204,
+      "grad_norm": 2.4411816596984863,
+      "learning_rate": 1.6612510154346062e-06,
+      "loss": 1.7378,
+      "step": 1934
+    },
+    {
+      "epoch": 2.1205479452054794,
+      "grad_norm": 2.6478421688079834,
+      "learning_rate": 1.6592201462225835e-06,
+      "loss": 1.7989,
+      "step": 1935
+    },
+    {
+      "epoch": 2.1216438356164384,
+      "grad_norm": 2.56093168258667,
+      "learning_rate": 1.6571892770105608e-06,
+      "loss": 1.7663,
+      "step": 1936
+    },
+    {
+      "epoch": 2.1227397260273975,
+      "grad_norm": 2.5598859786987305,
+      "learning_rate": 1.655158407798538e-06,
+      "loss": 1.7956,
+      "step": 1937
+    },
+    {
+      "epoch": 2.123835616438356,
+      "grad_norm": 2.4958112239837646,
+      "learning_rate": 1.653127538586515e-06,
+      "loss": 1.7482,
+      "step": 1938
+    },
+    {
+      "epoch": 2.124931506849315,
+      "grad_norm": 2.5133280754089355,
+      "learning_rate": 1.6510966693744924e-06,
+      "loss": 1.7811,
+      "step": 1939
+    },
+    {
+      "epoch": 2.126027397260274,
+      "grad_norm": 2.492892026901245,
+      "learning_rate": 1.6490658001624696e-06,
+      "loss": 1.7883,
+      "step": 1940
+    },
+    {
+      "epoch": 2.126027397260274,
+      "eval_loss": 1.9917834997177124,
+      "eval_runtime": 37.0222,
+      "eval_samples_per_second": 27.011,
+      "eval_steps_per_second": 1.134,
+      "step": 1940
+    },
+    {
+      "epoch": 2.1271232876712327,
+      "grad_norm": 2.560483455657959,
+      "learning_rate": 1.647034930950447e-06,
+      "loss": 1.7871,
+      "step": 1941
+    },
+    {
+      "epoch": 2.1282191780821917,
+      "grad_norm": 2.549100160598755,
+      "learning_rate": 1.6450040617384242e-06,
+      "loss": 1.7097,
+      "step": 1942
+    },
+    {
+      "epoch": 2.1293150684931508,
+      "grad_norm": 2.4832775592803955,
+      "learning_rate": 1.6429731925264015e-06,
+      "loss": 1.736,
+      "step": 1943
+    },
+    {
+      "epoch": 2.1304109589041094,
+      "grad_norm": 2.60006046295166,
+      "learning_rate": 1.6409423233143788e-06,
+      "loss": 1.765,
+      "step": 1944
+    },
+    {
+      "epoch": 2.1315068493150684,
+      "grad_norm": 2.4875144958496094,
+      "learning_rate": 1.638911454102356e-06,
+      "loss": 1.7317,
+      "step": 1945
+    },
+    {
+      "epoch": 2.1326027397260274,
+      "grad_norm": 2.5562970638275146,
+      "learning_rate": 1.6368805848903333e-06,
+      "loss": 1.6982,
+      "step": 1946
+    },
+    {
+      "epoch": 2.1336986301369865,
+      "grad_norm": 2.4928972721099854,
+      "learning_rate": 1.6348497156783106e-06,
+      "loss": 1.7821,
+      "step": 1947
+    },
+    {
+      "epoch": 2.134794520547945,
+      "grad_norm": 2.6159043312072754,
+      "learning_rate": 1.6328188464662879e-06,
+      "loss": 1.7223,
+      "step": 1948
+    },
+    {
+      "epoch": 2.135890410958904,
+      "grad_norm": 2.541520833969116,
+      "learning_rate": 1.6307879772542651e-06,
+      "loss": 1.7321,
+      "step": 1949
+    },
+    {
+      "epoch": 2.136986301369863,
+      "grad_norm": 2.5165696144104004,
+      "learning_rate": 1.628757108042242e-06,
+      "loss": 1.7571,
+      "step": 1950
+    },
+    {
+      "epoch": 2.136986301369863,
+      "eval_loss": 1.9868476390838623,
+      "eval_runtime": 37.0367,
+      "eval_samples_per_second": 27.0,
+      "eval_steps_per_second": 1.134,
+      "step": 1950
+    },
+    {
+      "epoch": 2.138082191780822,
+      "grad_norm": 2.5660274028778076,
+      "learning_rate": 1.6267262388302193e-06,
+      "loss": 1.7855,
+      "step": 1951
+    },
+    {
+      "epoch": 2.1391780821917807,
+      "grad_norm": 2.4606971740722656,
+      "learning_rate": 1.6246953696181965e-06,
+      "loss": 1.7324,
+      "step": 1952
+    },
+    {
+      "epoch": 2.1402739726027398,
+      "grad_norm": 2.6528608798980713,
+      "learning_rate": 1.6226645004061738e-06,
+      "loss": 1.7715,
+      "step": 1953
+    },
+    {
+      "epoch": 2.141369863013699,
+      "grad_norm": 2.7232985496520996,
+      "learning_rate": 1.620633631194151e-06,
+      "loss": 1.837,
+      "step": 1954
+    },
+    {
+      "epoch": 2.1424657534246574,
+      "grad_norm": 2.6264185905456543,
+      "learning_rate": 1.6186027619821284e-06,
+      "loss": 1.7678,
+      "step": 1955
+    },
+    {
+      "epoch": 2.1435616438356164,
+      "grad_norm": 2.6248252391815186,
+      "learning_rate": 1.6165718927701056e-06,
+      "loss": 1.7218,
+      "step": 1956
+    },
+    {
+      "epoch": 2.1446575342465755,
+      "grad_norm": 3.00539493560791,
+      "learning_rate": 1.614541023558083e-06,
+      "loss": 1.7678,
+      "step": 1957
+    },
+    {
+      "epoch": 2.145753424657534,
+      "grad_norm": 2.5368528366088867,
+      "learning_rate": 1.6125101543460602e-06,
+      "loss": 1.7737,
+      "step": 1958
+    },
+    {
+      "epoch": 2.146849315068493,
+      "grad_norm": 2.5805916786193848,
+      "learning_rate": 1.6104792851340375e-06,
+      "loss": 1.7394,
+      "step": 1959
+    },
+    {
+      "epoch": 2.147945205479452,
+      "grad_norm": 2.615856647491455,
+      "learning_rate": 1.6084484159220147e-06,
+      "loss": 1.766,
+      "step": 1960
+    },
+    {
+      "epoch": 2.147945205479452,
+      "eval_loss": 1.987140417098999,
+      "eval_runtime": 37.0593,
+      "eval_samples_per_second": 26.984,
+      "eval_steps_per_second": 1.133,
+      "step": 1960
+    },
+    {
+      "epoch": 2.149041095890411,
+      "grad_norm": 2.6550803184509277,
+      "learning_rate": 1.606417546709992e-06,
+      "loss": 1.7342,
+      "step": 1961
+    },
+    {
+      "epoch": 2.1501369863013697,
+      "grad_norm": 2.5999467372894287,
+      "learning_rate": 1.6043866774979693e-06,
+      "loss": 1.7175,
+      "step": 1962
+    },
+    {
+      "epoch": 2.1512328767123288,
+      "grad_norm": 2.561642646789551,
+      "learning_rate": 1.6023558082859464e-06,
+      "loss": 1.7409,
+      "step": 1963
+    },
+    {
+      "epoch": 2.152328767123288,
+      "grad_norm": 2.6062211990356445,
+      "learning_rate": 1.6003249390739236e-06,
+      "loss": 1.7052,
+      "step": 1964
+    },
+    {
+      "epoch": 2.1534246575342464,
+      "grad_norm": 2.5122230052948,
+      "learning_rate": 1.598294069861901e-06,
+      "loss": 1.7444,
+      "step": 1965
+    },
+    {
+      "epoch": 2.1545205479452054,
+      "grad_norm": 2.5838866233825684,
+      "learning_rate": 1.5962632006498782e-06,
+      "loss": 1.716,
+      "step": 1966
+    },
+    {
+      "epoch": 2.1556164383561645,
+      "grad_norm": 2.5454747676849365,
+      "learning_rate": 1.5942323314378555e-06,
+      "loss": 1.7902,
+      "step": 1967
+    },
+    {
+      "epoch": 2.1567123287671235,
+      "grad_norm": 2.57637619972229,
+      "learning_rate": 1.5922014622258327e-06,
+      "loss": 1.6963,
+      "step": 1968
+    },
+    {
+      "epoch": 2.157808219178082,
+      "grad_norm": 2.5542941093444824,
+      "learning_rate": 1.59017059301381e-06,
+      "loss": 1.7151,
+      "step": 1969
+    },
+    {
+      "epoch": 2.158904109589041,
+      "grad_norm": 2.4912102222442627,
+      "learning_rate": 1.5881397238017873e-06,
+      "loss": 1.7624,
+      "step": 1970
+    },
+    {
+      "epoch": 2.158904109589041,
+      "eval_loss": 1.9855093955993652,
+      "eval_runtime": 37.0796,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 1970
+    },
+    {
+      "epoch": 2.16,
+      "grad_norm": 2.4793057441711426,
+      "learning_rate": 1.5861088545897646e-06,
+      "loss": 1.7617,
+      "step": 1971
+    },
+    {
+      "epoch": 2.1610958904109587,
+      "grad_norm": 2.5231704711914062,
+      "learning_rate": 1.5840779853777418e-06,
+      "loss": 1.6845,
+      "step": 1972
+    },
+    {
+      "epoch": 2.1621917808219178,
+      "grad_norm": 2.5126073360443115,
+      "learning_rate": 1.5820471161657191e-06,
+      "loss": 1.7468,
+      "step": 1973
+    },
+    {
+      "epoch": 2.163287671232877,
+      "grad_norm": 2.5558865070343018,
+      "learning_rate": 1.5800162469536964e-06,
+      "loss": 1.7253,
+      "step": 1974
+    },
+    {
+      "epoch": 2.1643835616438354,
+      "grad_norm": 2.808987855911255,
+      "learning_rate": 1.5779853777416735e-06,
+      "loss": 1.7844,
+      "step": 1975
+    },
+    {
+      "epoch": 2.1654794520547944,
+      "grad_norm": 2.765275001525879,
+      "learning_rate": 1.5759545085296507e-06,
+      "loss": 1.7658,
+      "step": 1976
+    },
+    {
+      "epoch": 2.1665753424657535,
+      "grad_norm": 2.5721709728240967,
+      "learning_rate": 1.573923639317628e-06,
+      "loss": 1.74,
+      "step": 1977
+    },
+    {
+      "epoch": 2.1676712328767125,
+      "grad_norm": 2.4276933670043945,
+      "learning_rate": 1.5718927701056053e-06,
+      "loss": 1.7613,
+      "step": 1978
+    },
+    {
+      "epoch": 2.168767123287671,
+      "grad_norm": 2.5665223598480225,
+      "learning_rate": 1.5698619008935826e-06,
+      "loss": 1.7514,
+      "step": 1979
+    },
+    {
+      "epoch": 2.16986301369863,
+      "grad_norm": 2.5588295459747314,
+      "learning_rate": 1.5678310316815598e-06,
+      "loss": 1.7749,
+      "step": 1980
+    },
+    {
+      "epoch": 2.16986301369863,
+      "eval_loss": 1.9858319759368896,
+      "eval_runtime": 37.0705,
+      "eval_samples_per_second": 26.976,
+      "eval_steps_per_second": 1.133,
+      "step": 1980
+    },
+    {
+      "epoch": 2.170958904109589,
+      "grad_norm": 2.5413172245025635,
+      "learning_rate": 1.5658001624695371e-06,
+      "loss": 1.706,
+      "step": 1981
+    },
+    {
+      "epoch": 2.172054794520548,
+      "grad_norm": 2.5246310234069824,
+      "learning_rate": 1.5637692932575144e-06,
+      "loss": 1.7825,
+      "step": 1982
+    },
+    {
+      "epoch": 2.1731506849315068,
+      "grad_norm": 2.385380506515503,
+      "learning_rate": 1.5617384240454917e-06,
+      "loss": 1.7268,
+      "step": 1983
+    },
+    {
+      "epoch": 2.174246575342466,
+      "grad_norm": 2.4967970848083496,
+      "learning_rate": 1.559707554833469e-06,
+      "loss": 1.7424,
+      "step": 1984
+    },
+    {
+      "epoch": 2.175342465753425,
+      "grad_norm": 2.563683271408081,
+      "learning_rate": 1.5576766856214462e-06,
+      "loss": 1.7904,
+      "step": 1985
+    },
+    {
+      "epoch": 2.1764383561643834,
+      "grad_norm": 2.518751859664917,
+      "learning_rate": 1.5556458164094235e-06,
+      "loss": 1.6616,
+      "step": 1986
+    },
+    {
+      "epoch": 2.1775342465753424,
+      "grad_norm": 2.5789380073547363,
+      "learning_rate": 1.5536149471974008e-06,
+      "loss": 1.7363,
+      "step": 1987
+    },
+    {
+      "epoch": 2.1786301369863015,
+      "grad_norm": 2.567406415939331,
+      "learning_rate": 1.5515840779853778e-06,
+      "loss": 1.6835,
+      "step": 1988
+    },
+    {
+      "epoch": 2.17972602739726,
+      "grad_norm": 2.509427309036255,
+      "learning_rate": 1.549553208773355e-06,
+      "loss": 1.7493,
+      "step": 1989
+    },
+    {
+      "epoch": 2.180821917808219,
+      "grad_norm": 2.6355340480804443,
+      "learning_rate": 1.5475223395613324e-06,
+      "loss": 1.7292,
+      "step": 1990
+    },
+    {
+      "epoch": 2.180821917808219,
+      "eval_loss": 1.9893935918807983,
+      "eval_runtime": 37.0928,
+      "eval_samples_per_second": 26.959,
+      "eval_steps_per_second": 1.132,
+      "step": 1990
+    },
+    {
+      "epoch": 2.181917808219178,
+      "grad_norm": 2.564587116241455,
+      "learning_rate": 1.5454914703493097e-06,
+      "loss": 1.7248,
+      "step": 1991
+    },
+    {
+      "epoch": 2.183013698630137,
+      "grad_norm": 2.6708462238311768,
+      "learning_rate": 1.543460601137287e-06,
+      "loss": 1.8123,
+      "step": 1992
+    },
+    {
+      "epoch": 2.1841095890410958,
+      "grad_norm": 2.6396312713623047,
+      "learning_rate": 1.5414297319252642e-06,
+      "loss": 1.7742,
+      "step": 1993
+    },
+    {
+      "epoch": 2.185205479452055,
+      "grad_norm": 2.553544044494629,
+      "learning_rate": 1.5393988627132415e-06,
+      "loss": 1.7705,
+      "step": 1994
+    },
+    {
+      "epoch": 2.186301369863014,
+      "grad_norm": 2.554997682571411,
+      "learning_rate": 1.5373679935012188e-06,
+      "loss": 1.737,
+      "step": 1995
+    },
+    {
+      "epoch": 2.1873972602739724,
+      "grad_norm": 2.5650475025177,
+      "learning_rate": 1.535337124289196e-06,
+      "loss": 1.6865,
+      "step": 1996
+    },
+    {
+      "epoch": 2.1884931506849314,
+      "grad_norm": 2.8019111156463623,
+      "learning_rate": 1.5333062550771733e-06,
+      "loss": 1.7428,
+      "step": 1997
+    },
+    {
+      "epoch": 2.1895890410958905,
+      "grad_norm": 2.5704092979431152,
+      "learning_rate": 1.5312753858651506e-06,
+      "loss": 1.8291,
+      "step": 1998
+    },
+    {
+      "epoch": 2.1906849315068495,
+      "grad_norm": 2.5332226753234863,
+      "learning_rate": 1.5292445166531279e-06,
+      "loss": 1.7274,
+      "step": 1999
+    },
+    {
+      "epoch": 2.191780821917808,
+      "grad_norm": 2.555288553237915,
+      "learning_rate": 1.5272136474411047e-06,
+      "loss": 1.7497,
+      "step": 2000
+    },
+    {
+      "epoch": 2.191780821917808,
+      "eval_loss": 1.9881768226623535,
+      "eval_runtime": 37.0996,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 2000
+    },
+    {
+      "epoch": 2.192876712328767,
+      "grad_norm": 2.6068379878997803,
+      "learning_rate": 1.525182778229082e-06,
+      "loss": 1.7955,
+      "step": 2001
+    },
+    {
+      "epoch": 2.193972602739726,
+      "grad_norm": 2.479121208190918,
+      "learning_rate": 1.5231519090170593e-06,
+      "loss": 1.7342,
+      "step": 2002
+    },
+    {
+      "epoch": 2.1950684931506848,
+      "grad_norm": 2.495670795440674,
+      "learning_rate": 1.5211210398050365e-06,
+      "loss": 1.7125,
+      "step": 2003
+    },
+    {
+      "epoch": 2.196164383561644,
+      "grad_norm": 2.4988651275634766,
+      "learning_rate": 1.5190901705930138e-06,
+      "loss": 1.7137,
+      "step": 2004
+    },
+    {
+      "epoch": 2.197260273972603,
+      "grad_norm": 2.5191574096679688,
+      "learning_rate": 1.517059301380991e-06,
+      "loss": 1.7672,
+      "step": 2005
+    },
+    {
+      "epoch": 2.198356164383562,
+      "grad_norm": 2.5469892024993896,
+      "learning_rate": 1.5150284321689684e-06,
+      "loss": 1.7127,
+      "step": 2006
+    },
+    {
+      "epoch": 2.1994520547945204,
+      "grad_norm": 2.490896701812744,
+      "learning_rate": 1.5129975629569456e-06,
+      "loss": 1.7397,
+      "step": 2007
+    },
+    {
+      "epoch": 2.2005479452054795,
+      "grad_norm": 2.5142805576324463,
+      "learning_rate": 1.510966693744923e-06,
+      "loss": 1.7431,
+      "step": 2008
+    },
+    {
+      "epoch": 2.2016438356164385,
+      "grad_norm": 2.5397887229919434,
+      "learning_rate": 1.5089358245329002e-06,
+      "loss": 1.7439,
+      "step": 2009
+    },
+    {
+      "epoch": 2.202739726027397,
+      "grad_norm": 2.5063931941986084,
+      "learning_rate": 1.5069049553208775e-06,
+      "loss": 1.7259,
+      "step": 2010
+    },
+    {
+      "epoch": 2.202739726027397,
+      "eval_loss": 1.9853808879852295,
+      "eval_runtime": 36.8066,
+      "eval_samples_per_second": 27.169,
+      "eval_steps_per_second": 1.141,
+      "step": 2010
+    },
+    {
+      "epoch": 2.203835616438356,
+      "grad_norm": 2.6017589569091797,
+      "learning_rate": 1.5048740861088547e-06,
+      "loss": 1.7714,
+      "step": 2011
+    },
+    {
+      "epoch": 2.204931506849315,
+      "grad_norm": 2.5295040607452393,
+      "learning_rate": 1.502843216896832e-06,
+      "loss": 1.7561,
+      "step": 2012
+    },
+    {
+      "epoch": 2.206027397260274,
+      "grad_norm": 2.5187482833862305,
+      "learning_rate": 1.500812347684809e-06,
+      "loss": 1.6761,
+      "step": 2013
+    },
+    {
+      "epoch": 2.207123287671233,
+      "grad_norm": 2.5163650512695312,
+      "learning_rate": 1.4987814784727864e-06,
+      "loss": 1.7469,
+      "step": 2014
+    },
+    {
+      "epoch": 2.208219178082192,
+      "grad_norm": 2.731238603591919,
+      "learning_rate": 1.4967506092607636e-06,
+      "loss": 1.7527,
+      "step": 2015
+    },
+    {
+      "epoch": 2.209315068493151,
+      "grad_norm": 2.5298311710357666,
+      "learning_rate": 1.494719740048741e-06,
+      "loss": 1.6912,
+      "step": 2016
+    },
+    {
+      "epoch": 2.2104109589041094,
+      "grad_norm": 2.6385974884033203,
+      "learning_rate": 1.4926888708367182e-06,
+      "loss": 1.7664,
+      "step": 2017
+    },
+    {
+      "epoch": 2.2115068493150685,
+      "grad_norm": 2.6150033473968506,
+      "learning_rate": 1.4906580016246955e-06,
+      "loss": 1.7708,
+      "step": 2018
+    },
+    {
+      "epoch": 2.2126027397260275,
+      "grad_norm": 2.589329957962036,
+      "learning_rate": 1.4886271324126727e-06,
+      "loss": 1.7875,
+      "step": 2019
+    },
+    {
+      "epoch": 2.213698630136986,
+      "grad_norm": 2.6769933700561523,
+      "learning_rate": 1.48659626320065e-06,
+      "loss": 1.8136,
+      "step": 2020
+    },
+    {
+      "epoch": 2.213698630136986,
+      "eval_loss": 1.9827086925506592,
+      "eval_runtime": 37.0026,
+      "eval_samples_per_second": 27.025,
+      "eval_steps_per_second": 1.135,
+      "step": 2020
+    },
+    {
+      "epoch": 2.214794520547945,
+      "grad_norm": 2.7220497131347656,
+      "learning_rate": 1.4845653939886273e-06,
+      "loss": 1.6627,
+      "step": 2021
+    },
+    {
+      "epoch": 2.215890410958904,
+      "grad_norm": 2.6115968227386475,
+      "learning_rate": 1.4825345247766046e-06,
+      "loss": 1.6676,
+      "step": 2022
+    },
+    {
+      "epoch": 2.216986301369863,
+      "grad_norm": 2.4612581729888916,
+      "learning_rate": 1.4805036555645818e-06,
+      "loss": 1.6644,
+      "step": 2023
+    },
+    {
+      "epoch": 2.2180821917808218,
+      "grad_norm": 2.5708696842193604,
+      "learning_rate": 1.4784727863525591e-06,
+      "loss": 1.7357,
+      "step": 2024
+    },
+    {
+      "epoch": 2.219178082191781,
+      "grad_norm": 2.578158140182495,
+      "learning_rate": 1.4764419171405362e-06,
+      "loss": 1.7011,
+      "step": 2025
+    },
+    {
+      "epoch": 2.22027397260274,
+      "grad_norm": 2.547666072845459,
+      "learning_rate": 1.4744110479285135e-06,
+      "loss": 1.7787,
+      "step": 2026
+    },
+    {
+      "epoch": 2.2213698630136984,
+      "grad_norm": 2.617551326751709,
+      "learning_rate": 1.4723801787164907e-06,
+      "loss": 1.784,
+      "step": 2027
+    },
+    {
+      "epoch": 2.2224657534246575,
+      "grad_norm": 2.4385478496551514,
+      "learning_rate": 1.470349309504468e-06,
+      "loss": 1.7764,
+      "step": 2028
+    },
+    {
+      "epoch": 2.2235616438356165,
+      "grad_norm": 2.6088693141937256,
+      "learning_rate": 1.4683184402924453e-06,
+      "loss": 1.7477,
+      "step": 2029
+    },
+    {
+      "epoch": 2.2246575342465755,
+      "grad_norm": 2.6165895462036133,
+      "learning_rate": 1.4662875710804226e-06,
+      "loss": 1.7531,
+      "step": 2030
+    },
+    {
+      "epoch": 2.2246575342465755,
+      "eval_loss": 1.9826008081436157,
+      "eval_runtime": 37.085,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 2030
+    },
+    {
+      "epoch": 2.225753424657534,
+      "grad_norm": 2.6285006999969482,
+      "learning_rate": 1.4642567018683998e-06,
+      "loss": 1.8403,
+      "step": 2031
+    },
+    {
+      "epoch": 2.226849315068493,
+      "grad_norm": 2.549326181411743,
+      "learning_rate": 1.4622258326563771e-06,
+      "loss": 1.7304,
+      "step": 2032
+    },
+    {
+      "epoch": 2.227945205479452,
+      "grad_norm": 2.6198782920837402,
+      "learning_rate": 1.4601949634443544e-06,
+      "loss": 1.7147,
+      "step": 2033
+    },
+    {
+      "epoch": 2.2290410958904108,
+      "grad_norm": 2.536947727203369,
+      "learning_rate": 1.4581640942323317e-06,
+      "loss": 1.7639,
+      "step": 2034
+    },
+    {
+      "epoch": 2.23013698630137,
+      "grad_norm": 2.6757264137268066,
+      "learning_rate": 1.456133225020309e-06,
+      "loss": 1.8434,
+      "step": 2035
+    },
+    {
+      "epoch": 2.231232876712329,
+      "grad_norm": 2.598095655441284,
+      "learning_rate": 1.4541023558082862e-06,
+      "loss": 1.7593,
+      "step": 2036
+    },
+    {
+      "epoch": 2.232328767123288,
+      "grad_norm": 2.6104440689086914,
+      "learning_rate": 1.4520714865962635e-06,
+      "loss": 1.7473,
+      "step": 2037
+    },
+    {
+      "epoch": 2.2334246575342465,
+      "grad_norm": 2.6350162029266357,
+      "learning_rate": 1.4500406173842406e-06,
+      "loss": 1.777,
+      "step": 2038
+    },
+    {
+      "epoch": 2.2345205479452055,
+      "grad_norm": 2.593874216079712,
+      "learning_rate": 1.4480097481722178e-06,
+      "loss": 1.7506,
+      "step": 2039
+    },
+    {
+      "epoch": 2.2356164383561645,
+      "grad_norm": 2.6991093158721924,
+      "learning_rate": 1.4459788789601951e-06,
+      "loss": 1.7375,
+      "step": 2040
+    },
+    {
+      "epoch": 2.2356164383561645,
+      "eval_loss": 1.9815349578857422,
+      "eval_runtime": 37.1126,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 2040
+    },
+    {
+      "epoch": 2.236712328767123,
+      "grad_norm": 2.6339058876037598,
+      "learning_rate": 1.4439480097481724e-06,
+      "loss": 1.7045,
+      "step": 2041
+    },
+    {
+      "epoch": 2.237808219178082,
+      "grad_norm": 2.691631555557251,
+      "learning_rate": 1.4419171405361497e-06,
+      "loss": 1.7063,
+      "step": 2042
+    },
+    {
+      "epoch": 2.238904109589041,
+      "grad_norm": 3.021667003631592,
+      "learning_rate": 1.439886271324127e-06,
+      "loss": 1.7607,
+      "step": 2043
+    },
+    {
+      "epoch": 2.24,
+      "grad_norm": 2.690802812576294,
+      "learning_rate": 1.4378554021121042e-06,
+      "loss": 1.7594,
+      "step": 2044
+    },
+    {
+      "epoch": 2.241095890410959,
+      "grad_norm": 2.5371177196502686,
+      "learning_rate": 1.4358245329000815e-06,
+      "loss": 1.7177,
+      "step": 2045
+    },
+    {
+      "epoch": 2.242191780821918,
+      "grad_norm": 2.5586729049682617,
+      "learning_rate": 1.4337936636880588e-06,
+      "loss": 1.7864,
+      "step": 2046
+    },
+    {
+      "epoch": 2.243287671232877,
+      "grad_norm": 2.6621484756469727,
+      "learning_rate": 1.431762794476036e-06,
+      "loss": 1.7616,
+      "step": 2047
+    },
+    {
+      "epoch": 2.2443835616438355,
+      "grad_norm": 2.787508010864258,
+      "learning_rate": 1.4297319252640133e-06,
+      "loss": 1.7056,
+      "step": 2048
+    },
+    {
+      "epoch": 2.2454794520547945,
+      "grad_norm": 2.6896064281463623,
+      "learning_rate": 1.4277010560519906e-06,
+      "loss": 1.7413,
+      "step": 2049
+    },
+    {
+      "epoch": 2.2465753424657535,
+      "grad_norm": 2.540524959564209,
+      "learning_rate": 1.4256701868399674e-06,
+      "loss": 1.7096,
+      "step": 2050
+    },
+    {
+      "epoch": 2.2465753424657535,
+      "eval_loss": 1.9832353591918945,
+      "eval_runtime": 37.1202,
+      "eval_samples_per_second": 26.94,
+      "eval_steps_per_second": 1.131,
+      "step": 2050
+    },
+    {
+      "epoch": 2.247671232876712,
+      "grad_norm": 2.582171678543091,
+      "learning_rate": 1.4236393176279447e-06,
+      "loss": 1.8262,
+      "step": 2051
+    },
+    {
+      "epoch": 2.248767123287671,
+      "grad_norm": 2.659044027328491,
+      "learning_rate": 1.421608448415922e-06,
+      "loss": 1.7458,
+      "step": 2052
+    },
+    {
+      "epoch": 2.24986301369863,
+      "grad_norm": 2.7265775203704834,
+      "learning_rate": 1.4195775792038993e-06,
+      "loss": 1.8341,
+      "step": 2053
+    },
+    {
+      "epoch": 2.250958904109589,
+      "grad_norm": 2.922330379486084,
+      "learning_rate": 1.4175467099918765e-06,
+      "loss": 1.7103,
+      "step": 2054
+    },
+    {
+      "epoch": 2.252054794520548,
+      "grad_norm": 2.4665236473083496,
+      "learning_rate": 1.4155158407798538e-06,
+      "loss": 1.7461,
+      "step": 2055
+    },
+    {
+      "epoch": 2.253150684931507,
+      "grad_norm": 2.639538049697876,
+      "learning_rate": 1.413484971567831e-06,
+      "loss": 1.7683,
+      "step": 2056
+    },
+    {
+      "epoch": 2.254246575342466,
+      "grad_norm": 2.558363676071167,
+      "learning_rate": 1.4114541023558084e-06,
+      "loss": 1.6821,
+      "step": 2057
+    },
+    {
+      "epoch": 2.255342465753425,
+      "grad_norm": 2.6308321952819824,
+      "learning_rate": 1.4094232331437856e-06,
+      "loss": 1.7226,
+      "step": 2058
+    },
+    {
+      "epoch": 2.2564383561643835,
+      "grad_norm": 2.6325299739837646,
+      "learning_rate": 1.407392363931763e-06,
+      "loss": 1.783,
+      "step": 2059
+    },
+    {
+      "epoch": 2.2575342465753425,
+      "grad_norm": 2.6024723052978516,
+      "learning_rate": 1.4053614947197402e-06,
+      "loss": 1.8402,
+      "step": 2060
+    },
+    {
+      "epoch": 2.2575342465753425,
+      "eval_loss": 1.983332633972168,
+      "eval_runtime": 37.1133,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 2060
+    },
+    {
+      "epoch": 2.2586301369863016,
+      "grad_norm": 2.548522710800171,
+      "learning_rate": 1.4033306255077175e-06,
+      "loss": 1.7625,
+      "step": 2061
+    },
+    {
+      "epoch": 2.25972602739726,
+      "grad_norm": 2.5953831672668457,
+      "learning_rate": 1.4012997562956945e-06,
+      "loss": 1.6632,
+      "step": 2062
+    },
+    {
+      "epoch": 2.260821917808219,
+      "grad_norm": 2.5880215167999268,
+      "learning_rate": 1.3992688870836718e-06,
+      "loss": 1.7957,
+      "step": 2063
+    },
+    {
+      "epoch": 2.261917808219178,
+      "grad_norm": 2.439326047897339,
+      "learning_rate": 1.397238017871649e-06,
+      "loss": 1.76,
+      "step": 2064
+    },
+    {
+      "epoch": 2.263013698630137,
+      "grad_norm": 2.5185437202453613,
+      "learning_rate": 1.3952071486596264e-06,
+      "loss": 1.7385,
+      "step": 2065
+    },
+    {
+      "epoch": 2.264109589041096,
+      "grad_norm": 2.5943732261657715,
+      "learning_rate": 1.3931762794476036e-06,
+      "loss": 1.7066,
+      "step": 2066
+    },
+    {
+      "epoch": 2.265205479452055,
+      "grad_norm": 2.6375551223754883,
+      "learning_rate": 1.391145410235581e-06,
+      "loss": 1.7884,
+      "step": 2067
+    },
+    {
+      "epoch": 2.266301369863014,
+      "grad_norm": 2.4962515830993652,
+      "learning_rate": 1.3891145410235582e-06,
+      "loss": 1.7295,
+      "step": 2068
+    },
+    {
+      "epoch": 2.2673972602739725,
+      "grad_norm": 2.5873053073883057,
+      "learning_rate": 1.3870836718115355e-06,
+      "loss": 1.6697,
+      "step": 2069
+    },
+    {
+      "epoch": 2.2684931506849315,
+      "grad_norm": 2.573211908340454,
+      "learning_rate": 1.3850528025995127e-06,
+      "loss": 1.7198,
+      "step": 2070
+    },
+    {
+      "epoch": 2.2684931506849315,
+      "eval_loss": 1.9826557636260986,
+      "eval_runtime": 37.1271,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 2070
+    },
+    {
+      "epoch": 2.2695890410958905,
+      "grad_norm": 2.68015718460083,
+      "learning_rate": 1.38302193338749e-06,
+      "loss": 1.7294,
+      "step": 2071
+    },
+    {
+      "epoch": 2.270684931506849,
+      "grad_norm": 2.4882888793945312,
+      "learning_rate": 1.3809910641754673e-06,
+      "loss": 1.6592,
+      "step": 2072
+    },
+    {
+      "epoch": 2.271780821917808,
+      "grad_norm": 2.587048053741455,
+      "learning_rate": 1.3789601949634446e-06,
+      "loss": 1.7605,
+      "step": 2073
+    },
+    {
+      "epoch": 2.272876712328767,
+      "grad_norm": 2.6452062129974365,
+      "learning_rate": 1.3769293257514218e-06,
+      "loss": 1.6664,
+      "step": 2074
+    },
+    {
+      "epoch": 2.2739726027397262,
+      "grad_norm": 2.5663232803344727,
+      "learning_rate": 1.374898456539399e-06,
+      "loss": 1.771,
+      "step": 2075
+    },
+    {
+      "epoch": 2.275068493150685,
+      "grad_norm": 2.652989149093628,
+      "learning_rate": 1.3728675873273762e-06,
+      "loss": 1.7096,
+      "step": 2076
+    },
+    {
+      "epoch": 2.276164383561644,
+      "grad_norm": 2.538435220718384,
+      "learning_rate": 1.3708367181153535e-06,
+      "loss": 1.798,
+      "step": 2077
+    },
+    {
+      "epoch": 2.277260273972603,
+      "grad_norm": 2.4685096740722656,
+      "learning_rate": 1.3688058489033307e-06,
+      "loss": 1.7904,
+      "step": 2078
+    },
+    {
+      "epoch": 2.2783561643835615,
+      "grad_norm": 2.566723108291626,
+      "learning_rate": 1.366774979691308e-06,
+      "loss": 1.7569,
+      "step": 2079
+    },
+    {
+      "epoch": 2.2794520547945205,
+      "grad_norm": 2.560960531234741,
+      "learning_rate": 1.3647441104792853e-06,
+      "loss": 1.7734,
+      "step": 2080
+    },
+    {
+      "epoch": 2.2794520547945205,
+      "eval_loss": 1.9809880256652832,
+      "eval_runtime": 37.1137,
+      "eval_samples_per_second": 26.944,
+      "eval_steps_per_second": 1.132,
+      "step": 2080
+    },
+    {
+      "epoch": 2.2805479452054795,
+      "grad_norm": 2.576608419418335,
+      "learning_rate": 1.3627132412672626e-06,
+      "loss": 1.7431,
+      "step": 2081
+    },
+    {
+      "epoch": 2.281643835616438,
+      "grad_norm": 2.648301601409912,
+      "learning_rate": 1.3606823720552398e-06,
+      "loss": 1.6959,
+      "step": 2082
+    },
+    {
+      "epoch": 2.282739726027397,
+      "grad_norm": 2.516493082046509,
+      "learning_rate": 1.3586515028432171e-06,
+      "loss": 1.7519,
+      "step": 2083
+    },
+    {
+      "epoch": 2.283835616438356,
+      "grad_norm": 2.474616050720215,
+      "learning_rate": 1.3566206336311944e-06,
+      "loss": 1.6796,
+      "step": 2084
+    },
+    {
+      "epoch": 2.2849315068493152,
+      "grad_norm": 2.482044219970703,
+      "learning_rate": 1.3545897644191717e-06,
+      "loss": 1.7443,
+      "step": 2085
+    },
+    {
+      "epoch": 2.286027397260274,
+      "grad_norm": 2.6574456691741943,
+      "learning_rate": 1.352558895207149e-06,
+      "loss": 1.7145,
+      "step": 2086
+    },
+    {
+      "epoch": 2.287123287671233,
+      "grad_norm": 2.5350115299224854,
+      "learning_rate": 1.3505280259951258e-06,
+      "loss": 1.765,
+      "step": 2087
+    },
+    {
+      "epoch": 2.288219178082192,
+      "grad_norm": 2.6488730907440186,
+      "learning_rate": 1.348497156783103e-06,
+      "loss": 1.7407,
+      "step": 2088
+    },
+    {
+      "epoch": 2.289315068493151,
+      "grad_norm": 2.6020500659942627,
+      "learning_rate": 1.3464662875710804e-06,
+      "loss": 1.779,
+      "step": 2089
+    },
+    {
+      "epoch": 2.2904109589041095,
+      "grad_norm": 2.633013963699341,
+      "learning_rate": 1.3444354183590576e-06,
+      "loss": 1.6824,
+      "step": 2090
+    },
+    {
+      "epoch": 2.2904109589041095,
+      "eval_loss": 1.982380986213684,
+      "eval_runtime": 37.1166,
+      "eval_samples_per_second": 26.942,
+      "eval_steps_per_second": 1.132,
+      "step": 2090
+    },
+    {
+      "epoch": 2.2915068493150685,
+      "grad_norm": 2.6399149894714355,
+      "learning_rate": 1.342404549147035e-06,
+      "loss": 1.7634,
+      "step": 2091
+    },
+    {
+      "epoch": 2.2926027397260276,
+      "grad_norm": 2.580634832382202,
+      "learning_rate": 1.3403736799350122e-06,
+      "loss": 1.7099,
+      "step": 2092
+    },
+    {
+      "epoch": 2.293698630136986,
+      "grad_norm": 2.523400068283081,
+      "learning_rate": 1.3383428107229895e-06,
+      "loss": 1.7705,
+      "step": 2093
+    },
+    {
+      "epoch": 2.294794520547945,
+      "grad_norm": 2.5114550590515137,
+      "learning_rate": 1.3363119415109667e-06,
+      "loss": 1.7961,
+      "step": 2094
+    },
+    {
+      "epoch": 2.2958904109589042,
+      "grad_norm": 2.457369804382324,
+      "learning_rate": 1.334281072298944e-06,
+      "loss": 1.7507,
+      "step": 2095
+    },
+    {
+      "epoch": 2.296986301369863,
+      "grad_norm": 2.5671467781066895,
+      "learning_rate": 1.3322502030869213e-06,
+      "loss": 1.7607,
+      "step": 2096
+    },
+    {
+      "epoch": 2.298082191780822,
+      "grad_norm": 2.459540367126465,
+      "learning_rate": 1.3302193338748986e-06,
+      "loss": 1.7193,
+      "step": 2097
+    },
+    {
+      "epoch": 2.299178082191781,
+      "grad_norm": 2.4934308528900146,
+      "learning_rate": 1.3281884646628758e-06,
+      "loss": 1.7775,
+      "step": 2098
+    },
+    {
+      "epoch": 2.30027397260274,
+      "grad_norm": 2.520641565322876,
+      "learning_rate": 1.3261575954508531e-06,
+      "loss": 1.7847,
+      "step": 2099
+    },
+    {
+      "epoch": 2.3013698630136985,
+      "grad_norm": 2.4607160091400146,
+      "learning_rate": 1.3241267262388302e-06,
+      "loss": 1.7812,
+      "step": 2100
+    },
+    {
+      "epoch": 2.3013698630136985,
+      "eval_loss": 1.9836152791976929,
+      "eval_runtime": 37.1239,
+      "eval_samples_per_second": 26.937,
+      "eval_steps_per_second": 1.131,
+      "step": 2100
+    },
+    {
+      "epoch": 2.3024657534246575,
+      "grad_norm": 2.577761650085449,
+      "learning_rate": 1.3220958570268074e-06,
+      "loss": 1.7452,
+      "step": 2101
+    },
+    {
+      "epoch": 2.3035616438356166,
+      "grad_norm": 2.5331356525421143,
+      "learning_rate": 1.3200649878147847e-06,
+      "loss": 1.6989,
+      "step": 2102
+    },
+    {
+      "epoch": 2.304657534246575,
+      "grad_norm": 2.632581949234009,
+      "learning_rate": 1.318034118602762e-06,
+      "loss": 1.7368,
+      "step": 2103
+    },
+    {
+      "epoch": 2.305753424657534,
+      "grad_norm": 2.55940318107605,
+      "learning_rate": 1.3160032493907393e-06,
+      "loss": 1.7806,
+      "step": 2104
+    },
+    {
+      "epoch": 2.3068493150684932,
+      "grad_norm": 2.5680153369903564,
+      "learning_rate": 1.3139723801787166e-06,
+      "loss": 1.7482,
+      "step": 2105
+    },
+    {
+      "epoch": 2.3079452054794523,
+      "grad_norm": 2.7915756702423096,
+      "learning_rate": 1.3119415109666938e-06,
+      "loss": 1.832,
+      "step": 2106
+    },
+    {
+      "epoch": 2.309041095890411,
+      "grad_norm": 2.518031597137451,
+      "learning_rate": 1.309910641754671e-06,
+      "loss": 1.7401,
+      "step": 2107
+    },
+    {
+      "epoch": 2.31013698630137,
+      "grad_norm": 2.7410221099853516,
+      "learning_rate": 1.3078797725426484e-06,
+      "loss": 1.6992,
+      "step": 2108
+    },
+    {
+      "epoch": 2.311232876712329,
+      "grad_norm": 2.611825466156006,
+      "learning_rate": 1.3058489033306257e-06,
+      "loss": 1.7085,
+      "step": 2109
+    },
+    {
+      "epoch": 2.3123287671232875,
+      "grad_norm": 2.5833330154418945,
+      "learning_rate": 1.303818034118603e-06,
+      "loss": 1.7129,
+      "step": 2110
+    },
+    {
+      "epoch": 2.3123287671232875,
+      "eval_loss": 1.9824191331863403,
+      "eval_runtime": 36.773,
+      "eval_samples_per_second": 27.194,
+      "eval_steps_per_second": 1.142,
+      "step": 2110
+    },
+    {
+      "epoch": 2.3134246575342465,
+      "grad_norm": 2.587327003479004,
+      "learning_rate": 1.3017871649065802e-06,
+      "loss": 1.7439,
+      "step": 2111
+    },
+    {
+      "epoch": 2.3145205479452056,
+      "grad_norm": 2.574925422668457,
+      "learning_rate": 1.2997562956945573e-06,
+      "loss": 1.7173,
+      "step": 2112
+    },
+    {
+      "epoch": 2.315616438356164,
+      "grad_norm": 2.6404287815093994,
+      "learning_rate": 1.2977254264825345e-06,
+      "loss": 1.7052,
+      "step": 2113
+    },
+    {
+      "epoch": 2.316712328767123,
+      "grad_norm": 2.634584903717041,
+      "learning_rate": 1.2956945572705118e-06,
+      "loss": 1.741,
+      "step": 2114
+    },
+    {
+      "epoch": 2.317808219178082,
+      "grad_norm": 2.66159987449646,
+      "learning_rate": 1.293663688058489e-06,
+      "loss": 1.76,
+      "step": 2115
+    },
+    {
+      "epoch": 2.3189041095890413,
+      "grad_norm": 2.6337332725524902,
+      "learning_rate": 1.2916328188464664e-06,
+      "loss": 1.7035,
+      "step": 2116
+    },
+    {
+      "epoch": 2.32,
+      "grad_norm": 2.528524398803711,
+      "learning_rate": 1.2896019496344436e-06,
+      "loss": 1.6439,
+      "step": 2117
+    },
+    {
+      "epoch": 2.321095890410959,
+      "grad_norm": 2.5938384532928467,
+      "learning_rate": 1.287571080422421e-06,
+      "loss": 1.7134,
+      "step": 2118
+    },
+    {
+      "epoch": 2.322191780821918,
+      "grad_norm": 2.606271505355835,
+      "learning_rate": 1.2855402112103982e-06,
+      "loss": 1.7758,
+      "step": 2119
+    },
+    {
+      "epoch": 2.323287671232877,
+      "grad_norm": 2.530071258544922,
+      "learning_rate": 1.2835093419983755e-06,
+      "loss": 1.6927,
+      "step": 2120
+    },
+    {
+      "epoch": 2.323287671232877,
+      "eval_loss": 1.9856306314468384,
+      "eval_runtime": 36.9674,
+      "eval_samples_per_second": 27.051,
+      "eval_steps_per_second": 1.136,
+      "step": 2120
+    },
+    {
+      "epoch": 2.3243835616438355,
+      "grad_norm": 2.8125321865081787,
+      "learning_rate": 1.2814784727863528e-06,
+      "loss": 1.7225,
+      "step": 2121
+    },
+    {
+      "epoch": 2.3254794520547946,
+      "grad_norm": 2.57012939453125,
+      "learning_rate": 1.27944760357433e-06,
+      "loss": 1.7177,
+      "step": 2122
+    },
+    {
+      "epoch": 2.3265753424657536,
+      "grad_norm": 2.6595780849456787,
+      "learning_rate": 1.2774167343623073e-06,
+      "loss": 1.7807,
+      "step": 2123
+    },
+    {
+      "epoch": 2.327671232876712,
+      "grad_norm": 2.557115077972412,
+      "learning_rate": 1.2753858651502846e-06,
+      "loss": 1.7762,
+      "step": 2124
+    },
+    {
+      "epoch": 2.328767123287671,
+      "grad_norm": 2.763218879699707,
+      "learning_rate": 1.2733549959382616e-06,
+      "loss": 1.7638,
+      "step": 2125
+    },
+    {
+      "epoch": 2.3298630136986302,
+      "grad_norm": 2.7284505367279053,
+      "learning_rate": 1.271324126726239e-06,
+      "loss": 1.7387,
+      "step": 2126
+    },
+    {
+      "epoch": 2.330958904109589,
+      "grad_norm": 2.5893704891204834,
+      "learning_rate": 1.2692932575142162e-06,
+      "loss": 1.7322,
+      "step": 2127
+    },
+    {
+      "epoch": 2.332054794520548,
+      "grad_norm": 2.625582456588745,
+      "learning_rate": 1.2672623883021935e-06,
+      "loss": 1.6865,
+      "step": 2128
+    },
+    {
+      "epoch": 2.333150684931507,
+      "grad_norm": 2.5458641052246094,
+      "learning_rate": 1.2652315190901707e-06,
+      "loss": 1.6817,
+      "step": 2129
+    },
+    {
+      "epoch": 2.334246575342466,
+      "grad_norm": 2.5718271732330322,
+      "learning_rate": 1.263200649878148e-06,
+      "loss": 1.8076,
+      "step": 2130
+    },
+    {
+      "epoch": 2.334246575342466,
+      "eval_loss": 1.9806523323059082,
+      "eval_runtime": 36.9843,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 2130
+    },
+    {
+      "epoch": 2.3353424657534245,
+      "grad_norm": 2.4864919185638428,
+      "learning_rate": 1.2611697806661253e-06,
+      "loss": 1.7308,
+      "step": 2131
+    },
+    {
+      "epoch": 2.3364383561643836,
+      "grad_norm": 2.6007416248321533,
+      "learning_rate": 1.2591389114541026e-06,
+      "loss": 1.7523,
+      "step": 2132
+    },
+    {
+      "epoch": 2.3375342465753426,
+      "grad_norm": 2.600177049636841,
+      "learning_rate": 1.2571080422420798e-06,
+      "loss": 1.7422,
+      "step": 2133
+    },
+    {
+      "epoch": 2.338630136986301,
+      "grad_norm": 2.489025831222534,
+      "learning_rate": 1.2550771730300571e-06,
+      "loss": 1.7184,
+      "step": 2134
+    },
+    {
+      "epoch": 2.33972602739726,
+      "grad_norm": 2.5677363872528076,
+      "learning_rate": 1.2530463038180344e-06,
+      "loss": 1.7922,
+      "step": 2135
+    },
+    {
+      "epoch": 2.3408219178082192,
+      "grad_norm": 2.616976737976074,
+      "learning_rate": 1.2510154346060117e-06,
+      "loss": 1.7354,
+      "step": 2136
+    },
+    {
+      "epoch": 2.3419178082191783,
+      "grad_norm": 2.646456241607666,
+      "learning_rate": 1.2489845653939887e-06,
+      "loss": 1.7377,
+      "step": 2137
+    },
+    {
+      "epoch": 2.343013698630137,
+      "grad_norm": 2.550020933151245,
+      "learning_rate": 1.246953696181966e-06,
+      "loss": 1.7751,
+      "step": 2138
+    },
+    {
+      "epoch": 2.344109589041096,
+      "grad_norm": 2.634279251098633,
+      "learning_rate": 1.2449228269699433e-06,
+      "loss": 1.8157,
+      "step": 2139
+    },
+    {
+      "epoch": 2.345205479452055,
+      "grad_norm": 2.5621821880340576,
+      "learning_rate": 1.2428919577579204e-06,
+      "loss": 1.7383,
+      "step": 2140
+    },
+    {
+      "epoch": 2.345205479452055,
+      "eval_loss": 1.9809924364089966,
+      "eval_runtime": 37.004,
+      "eval_samples_per_second": 27.024,
+      "eval_steps_per_second": 1.135,
+      "step": 2140
+    },
+    {
+      "epoch": 2.3463013698630135,
+      "grad_norm": 2.638831615447998,
+      "learning_rate": 1.2408610885458976e-06,
+      "loss": 1.6988,
+      "step": 2141
+    },
+    {
+      "epoch": 2.3473972602739726,
+      "grad_norm": 2.634612798690796,
+      "learning_rate": 1.238830219333875e-06,
+      "loss": 1.7331,
+      "step": 2142
+    },
+    {
+      "epoch": 2.3484931506849316,
+      "grad_norm": 2.552955389022827,
+      "learning_rate": 1.2367993501218522e-06,
+      "loss": 1.7148,
+      "step": 2143
+    },
+    {
+      "epoch": 2.34958904109589,
+      "grad_norm": 2.6475753784179688,
+      "learning_rate": 1.2347684809098295e-06,
+      "loss": 1.7815,
+      "step": 2144
+    },
+    {
+      "epoch": 2.350684931506849,
+      "grad_norm": 2.7666079998016357,
+      "learning_rate": 1.2327376116978067e-06,
+      "loss": 1.7634,
+      "step": 2145
+    },
+    {
+      "epoch": 2.3517808219178082,
+      "grad_norm": 2.6810429096221924,
+      "learning_rate": 1.230706742485784e-06,
+      "loss": 1.6769,
+      "step": 2146
+    },
+    {
+      "epoch": 2.3528767123287673,
+      "grad_norm": 2.671382188796997,
+      "learning_rate": 1.2286758732737613e-06,
+      "loss": 1.6884,
+      "step": 2147
+    },
+    {
+      "epoch": 2.353972602739726,
+      "grad_norm": 2.614684820175171,
+      "learning_rate": 1.2266450040617386e-06,
+      "loss": 1.8226,
+      "step": 2148
+    },
+    {
+      "epoch": 2.355068493150685,
+      "grad_norm": 2.47189998626709,
+      "learning_rate": 1.2246141348497158e-06,
+      "loss": 1.819,
+      "step": 2149
+    },
+    {
+      "epoch": 2.356164383561644,
+      "grad_norm": 2.4641449451446533,
+      "learning_rate": 1.2225832656376931e-06,
+      "loss": 1.7494,
+      "step": 2150
+    },
+    {
+      "epoch": 2.356164383561644,
+      "eval_loss": 1.9806426763534546,
+      "eval_runtime": 37.0153,
+      "eval_samples_per_second": 27.016,
+      "eval_steps_per_second": 1.135,
+      "step": 2150
+    },
+    {
+      "epoch": 2.357260273972603,
+      "grad_norm": 2.7310373783111572,
+      "learning_rate": 1.2205523964256704e-06,
+      "loss": 1.7432,
+      "step": 2151
+    },
+    {
+      "epoch": 2.3583561643835615,
+      "grad_norm": 2.5554895401000977,
+      "learning_rate": 1.2185215272136477e-06,
+      "loss": 1.7213,
+      "step": 2152
+    },
+    {
+      "epoch": 2.3594520547945206,
+      "grad_norm": 2.5401952266693115,
+      "learning_rate": 1.2164906580016247e-06,
+      "loss": 1.7469,
+      "step": 2153
+    },
+    {
+      "epoch": 2.3605479452054796,
+      "grad_norm": 2.5213980674743652,
+      "learning_rate": 1.214459788789602e-06,
+      "loss": 1.7459,
+      "step": 2154
+    },
+    {
+      "epoch": 2.361643835616438,
+      "grad_norm": 2.5585155487060547,
+      "learning_rate": 1.2124289195775793e-06,
+      "loss": 1.6587,
+      "step": 2155
+    },
+    {
+      "epoch": 2.3627397260273972,
+      "grad_norm": 2.6329963207244873,
+      "learning_rate": 1.2103980503655566e-06,
+      "loss": 1.6925,
+      "step": 2156
+    },
+    {
+      "epoch": 2.3638356164383563,
+      "grad_norm": 2.678351640701294,
+      "learning_rate": 1.2083671811535338e-06,
+      "loss": 1.7322,
+      "step": 2157
+    },
+    {
+      "epoch": 2.364931506849315,
+      "grad_norm": 2.595313549041748,
+      "learning_rate": 1.2063363119415111e-06,
+      "loss": 1.7142,
+      "step": 2158
+    },
+    {
+      "epoch": 2.366027397260274,
+      "grad_norm": 2.5650861263275146,
+      "learning_rate": 1.2043054427294882e-06,
+      "loss": 1.7024,
+      "step": 2159
+    },
+    {
+      "epoch": 2.367123287671233,
+      "grad_norm": 2.495866537094116,
+      "learning_rate": 1.2022745735174654e-06,
+      "loss": 1.7205,
+      "step": 2160
+    },
+    {
+      "epoch": 2.367123287671233,
+      "eval_loss": 1.976922631263733,
+      "eval_runtime": 37.012,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 2160
+    },
+    {
+      "epoch": 2.368219178082192,
+      "grad_norm": 2.570307493209839,
+      "learning_rate": 1.2002437043054427e-06,
+      "loss": 1.7601,
+      "step": 2161
+    },
+    {
+      "epoch": 2.3693150684931505,
+      "grad_norm": 2.444373846054077,
+      "learning_rate": 1.19821283509342e-06,
+      "loss": 1.7857,
+      "step": 2162
+    },
+    {
+      "epoch": 2.3704109589041096,
+      "grad_norm": 2.525505781173706,
+      "learning_rate": 1.1961819658813973e-06,
+      "loss": 1.7875,
+      "step": 2163
+    },
+    {
+      "epoch": 2.3715068493150686,
+      "grad_norm": 2.622260093688965,
+      "learning_rate": 1.1941510966693746e-06,
+      "loss": 1.7596,
+      "step": 2164
+    },
+    {
+      "epoch": 2.3726027397260276,
+      "grad_norm": 2.542076349258423,
+      "learning_rate": 1.1921202274573518e-06,
+      "loss": 1.7248,
+      "step": 2165
+    },
+    {
+      "epoch": 2.3736986301369862,
+      "grad_norm": 2.510263681411743,
+      "learning_rate": 1.190089358245329e-06,
+      "loss": 1.7168,
+      "step": 2166
+    },
+    {
+      "epoch": 2.3747945205479453,
+      "grad_norm": 2.605023145675659,
+      "learning_rate": 1.1880584890333064e-06,
+      "loss": 1.798,
+      "step": 2167
+    },
+    {
+      "epoch": 2.3758904109589043,
+      "grad_norm": 2.593261480331421,
+      "learning_rate": 1.1860276198212837e-06,
+      "loss": 1.7247,
+      "step": 2168
+    },
+    {
+      "epoch": 2.376986301369863,
+      "grad_norm": 2.6010820865631104,
+      "learning_rate": 1.183996750609261e-06,
+      "loss": 1.7607,
+      "step": 2169
+    },
+    {
+      "epoch": 2.378082191780822,
+      "grad_norm": 2.62886118888855,
+      "learning_rate": 1.1819658813972382e-06,
+      "loss": 1.7641,
+      "step": 2170
+    },
+    {
+      "epoch": 2.378082191780822,
+      "eval_loss": 1.9790476560592651,
+      "eval_runtime": 37.0413,
+      "eval_samples_per_second": 26.997,
+      "eval_steps_per_second": 1.134,
+      "step": 2170
+    },
+    {
+      "epoch": 2.379178082191781,
+      "grad_norm": 2.5506582260131836,
+      "learning_rate": 1.1799350121852153e-06,
+      "loss": 1.7407,
+      "step": 2171
+    },
+    {
+      "epoch": 2.3802739726027395,
+      "grad_norm": 2.547445774078369,
+      "learning_rate": 1.1779041429731925e-06,
+      "loss": 1.7808,
+      "step": 2172
+    },
+    {
+      "epoch": 2.3813698630136986,
+      "grad_norm": 2.5443003177642822,
+      "learning_rate": 1.1758732737611698e-06,
+      "loss": 1.7842,
+      "step": 2173
+    },
+    {
+      "epoch": 2.3824657534246576,
+      "grad_norm": 2.5733375549316406,
+      "learning_rate": 1.173842404549147e-06,
+      "loss": 1.6902,
+      "step": 2174
+    },
+    {
+      "epoch": 2.383561643835616,
+      "grad_norm": 2.63702392578125,
+      "learning_rate": 1.1718115353371244e-06,
+      "loss": 1.8049,
+      "step": 2175
+    },
+    {
+      "epoch": 2.3846575342465752,
+      "grad_norm": 2.5577163696289062,
+      "learning_rate": 1.1697806661251016e-06,
+      "loss": 1.7162,
+      "step": 2176
+    },
+    {
+      "epoch": 2.3857534246575343,
+      "grad_norm": 2.4259274005889893,
+      "learning_rate": 1.167749796913079e-06,
+      "loss": 1.7883,
+      "step": 2177
+    },
+    {
+      "epoch": 2.3868493150684933,
+      "grad_norm": 2.537519931793213,
+      "learning_rate": 1.1657189277010562e-06,
+      "loss": 1.7882,
+      "step": 2178
+    },
+    {
+      "epoch": 2.387945205479452,
+      "grad_norm": 2.5480241775512695,
+      "learning_rate": 1.1636880584890335e-06,
+      "loss": 1.7245,
+      "step": 2179
+    },
+    {
+      "epoch": 2.389041095890411,
+      "grad_norm": 2.630337953567505,
+      "learning_rate": 1.1616571892770108e-06,
+      "loss": 1.7165,
+      "step": 2180
+    },
+    {
+      "epoch": 2.389041095890411,
+      "eval_loss": 1.9772392511367798,
+      "eval_runtime": 37.1283,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 2180
+    },
+    {
+      "epoch": 2.39013698630137,
+      "grad_norm": 2.5741171836853027,
+      "learning_rate": 1.159626320064988e-06,
+      "loss": 1.7806,
+      "step": 2181
+    },
+    {
+      "epoch": 2.391232876712329,
+      "grad_norm": 2.6166305541992188,
+      "learning_rate": 1.1575954508529653e-06,
+      "loss": 1.744,
+      "step": 2182
+    },
+    {
+      "epoch": 2.3923287671232876,
+      "grad_norm": 2.6965837478637695,
+      "learning_rate": 1.1555645816409426e-06,
+      "loss": 1.7736,
+      "step": 2183
+    },
+    {
+      "epoch": 2.3934246575342466,
+      "grad_norm": 2.586827516555786,
+      "learning_rate": 1.1535337124289196e-06,
+      "loss": 1.7455,
+      "step": 2184
+    },
+    {
+      "epoch": 2.3945205479452056,
+      "grad_norm": 2.587235689163208,
+      "learning_rate": 1.151502843216897e-06,
+      "loss": 1.6999,
+      "step": 2185
+    },
+    {
+      "epoch": 2.3956164383561642,
+      "grad_norm": 2.656740665435791,
+      "learning_rate": 1.1494719740048742e-06,
+      "loss": 1.6518,
+      "step": 2186
+    },
+    {
+      "epoch": 2.3967123287671233,
+      "grad_norm": 2.5353455543518066,
+      "learning_rate": 1.1474411047928515e-06,
+      "loss": 1.7389,
+      "step": 2187
+    },
+    {
+      "epoch": 2.3978082191780823,
+      "grad_norm": 2.622314453125,
+      "learning_rate": 1.1454102355808287e-06,
+      "loss": 1.7309,
+      "step": 2188
+    },
+    {
+      "epoch": 2.398904109589041,
+      "grad_norm": 2.6286797523498535,
+      "learning_rate": 1.143379366368806e-06,
+      "loss": 1.7094,
+      "step": 2189
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 2.5203921794891357,
+      "learning_rate": 1.141348497156783e-06,
+      "loss": 1.6905,
+      "step": 2190
+    },
+    {
+      "epoch": 2.4,
+      "eval_loss": 1.9822566509246826,
+      "eval_runtime": 37.0581,
+      "eval_samples_per_second": 26.985,
+      "eval_steps_per_second": 1.133,
+      "step": 2190
+    },
+    {
+      "epoch": 2.401095890410959,
+      "grad_norm": 2.6562373638153076,
+      "learning_rate": 1.1393176279447604e-06,
+      "loss": 1.8293,
+      "step": 2191
+    },
+    {
+      "epoch": 2.402191780821918,
+      "grad_norm": 2.5581510066986084,
+      "learning_rate": 1.1372867587327376e-06,
+      "loss": 1.7936,
+      "step": 2192
+    },
+    {
+      "epoch": 2.4032876712328766,
+      "grad_norm": 2.534675121307373,
+      "learning_rate": 1.135255889520715e-06,
+      "loss": 1.6648,
+      "step": 2193
+    },
+    {
+      "epoch": 2.4043835616438356,
+      "grad_norm": 2.7236168384552,
+      "learning_rate": 1.1332250203086922e-06,
+      "loss": 1.7654,
+      "step": 2194
+    },
+    {
+      "epoch": 2.4054794520547946,
+      "grad_norm": 2.5706138610839844,
+      "learning_rate": 1.1311941510966695e-06,
+      "loss": 1.7427,
+      "step": 2195
+    },
+    {
+      "epoch": 2.4065753424657537,
+      "grad_norm": 2.7030224800109863,
+      "learning_rate": 1.1291632818846467e-06,
+      "loss": 1.7216,
+      "step": 2196
+    },
+    {
+      "epoch": 2.4076712328767123,
+      "grad_norm": 2.4936161041259766,
+      "learning_rate": 1.127132412672624e-06,
+      "loss": 1.7033,
+      "step": 2197
+    },
+    {
+      "epoch": 2.4087671232876713,
+      "grad_norm": 2.5384793281555176,
+      "learning_rate": 1.1251015434606013e-06,
+      "loss": 1.7023,
+      "step": 2198
+    },
+    {
+      "epoch": 2.4098630136986303,
+      "grad_norm": 2.5271613597869873,
+      "learning_rate": 1.1230706742485786e-06,
+      "loss": 1.738,
+      "step": 2199
+    },
+    {
+      "epoch": 2.410958904109589,
+      "grad_norm": 2.557610034942627,
+      "learning_rate": 1.1210398050365558e-06,
+      "loss": 1.7927,
+      "step": 2200
+    },
+    {
+      "epoch": 2.410958904109589,
+      "eval_loss": 1.9795563220977783,
+      "eval_runtime": 37.0832,
+      "eval_samples_per_second": 26.966,
+      "eval_steps_per_second": 1.133,
+      "step": 2200
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2736,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 6.858080700596224e+17,
+  "train_batch_size": 3,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/training_args.bin b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..b7adbe5a2dbf27f05befec6cc65fbc93e90eefd0
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aaa882279512a8c278376c0c601206933f3c5662ddcd3c13c53bf32ae3680d14
+size 6011
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2200/zero_to_fp32.py
@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/config.json
@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "meta-llama/Llama-3.2-3B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "torch_dtype": "float16",
+  "transformers_version": "4.45.1",
+  "use_cache": true,
+  "vocab_size": 128256
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/generation_config.json
@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": 128001,
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.45.1"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/latest b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/latest
new file mode 100644
index 0000000000000000000000000000000000000000..78dc01ce4ce1498b753887dfb7af92409705a69b
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/latest
@@ -0,0 +1 @@
+global_step2300
\ No newline at end of file
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/model.safetensors.index.json
new file mode 100644
index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/model.safetensors.index.json
@@ -0,0 +1,262 @@
+{
+  "metadata": {
+    "total_size": 7213504512
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00002-of-00002.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/rng_state_0.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..cf9f7cc0f22e38cdfea7c36a09f25ba146158728
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b93ade60213134037a6e2a03060d84a67c5041245aff2ad374a5c97f337ee4c2
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/rng_state_1.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..22044561af392bd37b8a514ffc971741f1a025c2
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e17eca23ec632d21ab0115d8db1a9360104fa7b08d928954eee2eed9adf46b8a
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/rng_state_2.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..24228040a3cbfaaf0d0e62b9e12dbef736514695
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6614221a631cfb519fe3c961fa724d0b1c62628b10e8afa55fbb30d5380c88e3
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/scheduler.pt b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..cdc40707919775e0806d2d972b15c5d1bfeaf534
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:126d4d3556a09db5691c8b62e64036b220c95406a781431ac5270cca5cc5f1b3
+size 627
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/tokenizer.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb
+size 17210374
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128256": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|end_of_text|>",
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "[PAD]",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..c4a4e6e834f4bdd319ef13b159c597857a27b2ea
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/trainer_state.json
@@ -0,0 +1,17973 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.5205479452054793,
+  "eval_steps": 10,
+  "global_step": 2300,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001095890410958904,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0636,
+      "step": 1
+    },
+    {
+      "epoch": 0.002191780821917808,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0501,
+      "step": 2
+    },
+    {
+      "epoch": 0.0032876712328767125,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0238,
+      "step": 3
+    },
+    {
+      "epoch": 0.004383561643835616,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0346,
+      "step": 4
+    },
+    {
+      "epoch": 0.005479452054794521,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0108,
+      "step": 5
+    },
+    {
+      "epoch": 0.006575342465753425,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.131,
+      "step": 6
+    },
+    {
+      "epoch": 0.007671232876712329,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0488,
+      "step": 7
+    },
+    {
+      "epoch": 0.008767123287671232,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0905,
+      "step": 8
+    },
+    {
+      "epoch": 0.009863013698630137,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0493,
+      "step": 9
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.1038,
+      "step": 10
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "eval_loss": 3.100078821182251,
+      "eval_runtime": 37.5838,
+      "eval_samples_per_second": 26.607,
+      "eval_steps_per_second": 1.118,
+      "step": 10
+    },
+    {
+      "epoch": 0.012054794520547946,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0804,
+      "step": 11
+    },
+    {
+      "epoch": 0.01315068493150685,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0053,
+      "step": 12
+    },
+    {
+      "epoch": 0.014246575342465753,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0607,
+      "step": 13
+    },
+    {
+      "epoch": 0.015342465753424657,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0539,
+      "step": 14
+    },
+    {
+      "epoch": 0.01643835616438356,
+      "grad_norm": 5.420984745025635,
+      "learning_rate": 1.8248175182481753e-08,
+      "loss": 3.0941,
+      "step": 15
+    },
+    {
+      "epoch": 0.017534246575342465,
+      "grad_norm": 5.76840877532959,
+      "learning_rate": 3.6496350364963505e-08,
+      "loss": 3.036,
+      "step": 16
+    },
+    {
+      "epoch": 0.01863013698630137,
+      "grad_norm": 5.905357360839844,
+      "learning_rate": 5.4744525547445264e-08,
+      "loss": 3.0491,
+      "step": 17
+    },
+    {
+      "epoch": 0.019726027397260273,
+      "grad_norm": 5.981553554534912,
+      "learning_rate": 7.299270072992701e-08,
+      "loss": 3.0775,
+      "step": 18
+    },
+    {
+      "epoch": 0.020821917808219178,
+      "grad_norm": 6.240172863006592,
+      "learning_rate": 9.124087591240876e-08,
+      "loss": 3.0345,
+      "step": 19
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "grad_norm": 5.59375,
+      "learning_rate": 1.0948905109489053e-07,
+      "loss": 3.1063,
+      "step": 20
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "eval_loss": 3.098820686340332,
+      "eval_runtime": 37.2695,
+      "eval_samples_per_second": 26.832,
+      "eval_steps_per_second": 1.127,
+      "step": 20
+    },
+    {
+      "epoch": 0.023013698630136987,
+      "grad_norm": 5.819795608520508,
+      "learning_rate": 1.2773722627737227e-07,
+      "loss": 3.0808,
+      "step": 21
+    },
+    {
+      "epoch": 0.02410958904109589,
+      "grad_norm": 5.7714762687683105,
+      "learning_rate": 1.4598540145985402e-07,
+      "loss": 3.0055,
+      "step": 22
+    },
+    {
+      "epoch": 0.025205479452054796,
+      "grad_norm": 5.51552152633667,
+      "learning_rate": 1.642335766423358e-07,
+      "loss": 3.0396,
+      "step": 23
+    },
+    {
+      "epoch": 0.0263013698630137,
+      "grad_norm": 5.594012260437012,
+      "learning_rate": 1.8248175182481753e-07,
+      "loss": 3.0824,
+      "step": 24
+    },
+    {
+      "epoch": 0.0273972602739726,
+      "grad_norm": 5.1974029541015625,
+      "learning_rate": 2.0072992700729928e-07,
+      "loss": 3.08,
+      "step": 25
+    },
+    {
+      "epoch": 0.028493150684931506,
+      "grad_norm": 4.733968257904053,
+      "learning_rate": 2.1897810218978106e-07,
+      "loss": 3.0597,
+      "step": 26
+    },
+    {
+      "epoch": 0.02958904109589041,
+      "grad_norm": 4.717319011688232,
+      "learning_rate": 2.3722627737226278e-07,
+      "loss": 3.0696,
+      "step": 27
+    },
+    {
+      "epoch": 0.030684931506849315,
+      "grad_norm": 4.6303534507751465,
+      "learning_rate": 2.5547445255474454e-07,
+      "loss": 3.0273,
+      "step": 28
+    },
+    {
+      "epoch": 0.03178082191780822,
+      "grad_norm": 4.413999557495117,
+      "learning_rate": 2.737226277372263e-07,
+      "loss": 3.0202,
+      "step": 29
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "grad_norm": 4.516043186187744,
+      "learning_rate": 2.9197080291970804e-07,
+      "loss": 2.9926,
+      "step": 30
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "eval_loss": 3.0561797618865967,
+      "eval_runtime": 37.0351,
+      "eval_samples_per_second": 27.001,
+      "eval_steps_per_second": 1.134,
+      "step": 30
+    },
+    {
+      "epoch": 0.03397260273972603,
+      "grad_norm": 3.7179207801818848,
+      "learning_rate": 3.102189781021898e-07,
+      "loss": 3.0086,
+      "step": 31
+    },
+    {
+      "epoch": 0.03506849315068493,
+      "grad_norm": 3.7840425968170166,
+      "learning_rate": 3.284671532846716e-07,
+      "loss": 2.9986,
+      "step": 32
+    },
+    {
+      "epoch": 0.03616438356164384,
+      "grad_norm": 3.843116521835327,
+      "learning_rate": 3.467153284671533e-07,
+      "loss": 2.9677,
+      "step": 33
+    },
+    {
+      "epoch": 0.03726027397260274,
+      "grad_norm": 3.8528714179992676,
+      "learning_rate": 3.6496350364963505e-07,
+      "loss": 2.9732,
+      "step": 34
+    },
+    {
+      "epoch": 0.038356164383561646,
+      "grad_norm": 3.76963210105896,
+      "learning_rate": 3.8321167883211686e-07,
+      "loss": 2.9855,
+      "step": 35
+    },
+    {
+      "epoch": 0.03945205479452055,
+      "grad_norm": 3.6240015029907227,
+      "learning_rate": 4.0145985401459856e-07,
+      "loss": 3.0117,
+      "step": 36
+    },
+    {
+      "epoch": 0.040547945205479455,
+      "grad_norm": 3.1380624771118164,
+      "learning_rate": 4.197080291970803e-07,
+      "loss": 2.9481,
+      "step": 37
+    },
+    {
+      "epoch": 0.041643835616438356,
+      "grad_norm": 3.565329074859619,
+      "learning_rate": 4.379562043795621e-07,
+      "loss": 2.9175,
+      "step": 38
+    },
+    {
+      "epoch": 0.042739726027397264,
+      "grad_norm": 3.8062448501586914,
+      "learning_rate": 4.562043795620438e-07,
+      "loss": 3.0074,
+      "step": 39
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "grad_norm": 3.412046432495117,
+      "learning_rate": 4.7445255474452557e-07,
+      "loss": 2.9119,
+      "step": 40
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "eval_loss": 2.9874370098114014,
+      "eval_runtime": 37.0514,
+      "eval_samples_per_second": 26.99,
+      "eval_steps_per_second": 1.134,
+      "step": 40
+    },
+    {
+      "epoch": 0.044931506849315066,
+      "grad_norm": 3.227437973022461,
+      "learning_rate": 4.927007299270073e-07,
+      "loss": 2.9436,
+      "step": 41
+    },
+    {
+      "epoch": 0.046027397260273974,
+      "grad_norm": 2.9633443355560303,
+      "learning_rate": 5.109489051094891e-07,
+      "loss": 2.967,
+      "step": 42
+    },
+    {
+      "epoch": 0.047123287671232875,
+      "grad_norm": 2.942678928375244,
+      "learning_rate": 5.291970802919708e-07,
+      "loss": 2.9375,
+      "step": 43
+    },
+    {
+      "epoch": 0.04821917808219178,
+      "grad_norm": 2.9806888103485107,
+      "learning_rate": 5.474452554744526e-07,
+      "loss": 2.9474,
+      "step": 44
+    },
+    {
+      "epoch": 0.049315068493150684,
+      "grad_norm": 3.000063419342041,
+      "learning_rate": 5.656934306569343e-07,
+      "loss": 2.9289,
+      "step": 45
+    },
+    {
+      "epoch": 0.05041095890410959,
+      "grad_norm": 3.1481194496154785,
+      "learning_rate": 5.839416058394161e-07,
+      "loss": 2.8869,
+      "step": 46
+    },
+    {
+      "epoch": 0.05150684931506849,
+      "grad_norm": 2.949202060699463,
+      "learning_rate": 6.021897810218979e-07,
+      "loss": 2.8783,
+      "step": 47
+    },
+    {
+      "epoch": 0.0526027397260274,
+      "grad_norm": 2.706526279449463,
+      "learning_rate": 6.204379562043796e-07,
+      "loss": 2.9024,
+      "step": 48
+    },
+    {
+      "epoch": 0.0536986301369863,
+      "grad_norm": 2.7317655086517334,
+      "learning_rate": 6.386861313868614e-07,
+      "loss": 2.8325,
+      "step": 49
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "grad_norm": 2.25644588470459,
+      "learning_rate": 6.569343065693432e-07,
+      "loss": 2.9409,
+      "step": 50
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "eval_loss": 2.909320116043091,
+      "eval_runtime": 37.0714,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 50
+    },
+    {
+      "epoch": 0.05589041095890411,
+      "grad_norm": 2.8789613246917725,
+      "learning_rate": 6.751824817518248e-07,
+      "loss": 2.8459,
+      "step": 51
+    },
+    {
+      "epoch": 0.05698630136986301,
+      "grad_norm": 2.686873435974121,
+      "learning_rate": 6.934306569343066e-07,
+      "loss": 2.8653,
+      "step": 52
+    },
+    {
+      "epoch": 0.05808219178082192,
+      "grad_norm": 2.54382586479187,
+      "learning_rate": 7.116788321167883e-07,
+      "loss": 2.7973,
+      "step": 53
+    },
+    {
+      "epoch": 0.05917808219178082,
+      "grad_norm": 2.1871724128723145,
+      "learning_rate": 7.299270072992701e-07,
+      "loss": 2.7982,
+      "step": 54
+    },
+    {
+      "epoch": 0.06027397260273973,
+      "grad_norm": 2.145303249359131,
+      "learning_rate": 7.48175182481752e-07,
+      "loss": 2.8195,
+      "step": 55
+    },
+    {
+      "epoch": 0.06136986301369863,
+      "grad_norm": 2.1635069847106934,
+      "learning_rate": 7.664233576642337e-07,
+      "loss": 2.825,
+      "step": 56
+    },
+    {
+      "epoch": 0.06246575342465754,
+      "grad_norm": 2.3530752658843994,
+      "learning_rate": 7.846715328467154e-07,
+      "loss": 2.8407,
+      "step": 57
+    },
+    {
+      "epoch": 0.06356164383561644,
+      "grad_norm": 2.2455942630767822,
+      "learning_rate": 8.029197080291971e-07,
+      "loss": 2.8182,
+      "step": 58
+    },
+    {
+      "epoch": 0.06465753424657535,
+      "grad_norm": 1.8985673189163208,
+      "learning_rate": 8.211678832116789e-07,
+      "loss": 2.772,
+      "step": 59
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "grad_norm": 1.9699499607086182,
+      "learning_rate": 8.394160583941606e-07,
+      "loss": 2.7367,
+      "step": 60
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "eval_loss": 2.8451859951019287,
+      "eval_runtime": 37.0793,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 60
+    },
+    {
+      "epoch": 0.06684931506849315,
+      "grad_norm": 2.159547805786133,
+      "learning_rate": 8.576642335766425e-07,
+      "loss": 2.7635,
+      "step": 61
+    },
+    {
+      "epoch": 0.06794520547945206,
+      "grad_norm": 2.12705659866333,
+      "learning_rate": 8.759124087591242e-07,
+      "loss": 2.7624,
+      "step": 62
+    },
+    {
+      "epoch": 0.06904109589041096,
+      "grad_norm": 1.8393241167068481,
+      "learning_rate": 8.94160583941606e-07,
+      "loss": 2.7763,
+      "step": 63
+    },
+    {
+      "epoch": 0.07013698630136986,
+      "grad_norm": 1.683522343635559,
+      "learning_rate": 9.124087591240876e-07,
+      "loss": 2.7762,
+      "step": 64
+    },
+    {
+      "epoch": 0.07123287671232877,
+      "grad_norm": 1.9260376691818237,
+      "learning_rate": 9.306569343065694e-07,
+      "loss": 2.7418,
+      "step": 65
+    },
+    {
+      "epoch": 0.07232876712328767,
+      "grad_norm": 2.09322190284729,
+      "learning_rate": 9.489051094890511e-07,
+      "loss": 2.7625,
+      "step": 66
+    },
+    {
+      "epoch": 0.07342465753424658,
+      "grad_norm": 1.9183429479599,
+      "learning_rate": 9.67153284671533e-07,
+      "loss": 2.7453,
+      "step": 67
+    },
+    {
+      "epoch": 0.07452054794520548,
+      "grad_norm": 1.9792778491973877,
+      "learning_rate": 9.854014598540146e-07,
+      "loss": 2.7858,
+      "step": 68
+    },
+    {
+      "epoch": 0.07561643835616438,
+      "grad_norm": 2.1397311687469482,
+      "learning_rate": 1.0036496350364965e-06,
+      "loss": 2.7242,
+      "step": 69
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "grad_norm": 1.8556463718414307,
+      "learning_rate": 1.0218978102189781e-06,
+      "loss": 2.6853,
+      "step": 70
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "eval_loss": 2.7787046432495117,
+      "eval_runtime": 37.1017,
+      "eval_samples_per_second": 26.953,
+      "eval_steps_per_second": 1.132,
+      "step": 70
+    },
+    {
+      "epoch": 0.07780821917808219,
+      "grad_norm": 1.9776800870895386,
+      "learning_rate": 1.04014598540146e-06,
+      "loss": 2.749,
+      "step": 71
+    },
+    {
+      "epoch": 0.0789041095890411,
+      "grad_norm": 2.3972630500793457,
+      "learning_rate": 1.0583941605839416e-06,
+      "loss": 2.6457,
+      "step": 72
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 2.090210199356079,
+      "learning_rate": 1.0766423357664233e-06,
+      "loss": 2.6401,
+      "step": 73
+    },
+    {
+      "epoch": 0.08109589041095891,
+      "grad_norm": 2.4117701053619385,
+      "learning_rate": 1.0948905109489052e-06,
+      "loss": 2.6649,
+      "step": 74
+    },
+    {
+      "epoch": 0.0821917808219178,
+      "grad_norm": 2.4845173358917236,
+      "learning_rate": 1.113138686131387e-06,
+      "loss": 2.6855,
+      "step": 75
+    },
+    {
+      "epoch": 0.08328767123287671,
+      "grad_norm": 1.7779242992401123,
+      "learning_rate": 1.1313868613138687e-06,
+      "loss": 2.6926,
+      "step": 76
+    },
+    {
+      "epoch": 0.08438356164383562,
+      "grad_norm": 2.3163535594940186,
+      "learning_rate": 1.1496350364963505e-06,
+      "loss": 2.6302,
+      "step": 77
+    },
+    {
+      "epoch": 0.08547945205479453,
+      "grad_norm": 2.2296528816223145,
+      "learning_rate": 1.1678832116788322e-06,
+      "loss": 2.6054,
+      "step": 78
+    },
+    {
+      "epoch": 0.08657534246575342,
+      "grad_norm": 2.253431797027588,
+      "learning_rate": 1.1861313868613138e-06,
+      "loss": 2.6104,
+      "step": 79
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "grad_norm": 2.490609645843506,
+      "learning_rate": 1.2043795620437959e-06,
+      "loss": 2.6078,
+      "step": 80
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "eval_loss": 2.706103563308716,
+      "eval_runtime": 37.0892,
+      "eval_samples_per_second": 26.962,
+      "eval_steps_per_second": 1.132,
+      "step": 80
+    },
+    {
+      "epoch": 0.08876712328767124,
+      "grad_norm": 2.3048248291015625,
+      "learning_rate": 1.2226277372262775e-06,
+      "loss": 2.6096,
+      "step": 81
+    },
+    {
+      "epoch": 0.08986301369863013,
+      "grad_norm": 2.440666913986206,
+      "learning_rate": 1.2408759124087592e-06,
+      "loss": 2.6412,
+      "step": 82
+    },
+    {
+      "epoch": 0.09095890410958904,
+      "grad_norm": 1.9456374645233154,
+      "learning_rate": 1.259124087591241e-06,
+      "loss": 2.591,
+      "step": 83
+    },
+    {
+      "epoch": 0.09205479452054795,
+      "grad_norm": 2.598172664642334,
+      "learning_rate": 1.2773722627737229e-06,
+      "loss": 2.6093,
+      "step": 84
+    },
+    {
+      "epoch": 0.09315068493150686,
+      "grad_norm": 2.2587032318115234,
+      "learning_rate": 1.2956204379562045e-06,
+      "loss": 2.5949,
+      "step": 85
+    },
+    {
+      "epoch": 0.09424657534246575,
+      "grad_norm": 2.2927448749542236,
+      "learning_rate": 1.3138686131386864e-06,
+      "loss": 2.5464,
+      "step": 86
+    },
+    {
+      "epoch": 0.09534246575342466,
+      "grad_norm": 2.5940468311309814,
+      "learning_rate": 1.332116788321168e-06,
+      "loss": 2.5613,
+      "step": 87
+    },
+    {
+      "epoch": 0.09643835616438357,
+      "grad_norm": 2.073207378387451,
+      "learning_rate": 1.3503649635036497e-06,
+      "loss": 2.4955,
+      "step": 88
+    },
+    {
+      "epoch": 0.09753424657534246,
+      "grad_norm": 3.211766242980957,
+      "learning_rate": 1.3686131386861315e-06,
+      "loss": 2.5315,
+      "step": 89
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "grad_norm": 2.2590038776397705,
+      "learning_rate": 1.3868613138686132e-06,
+      "loss": 2.4641,
+      "step": 90
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "eval_loss": 2.625087261199951,
+      "eval_runtime": 37.0852,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 90
+    },
+    {
+      "epoch": 0.09972602739726028,
+      "grad_norm": 2.3959836959838867,
+      "learning_rate": 1.405109489051095e-06,
+      "loss": 2.5537,
+      "step": 91
+    },
+    {
+      "epoch": 0.10082191780821918,
+      "grad_norm": 2.596761465072632,
+      "learning_rate": 1.4233576642335767e-06,
+      "loss": 2.5135,
+      "step": 92
+    },
+    {
+      "epoch": 0.10191780821917808,
+      "grad_norm": 2.475759267807007,
+      "learning_rate": 1.4416058394160583e-06,
+      "loss": 2.5436,
+      "step": 93
+    },
+    {
+      "epoch": 0.10301369863013699,
+      "grad_norm": 3.0613298416137695,
+      "learning_rate": 1.4598540145985402e-06,
+      "loss": 2.5066,
+      "step": 94
+    },
+    {
+      "epoch": 0.10410958904109589,
+      "grad_norm": 2.406097650527954,
+      "learning_rate": 1.478102189781022e-06,
+      "loss": 2.4646,
+      "step": 95
+    },
+    {
+      "epoch": 0.1052054794520548,
+      "grad_norm": 2.5872504711151123,
+      "learning_rate": 1.496350364963504e-06,
+      "loss": 2.5351,
+      "step": 96
+    },
+    {
+      "epoch": 0.1063013698630137,
+      "grad_norm": 2.4294471740722656,
+      "learning_rate": 1.5145985401459856e-06,
+      "loss": 2.4953,
+      "step": 97
+    },
+    {
+      "epoch": 0.1073972602739726,
+      "grad_norm": 2.674884080886841,
+      "learning_rate": 1.5328467153284674e-06,
+      "loss": 2.5133,
+      "step": 98
+    },
+    {
+      "epoch": 0.10849315068493151,
+      "grad_norm": 2.5096852779388428,
+      "learning_rate": 1.551094890510949e-06,
+      "loss": 2.4693,
+      "step": 99
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "grad_norm": 2.7390334606170654,
+      "learning_rate": 1.5693430656934307e-06,
+      "loss": 2.4195,
+      "step": 100
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "eval_loss": 2.55997371673584,
+      "eval_runtime": 37.0732,
+      "eval_samples_per_second": 26.974,
+      "eval_steps_per_second": 1.133,
+      "step": 100
+    },
+    {
+      "epoch": 0.11068493150684931,
+      "grad_norm": 2.8887312412261963,
+      "learning_rate": 1.5875912408759126e-06,
+      "loss": 2.4193,
+      "step": 101
+    },
+    {
+      "epoch": 0.11178082191780822,
+      "grad_norm": 3.0721421241760254,
+      "learning_rate": 1.6058394160583942e-06,
+      "loss": 2.4709,
+      "step": 102
+    },
+    {
+      "epoch": 0.11287671232876713,
+      "grad_norm": 2.728156566619873,
+      "learning_rate": 1.624087591240876e-06,
+      "loss": 2.5105,
+      "step": 103
+    },
+    {
+      "epoch": 0.11397260273972602,
+      "grad_norm": 3.1214346885681152,
+      "learning_rate": 1.6423357664233577e-06,
+      "loss": 2.4316,
+      "step": 104
+    },
+    {
+      "epoch": 0.11506849315068493,
+      "grad_norm": 2.9249517917633057,
+      "learning_rate": 1.6605839416058394e-06,
+      "loss": 2.3737,
+      "step": 105
+    },
+    {
+      "epoch": 0.11616438356164384,
+      "grad_norm": 2.6646173000335693,
+      "learning_rate": 1.6788321167883212e-06,
+      "loss": 2.4165,
+      "step": 106
+    },
+    {
+      "epoch": 0.11726027397260275,
+      "grad_norm": 3.108067750930786,
+      "learning_rate": 1.6970802919708029e-06,
+      "loss": 2.4217,
+      "step": 107
+    },
+    {
+      "epoch": 0.11835616438356164,
+      "grad_norm": 3.142338991165161,
+      "learning_rate": 1.715328467153285e-06,
+      "loss": 2.3527,
+      "step": 108
+    },
+    {
+      "epoch": 0.11945205479452055,
+      "grad_norm": 2.9078011512756348,
+      "learning_rate": 1.7335766423357666e-06,
+      "loss": 2.3927,
+      "step": 109
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "grad_norm": 3.40766978263855,
+      "learning_rate": 1.7518248175182485e-06,
+      "loss": 2.4021,
+      "step": 110
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "eval_loss": 2.5031485557556152,
+      "eval_runtime": 36.8147,
+      "eval_samples_per_second": 27.163,
+      "eval_steps_per_second": 1.141,
+      "step": 110
+    },
+    {
+      "epoch": 0.12164383561643835,
+      "grad_norm": 4.396062850952148,
+      "learning_rate": 1.77007299270073e-06,
+      "loss": 2.4373,
+      "step": 111
+    },
+    {
+      "epoch": 0.12273972602739726,
+      "grad_norm": 4.0441365242004395,
+      "learning_rate": 1.788321167883212e-06,
+      "loss": 2.4162,
+      "step": 112
+    },
+    {
+      "epoch": 0.12383561643835617,
+      "grad_norm": 2.8869969844818115,
+      "learning_rate": 1.8065693430656936e-06,
+      "loss": 2.4515,
+      "step": 113
+    },
+    {
+      "epoch": 0.12493150684931507,
+      "grad_norm": 3.4833669662475586,
+      "learning_rate": 1.8248175182481753e-06,
+      "loss": 2.3965,
+      "step": 114
+    },
+    {
+      "epoch": 0.12602739726027398,
+      "grad_norm": 2.9996068477630615,
+      "learning_rate": 1.8430656934306571e-06,
+      "loss": 2.3374,
+      "step": 115
+    },
+    {
+      "epoch": 0.12712328767123288,
+      "grad_norm": 2.701371192932129,
+      "learning_rate": 1.8613138686131388e-06,
+      "loss": 2.3608,
+      "step": 116
+    },
+    {
+      "epoch": 0.12821917808219177,
+      "grad_norm": 2.9491288661956787,
+      "learning_rate": 1.8795620437956206e-06,
+      "loss": 2.355,
+      "step": 117
+    },
+    {
+      "epoch": 0.1293150684931507,
+      "grad_norm": 2.9266560077667236,
+      "learning_rate": 1.8978102189781023e-06,
+      "loss": 2.3475,
+      "step": 118
+    },
+    {
+      "epoch": 0.1304109589041096,
+      "grad_norm": 3.0419797897338867,
+      "learning_rate": 1.916058394160584e-06,
+      "loss": 2.3475,
+      "step": 119
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "grad_norm": 2.4531214237213135,
+      "learning_rate": 1.934306569343066e-06,
+      "loss": 2.3724,
+      "step": 120
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "eval_loss": 2.455423593521118,
+      "eval_runtime": 37.0304,
+      "eval_samples_per_second": 27.005,
+      "eval_steps_per_second": 1.134,
+      "step": 120
+    },
+    {
+      "epoch": 0.1326027397260274,
+      "grad_norm": 3.572502851486206,
+      "learning_rate": 1.952554744525548e-06,
+      "loss": 2.2671,
+      "step": 121
+    },
+    {
+      "epoch": 0.1336986301369863,
+      "grad_norm": 3.5408504009246826,
+      "learning_rate": 1.9708029197080293e-06,
+      "loss": 2.3133,
+      "step": 122
+    },
+    {
+      "epoch": 0.13479452054794522,
+      "grad_norm": 2.488725423812866,
+      "learning_rate": 1.989051094890511e-06,
+      "loss": 2.3141,
+      "step": 123
+    },
+    {
+      "epoch": 0.1358904109589041,
+      "grad_norm": 3.241276264190674,
+      "learning_rate": 2.007299270072993e-06,
+      "loss": 2.318,
+      "step": 124
+    },
+    {
+      "epoch": 0.136986301369863,
+      "grad_norm": 2.853959798812866,
+      "learning_rate": 2.025547445255475e-06,
+      "loss": 2.3712,
+      "step": 125
+    },
+    {
+      "epoch": 0.13808219178082193,
+      "grad_norm": 2.9798169136047363,
+      "learning_rate": 2.0437956204379563e-06,
+      "loss": 2.3072,
+      "step": 126
+    },
+    {
+      "epoch": 0.13917808219178082,
+      "grad_norm": 2.8529601097106934,
+      "learning_rate": 2.062043795620438e-06,
+      "loss": 2.318,
+      "step": 127
+    },
+    {
+      "epoch": 0.14027397260273972,
+      "grad_norm": 3.475337028503418,
+      "learning_rate": 2.08029197080292e-06,
+      "loss": 2.31,
+      "step": 128
+    },
+    {
+      "epoch": 0.14136986301369864,
+      "grad_norm": 4.213955879211426,
+      "learning_rate": 2.0985401459854014e-06,
+      "loss": 2.3734,
+      "step": 129
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "grad_norm": 3.0833330154418945,
+      "learning_rate": 2.1167883211678833e-06,
+      "loss": 2.2602,
+      "step": 130
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "eval_loss": 2.4205682277679443,
+      "eval_runtime": 37.0798,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 130
+    },
+    {
+      "epoch": 0.14356164383561643,
+      "grad_norm": 3.411235809326172,
+      "learning_rate": 2.135036496350365e-06,
+      "loss": 2.3255,
+      "step": 131
+    },
+    {
+      "epoch": 0.14465753424657535,
+      "grad_norm": 2.882053852081299,
+      "learning_rate": 2.1532846715328466e-06,
+      "loss": 2.3483,
+      "step": 132
+    },
+    {
+      "epoch": 0.14575342465753424,
+      "grad_norm": 3.2436420917510986,
+      "learning_rate": 2.171532846715329e-06,
+      "loss": 2.2492,
+      "step": 133
+    },
+    {
+      "epoch": 0.14684931506849316,
+      "grad_norm": 2.929466962814331,
+      "learning_rate": 2.1897810218978103e-06,
+      "loss": 2.2647,
+      "step": 134
+    },
+    {
+      "epoch": 0.14794520547945206,
+      "grad_norm": 2.9344966411590576,
+      "learning_rate": 2.208029197080292e-06,
+      "loss": 2.3061,
+      "step": 135
+    },
+    {
+      "epoch": 0.14904109589041095,
+      "grad_norm": 3.2567203044891357,
+      "learning_rate": 2.226277372262774e-06,
+      "loss": 2.2746,
+      "step": 136
+    },
+    {
+      "epoch": 0.15013698630136987,
+      "grad_norm": 3.1494054794311523,
+      "learning_rate": 2.244525547445256e-06,
+      "loss": 2.2866,
+      "step": 137
+    },
+    {
+      "epoch": 0.15123287671232877,
+      "grad_norm": 2.723390817642212,
+      "learning_rate": 2.2627737226277373e-06,
+      "loss": 2.3109,
+      "step": 138
+    },
+    {
+      "epoch": 0.15232876712328766,
+      "grad_norm": 2.8904967308044434,
+      "learning_rate": 2.281021897810219e-06,
+      "loss": 2.301,
+      "step": 139
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "grad_norm": 2.8899483680725098,
+      "learning_rate": 2.299270072992701e-06,
+      "loss": 2.2637,
+      "step": 140
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "eval_loss": 2.3952724933624268,
+      "eval_runtime": 37.1272,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 140
+    },
+    {
+      "epoch": 0.15452054794520548,
+      "grad_norm": 3.5309066772460938,
+      "learning_rate": 2.3175182481751825e-06,
+      "loss": 2.2572,
+      "step": 141
+    },
+    {
+      "epoch": 0.15561643835616437,
+      "grad_norm": 2.8369667530059814,
+      "learning_rate": 2.3357664233576643e-06,
+      "loss": 2.2712,
+      "step": 142
+    },
+    {
+      "epoch": 0.1567123287671233,
+      "grad_norm": 3.629161834716797,
+      "learning_rate": 2.354014598540146e-06,
+      "loss": 2.2301,
+      "step": 143
+    },
+    {
+      "epoch": 0.1578082191780822,
+      "grad_norm": 3.5313000679016113,
+      "learning_rate": 2.3722627737226276e-06,
+      "loss": 2.2712,
+      "step": 144
+    },
+    {
+      "epoch": 0.1589041095890411,
+      "grad_norm": 3.1328771114349365,
+      "learning_rate": 2.39051094890511e-06,
+      "loss": 2.2706,
+      "step": 145
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 4.6992950439453125,
+      "learning_rate": 2.4087591240875918e-06,
+      "loss": 2.219,
+      "step": 146
+    },
+    {
+      "epoch": 0.1610958904109589,
+      "grad_norm": 5.230221748352051,
+      "learning_rate": 2.427007299270073e-06,
+      "loss": 2.2396,
+      "step": 147
+    },
+    {
+      "epoch": 0.16219178082191782,
+      "grad_norm": 3.5680761337280273,
+      "learning_rate": 2.445255474452555e-06,
+      "loss": 2.2434,
+      "step": 148
+    },
+    {
+      "epoch": 0.1632876712328767,
+      "grad_norm": 3.1642441749572754,
+      "learning_rate": 2.463503649635037e-06,
+      "loss": 2.2856,
+      "step": 149
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "grad_norm": 2.694085121154785,
+      "learning_rate": 2.4817518248175183e-06,
+      "loss": 2.2006,
+      "step": 150
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "eval_loss": 2.37199068069458,
+      "eval_runtime": 37.1342,
+      "eval_samples_per_second": 26.929,
+      "eval_steps_per_second": 1.131,
+      "step": 150
+    },
+    {
+      "epoch": 0.16547945205479453,
+      "grad_norm": 2.9868853092193604,
+      "learning_rate": 2.5e-06,
+      "loss": 2.1694,
+      "step": 151
+    },
+    {
+      "epoch": 0.16657534246575342,
+      "grad_norm": 2.793532371520996,
+      "learning_rate": 2.518248175182482e-06,
+      "loss": 2.2965,
+      "step": 152
+    },
+    {
+      "epoch": 0.16767123287671232,
+      "grad_norm": 2.60152530670166,
+      "learning_rate": 2.5364963503649635e-06,
+      "loss": 2.1954,
+      "step": 153
+    },
+    {
+      "epoch": 0.16876712328767124,
+      "grad_norm": 2.6192219257354736,
+      "learning_rate": 2.5547445255474458e-06,
+      "loss": 2.1909,
+      "step": 154
+    },
+    {
+      "epoch": 0.16986301369863013,
+      "grad_norm": 2.467344045639038,
+      "learning_rate": 2.5729927007299272e-06,
+      "loss": 2.2438,
+      "step": 155
+    },
+    {
+      "epoch": 0.17095890410958905,
+      "grad_norm": 2.772012948989868,
+      "learning_rate": 2.591240875912409e-06,
+      "loss": 2.2213,
+      "step": 156
+    },
+    {
+      "epoch": 0.17205479452054795,
+      "grad_norm": 2.613675355911255,
+      "learning_rate": 2.6094890510948905e-06,
+      "loss": 2.2213,
+      "step": 157
+    },
+    {
+      "epoch": 0.17315068493150684,
+      "grad_norm": 2.849245071411133,
+      "learning_rate": 2.627737226277373e-06,
+      "loss": 2.255,
+      "step": 158
+    },
+    {
+      "epoch": 0.17424657534246576,
+      "grad_norm": 2.8388969898223877,
+      "learning_rate": 2.6459854014598542e-06,
+      "loss": 2.2572,
+      "step": 159
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "grad_norm": 2.5219321250915527,
+      "learning_rate": 2.664233576642336e-06,
+      "loss": 2.2253,
+      "step": 160
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "eval_loss": 2.3509061336517334,
+      "eval_runtime": 37.0953,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 160
+    },
+    {
+      "epoch": 0.17643835616438355,
+      "grad_norm": 2.8460891246795654,
+      "learning_rate": 2.6824817518248175e-06,
+      "loss": 2.2009,
+      "step": 161
+    },
+    {
+      "epoch": 0.17753424657534247,
+      "grad_norm": 2.673377275466919,
+      "learning_rate": 2.7007299270072994e-06,
+      "loss": 2.2211,
+      "step": 162
+    },
+    {
+      "epoch": 0.17863013698630137,
+      "grad_norm": 2.8514182567596436,
+      "learning_rate": 2.718978102189781e-06,
+      "loss": 2.1955,
+      "step": 163
+    },
+    {
+      "epoch": 0.17972602739726026,
+      "grad_norm": 3.530385971069336,
+      "learning_rate": 2.737226277372263e-06,
+      "loss": 2.2536,
+      "step": 164
+    },
+    {
+      "epoch": 0.18082191780821918,
+      "grad_norm": 2.733394145965576,
+      "learning_rate": 2.755474452554745e-06,
+      "loss": 2.2367,
+      "step": 165
+    },
+    {
+      "epoch": 0.18191780821917808,
+      "grad_norm": 3.871150016784668,
+      "learning_rate": 2.7737226277372264e-06,
+      "loss": 2.1949,
+      "step": 166
+    },
+    {
+      "epoch": 0.18301369863013697,
+      "grad_norm": 3.7782530784606934,
+      "learning_rate": 2.7919708029197082e-06,
+      "loss": 2.2979,
+      "step": 167
+    },
+    {
+      "epoch": 0.1841095890410959,
+      "grad_norm": 3.094599485397339,
+      "learning_rate": 2.81021897810219e-06,
+      "loss": 2.2512,
+      "step": 168
+    },
+    {
+      "epoch": 0.1852054794520548,
+      "grad_norm": 3.2166848182678223,
+      "learning_rate": 2.828467153284672e-06,
+      "loss": 2.1784,
+      "step": 169
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "grad_norm": 3.4127073287963867,
+      "learning_rate": 2.8467153284671534e-06,
+      "loss": 2.2241,
+      "step": 170
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "eval_loss": 2.333245038986206,
+      "eval_runtime": 37.0402,
+      "eval_samples_per_second": 26.998,
+      "eval_steps_per_second": 1.134,
+      "step": 170
+    },
+    {
+      "epoch": 0.1873972602739726,
+      "grad_norm": 2.677579879760742,
+      "learning_rate": 2.8649635036496353e-06,
+      "loss": 2.135,
+      "step": 171
+    },
+    {
+      "epoch": 0.1884931506849315,
+      "grad_norm": 2.9602150917053223,
+      "learning_rate": 2.8832116788321167e-06,
+      "loss": 2.1916,
+      "step": 172
+    },
+    {
+      "epoch": 0.18958904109589042,
+      "grad_norm": 2.580122232437134,
+      "learning_rate": 2.901459854014599e-06,
+      "loss": 2.1477,
+      "step": 173
+    },
+    {
+      "epoch": 0.19068493150684931,
+      "grad_norm": 2.652217388153076,
+      "learning_rate": 2.9197080291970804e-06,
+      "loss": 2.2394,
+      "step": 174
+    },
+    {
+      "epoch": 0.1917808219178082,
+      "grad_norm": 2.6887567043304443,
+      "learning_rate": 2.9379562043795623e-06,
+      "loss": 2.2245,
+      "step": 175
+    },
+    {
+      "epoch": 0.19287671232876713,
+      "grad_norm": 2.768165111541748,
+      "learning_rate": 2.956204379562044e-06,
+      "loss": 2.2017,
+      "step": 176
+    },
+    {
+      "epoch": 0.19397260273972602,
+      "grad_norm": 2.9375691413879395,
+      "learning_rate": 2.9744525547445256e-06,
+      "loss": 2.2464,
+      "step": 177
+    },
+    {
+      "epoch": 0.19506849315068492,
+      "grad_norm": 2.8364248275756836,
+      "learning_rate": 2.992700729927008e-06,
+      "loss": 2.2155,
+      "step": 178
+    },
+    {
+      "epoch": 0.19616438356164384,
+      "grad_norm": 2.843487024307251,
+      "learning_rate": 3.0109489051094893e-06,
+      "loss": 2.1577,
+      "step": 179
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "grad_norm": 3.005230665206909,
+      "learning_rate": 3.029197080291971e-06,
+      "loss": 2.1549,
+      "step": 180
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "eval_loss": 2.3209409713745117,
+      "eval_runtime": 37.022,
+      "eval_samples_per_second": 27.011,
+      "eval_steps_per_second": 1.134,
+      "step": 180
+    },
+    {
+      "epoch": 0.19835616438356166,
+      "grad_norm": 2.490260601043701,
+      "learning_rate": 3.0474452554744526e-06,
+      "loss": 2.2313,
+      "step": 181
+    },
+    {
+      "epoch": 0.19945205479452055,
+      "grad_norm": 2.973118543624878,
+      "learning_rate": 3.065693430656935e-06,
+      "loss": 2.2316,
+      "step": 182
+    },
+    {
+      "epoch": 0.20054794520547944,
+      "grad_norm": 3.3308680057525635,
+      "learning_rate": 3.0839416058394163e-06,
+      "loss": 2.2087,
+      "step": 183
+    },
+    {
+      "epoch": 0.20164383561643837,
+      "grad_norm": 3.1440727710723877,
+      "learning_rate": 3.102189781021898e-06,
+      "loss": 2.2256,
+      "step": 184
+    },
+    {
+      "epoch": 0.20273972602739726,
+      "grad_norm": 2.5696325302124023,
+      "learning_rate": 3.1204379562043796e-06,
+      "loss": 2.2093,
+      "step": 185
+    },
+    {
+      "epoch": 0.20383561643835615,
+      "grad_norm": 3.1225037574768066,
+      "learning_rate": 3.1386861313868614e-06,
+      "loss": 2.1425,
+      "step": 186
+    },
+    {
+      "epoch": 0.20493150684931508,
+      "grad_norm": 3.650743246078491,
+      "learning_rate": 3.156934306569343e-06,
+      "loss": 2.1339,
+      "step": 187
+    },
+    {
+      "epoch": 0.20602739726027397,
+      "grad_norm": 2.924072027206421,
+      "learning_rate": 3.175182481751825e-06,
+      "loss": 2.2363,
+      "step": 188
+    },
+    {
+      "epoch": 0.20712328767123286,
+      "grad_norm": 2.9561989307403564,
+      "learning_rate": 3.193430656934307e-06,
+      "loss": 2.1934,
+      "step": 189
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "grad_norm": 2.731393814086914,
+      "learning_rate": 3.2116788321167884e-06,
+      "loss": 2.1718,
+      "step": 190
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "eval_loss": 2.3087782859802246,
+      "eval_runtime": 37.0161,
+      "eval_samples_per_second": 27.015,
+      "eval_steps_per_second": 1.135,
+      "step": 190
+    },
+    {
+      "epoch": 0.20931506849315068,
+      "grad_norm": 2.774442434310913,
+      "learning_rate": 3.2299270072992707e-06,
+      "loss": 2.1841,
+      "step": 191
+    },
+    {
+      "epoch": 0.2104109589041096,
+      "grad_norm": 3.0579135417938232,
+      "learning_rate": 3.248175182481752e-06,
+      "loss": 2.1677,
+      "step": 192
+    },
+    {
+      "epoch": 0.2115068493150685,
+      "grad_norm": 2.732984781265259,
+      "learning_rate": 3.266423357664234e-06,
+      "loss": 2.0975,
+      "step": 193
+    },
+    {
+      "epoch": 0.2126027397260274,
+      "grad_norm": 3.442659854888916,
+      "learning_rate": 3.2846715328467155e-06,
+      "loss": 2.1923,
+      "step": 194
+    },
+    {
+      "epoch": 0.2136986301369863,
+      "grad_norm": 3.064223527908325,
+      "learning_rate": 3.3029197080291973e-06,
+      "loss": 2.1707,
+      "step": 195
+    },
+    {
+      "epoch": 0.2147945205479452,
+      "grad_norm": 2.5138468742370605,
+      "learning_rate": 3.3211678832116788e-06,
+      "loss": 2.1933,
+      "step": 196
+    },
+    {
+      "epoch": 0.2158904109589041,
+      "grad_norm": 3.1076464653015137,
+      "learning_rate": 3.339416058394161e-06,
+      "loss": 2.1725,
+      "step": 197
+    },
+    {
+      "epoch": 0.21698630136986302,
+      "grad_norm": 2.7291388511657715,
+      "learning_rate": 3.3576642335766425e-06,
+      "loss": 2.1722,
+      "step": 198
+    },
+    {
+      "epoch": 0.21808219178082192,
+      "grad_norm": 2.805506467819214,
+      "learning_rate": 3.3759124087591243e-06,
+      "loss": 2.1688,
+      "step": 199
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "grad_norm": 2.6190085411071777,
+      "learning_rate": 3.3941605839416058e-06,
+      "loss": 2.1708,
+      "step": 200
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "eval_loss": 2.295546054840088,
+      "eval_runtime": 37.0133,
+      "eval_samples_per_second": 27.017,
+      "eval_steps_per_second": 1.135,
+      "step": 200
+    },
+    {
+      "epoch": 0.22027397260273973,
+      "grad_norm": 3.1270673274993896,
+      "learning_rate": 3.412408759124088e-06,
+      "loss": 2.1759,
+      "step": 201
+    },
+    {
+      "epoch": 0.22136986301369863,
+      "grad_norm": 3.494178533554077,
+      "learning_rate": 3.43065693430657e-06,
+      "loss": 2.1795,
+      "step": 202
+    },
+    {
+      "epoch": 0.22246575342465755,
+      "grad_norm": 2.7398736476898193,
+      "learning_rate": 3.4489051094890513e-06,
+      "loss": 2.1201,
+      "step": 203
+    },
+    {
+      "epoch": 0.22356164383561644,
+      "grad_norm": 2.981506824493408,
+      "learning_rate": 3.467153284671533e-06,
+      "loss": 2.1693,
+      "step": 204
+    },
+    {
+      "epoch": 0.22465753424657534,
+      "grad_norm": 2.4129178524017334,
+      "learning_rate": 3.4854014598540146e-06,
+      "loss": 2.1125,
+      "step": 205
+    },
+    {
+      "epoch": 0.22575342465753426,
+      "grad_norm": 2.701197385787964,
+      "learning_rate": 3.503649635036497e-06,
+      "loss": 2.1565,
+      "step": 206
+    },
+    {
+      "epoch": 0.22684931506849315,
+      "grad_norm": 2.533306837081909,
+      "learning_rate": 3.5218978102189783e-06,
+      "loss": 2.1487,
+      "step": 207
+    },
+    {
+      "epoch": 0.22794520547945205,
+      "grad_norm": 2.8030788898468018,
+      "learning_rate": 3.54014598540146e-06,
+      "loss": 2.1801,
+      "step": 208
+    },
+    {
+      "epoch": 0.22904109589041097,
+      "grad_norm": 2.895414352416992,
+      "learning_rate": 3.5583941605839416e-06,
+      "loss": 2.189,
+      "step": 209
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "grad_norm": 2.508390426635742,
+      "learning_rate": 3.576642335766424e-06,
+      "loss": 2.2376,
+      "step": 210
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "eval_loss": 2.2900516986846924,
+      "eval_runtime": 36.7645,
+      "eval_samples_per_second": 27.2,
+      "eval_steps_per_second": 1.142,
+      "step": 210
+    },
+    {
+      "epoch": 0.23123287671232876,
+      "grad_norm": 2.6393816471099854,
+      "learning_rate": 3.5948905109489054e-06,
+      "loss": 2.2267,
+      "step": 211
+    },
+    {
+      "epoch": 0.23232876712328768,
+      "grad_norm": 3.0687549114227295,
+      "learning_rate": 3.6131386861313872e-06,
+      "loss": 2.0744,
+      "step": 212
+    },
+    {
+      "epoch": 0.23342465753424657,
+      "grad_norm": 2.6342105865478516,
+      "learning_rate": 3.6313868613138687e-06,
+      "loss": 2.0737,
+      "step": 213
+    },
+    {
+      "epoch": 0.2345205479452055,
+      "grad_norm": 3.5662875175476074,
+      "learning_rate": 3.6496350364963505e-06,
+      "loss": 2.1147,
+      "step": 214
+    },
+    {
+      "epoch": 0.2356164383561644,
+      "grad_norm": 2.5357561111450195,
+      "learning_rate": 3.667883211678833e-06,
+      "loss": 2.1268,
+      "step": 215
+    },
+    {
+      "epoch": 0.23671232876712328,
+      "grad_norm": 3.439711093902588,
+      "learning_rate": 3.6861313868613142e-06,
+      "loss": 2.1118,
+      "step": 216
+    },
+    {
+      "epoch": 0.2378082191780822,
+      "grad_norm": 2.9410204887390137,
+      "learning_rate": 3.704379562043796e-06,
+      "loss": 2.1087,
+      "step": 217
+    },
+    {
+      "epoch": 0.2389041095890411,
+      "grad_norm": 3.3182294368743896,
+      "learning_rate": 3.7226277372262775e-06,
+      "loss": 2.196,
+      "step": 218
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 3.3834011554718018,
+      "learning_rate": 3.7408759124087594e-06,
+      "loss": 2.1633,
+      "step": 219
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "grad_norm": 3.114952325820923,
+      "learning_rate": 3.7591240875912412e-06,
+      "loss": 2.1107,
+      "step": 220
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "eval_loss": 2.277580976486206,
+      "eval_runtime": 36.9846,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 220
+    },
+    {
+      "epoch": 0.2421917808219178,
+      "grad_norm": 3.1151392459869385,
+      "learning_rate": 3.777372262773723e-06,
+      "loss": 2.1471,
+      "step": 221
+    },
+    {
+      "epoch": 0.2432876712328767,
+      "grad_norm": 3.798450469970703,
+      "learning_rate": 3.7956204379562045e-06,
+      "loss": 2.1324,
+      "step": 222
+    },
+    {
+      "epoch": 0.24438356164383562,
+      "grad_norm": 2.6187098026275635,
+      "learning_rate": 3.8138686131386864e-06,
+      "loss": 2.1061,
+      "step": 223
+    },
+    {
+      "epoch": 0.24547945205479452,
+      "grad_norm": 4.014018535614014,
+      "learning_rate": 3.832116788321168e-06,
+      "loss": 2.166,
+      "step": 224
+    },
+    {
+      "epoch": 0.2465753424657534,
+      "grad_norm": 3.0043938159942627,
+      "learning_rate": 3.85036496350365e-06,
+      "loss": 2.1197,
+      "step": 225
+    },
+    {
+      "epoch": 0.24767123287671233,
+      "grad_norm": 3.507066249847412,
+      "learning_rate": 3.868613138686132e-06,
+      "loss": 2.0848,
+      "step": 226
+    },
+    {
+      "epoch": 0.24876712328767123,
+      "grad_norm": 3.5713281631469727,
+      "learning_rate": 3.886861313868613e-06,
+      "loss": 2.1239,
+      "step": 227
+    },
+    {
+      "epoch": 0.24986301369863015,
+      "grad_norm": 3.961082696914673,
+      "learning_rate": 3.905109489051096e-06,
+      "loss": 2.1406,
+      "step": 228
+    },
+    {
+      "epoch": 0.25095890410958904,
+      "grad_norm": 3.1514859199523926,
+      "learning_rate": 3.923357664233577e-06,
+      "loss": 2.1689,
+      "step": 229
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "grad_norm": 3.8486833572387695,
+      "learning_rate": 3.9416058394160585e-06,
+      "loss": 2.1321,
+      "step": 230
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "eval_loss": 2.2620601654052734,
+      "eval_runtime": 37.058,
+      "eval_samples_per_second": 26.985,
+      "eval_steps_per_second": 1.133,
+      "step": 230
+    },
+    {
+      "epoch": 0.25315068493150683,
+      "grad_norm": 2.5626907348632812,
+      "learning_rate": 3.95985401459854e-06,
+      "loss": 2.1777,
+      "step": 231
+    },
+    {
+      "epoch": 0.25424657534246575,
+      "grad_norm": 4.105278015136719,
+      "learning_rate": 3.978102189781022e-06,
+      "loss": 2.2099,
+      "step": 232
+    },
+    {
+      "epoch": 0.2553424657534247,
+      "grad_norm": 2.9818739891052246,
+      "learning_rate": 3.996350364963504e-06,
+      "loss": 2.0705,
+      "step": 233
+    },
+    {
+      "epoch": 0.25643835616438354,
+      "grad_norm": 2.87030291557312,
+      "learning_rate": 4.014598540145986e-06,
+      "loss": 2.1453,
+      "step": 234
+    },
+    {
+      "epoch": 0.25753424657534246,
+      "grad_norm": 2.4847915172576904,
+      "learning_rate": 4.032846715328467e-06,
+      "loss": 2.0461,
+      "step": 235
+    },
+    {
+      "epoch": 0.2586301369863014,
+      "grad_norm": 3.451420783996582,
+      "learning_rate": 4.05109489051095e-06,
+      "loss": 2.1133,
+      "step": 236
+    },
+    {
+      "epoch": 0.25972602739726025,
+      "grad_norm": 2.954723358154297,
+      "learning_rate": 4.069343065693431e-06,
+      "loss": 2.1316,
+      "step": 237
+    },
+    {
+      "epoch": 0.2608219178082192,
+      "grad_norm": 2.5295896530151367,
+      "learning_rate": 4.0875912408759126e-06,
+      "loss": 2.0901,
+      "step": 238
+    },
+    {
+      "epoch": 0.2619178082191781,
+      "grad_norm": 2.6359453201293945,
+      "learning_rate": 4.105839416058395e-06,
+      "loss": 2.117,
+      "step": 239
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "grad_norm": 2.5236105918884277,
+      "learning_rate": 4.124087591240876e-06,
+      "loss": 2.1958,
+      "step": 240
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "eval_loss": 2.250295400619507,
+      "eval_runtime": 37.0811,
+      "eval_samples_per_second": 26.968,
+      "eval_steps_per_second": 1.133,
+      "step": 240
+    },
+    {
+      "epoch": 0.2641095890410959,
+      "grad_norm": 2.581918954849243,
+      "learning_rate": 4.1423357664233586e-06,
+      "loss": 2.1054,
+      "step": 241
+    },
+    {
+      "epoch": 0.2652054794520548,
+      "grad_norm": 2.501332998275757,
+      "learning_rate": 4.16058394160584e-06,
+      "loss": 2.0878,
+      "step": 242
+    },
+    {
+      "epoch": 0.2663013698630137,
+      "grad_norm": 2.6477770805358887,
+      "learning_rate": 4.1788321167883214e-06,
+      "loss": 2.1656,
+      "step": 243
+    },
+    {
+      "epoch": 0.2673972602739726,
+      "grad_norm": 2.7252438068389893,
+      "learning_rate": 4.197080291970803e-06,
+      "loss": 2.0497,
+      "step": 244
+    },
+    {
+      "epoch": 0.2684931506849315,
+      "grad_norm": 2.6056301593780518,
+      "learning_rate": 4.215328467153285e-06,
+      "loss": 2.1411,
+      "step": 245
+    },
+    {
+      "epoch": 0.26958904109589044,
+      "grad_norm": 2.6581931114196777,
+      "learning_rate": 4.233576642335767e-06,
+      "loss": 2.0829,
+      "step": 246
+    },
+    {
+      "epoch": 0.2706849315068493,
+      "grad_norm": 2.6316752433776855,
+      "learning_rate": 4.251824817518249e-06,
+      "loss": 2.0882,
+      "step": 247
+    },
+    {
+      "epoch": 0.2717808219178082,
+      "grad_norm": 2.803823947906494,
+      "learning_rate": 4.27007299270073e-06,
+      "loss": 2.1203,
+      "step": 248
+    },
+    {
+      "epoch": 0.27287671232876715,
+      "grad_norm": 2.9551174640655518,
+      "learning_rate": 4.288321167883212e-06,
+      "loss": 2.0897,
+      "step": 249
+    },
+    {
+      "epoch": 0.273972602739726,
+      "grad_norm": 2.4847161769866943,
+      "learning_rate": 4.306569343065693e-06,
+      "loss": 2.0845,
+      "step": 250
+    },
+    {
+      "epoch": 0.273972602739726,
+      "eval_loss": 2.2491259574890137,
+      "eval_runtime": 37.1024,
+      "eval_samples_per_second": 26.952,
+      "eval_steps_per_second": 1.132,
+      "step": 250
+    },
+    {
+      "epoch": 0.27506849315068493,
+      "grad_norm": 2.6398351192474365,
+      "learning_rate": 4.3248175182481755e-06,
+      "loss": 2.0479,
+      "step": 251
+    },
+    {
+      "epoch": 0.27616438356164386,
+      "grad_norm": 3.0948562622070312,
+      "learning_rate": 4.343065693430658e-06,
+      "loss": 2.1184,
+      "step": 252
+    },
+    {
+      "epoch": 0.2772602739726027,
+      "grad_norm": 2.5378012657165527,
+      "learning_rate": 4.361313868613139e-06,
+      "loss": 2.1099,
+      "step": 253
+    },
+    {
+      "epoch": 0.27835616438356164,
+      "grad_norm": 3.0282225608825684,
+      "learning_rate": 4.379562043795621e-06,
+      "loss": 2.0982,
+      "step": 254
+    },
+    {
+      "epoch": 0.27945205479452057,
+      "grad_norm": 2.6205334663391113,
+      "learning_rate": 4.397810218978102e-06,
+      "loss": 2.1285,
+      "step": 255
+    },
+    {
+      "epoch": 0.28054794520547943,
+      "grad_norm": 3.2056095600128174,
+      "learning_rate": 4.416058394160584e-06,
+      "loss": 2.1221,
+      "step": 256
+    },
+    {
+      "epoch": 0.28164383561643835,
+      "grad_norm": 3.1370127201080322,
+      "learning_rate": 4.434306569343066e-06,
+      "loss": 2.1016,
+      "step": 257
+    },
+    {
+      "epoch": 0.2827397260273973,
+      "grad_norm": 2.494168519973755,
+      "learning_rate": 4.452554744525548e-06,
+      "loss": 2.0754,
+      "step": 258
+    },
+    {
+      "epoch": 0.28383561643835614,
+      "grad_norm": 2.7979183197021484,
+      "learning_rate": 4.4708029197080295e-06,
+      "loss": 2.1218,
+      "step": 259
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "grad_norm": 2.755682945251465,
+      "learning_rate": 4.489051094890512e-06,
+      "loss": 2.1136,
+      "step": 260
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "eval_loss": 2.2433817386627197,
+      "eval_runtime": 37.1121,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 260
+    },
+    {
+      "epoch": 0.286027397260274,
+      "grad_norm": 2.3641891479492188,
+      "learning_rate": 4.507299270072993e-06,
+      "loss": 2.1488,
+      "step": 261
+    },
+    {
+      "epoch": 0.28712328767123285,
+      "grad_norm": 2.741292953491211,
+      "learning_rate": 4.525547445255475e-06,
+      "loss": 2.1048,
+      "step": 262
+    },
+    {
+      "epoch": 0.2882191780821918,
+      "grad_norm": 2.811483383178711,
+      "learning_rate": 4.543795620437956e-06,
+      "loss": 2.0996,
+      "step": 263
+    },
+    {
+      "epoch": 0.2893150684931507,
+      "grad_norm": 2.6705479621887207,
+      "learning_rate": 4.562043795620438e-06,
+      "loss": 2.0652,
+      "step": 264
+    },
+    {
+      "epoch": 0.29041095890410956,
+      "grad_norm": 2.9383747577667236,
+      "learning_rate": 4.580291970802921e-06,
+      "loss": 2.1574,
+      "step": 265
+    },
+    {
+      "epoch": 0.2915068493150685,
+      "grad_norm": 2.591373920440674,
+      "learning_rate": 4.598540145985402e-06,
+      "loss": 2.1655,
+      "step": 266
+    },
+    {
+      "epoch": 0.2926027397260274,
+      "grad_norm": 3.16955828666687,
+      "learning_rate": 4.6167883211678835e-06,
+      "loss": 2.1285,
+      "step": 267
+    },
+    {
+      "epoch": 0.2936986301369863,
+      "grad_norm": 2.66941499710083,
+      "learning_rate": 4.635036496350365e-06,
+      "loss": 2.0777,
+      "step": 268
+    },
+    {
+      "epoch": 0.2947945205479452,
+      "grad_norm": 2.768139600753784,
+      "learning_rate": 4.653284671532847e-06,
+      "loss": 2.0748,
+      "step": 269
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "grad_norm": 2.5951924324035645,
+      "learning_rate": 4.671532846715329e-06,
+      "loss": 2.1289,
+      "step": 270
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "eval_loss": 2.2393550872802734,
+      "eval_runtime": 37.0993,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 270
+    },
+    {
+      "epoch": 0.29698630136986304,
+      "grad_norm": 2.5422160625457764,
+      "learning_rate": 4.689781021897811e-06,
+      "loss": 2.1342,
+      "step": 271
+    },
+    {
+      "epoch": 0.2980821917808219,
+      "grad_norm": 2.540555477142334,
+      "learning_rate": 4.708029197080292e-06,
+      "loss": 2.097,
+      "step": 272
+    },
+    {
+      "epoch": 0.2991780821917808,
+      "grad_norm": 2.995366334915161,
+      "learning_rate": 4.726277372262774e-06,
+      "loss": 2.1452,
+      "step": 273
+    },
+    {
+      "epoch": 0.30027397260273975,
+      "grad_norm": 2.2561020851135254,
+      "learning_rate": 4.744525547445255e-06,
+      "loss": 2.1069,
+      "step": 274
+    },
+    {
+      "epoch": 0.3013698630136986,
+      "grad_norm": 2.53476619720459,
+      "learning_rate": 4.7627737226277375e-06,
+      "loss": 2.0381,
+      "step": 275
+    },
+    {
+      "epoch": 0.30246575342465754,
+      "grad_norm": 2.3415749073028564,
+      "learning_rate": 4.78102189781022e-06,
+      "loss": 2.1027,
+      "step": 276
+    },
+    {
+      "epoch": 0.30356164383561646,
+      "grad_norm": 2.9756431579589844,
+      "learning_rate": 4.799270072992701e-06,
+      "loss": 2.1136,
+      "step": 277
+    },
+    {
+      "epoch": 0.3046575342465753,
+      "grad_norm": 2.606208324432373,
+      "learning_rate": 4.8175182481751835e-06,
+      "loss": 2.0797,
+      "step": 278
+    },
+    {
+      "epoch": 0.30575342465753425,
+      "grad_norm": 2.5089313983917236,
+      "learning_rate": 4.835766423357665e-06,
+      "loss": 2.1026,
+      "step": 279
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "grad_norm": 2.5003838539123535,
+      "learning_rate": 4.854014598540146e-06,
+      "loss": 2.0923,
+      "step": 280
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "eval_loss": 2.2351036071777344,
+      "eval_runtime": 37.111,
+      "eval_samples_per_second": 26.946,
+      "eval_steps_per_second": 1.132,
+      "step": 280
+    },
+    {
+      "epoch": 0.30794520547945203,
+      "grad_norm": 2.403602361679077,
+      "learning_rate": 4.872262773722628e-06,
+      "loss": 2.0388,
+      "step": 281
+    },
+    {
+      "epoch": 0.30904109589041096,
+      "grad_norm": 2.7775959968566895,
+      "learning_rate": 4.89051094890511e-06,
+      "loss": 2.0561,
+      "step": 282
+    },
+    {
+      "epoch": 0.3101369863013699,
+      "grad_norm": 2.7128653526306152,
+      "learning_rate": 4.9087591240875915e-06,
+      "loss": 2.042,
+      "step": 283
+    },
+    {
+      "epoch": 0.31123287671232874,
+      "grad_norm": 2.7029991149902344,
+      "learning_rate": 4.927007299270074e-06,
+      "loss": 2.0814,
+      "step": 284
+    },
+    {
+      "epoch": 0.31232876712328766,
+      "grad_norm": 2.3942933082580566,
+      "learning_rate": 4.945255474452555e-06,
+      "loss": 2.0906,
+      "step": 285
+    },
+    {
+      "epoch": 0.3134246575342466,
+      "grad_norm": 2.5086541175842285,
+      "learning_rate": 4.963503649635037e-06,
+      "loss": 2.0625,
+      "step": 286
+    },
+    {
+      "epoch": 0.31452054794520545,
+      "grad_norm": 2.6668760776519775,
+      "learning_rate": 4.981751824817518e-06,
+      "loss": 2.1538,
+      "step": 287
+    },
+    {
+      "epoch": 0.3156164383561644,
+      "grad_norm": 2.5066750049591064,
+      "learning_rate": 5e-06,
+      "loss": 2.1423,
+      "step": 288
+    },
+    {
+      "epoch": 0.3167123287671233,
+      "grad_norm": 2.84967041015625,
+      "learning_rate": 4.997969130787978e-06,
+      "loss": 2.0943,
+      "step": 289
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "grad_norm": 2.151946544647217,
+      "learning_rate": 4.995938261575955e-06,
+      "loss": 2.1098,
+      "step": 290
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "eval_loss": 2.223247528076172,
+      "eval_runtime": 37.1755,
+      "eval_samples_per_second": 26.899,
+      "eval_steps_per_second": 1.13,
+      "step": 290
+    },
+    {
+      "epoch": 0.3189041095890411,
+      "grad_norm": 2.887166976928711,
+      "learning_rate": 4.993907392363932e-06,
+      "loss": 2.0281,
+      "step": 291
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 2.5930211544036865,
+      "learning_rate": 4.9918765231519095e-06,
+      "loss": 2.1063,
+      "step": 292
+    },
+    {
+      "epoch": 0.32109589041095893,
+      "grad_norm": 2.5568413734436035,
+      "learning_rate": 4.989845653939887e-06,
+      "loss": 2.0313,
+      "step": 293
+    },
+    {
+      "epoch": 0.3221917808219178,
+      "grad_norm": 2.536095380783081,
+      "learning_rate": 4.987814784727864e-06,
+      "loss": 2.0749,
+      "step": 294
+    },
+    {
+      "epoch": 0.3232876712328767,
+      "grad_norm": 2.3385651111602783,
+      "learning_rate": 4.985783915515841e-06,
+      "loss": 2.0666,
+      "step": 295
+    },
+    {
+      "epoch": 0.32438356164383564,
+      "grad_norm": 2.504826068878174,
+      "learning_rate": 4.983753046303819e-06,
+      "loss": 2.0677,
+      "step": 296
+    },
+    {
+      "epoch": 0.3254794520547945,
+      "grad_norm": 2.5289409160614014,
+      "learning_rate": 4.981722177091796e-06,
+      "loss": 2.0828,
+      "step": 297
+    },
+    {
+      "epoch": 0.3265753424657534,
+      "grad_norm": 2.803795099258423,
+      "learning_rate": 4.979691307879773e-06,
+      "loss": 2.0733,
+      "step": 298
+    },
+    {
+      "epoch": 0.32767123287671235,
+      "grad_norm": 2.651273012161255,
+      "learning_rate": 4.9776604386677504e-06,
+      "loss": 2.0987,
+      "step": 299
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "grad_norm": 2.4715864658355713,
+      "learning_rate": 4.975629569455728e-06,
+      "loss": 2.0733,
+      "step": 300
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "eval_loss": 2.214858293533325,
+      "eval_runtime": 37.0559,
+      "eval_samples_per_second": 26.986,
+      "eval_steps_per_second": 1.133,
+      "step": 300
+    },
+    {
+      "epoch": 0.32986301369863014,
+      "grad_norm": 2.2573442459106445,
+      "learning_rate": 4.973598700243705e-06,
+      "loss": 2.1291,
+      "step": 301
+    },
+    {
+      "epoch": 0.33095890410958906,
+      "grad_norm": 2.645966053009033,
+      "learning_rate": 4.971567831031681e-06,
+      "loss": 2.046,
+      "step": 302
+    },
+    {
+      "epoch": 0.3320547945205479,
+      "grad_norm": 2.361392021179199,
+      "learning_rate": 4.9695369618196595e-06,
+      "loss": 2.0447,
+      "step": 303
+    },
+    {
+      "epoch": 0.33315068493150685,
+      "grad_norm": 2.272714376449585,
+      "learning_rate": 4.967506092607636e-06,
+      "loss": 1.9272,
+      "step": 304
+    },
+    {
+      "epoch": 0.33424657534246577,
+      "grad_norm": 2.5846943855285645,
+      "learning_rate": 4.965475223395614e-06,
+      "loss": 2.0199,
+      "step": 305
+    },
+    {
+      "epoch": 0.33534246575342463,
+      "grad_norm": 2.38985013961792,
+      "learning_rate": 4.9634443541835905e-06,
+      "loss": 2.0394,
+      "step": 306
+    },
+    {
+      "epoch": 0.33643835616438356,
+      "grad_norm": 2.3526251316070557,
+      "learning_rate": 4.961413484971569e-06,
+      "loss": 2.0403,
+      "step": 307
+    },
+    {
+      "epoch": 0.3375342465753425,
+      "grad_norm": 2.399531602859497,
+      "learning_rate": 4.959382615759545e-06,
+      "loss": 2.0607,
+      "step": 308
+    },
+    {
+      "epoch": 0.33863013698630134,
+      "grad_norm": 2.50803804397583,
+      "learning_rate": 4.957351746547523e-06,
+      "loss": 2.0688,
+      "step": 309
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "grad_norm": 2.278088092803955,
+      "learning_rate": 4.9553208773355e-06,
+      "loss": 2.0453,
+      "step": 310
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "eval_loss": 2.2185235023498535,
+      "eval_runtime": 36.9712,
+      "eval_samples_per_second": 27.048,
+      "eval_steps_per_second": 1.136,
+      "step": 310
+    },
+    {
+      "epoch": 0.3408219178082192,
+      "grad_norm": 2.555300712585449,
+      "learning_rate": 4.953290008123478e-06,
+      "loss": 2.1726,
+      "step": 311
+    },
+    {
+      "epoch": 0.3419178082191781,
+      "grad_norm": 2.614870548248291,
+      "learning_rate": 4.951259138911454e-06,
+      "loss": 2.0659,
+      "step": 312
+    },
+    {
+      "epoch": 0.343013698630137,
+      "grad_norm": 2.1932661533355713,
+      "learning_rate": 4.9492282696994315e-06,
+      "loss": 1.9981,
+      "step": 313
+    },
+    {
+      "epoch": 0.3441095890410959,
+      "grad_norm": 2.707451581954956,
+      "learning_rate": 4.947197400487409e-06,
+      "loss": 2.1321,
+      "step": 314
+    },
+    {
+      "epoch": 0.3452054794520548,
+      "grad_norm": 2.4317946434020996,
+      "learning_rate": 4.945166531275386e-06,
+      "loss": 2.0548,
+      "step": 315
+    },
+    {
+      "epoch": 0.3463013698630137,
+      "grad_norm": 2.639883279800415,
+      "learning_rate": 4.943135662063363e-06,
+      "loss": 2.0512,
+      "step": 316
+    },
+    {
+      "epoch": 0.3473972602739726,
+      "grad_norm": 3.1968822479248047,
+      "learning_rate": 4.9411047928513406e-06,
+      "loss": 2.0611,
+      "step": 317
+    },
+    {
+      "epoch": 0.34849315068493153,
+      "grad_norm": 2.4361510276794434,
+      "learning_rate": 4.939073923639318e-06,
+      "loss": 2.0974,
+      "step": 318
+    },
+    {
+      "epoch": 0.3495890410958904,
+      "grad_norm": 2.8278262615203857,
+      "learning_rate": 4.937043054427295e-06,
+      "loss": 2.0992,
+      "step": 319
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "grad_norm": 3.01247239112854,
+      "learning_rate": 4.935012185215272e-06,
+      "loss": 2.0372,
+      "step": 320
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "eval_loss": 2.206442356109619,
+      "eval_runtime": 36.8797,
+      "eval_samples_per_second": 27.115,
+      "eval_steps_per_second": 1.139,
+      "step": 320
+    },
+    {
+      "epoch": 0.35178082191780824,
+      "grad_norm": 2.3372156620025635,
+      "learning_rate": 4.93298131600325e-06,
+      "loss": 2.0211,
+      "step": 321
+    },
+    {
+      "epoch": 0.3528767123287671,
+      "grad_norm": 2.91377329826355,
+      "learning_rate": 4.930950446791227e-06,
+      "loss": 2.0879,
+      "step": 322
+    },
+    {
+      "epoch": 0.353972602739726,
+      "grad_norm": 2.3207666873931885,
+      "learning_rate": 4.928919577579204e-06,
+      "loss": 2.056,
+      "step": 323
+    },
+    {
+      "epoch": 0.35506849315068495,
+      "grad_norm": 2.3260514736175537,
+      "learning_rate": 4.9268887083671815e-06,
+      "loss": 2.0684,
+      "step": 324
+    },
+    {
+      "epoch": 0.3561643835616438,
+      "grad_norm": 2.3070554733276367,
+      "learning_rate": 4.924857839155159e-06,
+      "loss": 2.0494,
+      "step": 325
+    },
+    {
+      "epoch": 0.35726027397260274,
+      "grad_norm": 2.189871072769165,
+      "learning_rate": 4.922826969943136e-06,
+      "loss": 2.0531,
+      "step": 326
+    },
+    {
+      "epoch": 0.35835616438356166,
+      "grad_norm": 2.553520917892456,
+      "learning_rate": 4.920796100731113e-06,
+      "loss": 2.0749,
+      "step": 327
+    },
+    {
+      "epoch": 0.3594520547945205,
+      "grad_norm": 2.62296199798584,
+      "learning_rate": 4.918765231519091e-06,
+      "loss": 1.9983,
+      "step": 328
+    },
+    {
+      "epoch": 0.36054794520547945,
+      "grad_norm": 2.4612369537353516,
+      "learning_rate": 4.916734362307068e-06,
+      "loss": 2.0641,
+      "step": 329
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "grad_norm": 2.656157970428467,
+      "learning_rate": 4.914703493095045e-06,
+      "loss": 2.0148,
+      "step": 330
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "eval_loss": 2.2054214477539062,
+      "eval_runtime": 36.886,
+      "eval_samples_per_second": 27.111,
+      "eval_steps_per_second": 1.139,
+      "step": 330
+    },
+    {
+      "epoch": 0.36273972602739724,
+      "grad_norm": 2.2823736667633057,
+      "learning_rate": 4.912672623883022e-06,
+      "loss": 2.0221,
+      "step": 331
+    },
+    {
+      "epoch": 0.36383561643835616,
+      "grad_norm": 2.8403289318084717,
+      "learning_rate": 4.910641754671e-06,
+      "loss": 2.0149,
+      "step": 332
+    },
+    {
+      "epoch": 0.3649315068493151,
+      "grad_norm": 2.453887462615967,
+      "learning_rate": 4.908610885458977e-06,
+      "loss": 2.0411,
+      "step": 333
+    },
+    {
+      "epoch": 0.36602739726027395,
+      "grad_norm": 2.6094300746917725,
+      "learning_rate": 4.906580016246954e-06,
+      "loss": 2.0293,
+      "step": 334
+    },
+    {
+      "epoch": 0.36712328767123287,
+      "grad_norm": 2.4412384033203125,
+      "learning_rate": 4.9045491470349315e-06,
+      "loss": 2.0778,
+      "step": 335
+    },
+    {
+      "epoch": 0.3682191780821918,
+      "grad_norm": 2.4435057640075684,
+      "learning_rate": 4.902518277822909e-06,
+      "loss": 2.0597,
+      "step": 336
+    },
+    {
+      "epoch": 0.3693150684931507,
+      "grad_norm": 2.2390244007110596,
+      "learning_rate": 4.900487408610886e-06,
+      "loss": 2.0654,
+      "step": 337
+    },
+    {
+      "epoch": 0.3704109589041096,
+      "grad_norm": 2.4243953227996826,
+      "learning_rate": 4.898456539398863e-06,
+      "loss": 2.0696,
+      "step": 338
+    },
+    {
+      "epoch": 0.3715068493150685,
+      "grad_norm": 2.375720262527466,
+      "learning_rate": 4.89642567018684e-06,
+      "loss": 2.0866,
+      "step": 339
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "grad_norm": 2.6379668712615967,
+      "learning_rate": 4.894394800974818e-06,
+      "loss": 2.1552,
+      "step": 340
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "eval_loss": 2.196237802505493,
+      "eval_runtime": 36.8953,
+      "eval_samples_per_second": 27.104,
+      "eval_steps_per_second": 1.138,
+      "step": 340
+    },
+    {
+      "epoch": 0.3736986301369863,
+      "grad_norm": 2.417145252227783,
+      "learning_rate": 4.892363931762794e-06,
+      "loss": 2.0024,
+      "step": 341
+    },
+    {
+      "epoch": 0.3747945205479452,
+      "grad_norm": 2.234590768814087,
+      "learning_rate": 4.8903330625507725e-06,
+      "loss": 2.0837,
+      "step": 342
+    },
+    {
+      "epoch": 0.37589041095890413,
+      "grad_norm": 2.404266595840454,
+      "learning_rate": 4.888302193338749e-06,
+      "loss": 2.018,
+      "step": 343
+    },
+    {
+      "epoch": 0.376986301369863,
+      "grad_norm": 2.364802837371826,
+      "learning_rate": 4.886271324126727e-06,
+      "loss": 1.994,
+      "step": 344
+    },
+    {
+      "epoch": 0.3780821917808219,
+      "grad_norm": 2.2262251377105713,
+      "learning_rate": 4.8842404549147034e-06,
+      "loss": 2.0929,
+      "step": 345
+    },
+    {
+      "epoch": 0.37917808219178084,
+      "grad_norm": 2.4854862689971924,
+      "learning_rate": 4.8822095857026816e-06,
+      "loss": 2.0801,
+      "step": 346
+    },
+    {
+      "epoch": 0.3802739726027397,
+      "grad_norm": 2.3861470222473145,
+      "learning_rate": 4.880178716490658e-06,
+      "loss": 2.0496,
+      "step": 347
+    },
+    {
+      "epoch": 0.38136986301369863,
+      "grad_norm": 2.372326612472534,
+      "learning_rate": 4.878147847278636e-06,
+      "loss": 1.9939,
+      "step": 348
+    },
+    {
+      "epoch": 0.38246575342465755,
+      "grad_norm": 2.3593149185180664,
+      "learning_rate": 4.8761169780666125e-06,
+      "loss": 2.0108,
+      "step": 349
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "grad_norm": 2.1802141666412354,
+      "learning_rate": 4.874086108854591e-06,
+      "loss": 1.983,
+      "step": 350
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "eval_loss": 2.1947567462921143,
+      "eval_runtime": 36.8798,
+      "eval_samples_per_second": 27.115,
+      "eval_steps_per_second": 1.139,
+      "step": 350
+    },
+    {
+      "epoch": 0.38465753424657534,
+      "grad_norm": 2.47067928314209,
+      "learning_rate": 4.872055239642567e-06,
+      "loss": 2.0002,
+      "step": 351
+    },
+    {
+      "epoch": 0.38575342465753426,
+      "grad_norm": 2.359609365463257,
+      "learning_rate": 4.870024370430544e-06,
+      "loss": 1.9799,
+      "step": 352
+    },
+    {
+      "epoch": 0.3868493150684931,
+      "grad_norm": 2.3711702823638916,
+      "learning_rate": 4.867993501218522e-06,
+      "loss": 2.0341,
+      "step": 353
+    },
+    {
+      "epoch": 0.38794520547945205,
+      "grad_norm": 2.246647834777832,
+      "learning_rate": 4.865962632006499e-06,
+      "loss": 2.0687,
+      "step": 354
+    },
+    {
+      "epoch": 0.38904109589041097,
+      "grad_norm": 2.3031938076019287,
+      "learning_rate": 4.863931762794476e-06,
+      "loss": 1.9994,
+      "step": 355
+    },
+    {
+      "epoch": 0.39013698630136984,
+      "grad_norm": 2.492159605026245,
+      "learning_rate": 4.8619008935824535e-06,
+      "loss": 2.0195,
+      "step": 356
+    },
+    {
+      "epoch": 0.39123287671232876,
+      "grad_norm": 2.2878570556640625,
+      "learning_rate": 4.859870024370431e-06,
+      "loss": 2.0589,
+      "step": 357
+    },
+    {
+      "epoch": 0.3923287671232877,
+      "grad_norm": 2.444096088409424,
+      "learning_rate": 4.857839155158408e-06,
+      "loss": 2.0796,
+      "step": 358
+    },
+    {
+      "epoch": 0.3934246575342466,
+      "grad_norm": 2.463857650756836,
+      "learning_rate": 4.855808285946385e-06,
+      "loss": 2.0824,
+      "step": 359
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "grad_norm": 2.0893168449401855,
+      "learning_rate": 4.8537774167343626e-06,
+      "loss": 2.0651,
+      "step": 360
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "eval_loss": 2.188523054122925,
+      "eval_runtime": 36.8679,
+      "eval_samples_per_second": 27.124,
+      "eval_steps_per_second": 1.139,
+      "step": 360
+    },
+    {
+      "epoch": 0.3956164383561644,
+      "grad_norm": 2.602850914001465,
+      "learning_rate": 4.85174654752234e-06,
+      "loss": 2.0507,
+      "step": 361
+    },
+    {
+      "epoch": 0.3967123287671233,
+      "grad_norm": 2.4185140132904053,
+      "learning_rate": 4.849715678310317e-06,
+      "loss": 2.0386,
+      "step": 362
+    },
+    {
+      "epoch": 0.3978082191780822,
+      "grad_norm": 2.1492245197296143,
+      "learning_rate": 4.847684809098294e-06,
+      "loss": 2.0376,
+      "step": 363
+    },
+    {
+      "epoch": 0.3989041095890411,
+      "grad_norm": 2.5710153579711914,
+      "learning_rate": 4.845653939886272e-06,
+      "loss": 2.0383,
+      "step": 364
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 2.4986608028411865,
+      "learning_rate": 4.843623070674249e-06,
+      "loss": 2.0464,
+      "step": 365
+    },
+    {
+      "epoch": 0.4010958904109589,
+      "grad_norm": 2.260388135910034,
+      "learning_rate": 4.841592201462226e-06,
+      "loss": 2.0217,
+      "step": 366
+    },
+    {
+      "epoch": 0.4021917808219178,
+      "grad_norm": 2.6500394344329834,
+      "learning_rate": 4.8395613322502035e-06,
+      "loss": 2.0777,
+      "step": 367
+    },
+    {
+      "epoch": 0.40328767123287673,
+      "grad_norm": 2.159274101257324,
+      "learning_rate": 4.837530463038181e-06,
+      "loss": 1.9747,
+      "step": 368
+    },
+    {
+      "epoch": 0.4043835616438356,
+      "grad_norm": 2.4623043537139893,
+      "learning_rate": 4.835499593826158e-06,
+      "loss": 2.039,
+      "step": 369
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "grad_norm": 2.2265007495880127,
+      "learning_rate": 4.833468724614135e-06,
+      "loss": 2.0618,
+      "step": 370
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "eval_loss": 2.184051275253296,
+      "eval_runtime": 36.9284,
+      "eval_samples_per_second": 27.079,
+      "eval_steps_per_second": 1.137,
+      "step": 370
+    },
+    {
+      "epoch": 0.40657534246575344,
+      "grad_norm": 2.300276517868042,
+      "learning_rate": 4.831437855402113e-06,
+      "loss": 2.0382,
+      "step": 371
+    },
+    {
+      "epoch": 0.4076712328767123,
+      "grad_norm": 2.341156482696533,
+      "learning_rate": 4.82940698619009e-06,
+      "loss": 1.9361,
+      "step": 372
+    },
+    {
+      "epoch": 0.40876712328767123,
+      "grad_norm": 2.4196999073028564,
+      "learning_rate": 4.827376116978067e-06,
+      "loss": 2.0087,
+      "step": 373
+    },
+    {
+      "epoch": 0.40986301369863015,
+      "grad_norm": 2.4650514125823975,
+      "learning_rate": 4.8253452477660444e-06,
+      "loss": 2.0687,
+      "step": 374
+    },
+    {
+      "epoch": 0.410958904109589,
+      "grad_norm": 2.393739700317383,
+      "learning_rate": 4.823314378554022e-06,
+      "loss": 2.0421,
+      "step": 375
+    },
+    {
+      "epoch": 0.41205479452054794,
+      "grad_norm": 2.2436327934265137,
+      "learning_rate": 4.821283509341998e-06,
+      "loss": 2.0358,
+      "step": 376
+    },
+    {
+      "epoch": 0.41315068493150686,
+      "grad_norm": 2.1801650524139404,
+      "learning_rate": 4.819252640129976e-06,
+      "loss": 2.0081,
+      "step": 377
+    },
+    {
+      "epoch": 0.41424657534246573,
+      "grad_norm": 2.300882339477539,
+      "learning_rate": 4.817221770917953e-06,
+      "loss": 2.0321,
+      "step": 378
+    },
+    {
+      "epoch": 0.41534246575342465,
+      "grad_norm": 2.2136058807373047,
+      "learning_rate": 4.815190901705931e-06,
+      "loss": 1.9999,
+      "step": 379
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "grad_norm": 2.168034076690674,
+      "learning_rate": 4.813160032493907e-06,
+      "loss": 2.0336,
+      "step": 380
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "eval_loss": 2.1770179271698,
+      "eval_runtime": 37.0015,
+      "eval_samples_per_second": 27.026,
+      "eval_steps_per_second": 1.135,
+      "step": 380
+    },
+    {
+      "epoch": 0.41753424657534244,
+      "grad_norm": 2.1604163646698,
+      "learning_rate": 4.811129163281885e-06,
+      "loss": 2.0212,
+      "step": 381
+    },
+    {
+      "epoch": 0.41863013698630136,
+      "grad_norm": 2.23923659324646,
+      "learning_rate": 4.809098294069862e-06,
+      "loss": 2.0495,
+      "step": 382
+    },
+    {
+      "epoch": 0.4197260273972603,
+      "grad_norm": 2.149949312210083,
+      "learning_rate": 4.80706742485784e-06,
+      "loss": 2.041,
+      "step": 383
+    },
+    {
+      "epoch": 0.4208219178082192,
+      "grad_norm": 2.49120831489563,
+      "learning_rate": 4.805036555645816e-06,
+      "loss": 2.0931,
+      "step": 384
+    },
+    {
+      "epoch": 0.42191780821917807,
+      "grad_norm": 2.2123563289642334,
+      "learning_rate": 4.8030056864337945e-06,
+      "loss": 2.0087,
+      "step": 385
+    },
+    {
+      "epoch": 0.423013698630137,
+      "grad_norm": 2.4235594272613525,
+      "learning_rate": 4.800974817221771e-06,
+      "loss": 2.0656,
+      "step": 386
+    },
+    {
+      "epoch": 0.4241095890410959,
+      "grad_norm": 2.158252239227295,
+      "learning_rate": 4.798943948009749e-06,
+      "loss": 2.0181,
+      "step": 387
+    },
+    {
+      "epoch": 0.4252054794520548,
+      "grad_norm": 2.2412970066070557,
+      "learning_rate": 4.7969130787977254e-06,
+      "loss": 2.0323,
+      "step": 388
+    },
+    {
+      "epoch": 0.4263013698630137,
+      "grad_norm": 2.2520956993103027,
+      "learning_rate": 4.794882209585703e-06,
+      "loss": 2.026,
+      "step": 389
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "grad_norm": 2.168954849243164,
+      "learning_rate": 4.79285134037368e-06,
+      "loss": 2.0459,
+      "step": 390
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "eval_loss": 2.1705472469329834,
+      "eval_runtime": 37.0913,
+      "eval_samples_per_second": 26.96,
+      "eval_steps_per_second": 1.132,
+      "step": 390
+    },
+    {
+      "epoch": 0.4284931506849315,
+      "grad_norm": 2.1397407054901123,
+      "learning_rate": 4.790820471161657e-06,
+      "loss": 1.9995,
+      "step": 391
+    },
+    {
+      "epoch": 0.4295890410958904,
+      "grad_norm": 2.1747751235961914,
+      "learning_rate": 4.7887896019496345e-06,
+      "loss": 2.0088,
+      "step": 392
+    },
+    {
+      "epoch": 0.43068493150684933,
+      "grad_norm": 2.2860991954803467,
+      "learning_rate": 4.786758732737612e-06,
+      "loss": 2.0134,
+      "step": 393
+    },
+    {
+      "epoch": 0.4317808219178082,
+      "grad_norm": 2.2133519649505615,
+      "learning_rate": 4.784727863525589e-06,
+      "loss": 2.0586,
+      "step": 394
+    },
+    {
+      "epoch": 0.4328767123287671,
+      "grad_norm": 2.2147040367126465,
+      "learning_rate": 4.782696994313566e-06,
+      "loss": 1.9748,
+      "step": 395
+    },
+    {
+      "epoch": 0.43397260273972604,
+      "grad_norm": 2.343418836593628,
+      "learning_rate": 4.780666125101544e-06,
+      "loss": 2.0223,
+      "step": 396
+    },
+    {
+      "epoch": 0.4350684931506849,
+      "grad_norm": 2.198935031890869,
+      "learning_rate": 4.778635255889521e-06,
+      "loss": 1.9627,
+      "step": 397
+    },
+    {
+      "epoch": 0.43616438356164383,
+      "grad_norm": 2.1314704418182373,
+      "learning_rate": 4.776604386677498e-06,
+      "loss": 1.9938,
+      "step": 398
+    },
+    {
+      "epoch": 0.43726027397260275,
+      "grad_norm": 2.3005123138427734,
+      "learning_rate": 4.7745735174654755e-06,
+      "loss": 2.0029,
+      "step": 399
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "grad_norm": 2.2826664447784424,
+      "learning_rate": 4.772542648253453e-06,
+      "loss": 2.0585,
+      "step": 400
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "eval_loss": 2.1691415309906006,
+      "eval_runtime": 37.0852,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 400
+    },
+    {
+      "epoch": 0.43945205479452054,
+      "grad_norm": 2.0547878742218018,
+      "learning_rate": 4.77051177904143e-06,
+      "loss": 2.0582,
+      "step": 401
+    },
+    {
+      "epoch": 0.44054794520547946,
+      "grad_norm": 2.200042486190796,
+      "learning_rate": 4.768480909829407e-06,
+      "loss": 1.9936,
+      "step": 402
+    },
+    {
+      "epoch": 0.44164383561643833,
+      "grad_norm": 2.354637861251831,
+      "learning_rate": 4.766450040617385e-06,
+      "loss": 2.0327,
+      "step": 403
+    },
+    {
+      "epoch": 0.44273972602739725,
+      "grad_norm": 2.2332348823547363,
+      "learning_rate": 4.764419171405362e-06,
+      "loss": 2.0052,
+      "step": 404
+    },
+    {
+      "epoch": 0.4438356164383562,
+      "grad_norm": 2.288060426712036,
+      "learning_rate": 4.762388302193339e-06,
+      "loss": 2.0472,
+      "step": 405
+    },
+    {
+      "epoch": 0.4449315068493151,
+      "grad_norm": 2.1368887424468994,
+      "learning_rate": 4.760357432981316e-06,
+      "loss": 2.0305,
+      "step": 406
+    },
+    {
+      "epoch": 0.44602739726027396,
+      "grad_norm": 2.362630844116211,
+      "learning_rate": 4.758326563769294e-06,
+      "loss": 2.002,
+      "step": 407
+    },
+    {
+      "epoch": 0.4471232876712329,
+      "grad_norm": 2.283419370651245,
+      "learning_rate": 4.756295694557271e-06,
+      "loss": 2.0271,
+      "step": 408
+    },
+    {
+      "epoch": 0.4482191780821918,
+      "grad_norm": 2.0400819778442383,
+      "learning_rate": 4.754264825345248e-06,
+      "loss": 2.0256,
+      "step": 409
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "grad_norm": 2.588416814804077,
+      "learning_rate": 4.7522339561332255e-06,
+      "loss": 1.9756,
+      "step": 410
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "eval_loss": 2.17041015625,
+      "eval_runtime": 36.958,
+      "eval_samples_per_second": 27.058,
+      "eval_steps_per_second": 1.136,
+      "step": 410
+    },
+    {
+      "epoch": 0.4504109589041096,
+      "grad_norm": 2.2397494316101074,
+      "learning_rate": 4.750203086921203e-06,
+      "loss": 1.9876,
+      "step": 411
+    },
+    {
+      "epoch": 0.4515068493150685,
+      "grad_norm": 2.2085509300231934,
+      "learning_rate": 4.74817221770918e-06,
+      "loss": 1.9464,
+      "step": 412
+    },
+    {
+      "epoch": 0.4526027397260274,
+      "grad_norm": 2.140974760055542,
+      "learning_rate": 4.7461413484971565e-06,
+      "loss": 2.0179,
+      "step": 413
+    },
+    {
+      "epoch": 0.4536986301369863,
+      "grad_norm": 2.124366044998169,
+      "learning_rate": 4.744110479285135e-06,
+      "loss": 2.0732,
+      "step": 414
+    },
+    {
+      "epoch": 0.4547945205479452,
+      "grad_norm": 2.2463996410369873,
+      "learning_rate": 4.742079610073111e-06,
+      "loss": 1.9616,
+      "step": 415
+    },
+    {
+      "epoch": 0.4558904109589041,
+      "grad_norm": 2.1249961853027344,
+      "learning_rate": 4.740048740861089e-06,
+      "loss": 1.9889,
+      "step": 416
+    },
+    {
+      "epoch": 0.456986301369863,
+      "grad_norm": 2.25842022895813,
+      "learning_rate": 4.738017871649066e-06,
+      "loss": 2.0033,
+      "step": 417
+    },
+    {
+      "epoch": 0.45808219178082193,
+      "grad_norm": 2.271935224533081,
+      "learning_rate": 4.735987002437044e-06,
+      "loss": 2.0148,
+      "step": 418
+    },
+    {
+      "epoch": 0.4591780821917808,
+      "grad_norm": 2.4797163009643555,
+      "learning_rate": 4.73395613322502e-06,
+      "loss": 2.0736,
+      "step": 419
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "grad_norm": 2.1901395320892334,
+      "learning_rate": 4.731925264012998e-06,
+      "loss": 2.0763,
+      "step": 420
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "eval_loss": 2.161660671234131,
+      "eval_runtime": 36.9099,
+      "eval_samples_per_second": 27.093,
+      "eval_steps_per_second": 1.138,
+      "step": 420
+    },
+    {
+      "epoch": 0.46136986301369864,
+      "grad_norm": 2.083536386489868,
+      "learning_rate": 4.729894394800975e-06,
+      "loss": 2.0299,
+      "step": 421
+    },
+    {
+      "epoch": 0.4624657534246575,
+      "grad_norm": 2.1118173599243164,
+      "learning_rate": 4.727863525588953e-06,
+      "loss": 1.9742,
+      "step": 422
+    },
+    {
+      "epoch": 0.46356164383561643,
+      "grad_norm": 2.195802927017212,
+      "learning_rate": 4.725832656376929e-06,
+      "loss": 1.9318,
+      "step": 423
+    },
+    {
+      "epoch": 0.46465753424657535,
+      "grad_norm": 2.318098545074463,
+      "learning_rate": 4.723801787164907e-06,
+      "loss": 1.9875,
+      "step": 424
+    },
+    {
+      "epoch": 0.4657534246575342,
+      "grad_norm": 2.1698496341705322,
+      "learning_rate": 4.721770917952885e-06,
+      "loss": 2.031,
+      "step": 425
+    },
+    {
+      "epoch": 0.46684931506849314,
+      "grad_norm": 2.283935785293579,
+      "learning_rate": 4.719740048740861e-06,
+      "loss": 1.9766,
+      "step": 426
+    },
+    {
+      "epoch": 0.46794520547945206,
+      "grad_norm": 2.4180757999420166,
+      "learning_rate": 4.717709179528839e-06,
+      "loss": 1.9976,
+      "step": 427
+    },
+    {
+      "epoch": 0.469041095890411,
+      "grad_norm": 2.2689030170440674,
+      "learning_rate": 4.715678310316816e-06,
+      "loss": 2.03,
+      "step": 428
+    },
+    {
+      "epoch": 0.47013698630136985,
+      "grad_norm": 2.423708200454712,
+      "learning_rate": 4.713647441104794e-06,
+      "loss": 2.0202,
+      "step": 429
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "grad_norm": 2.184705972671509,
+      "learning_rate": 4.71161657189277e-06,
+      "loss": 1.9655,
+      "step": 430
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "eval_loss": 2.1548638343811035,
+      "eval_runtime": 36.9426,
+      "eval_samples_per_second": 27.069,
+      "eval_steps_per_second": 1.137,
+      "step": 430
+    },
+    {
+      "epoch": 0.4723287671232877,
+      "grad_norm": 2.295851945877075,
+      "learning_rate": 4.709585702680748e-06,
+      "loss": 1.9561,
+      "step": 431
+    },
+    {
+      "epoch": 0.47342465753424656,
+      "grad_norm": 2.1762046813964844,
+      "learning_rate": 4.707554833468725e-06,
+      "loss": 2.0341,
+      "step": 432
+    },
+    {
+      "epoch": 0.4745205479452055,
+      "grad_norm": 2.4684343338012695,
+      "learning_rate": 4.705523964256703e-06,
+      "loss": 2.0898,
+      "step": 433
+    },
+    {
+      "epoch": 0.4756164383561644,
+      "grad_norm": 2.293318271636963,
+      "learning_rate": 4.703493095044679e-06,
+      "loss": 1.9794,
+      "step": 434
+    },
+    {
+      "epoch": 0.4767123287671233,
+      "grad_norm": 2.1884210109710693,
+      "learning_rate": 4.701462225832657e-06,
+      "loss": 1.9979,
+      "step": 435
+    },
+    {
+      "epoch": 0.4778082191780822,
+      "grad_norm": 2.412504196166992,
+      "learning_rate": 4.699431356620634e-06,
+      "loss": 1.9237,
+      "step": 436
+    },
+    {
+      "epoch": 0.4789041095890411,
+      "grad_norm": 2.192824602127075,
+      "learning_rate": 4.697400487408612e-06,
+      "loss": 2.0756,
+      "step": 437
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 2.368100643157959,
+      "learning_rate": 4.695369618196588e-06,
+      "loss": 2.028,
+      "step": 438
+    },
+    {
+      "epoch": 0.4810958904109589,
+      "grad_norm": 2.218454122543335,
+      "learning_rate": 4.693338748984566e-06,
+      "loss": 2.0049,
+      "step": 439
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "grad_norm": 2.356168031692505,
+      "learning_rate": 4.691307879772543e-06,
+      "loss": 1.9958,
+      "step": 440
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "eval_loss": 2.1540727615356445,
+      "eval_runtime": 36.9557,
+      "eval_samples_per_second": 27.059,
+      "eval_steps_per_second": 1.136,
+      "step": 440
+    },
+    {
+      "epoch": 0.4832876712328767,
+      "grad_norm": 2.09956693649292,
+      "learning_rate": 4.68927701056052e-06,
+      "loss": 1.9388,
+      "step": 441
+    },
+    {
+      "epoch": 0.4843835616438356,
+      "grad_norm": 2.5746614933013916,
+      "learning_rate": 4.6872461413484975e-06,
+      "loss": 2.0425,
+      "step": 442
+    },
+    {
+      "epoch": 0.48547945205479454,
+      "grad_norm": 2.2064783573150635,
+      "learning_rate": 4.685215272136475e-06,
+      "loss": 2.0082,
+      "step": 443
+    },
+    {
+      "epoch": 0.4865753424657534,
+      "grad_norm": 2.159029006958008,
+      "learning_rate": 4.683184402924452e-06,
+      "loss": 2.0055,
+      "step": 444
+    },
+    {
+      "epoch": 0.4876712328767123,
+      "grad_norm": 2.5222008228302,
+      "learning_rate": 4.681153533712429e-06,
+      "loss": 2.0007,
+      "step": 445
+    },
+    {
+      "epoch": 0.48876712328767125,
+      "grad_norm": 2.163487672805786,
+      "learning_rate": 4.679122664500407e-06,
+      "loss": 2.0719,
+      "step": 446
+    },
+    {
+      "epoch": 0.4898630136986301,
+      "grad_norm": 2.3383803367614746,
+      "learning_rate": 4.677091795288384e-06,
+      "loss": 1.9623,
+      "step": 447
+    },
+    {
+      "epoch": 0.49095890410958903,
+      "grad_norm": 2.1056675910949707,
+      "learning_rate": 4.675060926076361e-06,
+      "loss": 1.9443,
+      "step": 448
+    },
+    {
+      "epoch": 0.49205479452054796,
+      "grad_norm": 2.145270824432373,
+      "learning_rate": 4.673030056864338e-06,
+      "loss": 2.0469,
+      "step": 449
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "grad_norm": 2.578706741333008,
+      "learning_rate": 4.670999187652316e-06,
+      "loss": 1.9569,
+      "step": 450
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "eval_loss": 2.1527743339538574,
+      "eval_runtime": 36.8961,
+      "eval_samples_per_second": 27.103,
+      "eval_steps_per_second": 1.138,
+      "step": 450
+    },
+    {
+      "epoch": 0.49424657534246574,
+      "grad_norm": 2.1328542232513428,
+      "learning_rate": 4.668968318440293e-06,
+      "loss": 1.9559,
+      "step": 451
+    },
+    {
+      "epoch": 0.49534246575342467,
+      "grad_norm": 2.148820638656616,
+      "learning_rate": 4.66693744922827e-06,
+      "loss": 1.9545,
+      "step": 452
+    },
+    {
+      "epoch": 0.4964383561643836,
+      "grad_norm": 2.250340461730957,
+      "learning_rate": 4.6649065800162475e-06,
+      "loss": 2.0358,
+      "step": 453
+    },
+    {
+      "epoch": 0.49753424657534245,
+      "grad_norm": 2.452397584915161,
+      "learning_rate": 4.662875710804225e-06,
+      "loss": 1.9294,
+      "step": 454
+    },
+    {
+      "epoch": 0.4986301369863014,
+      "grad_norm": 2.140615701675415,
+      "learning_rate": 4.660844841592202e-06,
+      "loss": 2.0084,
+      "step": 455
+    },
+    {
+      "epoch": 0.4997260273972603,
+      "grad_norm": 2.176806926727295,
+      "learning_rate": 4.658813972380179e-06,
+      "loss": 1.9174,
+      "step": 456
+    },
+    {
+      "epoch": 0.5008219178082192,
+      "grad_norm": 2.349031925201416,
+      "learning_rate": 4.656783103168157e-06,
+      "loss": 2.0377,
+      "step": 457
+    },
+    {
+      "epoch": 0.5019178082191781,
+      "grad_norm": 2.436889171600342,
+      "learning_rate": 4.654752233956134e-06,
+      "loss": 1.9133,
+      "step": 458
+    },
+    {
+      "epoch": 0.503013698630137,
+      "grad_norm": 2.116610288619995,
+      "learning_rate": 4.652721364744111e-06,
+      "loss": 1.9993,
+      "step": 459
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "grad_norm": 2.186915636062622,
+      "learning_rate": 4.6506904955320885e-06,
+      "loss": 1.9744,
+      "step": 460
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "eval_loss": 2.1580278873443604,
+      "eval_runtime": 36.9097,
+      "eval_samples_per_second": 27.093,
+      "eval_steps_per_second": 1.138,
+      "step": 460
+    },
+    {
+      "epoch": 0.5052054794520547,
+      "grad_norm": 2.2009212970733643,
+      "learning_rate": 4.648659626320066e-06,
+      "loss": 2.0173,
+      "step": 461
+    },
+    {
+      "epoch": 0.5063013698630137,
+      "grad_norm": 2.117286443710327,
+      "learning_rate": 4.646628757108043e-06,
+      "loss": 2.0238,
+      "step": 462
+    },
+    {
+      "epoch": 0.5073972602739726,
+      "grad_norm": 2.238790273666382,
+      "learning_rate": 4.6445978878960194e-06,
+      "loss": 1.9885,
+      "step": 463
+    },
+    {
+      "epoch": 0.5084931506849315,
+      "grad_norm": 2.3493096828460693,
+      "learning_rate": 4.6425670186839976e-06,
+      "loss": 1.9593,
+      "step": 464
+    },
+    {
+      "epoch": 0.5095890410958904,
+      "grad_norm": 2.3276610374450684,
+      "learning_rate": 4.640536149471974e-06,
+      "loss": 1.9993,
+      "step": 465
+    },
+    {
+      "epoch": 0.5106849315068493,
+      "grad_norm": 2.3671715259552,
+      "learning_rate": 4.638505280259952e-06,
+      "loss": 2.0256,
+      "step": 466
+    },
+    {
+      "epoch": 0.5117808219178083,
+      "grad_norm": 2.251803159713745,
+      "learning_rate": 4.6364744110479285e-06,
+      "loss": 2.0436,
+      "step": 467
+    },
+    {
+      "epoch": 0.5128767123287671,
+      "grad_norm": 2.3023312091827393,
+      "learning_rate": 4.634443541835907e-06,
+      "loss": 1.9968,
+      "step": 468
+    },
+    {
+      "epoch": 0.513972602739726,
+      "grad_norm": 2.3660333156585693,
+      "learning_rate": 4.632412672623883e-06,
+      "loss": 2.0492,
+      "step": 469
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "grad_norm": 2.262179136276245,
+      "learning_rate": 4.630381803411861e-06,
+      "loss": 1.9698,
+      "step": 470
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "eval_loss": 2.1516528129577637,
+      "eval_runtime": 36.7956,
+      "eval_samples_per_second": 27.177,
+      "eval_steps_per_second": 1.141,
+      "step": 470
+    },
+    {
+      "epoch": 0.5161643835616438,
+      "grad_norm": 2.450770616531372,
+      "learning_rate": 4.628350934199838e-06,
+      "loss": 1.8799,
+      "step": 471
+    },
+    {
+      "epoch": 0.5172602739726028,
+      "grad_norm": 2.187458038330078,
+      "learning_rate": 4.626320064987816e-06,
+      "loss": 1.9829,
+      "step": 472
+    },
+    {
+      "epoch": 0.5183561643835617,
+      "grad_norm": 2.1726722717285156,
+      "learning_rate": 4.624289195775792e-06,
+      "loss": 2.012,
+      "step": 473
+    },
+    {
+      "epoch": 0.5194520547945205,
+      "grad_norm": 2.4177136421203613,
+      "learning_rate": 4.62225832656377e-06,
+      "loss": 1.9796,
+      "step": 474
+    },
+    {
+      "epoch": 0.5205479452054794,
+      "grad_norm": 2.2353951930999756,
+      "learning_rate": 4.620227457351747e-06,
+      "loss": 1.99,
+      "step": 475
+    },
+    {
+      "epoch": 0.5216438356164383,
+      "grad_norm": 2.369591236114502,
+      "learning_rate": 4.618196588139724e-06,
+      "loss": 1.9847,
+      "step": 476
+    },
+    {
+      "epoch": 0.5227397260273973,
+      "grad_norm": 2.546703577041626,
+      "learning_rate": 4.616165718927701e-06,
+      "loss": 2.0047,
+      "step": 477
+    },
+    {
+      "epoch": 0.5238356164383562,
+      "grad_norm": 2.3509185314178467,
+      "learning_rate": 4.6141348497156786e-06,
+      "loss": 1.9589,
+      "step": 478
+    },
+    {
+      "epoch": 0.5249315068493151,
+      "grad_norm": 2.3063488006591797,
+      "learning_rate": 4.612103980503656e-06,
+      "loss": 2.045,
+      "step": 479
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "grad_norm": 2.285572052001953,
+      "learning_rate": 4.610073111291633e-06,
+      "loss": 2.0218,
+      "step": 480
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "eval_loss": 2.1472301483154297,
+      "eval_runtime": 36.7689,
+      "eval_samples_per_second": 27.197,
+      "eval_steps_per_second": 1.142,
+      "step": 480
+    },
+    {
+      "epoch": 0.5271232876712328,
+      "grad_norm": 2.22987961769104,
+      "learning_rate": 4.60804224207961e-06,
+      "loss": 1.9955,
+      "step": 481
+    },
+    {
+      "epoch": 0.5282191780821918,
+      "grad_norm": 2.200918197631836,
+      "learning_rate": 4.606011372867588e-06,
+      "loss": 1.9768,
+      "step": 482
+    },
+    {
+      "epoch": 0.5293150684931507,
+      "grad_norm": 2.2441930770874023,
+      "learning_rate": 4.603980503655565e-06,
+      "loss": 2.0088,
+      "step": 483
+    },
+    {
+      "epoch": 0.5304109589041096,
+      "grad_norm": 2.380383014678955,
+      "learning_rate": 4.601949634443542e-06,
+      "loss": 1.9853,
+      "step": 484
+    },
+    {
+      "epoch": 0.5315068493150685,
+      "grad_norm": 2.317995071411133,
+      "learning_rate": 4.5999187652315195e-06,
+      "loss": 1.9183,
+      "step": 485
+    },
+    {
+      "epoch": 0.5326027397260275,
+      "grad_norm": 2.2462668418884277,
+      "learning_rate": 4.597887896019497e-06,
+      "loss": 1.9201,
+      "step": 486
+    },
+    {
+      "epoch": 0.5336986301369863,
+      "grad_norm": 2.1482582092285156,
+      "learning_rate": 4.595857026807474e-06,
+      "loss": 1.9722,
+      "step": 487
+    },
+    {
+      "epoch": 0.5347945205479452,
+      "grad_norm": 2.079831123352051,
+      "learning_rate": 4.593826157595451e-06,
+      "loss": 1.977,
+      "step": 488
+    },
+    {
+      "epoch": 0.5358904109589041,
+      "grad_norm": 2.2169578075408936,
+      "learning_rate": 4.591795288383429e-06,
+      "loss": 1.9615,
+      "step": 489
+    },
+    {
+      "epoch": 0.536986301369863,
+      "grad_norm": 2.087846279144287,
+      "learning_rate": 4.589764419171406e-06,
+      "loss": 2.0402,
+      "step": 490
+    },
+    {
+      "epoch": 0.536986301369863,
+      "eval_loss": 2.1416702270507812,
+      "eval_runtime": 37.1191,
+      "eval_samples_per_second": 26.94,
+      "eval_steps_per_second": 1.131,
+      "step": 490
+    },
+    {
+      "epoch": 0.538082191780822,
+      "grad_norm": 2.2161977291107178,
+      "learning_rate": 4.587733549959383e-06,
+      "loss": 1.9959,
+      "step": 491
+    },
+    {
+      "epoch": 0.5391780821917809,
+      "grad_norm": 2.19530987739563,
+      "learning_rate": 4.5857026807473604e-06,
+      "loss": 1.961,
+      "step": 492
+    },
+    {
+      "epoch": 0.5402739726027397,
+      "grad_norm": 2.1150615215301514,
+      "learning_rate": 4.583671811535338e-06,
+      "loss": 1.946,
+      "step": 493
+    },
+    {
+      "epoch": 0.5413698630136986,
+      "grad_norm": 2.186352491378784,
+      "learning_rate": 4.581640942323315e-06,
+      "loss": 1.9137,
+      "step": 494
+    },
+    {
+      "epoch": 0.5424657534246575,
+      "grad_norm": 2.233370780944824,
+      "learning_rate": 4.579610073111292e-06,
+      "loss": 2.021,
+      "step": 495
+    },
+    {
+      "epoch": 0.5435616438356164,
+      "grad_norm": 2.1326382160186768,
+      "learning_rate": 4.5775792038992695e-06,
+      "loss": 2.0242,
+      "step": 496
+    },
+    {
+      "epoch": 0.5446575342465754,
+      "grad_norm": 2.536837577819824,
+      "learning_rate": 4.575548334687247e-06,
+      "loss": 2.0185,
+      "step": 497
+    },
+    {
+      "epoch": 0.5457534246575343,
+      "grad_norm": 2.193598747253418,
+      "learning_rate": 4.573517465475224e-06,
+      "loss": 2.0533,
+      "step": 498
+    },
+    {
+      "epoch": 0.5468493150684931,
+      "grad_norm": 2.1346311569213867,
+      "learning_rate": 4.571486596263201e-06,
+      "loss": 1.9564,
+      "step": 499
+    },
+    {
+      "epoch": 0.547945205479452,
+      "grad_norm": 2.4294722080230713,
+      "learning_rate": 4.569455727051178e-06,
+      "loss": 2.0003,
+      "step": 500
+    },
+    {
+      "epoch": 0.547945205479452,
+      "eval_loss": 2.1430764198303223,
+      "eval_runtime": 37.0591,
+      "eval_samples_per_second": 26.984,
+      "eval_steps_per_second": 1.133,
+      "step": 500
+    },
+    {
+      "epoch": 0.549041095890411,
+      "grad_norm": 2.2394182682037354,
+      "learning_rate": 4.567424857839156e-06,
+      "loss": 2.0055,
+      "step": 501
+    },
+    {
+      "epoch": 0.5501369863013699,
+      "grad_norm": 2.2543067932128906,
+      "learning_rate": 4.565393988627132e-06,
+      "loss": 1.9837,
+      "step": 502
+    },
+    {
+      "epoch": 0.5512328767123288,
+      "grad_norm": 2.2805464267730713,
+      "learning_rate": 4.5633631194151105e-06,
+      "loss": 2.0495,
+      "step": 503
+    },
+    {
+      "epoch": 0.5523287671232877,
+      "grad_norm": 2.2838146686553955,
+      "learning_rate": 4.561332250203087e-06,
+      "loss": 1.9866,
+      "step": 504
+    },
+    {
+      "epoch": 0.5534246575342465,
+      "grad_norm": 2.286147356033325,
+      "learning_rate": 4.559301380991065e-06,
+      "loss": 1.9805,
+      "step": 505
+    },
+    {
+      "epoch": 0.5545205479452054,
+      "grad_norm": 2.064378023147583,
+      "learning_rate": 4.5572705117790414e-06,
+      "loss": 1.9259,
+      "step": 506
+    },
+    {
+      "epoch": 0.5556164383561644,
+      "grad_norm": 2.4020493030548096,
+      "learning_rate": 4.5552396425670196e-06,
+      "loss": 1.9731,
+      "step": 507
+    },
+    {
+      "epoch": 0.5567123287671233,
+      "grad_norm": 2.11594557762146,
+      "learning_rate": 4.553208773354996e-06,
+      "loss": 2.0126,
+      "step": 508
+    },
+    {
+      "epoch": 0.5578082191780822,
+      "grad_norm": 2.4520673751831055,
+      "learning_rate": 4.551177904142974e-06,
+      "loss": 1.9396,
+      "step": 509
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "grad_norm": 2.302771806716919,
+      "learning_rate": 4.5491470349309505e-06,
+      "loss": 2.0361,
+      "step": 510
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "eval_loss": 2.1408755779266357,
+      "eval_runtime": 36.9881,
+      "eval_samples_per_second": 27.036,
+      "eval_steps_per_second": 1.136,
+      "step": 510
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 2.250927686691284,
+      "learning_rate": 4.547116165718929e-06,
+      "loss": 1.9616,
+      "step": 511
+    },
+    {
+      "epoch": 0.5610958904109589,
+      "grad_norm": 2.4257938861846924,
+      "learning_rate": 4.545085296506905e-06,
+      "loss": 1.9836,
+      "step": 512
+    },
+    {
+      "epoch": 0.5621917808219178,
+      "grad_norm": 2.08693265914917,
+      "learning_rate": 4.543054427294882e-06,
+      "loss": 1.9541,
+      "step": 513
+    },
+    {
+      "epoch": 0.5632876712328767,
+      "grad_norm": 2.133875846862793,
+      "learning_rate": 4.54102355808286e-06,
+      "loss": 1.9377,
+      "step": 514
+    },
+    {
+      "epoch": 0.5643835616438356,
+      "grad_norm": 2.1217594146728516,
+      "learning_rate": 4.538992688870837e-06,
+      "loss": 1.9369,
+      "step": 515
+    },
+    {
+      "epoch": 0.5654794520547946,
+      "grad_norm": 2.117994546890259,
+      "learning_rate": 4.536961819658814e-06,
+      "loss": 1.9666,
+      "step": 516
+    },
+    {
+      "epoch": 0.5665753424657535,
+      "grad_norm": 2.1514523029327393,
+      "learning_rate": 4.5349309504467915e-06,
+      "loss": 1.9704,
+      "step": 517
+    },
+    {
+      "epoch": 0.5676712328767123,
+      "grad_norm": 2.2309377193450928,
+      "learning_rate": 4.532900081234769e-06,
+      "loss": 1.9858,
+      "step": 518
+    },
+    {
+      "epoch": 0.5687671232876712,
+      "grad_norm": 2.100398540496826,
+      "learning_rate": 4.530869212022746e-06,
+      "loss": 1.9573,
+      "step": 519
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "grad_norm": 2.122840166091919,
+      "learning_rate": 4.528838342810723e-06,
+      "loss": 1.9813,
+      "step": 520
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "eval_loss": 2.135481357574463,
+      "eval_runtime": 36.893,
+      "eval_samples_per_second": 27.105,
+      "eval_steps_per_second": 1.138,
+      "step": 520
+    },
+    {
+      "epoch": 0.570958904109589,
+      "grad_norm": 2.1103999614715576,
+      "learning_rate": 4.526807473598701e-06,
+      "loss": 1.9691,
+      "step": 521
+    },
+    {
+      "epoch": 0.572054794520548,
+      "grad_norm": 2.059232473373413,
+      "learning_rate": 4.524776604386678e-06,
+      "loss": 1.9687,
+      "step": 522
+    },
+    {
+      "epoch": 0.5731506849315069,
+      "grad_norm": 2.207000732421875,
+      "learning_rate": 4.522745735174655e-06,
+      "loss": 1.949,
+      "step": 523
+    },
+    {
+      "epoch": 0.5742465753424657,
+      "grad_norm": 2.264000177383423,
+      "learning_rate": 4.520714865962632e-06,
+      "loss": 2.0066,
+      "step": 524
+    },
+    {
+      "epoch": 0.5753424657534246,
+      "grad_norm": 2.121701717376709,
+      "learning_rate": 4.51868399675061e-06,
+      "loss": 1.9879,
+      "step": 525
+    },
+    {
+      "epoch": 0.5764383561643835,
+      "grad_norm": 2.0745644569396973,
+      "learning_rate": 4.516653127538587e-06,
+      "loss": 1.9924,
+      "step": 526
+    },
+    {
+      "epoch": 0.5775342465753425,
+      "grad_norm": 2.2257349491119385,
+      "learning_rate": 4.514622258326564e-06,
+      "loss": 1.9525,
+      "step": 527
+    },
+    {
+      "epoch": 0.5786301369863014,
+      "grad_norm": 2.1678507328033447,
+      "learning_rate": 4.5125913891145415e-06,
+      "loss": 1.9378,
+      "step": 528
+    },
+    {
+      "epoch": 0.5797260273972603,
+      "grad_norm": 2.2800841331481934,
+      "learning_rate": 4.510560519902519e-06,
+      "loss": 1.9685,
+      "step": 529
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "grad_norm": 2.1471869945526123,
+      "learning_rate": 4.508529650690496e-06,
+      "loss": 2.0027,
+      "step": 530
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "eval_loss": 2.1354892253875732,
+      "eval_runtime": 36.8434,
+      "eval_samples_per_second": 27.142,
+      "eval_steps_per_second": 1.14,
+      "step": 530
+    },
+    {
+      "epoch": 0.581917808219178,
+      "grad_norm": 2.2238621711730957,
+      "learning_rate": 4.506498781478473e-06,
+      "loss": 1.9258,
+      "step": 531
+    },
+    {
+      "epoch": 0.583013698630137,
+      "grad_norm": 2.2224624156951904,
+      "learning_rate": 4.504467912266451e-06,
+      "loss": 1.9168,
+      "step": 532
+    },
+    {
+      "epoch": 0.5841095890410959,
+      "grad_norm": 1.9850294589996338,
+      "learning_rate": 4.502437043054428e-06,
+      "loss": 1.9656,
+      "step": 533
+    },
+    {
+      "epoch": 0.5852054794520548,
+      "grad_norm": 2.1303517818450928,
+      "learning_rate": 4.500406173842405e-06,
+      "loss": 2.0349,
+      "step": 534
+    },
+    {
+      "epoch": 0.5863013698630137,
+      "grad_norm": 2.2114832401275635,
+      "learning_rate": 4.4983753046303824e-06,
+      "loss": 1.9878,
+      "step": 535
+    },
+    {
+      "epoch": 0.5873972602739727,
+      "grad_norm": 1.9375886917114258,
+      "learning_rate": 4.49634443541836e-06,
+      "loss": 1.9724,
+      "step": 536
+    },
+    {
+      "epoch": 0.5884931506849315,
+      "grad_norm": 2.185518741607666,
+      "learning_rate": 4.494313566206336e-06,
+      "loss": 1.9584,
+      "step": 537
+    },
+    {
+      "epoch": 0.5895890410958904,
+      "grad_norm": 2.125638484954834,
+      "learning_rate": 4.492282696994314e-06,
+      "loss": 2.0203,
+      "step": 538
+    },
+    {
+      "epoch": 0.5906849315068493,
+      "grad_norm": 1.9456262588500977,
+      "learning_rate": 4.490251827782291e-06,
+      "loss": 1.9523,
+      "step": 539
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "grad_norm": 2.088125705718994,
+      "learning_rate": 4.488220958570269e-06,
+      "loss": 1.9271,
+      "step": 540
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "eval_loss": 2.134958505630493,
+      "eval_runtime": 36.8342,
+      "eval_samples_per_second": 27.149,
+      "eval_steps_per_second": 1.14,
+      "step": 540
+    },
+    {
+      "epoch": 0.5928767123287672,
+      "grad_norm": 2.1423559188842773,
+      "learning_rate": 4.486190089358245e-06,
+      "loss": 1.9467,
+      "step": 541
+    },
+    {
+      "epoch": 0.5939726027397261,
+      "grad_norm": 2.2169737815856934,
+      "learning_rate": 4.484159220146223e-06,
+      "loss": 1.8901,
+      "step": 542
+    },
+    {
+      "epoch": 0.5950684931506849,
+      "grad_norm": 2.419947385787964,
+      "learning_rate": 4.4821283509342e-06,
+      "loss": 2.0064,
+      "step": 543
+    },
+    {
+      "epoch": 0.5961643835616438,
+      "grad_norm": 2.0396435260772705,
+      "learning_rate": 4.480097481722178e-06,
+      "loss": 1.9819,
+      "step": 544
+    },
+    {
+      "epoch": 0.5972602739726027,
+      "grad_norm": 2.210700511932373,
+      "learning_rate": 4.478066612510154e-06,
+      "loss": 1.9578,
+      "step": 545
+    },
+    {
+      "epoch": 0.5983561643835617,
+      "grad_norm": 2.159698009490967,
+      "learning_rate": 4.4760357432981325e-06,
+      "loss": 1.9313,
+      "step": 546
+    },
+    {
+      "epoch": 0.5994520547945206,
+      "grad_norm": 2.125058650970459,
+      "learning_rate": 4.474004874086109e-06,
+      "loss": 1.9549,
+      "step": 547
+    },
+    {
+      "epoch": 0.6005479452054795,
+      "grad_norm": 2.477320432662964,
+      "learning_rate": 4.471974004874087e-06,
+      "loss": 2.0883,
+      "step": 548
+    },
+    {
+      "epoch": 0.6016438356164383,
+      "grad_norm": 2.214901924133301,
+      "learning_rate": 4.4699431356620635e-06,
+      "loss": 1.9184,
+      "step": 549
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "grad_norm": 2.1001312732696533,
+      "learning_rate": 4.467912266450041e-06,
+      "loss": 2.0231,
+      "step": 550
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "eval_loss": 2.1315157413482666,
+      "eval_runtime": 36.8131,
+      "eval_samples_per_second": 27.164,
+      "eval_steps_per_second": 1.141,
+      "step": 550
+    },
+    {
+      "epoch": 0.6038356164383561,
+      "grad_norm": 2.1030538082122803,
+      "learning_rate": 4.465881397238018e-06,
+      "loss": 1.967,
+      "step": 551
+    },
+    {
+      "epoch": 0.6049315068493151,
+      "grad_norm": 2.0390465259552,
+      "learning_rate": 4.463850528025995e-06,
+      "loss": 1.9615,
+      "step": 552
+    },
+    {
+      "epoch": 0.606027397260274,
+      "grad_norm": 2.1331627368927,
+      "learning_rate": 4.4618196588139726e-06,
+      "loss": 1.9667,
+      "step": 553
+    },
+    {
+      "epoch": 0.6071232876712329,
+      "grad_norm": 2.1987671852111816,
+      "learning_rate": 4.45978878960195e-06,
+      "loss": 1.9922,
+      "step": 554
+    },
+    {
+      "epoch": 0.6082191780821918,
+      "grad_norm": 2.127434253692627,
+      "learning_rate": 4.457757920389927e-06,
+      "loss": 1.9913,
+      "step": 555
+    },
+    {
+      "epoch": 0.6093150684931506,
+      "grad_norm": 2.2486324310302734,
+      "learning_rate": 4.455727051177904e-06,
+      "loss": 1.9239,
+      "step": 556
+    },
+    {
+      "epoch": 0.6104109589041096,
+      "grad_norm": 2.127575397491455,
+      "learning_rate": 4.453696181965882e-06,
+      "loss": 1.9521,
+      "step": 557
+    },
+    {
+      "epoch": 0.6115068493150685,
+      "grad_norm": 2.0798254013061523,
+      "learning_rate": 4.451665312753859e-06,
+      "loss": 2.0085,
+      "step": 558
+    },
+    {
+      "epoch": 0.6126027397260274,
+      "grad_norm": 2.287574291229248,
+      "learning_rate": 4.449634443541836e-06,
+      "loss": 1.98,
+      "step": 559
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "grad_norm": 2.2056541442871094,
+      "learning_rate": 4.4476035743298135e-06,
+      "loss": 2.0282,
+      "step": 560
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "eval_loss": 2.1287026405334473,
+      "eval_runtime": 36.7788,
+      "eval_samples_per_second": 27.19,
+      "eval_steps_per_second": 1.142,
+      "step": 560
+    },
+    {
+      "epoch": 0.6147945205479453,
+      "grad_norm": 2.169157028198242,
+      "learning_rate": 4.445572705117791e-06,
+      "loss": 1.9308,
+      "step": 561
+    },
+    {
+      "epoch": 0.6158904109589041,
+      "grad_norm": 2.1035149097442627,
+      "learning_rate": 4.443541835905768e-06,
+      "loss": 1.9507,
+      "step": 562
+    },
+    {
+      "epoch": 0.616986301369863,
+      "grad_norm": 2.1174328327178955,
+      "learning_rate": 4.441510966693745e-06,
+      "loss": 2.0162,
+      "step": 563
+    },
+    {
+      "epoch": 0.6180821917808219,
+      "grad_norm": 2.201641321182251,
+      "learning_rate": 4.439480097481723e-06,
+      "loss": 1.9573,
+      "step": 564
+    },
+    {
+      "epoch": 0.6191780821917808,
+      "grad_norm": 2.1763603687286377,
+      "learning_rate": 4.4374492282697e-06,
+      "loss": 2.0149,
+      "step": 565
+    },
+    {
+      "epoch": 0.6202739726027398,
+      "grad_norm": 2.2866439819335938,
+      "learning_rate": 4.435418359057677e-06,
+      "loss": 1.8836,
+      "step": 566
+    },
+    {
+      "epoch": 0.6213698630136987,
+      "grad_norm": 2.2060160636901855,
+      "learning_rate": 4.433387489845654e-06,
+      "loss": 1.9806,
+      "step": 567
+    },
+    {
+      "epoch": 0.6224657534246575,
+      "grad_norm": 2.0614662170410156,
+      "learning_rate": 4.431356620633632e-06,
+      "loss": 1.9821,
+      "step": 568
+    },
+    {
+      "epoch": 0.6235616438356164,
+      "grad_norm": 2.1055402755737305,
+      "learning_rate": 4.429325751421609e-06,
+      "loss": 1.9365,
+      "step": 569
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "grad_norm": 1.9907679557800293,
+      "learning_rate": 4.427294882209586e-06,
+      "loss": 1.9429,
+      "step": 570
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "eval_loss": 2.1286206245422363,
+      "eval_runtime": 36.7748,
+      "eval_samples_per_second": 27.193,
+      "eval_steps_per_second": 1.142,
+      "step": 570
+    },
+    {
+      "epoch": 0.6257534246575343,
+      "grad_norm": 2.0169243812561035,
+      "learning_rate": 4.4252640129975635e-06,
+      "loss": 1.9643,
+      "step": 571
+    },
+    {
+      "epoch": 0.6268493150684932,
+      "grad_norm": 2.1082401275634766,
+      "learning_rate": 4.423233143785541e-06,
+      "loss": 1.9009,
+      "step": 572
+    },
+    {
+      "epoch": 0.6279452054794521,
+      "grad_norm": 2.202542543411255,
+      "learning_rate": 4.421202274573518e-06,
+      "loss": 1.9694,
+      "step": 573
+    },
+    {
+      "epoch": 0.6290410958904109,
+      "grad_norm": 2.028489828109741,
+      "learning_rate": 4.4191714053614945e-06,
+      "loss": 1.9697,
+      "step": 574
+    },
+    {
+      "epoch": 0.6301369863013698,
+      "grad_norm": 2.134274482727051,
+      "learning_rate": 4.417140536149473e-06,
+      "loss": 1.8672,
+      "step": 575
+    },
+    {
+      "epoch": 0.6312328767123287,
+      "grad_norm": 2.063542366027832,
+      "learning_rate": 4.415109666937449e-06,
+      "loss": 1.909,
+      "step": 576
+    },
+    {
+      "epoch": 0.6323287671232877,
+      "grad_norm": 1.9747039079666138,
+      "learning_rate": 4.413078797725427e-06,
+      "loss": 1.9795,
+      "step": 577
+    },
+    {
+      "epoch": 0.6334246575342466,
+      "grad_norm": 2.063673496246338,
+      "learning_rate": 4.411047928513404e-06,
+      "loss": 1.9283,
+      "step": 578
+    },
+    {
+      "epoch": 0.6345205479452055,
+      "grad_norm": 2.024611711502075,
+      "learning_rate": 4.409017059301382e-06,
+      "loss": 1.9682,
+      "step": 579
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "grad_norm": 2.0471599102020264,
+      "learning_rate": 4.406986190089358e-06,
+      "loss": 1.969,
+      "step": 580
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "eval_loss": 2.1233692169189453,
+      "eval_runtime": 36.8185,
+      "eval_samples_per_second": 27.16,
+      "eval_steps_per_second": 1.141,
+      "step": 580
+    },
+    {
+      "epoch": 0.6367123287671232,
+      "grad_norm": 2.1038553714752197,
+      "learning_rate": 4.404955320877336e-06,
+      "loss": 1.9915,
+      "step": 581
+    },
+    {
+      "epoch": 0.6378082191780822,
+      "grad_norm": 2.1222152709960938,
+      "learning_rate": 4.402924451665313e-06,
+      "loss": 1.9892,
+      "step": 582
+    },
+    {
+      "epoch": 0.6389041095890411,
+      "grad_norm": 2.09137225151062,
+      "learning_rate": 4.400893582453291e-06,
+      "loss": 1.934,
+      "step": 583
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 2.2158281803131104,
+      "learning_rate": 4.398862713241267e-06,
+      "loss": 1.9188,
+      "step": 584
+    },
+    {
+      "epoch": 0.6410958904109589,
+      "grad_norm": 2.34619140625,
+      "learning_rate": 4.396831844029245e-06,
+      "loss": 1.9495,
+      "step": 585
+    },
+    {
+      "epoch": 0.6421917808219179,
+      "grad_norm": 2.2732315063476562,
+      "learning_rate": 4.394800974817222e-06,
+      "loss": 1.9632,
+      "step": 586
+    },
+    {
+      "epoch": 0.6432876712328767,
+      "grad_norm": 2.153338670730591,
+      "learning_rate": 4.392770105605199e-06,
+      "loss": 1.9551,
+      "step": 587
+    },
+    {
+      "epoch": 0.6443835616438356,
+      "grad_norm": 2.2201521396636963,
+      "learning_rate": 4.390739236393176e-06,
+      "loss": 1.9677,
+      "step": 588
+    },
+    {
+      "epoch": 0.6454794520547945,
+      "grad_norm": 1.9894570112228394,
+      "learning_rate": 4.388708367181154e-06,
+      "loss": 1.9826,
+      "step": 589
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "grad_norm": 2.145061492919922,
+      "learning_rate": 4.386677497969131e-06,
+      "loss": 1.9826,
+      "step": 590
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "eval_loss": 2.122312545776367,
+      "eval_runtime": 37.1153,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 590
+    },
+    {
+      "epoch": 0.6476712328767124,
+      "grad_norm": 2.257237672805786,
+      "learning_rate": 4.384646628757108e-06,
+      "loss": 1.9722,
+      "step": 591
+    },
+    {
+      "epoch": 0.6487671232876713,
+      "grad_norm": 2.2406375408172607,
+      "learning_rate": 4.3826157595450855e-06,
+      "loss": 2.0141,
+      "step": 592
+    },
+    {
+      "epoch": 0.6498630136986301,
+      "grad_norm": 1.9656853675842285,
+      "learning_rate": 4.380584890333063e-06,
+      "loss": 1.9435,
+      "step": 593
+    },
+    {
+      "epoch": 0.650958904109589,
+      "grad_norm": 2.1063120365142822,
+      "learning_rate": 4.37855402112104e-06,
+      "loss": 1.9001,
+      "step": 594
+    },
+    {
+      "epoch": 0.6520547945205479,
+      "grad_norm": 2.1594269275665283,
+      "learning_rate": 4.376523151909017e-06,
+      "loss": 1.8596,
+      "step": 595
+    },
+    {
+      "epoch": 0.6531506849315069,
+      "grad_norm": 2.2261908054351807,
+      "learning_rate": 4.3744922826969946e-06,
+      "loss": 2.016,
+      "step": 596
+    },
+    {
+      "epoch": 0.6542465753424658,
+      "grad_norm": 1.9940093755722046,
+      "learning_rate": 4.372461413484972e-06,
+      "loss": 1.9293,
+      "step": 597
+    },
+    {
+      "epoch": 0.6553424657534247,
+      "grad_norm": 2.62186861038208,
+      "learning_rate": 4.370430544272949e-06,
+      "loss": 1.9656,
+      "step": 598
+    },
+    {
+      "epoch": 0.6564383561643835,
+      "grad_norm": 2.3531494140625,
+      "learning_rate": 4.368399675060926e-06,
+      "loss": 1.9984,
+      "step": 599
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "grad_norm": 2.362769365310669,
+      "learning_rate": 4.366368805848904e-06,
+      "loss": 1.9971,
+      "step": 600
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "eval_loss": 2.1172940731048584,
+      "eval_runtime": 37.0498,
+      "eval_samples_per_second": 26.991,
+      "eval_steps_per_second": 1.134,
+      "step": 600
+    },
+    {
+      "epoch": 0.6586301369863014,
+      "grad_norm": 2.3806607723236084,
+      "learning_rate": 4.364337936636881e-06,
+      "loss": 1.9566,
+      "step": 601
+    },
+    {
+      "epoch": 0.6597260273972603,
+      "grad_norm": 2.5531814098358154,
+      "learning_rate": 4.362307067424858e-06,
+      "loss": 1.9002,
+      "step": 602
+    },
+    {
+      "epoch": 0.6608219178082192,
+      "grad_norm": 2.283364772796631,
+      "learning_rate": 4.3602761982128355e-06,
+      "loss": 1.9633,
+      "step": 603
+    },
+    {
+      "epoch": 0.6619178082191781,
+      "grad_norm": 2.180171489715576,
+      "learning_rate": 4.358245329000813e-06,
+      "loss": 1.9277,
+      "step": 604
+    },
+    {
+      "epoch": 0.663013698630137,
+      "grad_norm": 2.0771350860595703,
+      "learning_rate": 4.35621445978879e-06,
+      "loss": 1.9484,
+      "step": 605
+    },
+    {
+      "epoch": 0.6641095890410958,
+      "grad_norm": 2.3772811889648438,
+      "learning_rate": 4.354183590576767e-06,
+      "loss": 1.9564,
+      "step": 606
+    },
+    {
+      "epoch": 0.6652054794520548,
+      "grad_norm": 2.205486297607422,
+      "learning_rate": 4.352152721364745e-06,
+      "loss": 1.9389,
+      "step": 607
+    },
+    {
+      "epoch": 0.6663013698630137,
+      "grad_norm": 2.085190773010254,
+      "learning_rate": 4.350121852152722e-06,
+      "loss": 1.8948,
+      "step": 608
+    },
+    {
+      "epoch": 0.6673972602739726,
+      "grad_norm": 2.4398982524871826,
+      "learning_rate": 4.348090982940699e-06,
+      "loss": 1.9736,
+      "step": 609
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "grad_norm": 2.194087028503418,
+      "learning_rate": 4.3460601137286764e-06,
+      "loss": 1.9607,
+      "step": 610
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "eval_loss": 2.115041971206665,
+      "eval_runtime": 36.9738,
+      "eval_samples_per_second": 27.046,
+      "eval_steps_per_second": 1.136,
+      "step": 610
+    },
+    {
+      "epoch": 0.6695890410958905,
+      "grad_norm": 2.143054485321045,
+      "learning_rate": 4.344029244516654e-06,
+      "loss": 1.9733,
+      "step": 611
+    },
+    {
+      "epoch": 0.6706849315068493,
+      "grad_norm": 2.2484617233276367,
+      "learning_rate": 4.341998375304631e-06,
+      "loss": 1.9417,
+      "step": 612
+    },
+    {
+      "epoch": 0.6717808219178082,
+      "grad_norm": 2.0430805683135986,
+      "learning_rate": 4.339967506092607e-06,
+      "loss": 2.0192,
+      "step": 613
+    },
+    {
+      "epoch": 0.6728767123287671,
+      "grad_norm": 2.061089277267456,
+      "learning_rate": 4.3379366368805855e-06,
+      "loss": 1.9543,
+      "step": 614
+    },
+    {
+      "epoch": 0.673972602739726,
+      "grad_norm": 2.298022747039795,
+      "learning_rate": 4.335905767668562e-06,
+      "loss": 1.9579,
+      "step": 615
+    },
+    {
+      "epoch": 0.675068493150685,
+      "grad_norm": 2.1418962478637695,
+      "learning_rate": 4.33387489845654e-06,
+      "loss": 1.938,
+      "step": 616
+    },
+    {
+      "epoch": 0.6761643835616439,
+      "grad_norm": 2.1005501747131348,
+      "learning_rate": 4.3318440292445165e-06,
+      "loss": 1.972,
+      "step": 617
+    },
+    {
+      "epoch": 0.6772602739726027,
+      "grad_norm": 2.1665523052215576,
+      "learning_rate": 4.329813160032495e-06,
+      "loss": 1.9535,
+      "step": 618
+    },
+    {
+      "epoch": 0.6783561643835616,
+      "grad_norm": 2.144580841064453,
+      "learning_rate": 4.327782290820471e-06,
+      "loss": 1.9188,
+      "step": 619
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "grad_norm": 1.9855908155441284,
+      "learning_rate": 4.325751421608449e-06,
+      "loss": 1.912,
+      "step": 620
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "eval_loss": 2.118319272994995,
+      "eval_runtime": 36.9677,
+      "eval_samples_per_second": 27.051,
+      "eval_steps_per_second": 1.136,
+      "step": 620
+    },
+    {
+      "epoch": 0.6805479452054795,
+      "grad_norm": 2.395538091659546,
+      "learning_rate": 4.323720552396426e-06,
+      "loss": 1.9886,
+      "step": 621
+    },
+    {
+      "epoch": 0.6816438356164384,
+      "grad_norm": 2.190574884414673,
+      "learning_rate": 4.321689683184404e-06,
+      "loss": 1.9326,
+      "step": 622
+    },
+    {
+      "epoch": 0.6827397260273973,
+      "grad_norm": 2.1826560497283936,
+      "learning_rate": 4.31965881397238e-06,
+      "loss": 1.9278,
+      "step": 623
+    },
+    {
+      "epoch": 0.6838356164383562,
+      "grad_norm": 2.151097536087036,
+      "learning_rate": 4.3176279447603574e-06,
+      "loss": 1.9359,
+      "step": 624
+    },
+    {
+      "epoch": 0.684931506849315,
+      "grad_norm": 2.1830122470855713,
+      "learning_rate": 4.315597075548335e-06,
+      "loss": 1.9728,
+      "step": 625
+    },
+    {
+      "epoch": 0.686027397260274,
+      "grad_norm": 2.203325033187866,
+      "learning_rate": 4.313566206336312e-06,
+      "loss": 1.9638,
+      "step": 626
+    },
+    {
+      "epoch": 0.6871232876712329,
+      "grad_norm": 2.7240309715270996,
+      "learning_rate": 4.311535337124289e-06,
+      "loss": 1.9907,
+      "step": 627
+    },
+    {
+      "epoch": 0.6882191780821918,
+      "grad_norm": 2.133599281311035,
+      "learning_rate": 4.3095044679122665e-06,
+      "loss": 1.9352,
+      "step": 628
+    },
+    {
+      "epoch": 0.6893150684931507,
+      "grad_norm": 2.0318551063537598,
+      "learning_rate": 4.307473598700244e-06,
+      "loss": 1.9432,
+      "step": 629
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "grad_norm": 2.502101182937622,
+      "learning_rate": 4.305442729488221e-06,
+      "loss": 1.9641,
+      "step": 630
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "eval_loss": 2.1158926486968994,
+      "eval_runtime": 36.8818,
+      "eval_samples_per_second": 27.114,
+      "eval_steps_per_second": 1.139,
+      "step": 630
+    },
+    {
+      "epoch": 0.6915068493150685,
+      "grad_norm": 2.3620944023132324,
+      "learning_rate": 4.303411860276198e-06,
+      "loss": 1.8841,
+      "step": 631
+    },
+    {
+      "epoch": 0.6926027397260274,
+      "grad_norm": 2.3139240741729736,
+      "learning_rate": 4.301380991064176e-06,
+      "loss": 1.9896,
+      "step": 632
+    },
+    {
+      "epoch": 0.6936986301369863,
+      "grad_norm": 2.314877510070801,
+      "learning_rate": 4.299350121852153e-06,
+      "loss": 1.9655,
+      "step": 633
+    },
+    {
+      "epoch": 0.6947945205479452,
+      "grad_norm": 2.036371946334839,
+      "learning_rate": 4.29731925264013e-06,
+      "loss": 1.9917,
+      "step": 634
+    },
+    {
+      "epoch": 0.6958904109589041,
+      "grad_norm": 2.1859662532806396,
+      "learning_rate": 4.2952883834281075e-06,
+      "loss": 1.9805,
+      "step": 635
+    },
+    {
+      "epoch": 0.6969863013698631,
+      "grad_norm": 2.2317097187042236,
+      "learning_rate": 4.293257514216085e-06,
+      "loss": 1.9402,
+      "step": 636
+    },
+    {
+      "epoch": 0.6980821917808219,
+      "grad_norm": 2.3249905109405518,
+      "learning_rate": 4.291226645004062e-06,
+      "loss": 1.9013,
+      "step": 637
+    },
+    {
+      "epoch": 0.6991780821917808,
+      "grad_norm": 2.199758768081665,
+      "learning_rate": 4.289195775792039e-06,
+      "loss": 1.9783,
+      "step": 638
+    },
+    {
+      "epoch": 0.7002739726027397,
+      "grad_norm": 2.3598549365997314,
+      "learning_rate": 4.287164906580017e-06,
+      "loss": 1.9936,
+      "step": 639
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "grad_norm": 2.1705522537231445,
+      "learning_rate": 4.285134037367994e-06,
+      "loss": 1.9304,
+      "step": 640
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "eval_loss": 2.1055190563201904,
+      "eval_runtime": 36.8673,
+      "eval_samples_per_second": 27.124,
+      "eval_steps_per_second": 1.139,
+      "step": 640
+    },
+    {
+      "epoch": 0.7024657534246576,
+      "grad_norm": 2.3253612518310547,
+      "learning_rate": 4.283103168155971e-06,
+      "loss": 1.9196,
+      "step": 641
+    },
+    {
+      "epoch": 0.7035616438356165,
+      "grad_norm": 2.3054087162017822,
+      "learning_rate": 4.281072298943948e-06,
+      "loss": 1.9053,
+      "step": 642
+    },
+    {
+      "epoch": 0.7046575342465753,
+      "grad_norm": 2.262057304382324,
+      "learning_rate": 4.279041429731926e-06,
+      "loss": 1.8999,
+      "step": 643
+    },
+    {
+      "epoch": 0.7057534246575342,
+      "grad_norm": 2.421391248703003,
+      "learning_rate": 4.277010560519903e-06,
+      "loss": 1.9438,
+      "step": 644
+    },
+    {
+      "epoch": 0.7068493150684931,
+      "grad_norm": 2.2841498851776123,
+      "learning_rate": 4.27497969130788e-06,
+      "loss": 1.9187,
+      "step": 645
+    },
+    {
+      "epoch": 0.707945205479452,
+      "grad_norm": 2.3226938247680664,
+      "learning_rate": 4.2729488220958575e-06,
+      "loss": 1.9275,
+      "step": 646
+    },
+    {
+      "epoch": 0.709041095890411,
+      "grad_norm": 2.2306301593780518,
+      "learning_rate": 4.270917952883835e-06,
+      "loss": 1.9684,
+      "step": 647
+    },
+    {
+      "epoch": 0.7101369863013699,
+      "grad_norm": 2.0551652908325195,
+      "learning_rate": 4.268887083671812e-06,
+      "loss": 1.9849,
+      "step": 648
+    },
+    {
+      "epoch": 0.7112328767123288,
+      "grad_norm": 2.2952210903167725,
+      "learning_rate": 4.266856214459789e-06,
+      "loss": 1.9026,
+      "step": 649
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "grad_norm": 2.1630797386169434,
+      "learning_rate": 4.264825345247766e-06,
+      "loss": 2.0184,
+      "step": 650
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "eval_loss": 2.1037871837615967,
+      "eval_runtime": 36.8221,
+      "eval_samples_per_second": 27.158,
+      "eval_steps_per_second": 1.141,
+      "step": 650
+    },
+    {
+      "epoch": 0.7134246575342466,
+      "grad_norm": 1.964150309562683,
+      "learning_rate": 4.262794476035744e-06,
+      "loss": 1.8932,
+      "step": 651
+    },
+    {
+      "epoch": 0.7145205479452055,
+      "grad_norm": 2.1474685668945312,
+      "learning_rate": 4.26076360682372e-06,
+      "loss": 1.9383,
+      "step": 652
+    },
+    {
+      "epoch": 0.7156164383561644,
+      "grad_norm": 2.199824094772339,
+      "learning_rate": 4.2587327376116984e-06,
+      "loss": 1.8599,
+      "step": 653
+    },
+    {
+      "epoch": 0.7167123287671233,
+      "grad_norm": 2.12156343460083,
+      "learning_rate": 4.256701868399675e-06,
+      "loss": 1.9162,
+      "step": 654
+    },
+    {
+      "epoch": 0.7178082191780822,
+      "grad_norm": 2.3832154273986816,
+      "learning_rate": 4.254670999187653e-06,
+      "loss": 1.9613,
+      "step": 655
+    },
+    {
+      "epoch": 0.718904109589041,
+      "grad_norm": 2.063906192779541,
+      "learning_rate": 4.252640129975629e-06,
+      "loss": 1.9421,
+      "step": 656
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 2.184842586517334,
+      "learning_rate": 4.2506092607636075e-06,
+      "loss": 1.9471,
+      "step": 657
+    },
+    {
+      "epoch": 0.7210958904109589,
+      "grad_norm": 2.1340036392211914,
+      "learning_rate": 4.248578391551584e-06,
+      "loss": 1.931,
+      "step": 658
+    },
+    {
+      "epoch": 0.7221917808219178,
+      "grad_norm": 2.117544651031494,
+      "learning_rate": 4.246547522339562e-06,
+      "loss": 1.8698,
+      "step": 659
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "grad_norm": 2.308344841003418,
+      "learning_rate": 4.2445166531275385e-06,
+      "loss": 1.9217,
+      "step": 660
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "eval_loss": 2.1086323261260986,
+      "eval_runtime": 36.8082,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 660
+    },
+    {
+      "epoch": 0.7243835616438357,
+      "grad_norm": 2.2239017486572266,
+      "learning_rate": 4.242485783915517e-06,
+      "loss": 1.9286,
+      "step": 661
+    },
+    {
+      "epoch": 0.7254794520547945,
+      "grad_norm": 2.263697624206543,
+      "learning_rate": 4.240454914703493e-06,
+      "loss": 1.8821,
+      "step": 662
+    },
+    {
+      "epoch": 0.7265753424657534,
+      "grad_norm": 2.4305059909820557,
+      "learning_rate": 4.23842404549147e-06,
+      "loss": 1.8996,
+      "step": 663
+    },
+    {
+      "epoch": 0.7276712328767123,
+      "grad_norm": 2.2333695888519287,
+      "learning_rate": 4.236393176279448e-06,
+      "loss": 1.8906,
+      "step": 664
+    },
+    {
+      "epoch": 0.7287671232876712,
+      "grad_norm": 2.160982608795166,
+      "learning_rate": 4.234362307067425e-06,
+      "loss": 1.879,
+      "step": 665
+    },
+    {
+      "epoch": 0.7298630136986302,
+      "grad_norm": 2.3707094192504883,
+      "learning_rate": 4.232331437855402e-06,
+      "loss": 1.9577,
+      "step": 666
+    },
+    {
+      "epoch": 0.7309589041095891,
+      "grad_norm": 2.0502827167510986,
+      "learning_rate": 4.2303005686433795e-06,
+      "loss": 1.9801,
+      "step": 667
+    },
+    {
+      "epoch": 0.7320547945205479,
+      "grad_norm": 2.0648176670074463,
+      "learning_rate": 4.228269699431357e-06,
+      "loss": 1.978,
+      "step": 668
+    },
+    {
+      "epoch": 0.7331506849315068,
+      "grad_norm": 2.346316337585449,
+      "learning_rate": 4.226238830219334e-06,
+      "loss": 1.9387,
+      "step": 669
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "grad_norm": 1.9734236001968384,
+      "learning_rate": 4.224207961007311e-06,
+      "loss": 1.9396,
+      "step": 670
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "eval_loss": 2.1104705333709717,
+      "eval_runtime": 36.8144,
+      "eval_samples_per_second": 27.163,
+      "eval_steps_per_second": 1.141,
+      "step": 670
+    },
+    {
+      "epoch": 0.7353424657534247,
+      "grad_norm": 2.3170549869537354,
+      "learning_rate": 4.2221770917952886e-06,
+      "loss": 1.9398,
+      "step": 671
+    },
+    {
+      "epoch": 0.7364383561643836,
+      "grad_norm": 2.0437264442443848,
+      "learning_rate": 4.220146222583266e-06,
+      "loss": 1.9284,
+      "step": 672
+    },
+    {
+      "epoch": 0.7375342465753425,
+      "grad_norm": 2.0808510780334473,
+      "learning_rate": 4.218115353371243e-06,
+      "loss": 1.9425,
+      "step": 673
+    },
+    {
+      "epoch": 0.7386301369863014,
+      "grad_norm": 2.0854742527008057,
+      "learning_rate": 4.21608448415922e-06,
+      "loss": 1.9358,
+      "step": 674
+    },
+    {
+      "epoch": 0.7397260273972602,
+      "grad_norm": 2.131622552871704,
+      "learning_rate": 4.214053614947198e-06,
+      "loss": 1.9048,
+      "step": 675
+    },
+    {
+      "epoch": 0.7408219178082192,
+      "grad_norm": 2.276359796524048,
+      "learning_rate": 4.212022745735175e-06,
+      "loss": 1.9871,
+      "step": 676
+    },
+    {
+      "epoch": 0.7419178082191781,
+      "grad_norm": 2.2194812297821045,
+      "learning_rate": 4.209991876523152e-06,
+      "loss": 1.882,
+      "step": 677
+    },
+    {
+      "epoch": 0.743013698630137,
+      "grad_norm": 2.181891679763794,
+      "learning_rate": 4.2079610073111295e-06,
+      "loss": 1.9546,
+      "step": 678
+    },
+    {
+      "epoch": 0.7441095890410959,
+      "grad_norm": 2.0956673622131348,
+      "learning_rate": 4.205930138099107e-06,
+      "loss": 1.8836,
+      "step": 679
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "grad_norm": 2.057894229888916,
+      "learning_rate": 4.203899268887084e-06,
+      "loss": 1.9842,
+      "step": 680
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "eval_loss": 2.104182720184326,
+      "eval_runtime": 36.8189,
+      "eval_samples_per_second": 27.16,
+      "eval_steps_per_second": 1.141,
+      "step": 680
+    },
+    {
+      "epoch": 0.7463013698630137,
+      "grad_norm": 2.0965662002563477,
+      "learning_rate": 4.201868399675061e-06,
+      "loss": 1.9378,
+      "step": 681
+    },
+    {
+      "epoch": 0.7473972602739726,
+      "grad_norm": 2.1843457221984863,
+      "learning_rate": 4.199837530463039e-06,
+      "loss": 1.9784,
+      "step": 682
+    },
+    {
+      "epoch": 0.7484931506849315,
+      "grad_norm": 2.2044246196746826,
+      "learning_rate": 4.197806661251016e-06,
+      "loss": 1.8642,
+      "step": 683
+    },
+    {
+      "epoch": 0.7495890410958904,
+      "grad_norm": 2.202115058898926,
+      "learning_rate": 4.195775792038993e-06,
+      "loss": 1.929,
+      "step": 684
+    },
+    {
+      "epoch": 0.7506849315068493,
+      "grad_norm": 2.3671462535858154,
+      "learning_rate": 4.19374492282697e-06,
+      "loss": 1.9172,
+      "step": 685
+    },
+    {
+      "epoch": 0.7517808219178083,
+      "grad_norm": 2.0602803230285645,
+      "learning_rate": 4.191714053614948e-06,
+      "loss": 1.8503,
+      "step": 686
+    },
+    {
+      "epoch": 0.7528767123287671,
+      "grad_norm": 2.12145733833313,
+      "learning_rate": 4.189683184402924e-06,
+      "loss": 1.9511,
+      "step": 687
+    },
+    {
+      "epoch": 0.753972602739726,
+      "grad_norm": 2.1049914360046387,
+      "learning_rate": 4.187652315190902e-06,
+      "loss": 1.9043,
+      "step": 688
+    },
+    {
+      "epoch": 0.7550684931506849,
+      "grad_norm": 2.235497236251831,
+      "learning_rate": 4.185621445978879e-06,
+      "loss": 1.9784,
+      "step": 689
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "grad_norm": 2.173222303390503,
+      "learning_rate": 4.183590576766857e-06,
+      "loss": 1.9052,
+      "step": 690
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "eval_loss": 2.0982251167297363,
+      "eval_runtime": 37.1446,
+      "eval_samples_per_second": 26.922,
+      "eval_steps_per_second": 1.131,
+      "step": 690
+    },
+    {
+      "epoch": 0.7572602739726028,
+      "grad_norm": 2.0693562030792236,
+      "learning_rate": 4.181559707554833e-06,
+      "loss": 1.9223,
+      "step": 691
+    },
+    {
+      "epoch": 0.7583561643835617,
+      "grad_norm": 2.209419012069702,
+      "learning_rate": 4.179528838342811e-06,
+      "loss": 1.9709,
+      "step": 692
+    },
+    {
+      "epoch": 0.7594520547945205,
+      "grad_norm": 2.066041946411133,
+      "learning_rate": 4.177497969130788e-06,
+      "loss": 1.889,
+      "step": 693
+    },
+    {
+      "epoch": 0.7605479452054794,
+      "grad_norm": 2.4114325046539307,
+      "learning_rate": 4.175467099918766e-06,
+      "loss": 1.9059,
+      "step": 694
+    },
+    {
+      "epoch": 0.7616438356164383,
+      "grad_norm": 2.0185043811798096,
+      "learning_rate": 4.173436230706742e-06,
+      "loss": 1.9877,
+      "step": 695
+    },
+    {
+      "epoch": 0.7627397260273973,
+      "grad_norm": 2.23392915725708,
+      "learning_rate": 4.1714053614947205e-06,
+      "loss": 1.9199,
+      "step": 696
+    },
+    {
+      "epoch": 0.7638356164383562,
+      "grad_norm": 2.2158737182617188,
+      "learning_rate": 4.169374492282697e-06,
+      "loss": 1.9001,
+      "step": 697
+    },
+    {
+      "epoch": 0.7649315068493151,
+      "grad_norm": 2.1511690616607666,
+      "learning_rate": 4.167343623070675e-06,
+      "loss": 1.9166,
+      "step": 698
+    },
+    {
+      "epoch": 0.766027397260274,
+      "grad_norm": 2.374741554260254,
+      "learning_rate": 4.165312753858652e-06,
+      "loss": 1.9684,
+      "step": 699
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "grad_norm": 2.0409998893737793,
+      "learning_rate": 4.163281884646629e-06,
+      "loss": 1.9177,
+      "step": 700
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "eval_loss": 2.096895933151245,
+      "eval_runtime": 37.066,
+      "eval_samples_per_second": 26.979,
+      "eval_steps_per_second": 1.133,
+      "step": 700
+    },
+    {
+      "epoch": 0.7682191780821918,
+      "grad_norm": 2.3173422813415527,
+      "learning_rate": 4.161251015434607e-06,
+      "loss": 1.9071,
+      "step": 701
+    },
+    {
+      "epoch": 0.7693150684931507,
+      "grad_norm": 2.077842950820923,
+      "learning_rate": 4.159220146222583e-06,
+      "loss": 1.935,
+      "step": 702
+    },
+    {
+      "epoch": 0.7704109589041096,
+      "grad_norm": 2.0518617630004883,
+      "learning_rate": 4.157189277010561e-06,
+      "loss": 1.8942,
+      "step": 703
+    },
+    {
+      "epoch": 0.7715068493150685,
+      "grad_norm": 1.9950882196426392,
+      "learning_rate": 4.155158407798538e-06,
+      "loss": 1.917,
+      "step": 704
+    },
+    {
+      "epoch": 0.7726027397260274,
+      "grad_norm": 2.3222219944000244,
+      "learning_rate": 4.153127538586516e-06,
+      "loss": 1.9551,
+      "step": 705
+    },
+    {
+      "epoch": 0.7736986301369863,
+      "grad_norm": 2.129333972930908,
+      "learning_rate": 4.151096669374492e-06,
+      "loss": 1.9204,
+      "step": 706
+    },
+    {
+      "epoch": 0.7747945205479452,
+      "grad_norm": 2.0357553958892822,
+      "learning_rate": 4.1490658001624705e-06,
+      "loss": 1.9399,
+      "step": 707
+    },
+    {
+      "epoch": 0.7758904109589041,
+      "grad_norm": 2.233694553375244,
+      "learning_rate": 4.147034930950447e-06,
+      "loss": 1.9182,
+      "step": 708
+    },
+    {
+      "epoch": 0.776986301369863,
+      "grad_norm": 2.1391239166259766,
+      "learning_rate": 4.145004061738425e-06,
+      "loss": 1.9157,
+      "step": 709
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "grad_norm": 2.162320613861084,
+      "learning_rate": 4.1429731925264015e-06,
+      "loss": 1.9115,
+      "step": 710
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "eval_loss": 2.0954039096832275,
+      "eval_runtime": 36.9862,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 710
+    },
+    {
+      "epoch": 0.7791780821917809,
+      "grad_norm": 2.2382194995880127,
+      "learning_rate": 4.140942323314379e-06,
+      "loss": 1.8428,
+      "step": 711
+    },
+    {
+      "epoch": 0.7802739726027397,
+      "grad_norm": 2.3286185264587402,
+      "learning_rate": 4.138911454102356e-06,
+      "loss": 1.8906,
+      "step": 712
+    },
+    {
+      "epoch": 0.7813698630136986,
+      "grad_norm": 2.182708501815796,
+      "learning_rate": 4.136880584890333e-06,
+      "loss": 1.9132,
+      "step": 713
+    },
+    {
+      "epoch": 0.7824657534246575,
+      "grad_norm": 1.9978337287902832,
+      "learning_rate": 4.1348497156783106e-06,
+      "loss": 1.9433,
+      "step": 714
+    },
+    {
+      "epoch": 0.7835616438356164,
+      "grad_norm": 2.3092565536499023,
+      "learning_rate": 4.132818846466288e-06,
+      "loss": 1.9374,
+      "step": 715
+    },
+    {
+      "epoch": 0.7846575342465754,
+      "grad_norm": 2.0351054668426514,
+      "learning_rate": 4.130787977254265e-06,
+      "loss": 1.9361,
+      "step": 716
+    },
+    {
+      "epoch": 0.7857534246575343,
+      "grad_norm": 2.14770770072937,
+      "learning_rate": 4.128757108042242e-06,
+      "loss": 1.9553,
+      "step": 717
+    },
+    {
+      "epoch": 0.7868493150684932,
+      "grad_norm": 2.0613555908203125,
+      "learning_rate": 4.12672623883022e-06,
+      "loss": 1.9506,
+      "step": 718
+    },
+    {
+      "epoch": 0.787945205479452,
+      "grad_norm": 2.1312081813812256,
+      "learning_rate": 4.124695369618197e-06,
+      "loss": 2.0079,
+      "step": 719
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "grad_norm": 2.088552474975586,
+      "learning_rate": 4.122664500406174e-06,
+      "loss": 1.9166,
+      "step": 720
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "eval_loss": 2.093032121658325,
+      "eval_runtime": 36.8821,
+      "eval_samples_per_second": 27.113,
+      "eval_steps_per_second": 1.139,
+      "step": 720
+    },
+    {
+      "epoch": 0.7901369863013699,
+      "grad_norm": 2.1748929023742676,
+      "learning_rate": 4.1206336311941515e-06,
+      "loss": 1.9212,
+      "step": 721
+    },
+    {
+      "epoch": 0.7912328767123288,
+      "grad_norm": 2.1492016315460205,
+      "learning_rate": 4.118602761982129e-06,
+      "loss": 1.9452,
+      "step": 722
+    },
+    {
+      "epoch": 0.7923287671232877,
+      "grad_norm": 2.3364970684051514,
+      "learning_rate": 4.116571892770106e-06,
+      "loss": 1.8821,
+      "step": 723
+    },
+    {
+      "epoch": 0.7934246575342466,
+      "grad_norm": 2.318488597869873,
+      "learning_rate": 4.114541023558083e-06,
+      "loss": 1.8597,
+      "step": 724
+    },
+    {
+      "epoch": 0.7945205479452054,
+      "grad_norm": 2.153841257095337,
+      "learning_rate": 4.112510154346061e-06,
+      "loss": 1.9843,
+      "step": 725
+    },
+    {
+      "epoch": 0.7956164383561644,
+      "grad_norm": 2.3464651107788086,
+      "learning_rate": 4.110479285134038e-06,
+      "loss": 1.9105,
+      "step": 726
+    },
+    {
+      "epoch": 0.7967123287671233,
+      "grad_norm": 2.34413743019104,
+      "learning_rate": 4.108448415922015e-06,
+      "loss": 1.9285,
+      "step": 727
+    },
+    {
+      "epoch": 0.7978082191780822,
+      "grad_norm": 2.0822196006774902,
+      "learning_rate": 4.1064175467099924e-06,
+      "loss": 1.9536,
+      "step": 728
+    },
+    {
+      "epoch": 0.7989041095890411,
+      "grad_norm": 2.1978065967559814,
+      "learning_rate": 4.10438667749797e-06,
+      "loss": 1.9487,
+      "step": 729
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 2.4218783378601074,
+      "learning_rate": 4.102355808285947e-06,
+      "loss": 1.8113,
+      "step": 730
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 2.0963101387023926,
+      "eval_runtime": 36.8853,
+      "eval_samples_per_second": 27.111,
+      "eval_steps_per_second": 1.139,
+      "step": 730
+    },
+    {
+      "epoch": 0.8010958904109589,
+      "grad_norm": 2.3887739181518555,
+      "learning_rate": 4.100324939073924e-06,
+      "loss": 2.0024,
+      "step": 731
+    },
+    {
+      "epoch": 0.8021917808219178,
+      "grad_norm": 2.296579360961914,
+      "learning_rate": 4.0982940698619015e-06,
+      "loss": 1.9092,
+      "step": 732
+    },
+    {
+      "epoch": 0.8032876712328767,
+      "grad_norm": 2.3881773948669434,
+      "learning_rate": 4.096263200649879e-06,
+      "loss": 1.9112,
+      "step": 733
+    },
+    {
+      "epoch": 0.8043835616438356,
+      "grad_norm": 2.40120530128479,
+      "learning_rate": 4.094232331437856e-06,
+      "loss": 1.9191,
+      "step": 734
+    },
+    {
+      "epoch": 0.8054794520547945,
+      "grad_norm": 2.184706211090088,
+      "learning_rate": 4.092201462225833e-06,
+      "loss": 1.8857,
+      "step": 735
+    },
+    {
+      "epoch": 0.8065753424657535,
+      "grad_norm": 2.3338732719421387,
+      "learning_rate": 4.090170593013811e-06,
+      "loss": 1.8889,
+      "step": 736
+    },
+    {
+      "epoch": 0.8076712328767123,
+      "grad_norm": 2.4065468311309814,
+      "learning_rate": 4.088139723801787e-06,
+      "loss": 1.9309,
+      "step": 737
+    },
+    {
+      "epoch": 0.8087671232876712,
+      "grad_norm": 2.044992446899414,
+      "learning_rate": 4.086108854589765e-06,
+      "loss": 1.8727,
+      "step": 738
+    },
+    {
+      "epoch": 0.8098630136986301,
+      "grad_norm": 2.23860239982605,
+      "learning_rate": 4.084077985377742e-06,
+      "loss": 1.8936,
+      "step": 739
+    },
+    {
+      "epoch": 0.810958904109589,
+      "grad_norm": 2.1324410438537598,
+      "learning_rate": 4.08204711616572e-06,
+      "loss": 1.9035,
+      "step": 740
+    },
+    {
+      "epoch": 0.810958904109589,
+      "eval_loss": 2.092885732650757,
+      "eval_runtime": 36.7991,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 740
+    },
+    {
+      "epoch": 0.812054794520548,
+      "grad_norm": 2.337340831756592,
+      "learning_rate": 4.080016246953696e-06,
+      "loss": 1.9747,
+      "step": 741
+    },
+    {
+      "epoch": 0.8131506849315069,
+      "grad_norm": 2.262801170349121,
+      "learning_rate": 4.077985377741674e-06,
+      "loss": 1.8817,
+      "step": 742
+    },
+    {
+      "epoch": 0.8142465753424658,
+      "grad_norm": 2.0494258403778076,
+      "learning_rate": 4.075954508529651e-06,
+      "loss": 1.9493,
+      "step": 743
+    },
+    {
+      "epoch": 0.8153424657534246,
+      "grad_norm": 2.085829019546509,
+      "learning_rate": 4.073923639317629e-06,
+      "loss": 1.9611,
+      "step": 744
+    },
+    {
+      "epoch": 0.8164383561643835,
+      "grad_norm": 2.105806827545166,
+      "learning_rate": 4.071892770105605e-06,
+      "loss": 1.9779,
+      "step": 745
+    },
+    {
+      "epoch": 0.8175342465753425,
+      "grad_norm": 2.2664732933044434,
+      "learning_rate": 4.069861900893583e-06,
+      "loss": 1.9009,
+      "step": 746
+    },
+    {
+      "epoch": 0.8186301369863014,
+      "grad_norm": 1.9899768829345703,
+      "learning_rate": 4.06783103168156e-06,
+      "loss": 1.8585,
+      "step": 747
+    },
+    {
+      "epoch": 0.8197260273972603,
+      "grad_norm": 2.242236614227295,
+      "learning_rate": 4.065800162469538e-06,
+      "loss": 1.8904,
+      "step": 748
+    },
+    {
+      "epoch": 0.8208219178082192,
+      "grad_norm": 2.4481139183044434,
+      "learning_rate": 4.063769293257514e-06,
+      "loss": 1.8922,
+      "step": 749
+    },
+    {
+      "epoch": 0.821917808219178,
+      "grad_norm": 2.231722354888916,
+      "learning_rate": 4.061738424045492e-06,
+      "loss": 1.9292,
+      "step": 750
+    },
+    {
+      "epoch": 0.821917808219178,
+      "eval_loss": 2.0909852981567383,
+      "eval_runtime": 36.8255,
+      "eval_samples_per_second": 27.155,
+      "eval_steps_per_second": 1.141,
+      "step": 750
+    },
+    {
+      "epoch": 0.823013698630137,
+      "grad_norm": 2.323070764541626,
+      "learning_rate": 4.059707554833469e-06,
+      "loss": 1.9053,
+      "step": 751
+    },
+    {
+      "epoch": 0.8241095890410959,
+      "grad_norm": 2.137535572052002,
+      "learning_rate": 4.057676685621446e-06,
+      "loss": 1.9272,
+      "step": 752
+    },
+    {
+      "epoch": 0.8252054794520548,
+      "grad_norm": 2.2087643146514893,
+      "learning_rate": 4.0556458164094235e-06,
+      "loss": 1.9541,
+      "step": 753
+    },
+    {
+      "epoch": 0.8263013698630137,
+      "grad_norm": 2.04811692237854,
+      "learning_rate": 4.053614947197401e-06,
+      "loss": 1.922,
+      "step": 754
+    },
+    {
+      "epoch": 0.8273972602739726,
+      "grad_norm": 2.062622308731079,
+      "learning_rate": 4.051584077985378e-06,
+      "loss": 1.8388,
+      "step": 755
+    },
+    {
+      "epoch": 0.8284931506849315,
+      "grad_norm": 2.14631986618042,
+      "learning_rate": 4.049553208773355e-06,
+      "loss": 1.9153,
+      "step": 756
+    },
+    {
+      "epoch": 0.8295890410958904,
+      "grad_norm": 2.353529214859009,
+      "learning_rate": 4.047522339561333e-06,
+      "loss": 1.9623,
+      "step": 757
+    },
+    {
+      "epoch": 0.8306849315068493,
+      "grad_norm": 2.122342348098755,
+      "learning_rate": 4.04549147034931e-06,
+      "loss": 1.9095,
+      "step": 758
+    },
+    {
+      "epoch": 0.8317808219178082,
+      "grad_norm": 2.2540230751037598,
+      "learning_rate": 4.043460601137287e-06,
+      "loss": 1.9068,
+      "step": 759
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "grad_norm": 2.164384365081787,
+      "learning_rate": 4.041429731925264e-06,
+      "loss": 2.0239,
+      "step": 760
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "eval_loss": 2.0904648303985596,
+      "eval_runtime": 36.7617,
+      "eval_samples_per_second": 27.202,
+      "eval_steps_per_second": 1.142,
+      "step": 760
+    },
+    {
+      "epoch": 0.8339726027397261,
+      "grad_norm": 2.2487905025482178,
+      "learning_rate": 4.039398862713242e-06,
+      "loss": 1.9109,
+      "step": 761
+    },
+    {
+      "epoch": 0.8350684931506849,
+      "grad_norm": 2.0947930812835693,
+      "learning_rate": 4.037367993501219e-06,
+      "loss": 1.9133,
+      "step": 762
+    },
+    {
+      "epoch": 0.8361643835616438,
+      "grad_norm": 2.2259862422943115,
+      "learning_rate": 4.035337124289196e-06,
+      "loss": 1.9821,
+      "step": 763
+    },
+    {
+      "epoch": 0.8372602739726027,
+      "grad_norm": 2.0616455078125,
+      "learning_rate": 4.0333062550771735e-06,
+      "loss": 1.9037,
+      "step": 764
+    },
+    {
+      "epoch": 0.8383561643835616,
+      "grad_norm": 2.2704474925994873,
+      "learning_rate": 4.031275385865151e-06,
+      "loss": 1.8664,
+      "step": 765
+    },
+    {
+      "epoch": 0.8394520547945206,
+      "grad_norm": 2.0059332847595215,
+      "learning_rate": 4.029244516653128e-06,
+      "loss": 1.908,
+      "step": 766
+    },
+    {
+      "epoch": 0.8405479452054795,
+      "grad_norm": 2.0462095737457275,
+      "learning_rate": 4.027213647441105e-06,
+      "loss": 1.9397,
+      "step": 767
+    },
+    {
+      "epoch": 0.8416438356164384,
+      "grad_norm": 2.150207996368408,
+      "learning_rate": 4.025182778229083e-06,
+      "loss": 1.8505,
+      "step": 768
+    },
+    {
+      "epoch": 0.8427397260273972,
+      "grad_norm": 2.174863815307617,
+      "learning_rate": 4.02315190901706e-06,
+      "loss": 1.8895,
+      "step": 769
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "grad_norm": 2.1907355785369873,
+      "learning_rate": 4.021121039805037e-06,
+      "loss": 1.9156,
+      "step": 770
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "eval_loss": 2.0900955200195312,
+      "eval_runtime": 36.8004,
+      "eval_samples_per_second": 27.174,
+      "eval_steps_per_second": 1.141,
+      "step": 770
+    },
+    {
+      "epoch": 0.8449315068493151,
+      "grad_norm": 2.2831835746765137,
+      "learning_rate": 4.0190901705930144e-06,
+      "loss": 1.9115,
+      "step": 771
+    },
+    {
+      "epoch": 0.846027397260274,
+      "grad_norm": 2.039520263671875,
+      "learning_rate": 4.017059301380992e-06,
+      "loss": 2.0047,
+      "step": 772
+    },
+    {
+      "epoch": 0.8471232876712329,
+      "grad_norm": 2.147209405899048,
+      "learning_rate": 4.015028432168969e-06,
+      "loss": 1.8845,
+      "step": 773
+    },
+    {
+      "epoch": 0.8482191780821918,
+      "grad_norm": 2.168997287750244,
+      "learning_rate": 4.012997562956945e-06,
+      "loss": 1.9606,
+      "step": 774
+    },
+    {
+      "epoch": 0.8493150684931506,
+      "grad_norm": 2.038966417312622,
+      "learning_rate": 4.0109666937449235e-06,
+      "loss": 1.8734,
+      "step": 775
+    },
+    {
+      "epoch": 0.8504109589041096,
+      "grad_norm": 2.0531768798828125,
+      "learning_rate": 4.0089358245329e-06,
+      "loss": 1.889,
+      "step": 776
+    },
+    {
+      "epoch": 0.8515068493150685,
+      "grad_norm": 2.348008155822754,
+      "learning_rate": 4.006904955320878e-06,
+      "loss": 1.8972,
+      "step": 777
+    },
+    {
+      "epoch": 0.8526027397260274,
+      "grad_norm": 2.1647253036499023,
+      "learning_rate": 4.0048740861088545e-06,
+      "loss": 1.8652,
+      "step": 778
+    },
+    {
+      "epoch": 0.8536986301369863,
+      "grad_norm": 2.1971640586853027,
+      "learning_rate": 4.002843216896833e-06,
+      "loss": 1.9495,
+      "step": 779
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "grad_norm": 2.169632911682129,
+      "learning_rate": 4.000812347684809e-06,
+      "loss": 1.9307,
+      "step": 780
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "eval_loss": 2.087860107421875,
+      "eval_runtime": 36.7982,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 780
+    },
+    {
+      "epoch": 0.8558904109589041,
+      "grad_norm": 2.4424400329589844,
+      "learning_rate": 3.998781478472787e-06,
+      "loss": 1.8849,
+      "step": 781
+    },
+    {
+      "epoch": 0.856986301369863,
+      "grad_norm": 2.2226672172546387,
+      "learning_rate": 3.996750609260764e-06,
+      "loss": 1.8975,
+      "step": 782
+    },
+    {
+      "epoch": 0.8580821917808219,
+      "grad_norm": 2.1898491382598877,
+      "learning_rate": 3.994719740048742e-06,
+      "loss": 1.8751,
+      "step": 783
+    },
+    {
+      "epoch": 0.8591780821917808,
+      "grad_norm": 2.466963052749634,
+      "learning_rate": 3.992688870836718e-06,
+      "loss": 2.0155,
+      "step": 784
+    },
+    {
+      "epoch": 0.8602739726027397,
+      "grad_norm": 2.184039831161499,
+      "learning_rate": 3.990658001624696e-06,
+      "loss": 1.9133,
+      "step": 785
+    },
+    {
+      "epoch": 0.8613698630136987,
+      "grad_norm": 2.089228868484497,
+      "learning_rate": 3.988627132412673e-06,
+      "loss": 1.8707,
+      "step": 786
+    },
+    {
+      "epoch": 0.8624657534246576,
+      "grad_norm": 2.0700666904449463,
+      "learning_rate": 3.98659626320065e-06,
+      "loss": 1.9487,
+      "step": 787
+    },
+    {
+      "epoch": 0.8635616438356164,
+      "grad_norm": 2.1495554447174072,
+      "learning_rate": 3.984565393988627e-06,
+      "loss": 1.9037,
+      "step": 788
+    },
+    {
+      "epoch": 0.8646575342465753,
+      "grad_norm": 2.0916988849639893,
+      "learning_rate": 3.9825345247766046e-06,
+      "loss": 1.9096,
+      "step": 789
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "grad_norm": 2.072387933731079,
+      "learning_rate": 3.980503655564582e-06,
+      "loss": 1.9139,
+      "step": 790
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "eval_loss": 2.0876142978668213,
+      "eval_runtime": 37.1068,
+      "eval_samples_per_second": 26.949,
+      "eval_steps_per_second": 1.132,
+      "step": 790
+    },
+    {
+      "epoch": 0.8668493150684932,
+      "grad_norm": 2.1206343173980713,
+      "learning_rate": 3.978472786352559e-06,
+      "loss": 1.9331,
+      "step": 791
+    },
+    {
+      "epoch": 0.8679452054794521,
+      "grad_norm": 2.20263934135437,
+      "learning_rate": 3.976441917140536e-06,
+      "loss": 1.9297,
+      "step": 792
+    },
+    {
+      "epoch": 0.869041095890411,
+      "grad_norm": 2.027554750442505,
+      "learning_rate": 3.974411047928514e-06,
+      "loss": 1.919,
+      "step": 793
+    },
+    {
+      "epoch": 0.8701369863013698,
+      "grad_norm": 2.0291616916656494,
+      "learning_rate": 3.972380178716491e-06,
+      "loss": 1.9682,
+      "step": 794
+    },
+    {
+      "epoch": 0.8712328767123287,
+      "grad_norm": 2.2719762325286865,
+      "learning_rate": 3.970349309504468e-06,
+      "loss": 1.9407,
+      "step": 795
+    },
+    {
+      "epoch": 0.8723287671232877,
+      "grad_norm": 2.0452375411987305,
+      "learning_rate": 3.9683184402924455e-06,
+      "loss": 1.9487,
+      "step": 796
+    },
+    {
+      "epoch": 0.8734246575342466,
+      "grad_norm": 2.170441150665283,
+      "learning_rate": 3.966287571080423e-06,
+      "loss": 1.8651,
+      "step": 797
+    },
+    {
+      "epoch": 0.8745205479452055,
+      "grad_norm": 2.3370516300201416,
+      "learning_rate": 3.9642567018684e-06,
+      "loss": 1.8599,
+      "step": 798
+    },
+    {
+      "epoch": 0.8756164383561644,
+      "grad_norm": 2.336737632751465,
+      "learning_rate": 3.962225832656377e-06,
+      "loss": 1.9304,
+      "step": 799
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "grad_norm": 1.9708369970321655,
+      "learning_rate": 3.960194963444355e-06,
+      "loss": 1.9125,
+      "step": 800
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "eval_loss": 2.0812273025512695,
+      "eval_runtime": 37.0752,
+      "eval_samples_per_second": 26.972,
+      "eval_steps_per_second": 1.133,
+      "step": 800
+    },
+    {
+      "epoch": 0.8778082191780822,
+      "grad_norm": 2.154219627380371,
+      "learning_rate": 3.958164094232332e-06,
+      "loss": 1.8304,
+      "step": 801
+    },
+    {
+      "epoch": 0.8789041095890411,
+      "grad_norm": 2.237786054611206,
+      "learning_rate": 3.956133225020309e-06,
+      "loss": 1.9147,
+      "step": 802
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 2.07143235206604,
+      "learning_rate": 3.954102355808286e-06,
+      "loss": 1.9128,
+      "step": 803
+    },
+    {
+      "epoch": 0.8810958904109589,
+      "grad_norm": 2.2780392169952393,
+      "learning_rate": 3.952071486596264e-06,
+      "loss": 1.9145,
+      "step": 804
+    },
+    {
+      "epoch": 0.8821917808219178,
+      "grad_norm": 2.2396292686462402,
+      "learning_rate": 3.950040617384241e-06,
+      "loss": 1.9049,
+      "step": 805
+    },
+    {
+      "epoch": 0.8832876712328767,
+      "grad_norm": 2.426511287689209,
+      "learning_rate": 3.948009748172218e-06,
+      "loss": 1.8984,
+      "step": 806
+    },
+    {
+      "epoch": 0.8843835616438356,
+      "grad_norm": 2.1874325275421143,
+      "learning_rate": 3.9459788789601955e-06,
+      "loss": 1.8888,
+      "step": 807
+    },
+    {
+      "epoch": 0.8854794520547945,
+      "grad_norm": 2.038098096847534,
+      "learning_rate": 3.943948009748173e-06,
+      "loss": 1.9582,
+      "step": 808
+    },
+    {
+      "epoch": 0.8865753424657534,
+      "grad_norm": 2.3748130798339844,
+      "learning_rate": 3.94191714053615e-06,
+      "loss": 1.9091,
+      "step": 809
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "grad_norm": 2.2457494735717773,
+      "learning_rate": 3.939886271324127e-06,
+      "loss": 1.9067,
+      "step": 810
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "eval_loss": 2.0766079425811768,
+      "eval_runtime": 36.9896,
+      "eval_samples_per_second": 27.035,
+      "eval_steps_per_second": 1.135,
+      "step": 810
+    },
+    {
+      "epoch": 0.8887671232876713,
+      "grad_norm": 2.156644821166992,
+      "learning_rate": 3.937855402112104e-06,
+      "loss": 1.9347,
+      "step": 811
+    },
+    {
+      "epoch": 0.8898630136986302,
+      "grad_norm": 2.3991472721099854,
+      "learning_rate": 3.935824532900082e-06,
+      "loss": 1.8645,
+      "step": 812
+    },
+    {
+      "epoch": 0.890958904109589,
+      "grad_norm": 2.252729654312134,
+      "learning_rate": 3.933793663688058e-06,
+      "loss": 1.8537,
+      "step": 813
+    },
+    {
+      "epoch": 0.8920547945205479,
+      "grad_norm": 2.2917964458465576,
+      "learning_rate": 3.9317627944760365e-06,
+      "loss": 1.9271,
+      "step": 814
+    },
+    {
+      "epoch": 0.8931506849315068,
+      "grad_norm": 2.217310667037964,
+      "learning_rate": 3.929731925264013e-06,
+      "loss": 1.9406,
+      "step": 815
+    },
+    {
+      "epoch": 0.8942465753424658,
+      "grad_norm": 2.167931079864502,
+      "learning_rate": 3.927701056051991e-06,
+      "loss": 1.8554,
+      "step": 816
+    },
+    {
+      "epoch": 0.8953424657534247,
+      "grad_norm": 2.18337082862854,
+      "learning_rate": 3.9256701868399674e-06,
+      "loss": 1.8795,
+      "step": 817
+    },
+    {
+      "epoch": 0.8964383561643836,
+      "grad_norm": 2.2348973751068115,
+      "learning_rate": 3.9236393176279456e-06,
+      "loss": 1.8831,
+      "step": 818
+    },
+    {
+      "epoch": 0.8975342465753424,
+      "grad_norm": 2.1822657585144043,
+      "learning_rate": 3.921608448415922e-06,
+      "loss": 1.8728,
+      "step": 819
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "grad_norm": 2.5253243446350098,
+      "learning_rate": 3.9195775792039e-06,
+      "loss": 1.8633,
+      "step": 820
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "eval_loss": 2.0769834518432617,
+      "eval_runtime": 36.838,
+      "eval_samples_per_second": 27.146,
+      "eval_steps_per_second": 1.14,
+      "step": 820
+    },
+    {
+      "epoch": 0.8997260273972603,
+      "grad_norm": 2.2109973430633545,
+      "learning_rate": 3.9175467099918765e-06,
+      "loss": 1.912,
+      "step": 821
+    },
+    {
+      "epoch": 0.9008219178082192,
+      "grad_norm": 2.053692102432251,
+      "learning_rate": 3.915515840779855e-06,
+      "loss": 1.8928,
+      "step": 822
+    },
+    {
+      "epoch": 0.9019178082191781,
+      "grad_norm": 2.1418213844299316,
+      "learning_rate": 3.913484971567831e-06,
+      "loss": 1.9399,
+      "step": 823
+    },
+    {
+      "epoch": 0.903013698630137,
+      "grad_norm": 2.3615458011627197,
+      "learning_rate": 3.911454102355808e-06,
+      "loss": 1.9293,
+      "step": 824
+    },
+    {
+      "epoch": 0.9041095890410958,
+      "grad_norm": 2.098254680633545,
+      "learning_rate": 3.909423233143786e-06,
+      "loss": 1.8301,
+      "step": 825
+    },
+    {
+      "epoch": 0.9052054794520548,
+      "grad_norm": 2.4473679065704346,
+      "learning_rate": 3.907392363931763e-06,
+      "loss": 1.9254,
+      "step": 826
+    },
+    {
+      "epoch": 0.9063013698630137,
+      "grad_norm": 2.252843141555786,
+      "learning_rate": 3.90536149471974e-06,
+      "loss": 1.9155,
+      "step": 827
+    },
+    {
+      "epoch": 0.9073972602739726,
+      "grad_norm": 2.1574337482452393,
+      "learning_rate": 3.9033306255077175e-06,
+      "loss": 1.9809,
+      "step": 828
+    },
+    {
+      "epoch": 0.9084931506849315,
+      "grad_norm": 2.2899720668792725,
+      "learning_rate": 3.901299756295695e-06,
+      "loss": 2.0045,
+      "step": 829
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "grad_norm": 2.227285146713257,
+      "learning_rate": 3.899268887083672e-06,
+      "loss": 2.026,
+      "step": 830
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "eval_loss": 2.0764760971069336,
+      "eval_runtime": 36.9195,
+      "eval_samples_per_second": 27.086,
+      "eval_steps_per_second": 1.138,
+      "step": 830
+    },
+    {
+      "epoch": 0.9106849315068493,
+      "grad_norm": 2.1935312747955322,
+      "learning_rate": 3.897238017871649e-06,
+      "loss": 1.8922,
+      "step": 831
+    },
+    {
+      "epoch": 0.9117808219178082,
+      "grad_norm": 2.129692316055298,
+      "learning_rate": 3.8952071486596266e-06,
+      "loss": 1.8981,
+      "step": 832
+    },
+    {
+      "epoch": 0.9128767123287671,
+      "grad_norm": 2.4435646533966064,
+      "learning_rate": 3.893176279447604e-06,
+      "loss": 1.91,
+      "step": 833
+    },
+    {
+      "epoch": 0.913972602739726,
+      "grad_norm": 2.1556732654571533,
+      "learning_rate": 3.891145410235581e-06,
+      "loss": 1.8688,
+      "step": 834
+    },
+    {
+      "epoch": 0.915068493150685,
+      "grad_norm": 2.175891160964966,
+      "learning_rate": 3.889114541023558e-06,
+      "loss": 1.9777,
+      "step": 835
+    },
+    {
+      "epoch": 0.9161643835616439,
+      "grad_norm": 2.220466375350952,
+      "learning_rate": 3.887083671811536e-06,
+      "loss": 1.9152,
+      "step": 836
+    },
+    {
+      "epoch": 0.9172602739726028,
+      "grad_norm": 2.331397294998169,
+      "learning_rate": 3.885052802599513e-06,
+      "loss": 1.892,
+      "step": 837
+    },
+    {
+      "epoch": 0.9183561643835616,
+      "grad_norm": 2.2434773445129395,
+      "learning_rate": 3.88302193338749e-06,
+      "loss": 1.841,
+      "step": 838
+    },
+    {
+      "epoch": 0.9194520547945205,
+      "grad_norm": 2.3248002529144287,
+      "learning_rate": 3.8809910641754675e-06,
+      "loss": 1.9657,
+      "step": 839
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "grad_norm": 2.090252637863159,
+      "learning_rate": 3.878960194963445e-06,
+      "loss": 1.9121,
+      "step": 840
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "eval_loss": 2.0798065662384033,
+      "eval_runtime": 36.8557,
+      "eval_samples_per_second": 27.133,
+      "eval_steps_per_second": 1.14,
+      "step": 840
+    },
+    {
+      "epoch": 0.9216438356164384,
+      "grad_norm": 2.6968672275543213,
+      "learning_rate": 3.876929325751422e-06,
+      "loss": 1.9003,
+      "step": 841
+    },
+    {
+      "epoch": 0.9227397260273973,
+      "grad_norm": 2.356321096420288,
+      "learning_rate": 3.874898456539399e-06,
+      "loss": 1.9471,
+      "step": 842
+    },
+    {
+      "epoch": 0.9238356164383562,
+      "grad_norm": 2.2678606510162354,
+      "learning_rate": 3.872867587327377e-06,
+      "loss": 1.8488,
+      "step": 843
+    },
+    {
+      "epoch": 0.924931506849315,
+      "grad_norm": 2.1122310161590576,
+      "learning_rate": 3.870836718115354e-06,
+      "loss": 1.9247,
+      "step": 844
+    },
+    {
+      "epoch": 0.9260273972602739,
+      "grad_norm": 2.1219136714935303,
+      "learning_rate": 3.868805848903331e-06,
+      "loss": 1.8829,
+      "step": 845
+    },
+    {
+      "epoch": 0.9271232876712329,
+      "grad_norm": 2.175607204437256,
+      "learning_rate": 3.8667749796913084e-06,
+      "loss": 1.887,
+      "step": 846
+    },
+    {
+      "epoch": 0.9282191780821918,
+      "grad_norm": 2.156214714050293,
+      "learning_rate": 3.864744110479286e-06,
+      "loss": 1.9119,
+      "step": 847
+    },
+    {
+      "epoch": 0.9293150684931507,
+      "grad_norm": 2.158461570739746,
+      "learning_rate": 3.862713241267262e-06,
+      "loss": 1.9092,
+      "step": 848
+    },
+    {
+      "epoch": 0.9304109589041096,
+      "grad_norm": 2.180551767349243,
+      "learning_rate": 3.86068237205524e-06,
+      "loss": 1.946,
+      "step": 849
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "grad_norm": 2.1030542850494385,
+      "learning_rate": 3.858651502843217e-06,
+      "loss": 1.9484,
+      "step": 850
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "eval_loss": 2.0767791271209717,
+      "eval_runtime": 36.8047,
+      "eval_samples_per_second": 27.17,
+      "eval_steps_per_second": 1.141,
+      "step": 850
+    },
+    {
+      "epoch": 0.9326027397260274,
+      "grad_norm": 2.240769386291504,
+      "learning_rate": 3.856620633631195e-06,
+      "loss": 1.9114,
+      "step": 851
+    },
+    {
+      "epoch": 0.9336986301369863,
+      "grad_norm": 2.1964619159698486,
+      "learning_rate": 3.854589764419171e-06,
+      "loss": 1.8339,
+      "step": 852
+    },
+    {
+      "epoch": 0.9347945205479452,
+      "grad_norm": 2.2875845432281494,
+      "learning_rate": 3.852558895207149e-06,
+      "loss": 1.8942,
+      "step": 853
+    },
+    {
+      "epoch": 0.9358904109589041,
+      "grad_norm": 2.1462507247924805,
+      "learning_rate": 3.850528025995126e-06,
+      "loss": 1.9838,
+      "step": 854
+    },
+    {
+      "epoch": 0.936986301369863,
+      "grad_norm": 2.0982913970947266,
+      "learning_rate": 3.848497156783104e-06,
+      "loss": 1.9351,
+      "step": 855
+    },
+    {
+      "epoch": 0.938082191780822,
+      "grad_norm": 2.260263442993164,
+      "learning_rate": 3.84646628757108e-06,
+      "loss": 1.8745,
+      "step": 856
+    },
+    {
+      "epoch": 0.9391780821917808,
+      "grad_norm": 2.091778039932251,
+      "learning_rate": 3.8444354183590585e-06,
+      "loss": 1.8888,
+      "step": 857
+    },
+    {
+      "epoch": 0.9402739726027397,
+      "grad_norm": 2.054974317550659,
+      "learning_rate": 3.842404549147035e-06,
+      "loss": 1.8801,
+      "step": 858
+    },
+    {
+      "epoch": 0.9413698630136986,
+      "grad_norm": 2.079641580581665,
+      "learning_rate": 3.840373679935013e-06,
+      "loss": 1.8879,
+      "step": 859
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "grad_norm": 2.3441810607910156,
+      "learning_rate": 3.8383428107229894e-06,
+      "loss": 1.8571,
+      "step": 860
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "eval_loss": 2.0766854286193848,
+      "eval_runtime": 36.8414,
+      "eval_samples_per_second": 27.143,
+      "eval_steps_per_second": 1.14,
+      "step": 860
+    },
+    {
+      "epoch": 0.9435616438356165,
+      "grad_norm": 2.123061180114746,
+      "learning_rate": 3.836311941510967e-06,
+      "loss": 1.8922,
+      "step": 861
+    },
+    {
+      "epoch": 0.9446575342465754,
+      "grad_norm": 2.1573386192321777,
+      "learning_rate": 3.834281072298944e-06,
+      "loss": 1.8713,
+      "step": 862
+    },
+    {
+      "epoch": 0.9457534246575342,
+      "grad_norm": 2.225186824798584,
+      "learning_rate": 3.832250203086921e-06,
+      "loss": 1.925,
+      "step": 863
+    },
+    {
+      "epoch": 0.9468493150684931,
+      "grad_norm": 2.1539998054504395,
+      "learning_rate": 3.8302193338748985e-06,
+      "loss": 1.8721,
+      "step": 864
+    },
+    {
+      "epoch": 0.947945205479452,
+      "grad_norm": 2.075580358505249,
+      "learning_rate": 3.828188464662876e-06,
+      "loss": 1.9489,
+      "step": 865
+    },
+    {
+      "epoch": 0.949041095890411,
+      "grad_norm": 2.1789908409118652,
+      "learning_rate": 3.826157595450853e-06,
+      "loss": 1.9883,
+      "step": 866
+    },
+    {
+      "epoch": 0.9501369863013699,
+      "grad_norm": 2.0838615894317627,
+      "learning_rate": 3.82412672623883e-06,
+      "loss": 1.9149,
+      "step": 867
+    },
+    {
+      "epoch": 0.9512328767123288,
+      "grad_norm": 2.1321213245391846,
+      "learning_rate": 3.822095857026808e-06,
+      "loss": 1.8505,
+      "step": 868
+    },
+    {
+      "epoch": 0.9523287671232876,
+      "grad_norm": 2.1419479846954346,
+      "learning_rate": 3.820064987814785e-06,
+      "loss": 1.9226,
+      "step": 869
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "grad_norm": 2.290163040161133,
+      "learning_rate": 3.818034118602762e-06,
+      "loss": 1.8737,
+      "step": 870
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "eval_loss": 2.073598623275757,
+      "eval_runtime": 36.7896,
+      "eval_samples_per_second": 27.182,
+      "eval_steps_per_second": 1.142,
+      "step": 870
+    },
+    {
+      "epoch": 0.9545205479452055,
+      "grad_norm": 2.1951074600219727,
+      "learning_rate": 3.8160032493907395e-06,
+      "loss": 1.8699,
+      "step": 871
+    },
+    {
+      "epoch": 0.9556164383561644,
+      "grad_norm": 2.25960111618042,
+      "learning_rate": 3.813972380178717e-06,
+      "loss": 1.8813,
+      "step": 872
+    },
+    {
+      "epoch": 0.9567123287671233,
+      "grad_norm": 2.4376907348632812,
+      "learning_rate": 3.811941510966694e-06,
+      "loss": 1.8782,
+      "step": 873
+    },
+    {
+      "epoch": 0.9578082191780822,
+      "grad_norm": 1.980772614479065,
+      "learning_rate": 3.809910641754671e-06,
+      "loss": 1.9205,
+      "step": 874
+    },
+    {
+      "epoch": 0.958904109589041,
+      "grad_norm": 2.0699028968811035,
+      "learning_rate": 3.8078797725426486e-06,
+      "loss": 1.8674,
+      "step": 875
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 2.2765867710113525,
+      "learning_rate": 3.8058489033306254e-06,
+      "loss": 1.81,
+      "step": 876
+    },
+    {
+      "epoch": 0.9610958904109589,
+      "grad_norm": 2.0998799800872803,
+      "learning_rate": 3.803818034118603e-06,
+      "loss": 1.9061,
+      "step": 877
+    },
+    {
+      "epoch": 0.9621917808219178,
+      "grad_norm": 2.2362914085388184,
+      "learning_rate": 3.80178716490658e-06,
+      "loss": 1.8798,
+      "step": 878
+    },
+    {
+      "epoch": 0.9632876712328767,
+      "grad_norm": 2.0789835453033447,
+      "learning_rate": 3.7997562956945577e-06,
+      "loss": 1.9491,
+      "step": 879
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "grad_norm": 2.0458688735961914,
+      "learning_rate": 3.7977254264825345e-06,
+      "loss": 1.931,
+      "step": 880
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "eval_loss": 2.0672221183776855,
+      "eval_runtime": 36.8218,
+      "eval_samples_per_second": 27.158,
+      "eval_steps_per_second": 1.141,
+      "step": 880
+    },
+    {
+      "epoch": 0.9654794520547946,
+      "grad_norm": 2.361626625061035,
+      "learning_rate": 3.7956945572705122e-06,
+      "loss": 1.8815,
+      "step": 881
+    },
+    {
+      "epoch": 0.9665753424657534,
+      "grad_norm": 2.294160842895508,
+      "learning_rate": 3.793663688058489e-06,
+      "loss": 1.8649,
+      "step": 882
+    },
+    {
+      "epoch": 0.9676712328767123,
+      "grad_norm": 2.2242612838745117,
+      "learning_rate": 3.791632818846467e-06,
+      "loss": 1.9535,
+      "step": 883
+    },
+    {
+      "epoch": 0.9687671232876712,
+      "grad_norm": 2.220269203186035,
+      "learning_rate": 3.7896019496344436e-06,
+      "loss": 1.9506,
+      "step": 884
+    },
+    {
+      "epoch": 0.9698630136986301,
+      "grad_norm": 2.112518072128296,
+      "learning_rate": 3.787571080422421e-06,
+      "loss": 1.9487,
+      "step": 885
+    },
+    {
+      "epoch": 0.9709589041095891,
+      "grad_norm": 2.17695951461792,
+      "learning_rate": 3.785540211210398e-06,
+      "loss": 1.9467,
+      "step": 886
+    },
+    {
+      "epoch": 0.972054794520548,
+      "grad_norm": 2.0990567207336426,
+      "learning_rate": 3.7835093419983755e-06,
+      "loss": 1.9216,
+      "step": 887
+    },
+    {
+      "epoch": 0.9731506849315068,
+      "grad_norm": 2.222566604614258,
+      "learning_rate": 3.7814784727863527e-06,
+      "loss": 1.9043,
+      "step": 888
+    },
+    {
+      "epoch": 0.9742465753424657,
+      "grad_norm": 2.09890079498291,
+      "learning_rate": 3.77944760357433e-06,
+      "loss": 1.884,
+      "step": 889
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "grad_norm": 2.136141061782837,
+      "learning_rate": 3.7774167343623073e-06,
+      "loss": 2.0032,
+      "step": 890
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "eval_loss": 2.0665438175201416,
+      "eval_runtime": 37.1207,
+      "eval_samples_per_second": 26.939,
+      "eval_steps_per_second": 1.131,
+      "step": 890
+    },
+    {
+      "epoch": 0.9764383561643836,
+      "grad_norm": 2.267444372177124,
+      "learning_rate": 3.7753858651502846e-06,
+      "loss": 1.947,
+      "step": 891
+    },
+    {
+      "epoch": 0.9775342465753425,
+      "grad_norm": 2.1141815185546875,
+      "learning_rate": 3.773354995938262e-06,
+      "loss": 1.8705,
+      "step": 892
+    },
+    {
+      "epoch": 0.9786301369863014,
+      "grad_norm": 2.1036782264709473,
+      "learning_rate": 3.771324126726239e-06,
+      "loss": 1.9381,
+      "step": 893
+    },
+    {
+      "epoch": 0.9797260273972602,
+      "grad_norm": 2.081982135772705,
+      "learning_rate": 3.7692932575142164e-06,
+      "loss": 1.9079,
+      "step": 894
+    },
+    {
+      "epoch": 0.9808219178082191,
+      "grad_norm": 2.0815446376800537,
+      "learning_rate": 3.7672623883021937e-06,
+      "loss": 1.9417,
+      "step": 895
+    },
+    {
+      "epoch": 0.9819178082191781,
+      "grad_norm": 2.0531511306762695,
+      "learning_rate": 3.765231519090171e-06,
+      "loss": 1.876,
+      "step": 896
+    },
+    {
+      "epoch": 0.983013698630137,
+      "grad_norm": 2.212940216064453,
+      "learning_rate": 3.7632006498781482e-06,
+      "loss": 1.8008,
+      "step": 897
+    },
+    {
+      "epoch": 0.9841095890410959,
+      "grad_norm": 2.1541531085968018,
+      "learning_rate": 3.761169780666125e-06,
+      "loss": 1.9039,
+      "step": 898
+    },
+    {
+      "epoch": 0.9852054794520548,
+      "grad_norm": 2.113884925842285,
+      "learning_rate": 3.7591389114541028e-06,
+      "loss": 1.8562,
+      "step": 899
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "grad_norm": 2.1503701210021973,
+      "learning_rate": 3.7571080422420796e-06,
+      "loss": 1.9052,
+      "step": 900
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "eval_loss": 2.0664408206939697,
+      "eval_runtime": 37.0876,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 900
+    },
+    {
+      "epoch": 0.9873972602739726,
+      "grad_norm": 2.1584348678588867,
+      "learning_rate": 3.7550771730300573e-06,
+      "loss": 1.9296,
+      "step": 901
+    },
+    {
+      "epoch": 0.9884931506849315,
+      "grad_norm": 2.164318084716797,
+      "learning_rate": 3.753046303818034e-06,
+      "loss": 1.8834,
+      "step": 902
+    },
+    {
+      "epoch": 0.9895890410958904,
+      "grad_norm": 2.0222678184509277,
+      "learning_rate": 3.751015434606012e-06,
+      "loss": 1.9426,
+      "step": 903
+    },
+    {
+      "epoch": 0.9906849315068493,
+      "grad_norm": 2.116835832595825,
+      "learning_rate": 3.7489845653939887e-06,
+      "loss": 1.9195,
+      "step": 904
+    },
+    {
+      "epoch": 0.9917808219178083,
+      "grad_norm": 2.2708873748779297,
+      "learning_rate": 3.7469536961819664e-06,
+      "loss": 1.9631,
+      "step": 905
+    },
+    {
+      "epoch": 0.9928767123287672,
+      "grad_norm": 2.1174001693725586,
+      "learning_rate": 3.7449228269699433e-06,
+      "loss": 1.8602,
+      "step": 906
+    },
+    {
+      "epoch": 0.993972602739726,
+      "grad_norm": 2.091830015182495,
+      "learning_rate": 3.742891957757921e-06,
+      "loss": 1.8906,
+      "step": 907
+    },
+    {
+      "epoch": 0.9950684931506849,
+      "grad_norm": 2.007119655609131,
+      "learning_rate": 3.740861088545898e-06,
+      "loss": 1.9122,
+      "step": 908
+    },
+    {
+      "epoch": 0.9961643835616438,
+      "grad_norm": 2.070396661758423,
+      "learning_rate": 3.7388302193338755e-06,
+      "loss": 1.9106,
+      "step": 909
+    },
+    {
+      "epoch": 0.9972602739726028,
+      "grad_norm": 2.080331325531006,
+      "learning_rate": 3.7367993501218524e-06,
+      "loss": 1.934,
+      "step": 910
+    },
+    {
+      "epoch": 0.9972602739726028,
+      "eval_loss": 2.0694408416748047,
+      "eval_runtime": 36.9953,
+      "eval_samples_per_second": 27.03,
+      "eval_steps_per_second": 1.135,
+      "step": 910
+    },
+    {
+      "epoch": 0.9983561643835617,
+      "grad_norm": 2.0920615196228027,
+      "learning_rate": 3.7347684809098297e-06,
+      "loss": 1.8679,
+      "step": 911
+    },
+    {
+      "epoch": 0.9994520547945206,
+      "grad_norm": 2.112732172012329,
+      "learning_rate": 3.732737611697807e-06,
+      "loss": 1.8723,
+      "step": 912
+    },
+    {
+      "epoch": 1.0005479452054795,
+      "grad_norm": 2.1449220180511475,
+      "learning_rate": 3.7307067424857842e-06,
+      "loss": 1.8509,
+      "step": 913
+    },
+    {
+      "epoch": 1.0016438356164383,
+      "grad_norm": 2.2179951667785645,
+      "learning_rate": 3.7286758732737615e-06,
+      "loss": 1.7911,
+      "step": 914
+    },
+    {
+      "epoch": 1.0027397260273974,
+      "grad_norm": 2.1392273902893066,
+      "learning_rate": 3.7266450040617388e-06,
+      "loss": 1.8475,
+      "step": 915
+    },
+    {
+      "epoch": 1.0038356164383562,
+      "grad_norm": 2.190265417098999,
+      "learning_rate": 3.724614134849716e-06,
+      "loss": 1.8394,
+      "step": 916
+    },
+    {
+      "epoch": 1.004931506849315,
+      "grad_norm": 2.1833760738372803,
+      "learning_rate": 3.7225832656376933e-06,
+      "loss": 1.9696,
+      "step": 917
+    },
+    {
+      "epoch": 1.006027397260274,
+      "grad_norm": 2.2504308223724365,
+      "learning_rate": 3.7205523964256706e-06,
+      "loss": 1.8135,
+      "step": 918
+    },
+    {
+      "epoch": 1.0071232876712328,
+      "grad_norm": 2.3909389972686768,
+      "learning_rate": 3.718521527213648e-06,
+      "loss": 1.9331,
+      "step": 919
+    },
+    {
+      "epoch": 1.0082191780821919,
+      "grad_norm": 2.3367602825164795,
+      "learning_rate": 3.716490658001625e-06,
+      "loss": 1.9026,
+      "step": 920
+    },
+    {
+      "epoch": 1.0082191780821919,
+      "eval_loss": 2.066087484359741,
+      "eval_runtime": 36.8701,
+      "eval_samples_per_second": 27.122,
+      "eval_steps_per_second": 1.139,
+      "step": 920
+    },
+    {
+      "epoch": 1.0093150684931507,
+      "grad_norm": 2.1732144355773926,
+      "learning_rate": 3.7144597887896024e-06,
+      "loss": 1.8888,
+      "step": 921
+    },
+    {
+      "epoch": 1.0104109589041095,
+      "grad_norm": 2.4632201194763184,
+      "learning_rate": 3.7124289195775797e-06,
+      "loss": 1.7725,
+      "step": 922
+    },
+    {
+      "epoch": 1.0115068493150685,
+      "grad_norm": 2.4887173175811768,
+      "learning_rate": 3.710398050365557e-06,
+      "loss": 1.9125,
+      "step": 923
+    },
+    {
+      "epoch": 1.0126027397260273,
+      "grad_norm": 2.4684619903564453,
+      "learning_rate": 3.708367181153534e-06,
+      "loss": 1.8829,
+      "step": 924
+    },
+    {
+      "epoch": 1.0136986301369864,
+      "grad_norm": 2.348788022994995,
+      "learning_rate": 3.7063363119415115e-06,
+      "loss": 1.8542,
+      "step": 925
+    },
+    {
+      "epoch": 1.0147945205479452,
+      "grad_norm": 2.4432766437530518,
+      "learning_rate": 3.7043054427294884e-06,
+      "loss": 1.903,
+      "step": 926
+    },
+    {
+      "epoch": 1.0158904109589042,
+      "grad_norm": 2.425910234451294,
+      "learning_rate": 3.702274573517466e-06,
+      "loss": 1.8532,
+      "step": 927
+    },
+    {
+      "epoch": 1.016986301369863,
+      "grad_norm": 2.447777032852173,
+      "learning_rate": 3.700243704305443e-06,
+      "loss": 1.8277,
+      "step": 928
+    },
+    {
+      "epoch": 1.0180821917808218,
+      "grad_norm": 2.1830408573150635,
+      "learning_rate": 3.6982128350934206e-06,
+      "loss": 1.8546,
+      "step": 929
+    },
+    {
+      "epoch": 1.0191780821917809,
+      "grad_norm": 2.2692818641662598,
+      "learning_rate": 3.6961819658813975e-06,
+      "loss": 1.8356,
+      "step": 930
+    },
+    {
+      "epoch": 1.0191780821917809,
+      "eval_loss": 2.0714614391326904,
+      "eval_runtime": 36.8661,
+      "eval_samples_per_second": 27.125,
+      "eval_steps_per_second": 1.139,
+      "step": 930
+    },
+    {
+      "epoch": 1.0202739726027397,
+      "grad_norm": 2.5044169425964355,
+      "learning_rate": 3.694151096669375e-06,
+      "loss": 1.8173,
+      "step": 931
+    },
+    {
+      "epoch": 1.0213698630136987,
+      "grad_norm": 2.419637680053711,
+      "learning_rate": 3.692120227457352e-06,
+      "loss": 1.9418,
+      "step": 932
+    },
+    {
+      "epoch": 1.0224657534246575,
+      "grad_norm": 2.282288074493408,
+      "learning_rate": 3.6900893582453297e-06,
+      "loss": 1.8698,
+      "step": 933
+    },
+    {
+      "epoch": 1.0235616438356165,
+      "grad_norm": 2.3215551376342773,
+      "learning_rate": 3.6880584890333066e-06,
+      "loss": 1.8935,
+      "step": 934
+    },
+    {
+      "epoch": 1.0246575342465754,
+      "grad_norm": 2.5938258171081543,
+      "learning_rate": 3.6860276198212834e-06,
+      "loss": 1.79,
+      "step": 935
+    },
+    {
+      "epoch": 1.0257534246575342,
+      "grad_norm": 2.215024471282959,
+      "learning_rate": 3.683996750609261e-06,
+      "loss": 1.8768,
+      "step": 936
+    },
+    {
+      "epoch": 1.0268493150684932,
+      "grad_norm": 2.239609956741333,
+      "learning_rate": 3.681965881397238e-06,
+      "loss": 1.8887,
+      "step": 937
+    },
+    {
+      "epoch": 1.027945205479452,
+      "grad_norm": 2.3134827613830566,
+      "learning_rate": 3.6799350121852157e-06,
+      "loss": 1.8897,
+      "step": 938
+    },
+    {
+      "epoch": 1.029041095890411,
+      "grad_norm": 2.4100260734558105,
+      "learning_rate": 3.6779041429731925e-06,
+      "loss": 1.909,
+      "step": 939
+    },
+    {
+      "epoch": 1.0301369863013699,
+      "grad_norm": 2.26204776763916,
+      "learning_rate": 3.6758732737611702e-06,
+      "loss": 1.8514,
+      "step": 940
+    },
+    {
+      "epoch": 1.0301369863013699,
+      "eval_loss": 2.0666027069091797,
+      "eval_runtime": 36.8336,
+      "eval_samples_per_second": 27.149,
+      "eval_steps_per_second": 1.14,
+      "step": 940
+    },
+    {
+      "epoch": 1.0312328767123287,
+      "grad_norm": 2.22210955619812,
+      "learning_rate": 3.673842404549147e-06,
+      "loss": 1.8743,
+      "step": 941
+    },
+    {
+      "epoch": 1.0323287671232877,
+      "grad_norm": 2.2371468544006348,
+      "learning_rate": 3.671811535337125e-06,
+      "loss": 1.8573,
+      "step": 942
+    },
+    {
+      "epoch": 1.0334246575342465,
+      "grad_norm": 2.346745014190674,
+      "learning_rate": 3.6697806661251016e-06,
+      "loss": 1.8614,
+      "step": 943
+    },
+    {
+      "epoch": 1.0345205479452055,
+      "grad_norm": 2.4765074253082275,
+      "learning_rate": 3.6677497969130793e-06,
+      "loss": 1.8654,
+      "step": 944
+    },
+    {
+      "epoch": 1.0356164383561643,
+      "grad_norm": 2.3023204803466797,
+      "learning_rate": 3.665718927701056e-06,
+      "loss": 1.9901,
+      "step": 945
+    },
+    {
+      "epoch": 1.0367123287671234,
+      "grad_norm": 2.322421073913574,
+      "learning_rate": 3.663688058489034e-06,
+      "loss": 1.8775,
+      "step": 946
+    },
+    {
+      "epoch": 1.0378082191780822,
+      "grad_norm": 2.2324275970458984,
+      "learning_rate": 3.6616571892770107e-06,
+      "loss": 1.9303,
+      "step": 947
+    },
+    {
+      "epoch": 1.038904109589041,
+      "grad_norm": 2.372067928314209,
+      "learning_rate": 3.659626320064988e-06,
+      "loss": 1.8673,
+      "step": 948
+    },
+    {
+      "epoch": 1.04,
+      "grad_norm": 2.2404983043670654,
+      "learning_rate": 3.6575954508529653e-06,
+      "loss": 1.8025,
+      "step": 949
+    },
+    {
+      "epoch": 1.0410958904109588,
+      "grad_norm": 2.180673599243164,
+      "learning_rate": 3.6555645816409426e-06,
+      "loss": 1.8563,
+      "step": 950
+    },
+    {
+      "epoch": 1.0410958904109588,
+      "eval_loss": 2.061262845993042,
+      "eval_runtime": 36.7727,
+      "eval_samples_per_second": 27.194,
+      "eval_steps_per_second": 1.142,
+      "step": 950
+    },
+    {
+      "epoch": 1.0421917808219179,
+      "grad_norm": 2.299043893814087,
+      "learning_rate": 3.65353371242892e-06,
+      "loss": 1.8453,
+      "step": 951
+    },
+    {
+      "epoch": 1.0432876712328767,
+      "grad_norm": 2.4878089427948,
+      "learning_rate": 3.651502843216897e-06,
+      "loss": 1.8901,
+      "step": 952
+    },
+    {
+      "epoch": 1.0443835616438357,
+      "grad_norm": 2.150404930114746,
+      "learning_rate": 3.6494719740048744e-06,
+      "loss": 1.7371,
+      "step": 953
+    },
+    {
+      "epoch": 1.0454794520547945,
+      "grad_norm": 2.177750587463379,
+      "learning_rate": 3.6474411047928517e-06,
+      "loss": 1.8814,
+      "step": 954
+    },
+    {
+      "epoch": 1.0465753424657533,
+      "grad_norm": 2.230797052383423,
+      "learning_rate": 3.645410235580829e-06,
+      "loss": 1.8323,
+      "step": 955
+    },
+    {
+      "epoch": 1.0476712328767124,
+      "grad_norm": 2.2484378814697266,
+      "learning_rate": 3.6433793663688062e-06,
+      "loss": 1.8749,
+      "step": 956
+    },
+    {
+      "epoch": 1.0487671232876712,
+      "grad_norm": 2.201279640197754,
+      "learning_rate": 3.6413484971567835e-06,
+      "loss": 1.8518,
+      "step": 957
+    },
+    {
+      "epoch": 1.0498630136986302,
+      "grad_norm": 2.1125736236572266,
+      "learning_rate": 3.6393176279447608e-06,
+      "loss": 1.846,
+      "step": 958
+    },
+    {
+      "epoch": 1.050958904109589,
+      "grad_norm": 2.3078789710998535,
+      "learning_rate": 3.637286758732738e-06,
+      "loss": 1.8791,
+      "step": 959
+    },
+    {
+      "epoch": 1.0520547945205478,
+      "grad_norm": 2.2911148071289062,
+      "learning_rate": 3.6352558895207153e-06,
+      "loss": 1.8901,
+      "step": 960
+    },
+    {
+      "epoch": 1.0520547945205478,
+      "eval_loss": 2.0583066940307617,
+      "eval_runtime": 36.7535,
+      "eval_samples_per_second": 27.208,
+      "eval_steps_per_second": 1.143,
+      "step": 960
+    },
+    {
+      "epoch": 1.0531506849315069,
+      "grad_norm": 2.3990273475646973,
+      "learning_rate": 3.633225020308692e-06,
+      "loss": 1.9071,
+      "step": 961
+    },
+    {
+      "epoch": 1.0542465753424657,
+      "grad_norm": 2.358685255050659,
+      "learning_rate": 3.63119415109667e-06,
+      "loss": 1.8982,
+      "step": 962
+    },
+    {
+      "epoch": 1.0553424657534247,
+      "grad_norm": 2.1145992279052734,
+      "learning_rate": 3.6291632818846467e-06,
+      "loss": 1.8649,
+      "step": 963
+    },
+    {
+      "epoch": 1.0564383561643835,
+      "grad_norm": 2.2302935123443604,
+      "learning_rate": 3.6271324126726244e-06,
+      "loss": 1.8911,
+      "step": 964
+    },
+    {
+      "epoch": 1.0575342465753426,
+      "grad_norm": 2.344797372817993,
+      "learning_rate": 3.6251015434606013e-06,
+      "loss": 1.8,
+      "step": 965
+    },
+    {
+      "epoch": 1.0586301369863014,
+      "grad_norm": 2.2514073848724365,
+      "learning_rate": 3.623070674248579e-06,
+      "loss": 1.8128,
+      "step": 966
+    },
+    {
+      "epoch": 1.0597260273972602,
+      "grad_norm": 2.541221857070923,
+      "learning_rate": 3.621039805036556e-06,
+      "loss": 1.8504,
+      "step": 967
+    },
+    {
+      "epoch": 1.0608219178082192,
+      "grad_norm": 2.3211770057678223,
+      "learning_rate": 3.6190089358245335e-06,
+      "loss": 1.8329,
+      "step": 968
+    },
+    {
+      "epoch": 1.061917808219178,
+      "grad_norm": 2.1008756160736084,
+      "learning_rate": 3.6169780666125104e-06,
+      "loss": 1.8972,
+      "step": 969
+    },
+    {
+      "epoch": 1.063013698630137,
+      "grad_norm": 2.4148898124694824,
+      "learning_rate": 3.614947197400488e-06,
+      "loss": 1.8748,
+      "step": 970
+    },
+    {
+      "epoch": 1.063013698630137,
+      "eval_loss": 2.0592916011810303,
+      "eval_runtime": 36.7663,
+      "eval_samples_per_second": 27.199,
+      "eval_steps_per_second": 1.142,
+      "step": 970
+    },
+    {
+      "epoch": 1.0641095890410959,
+      "grad_norm": 2.4116220474243164,
+      "learning_rate": 3.612916328188465e-06,
+      "loss": 1.8915,
+      "step": 971
+    },
+    {
+      "epoch": 1.0652054794520547,
+      "grad_norm": 2.257411241531372,
+      "learning_rate": 3.610885458976442e-06,
+      "loss": 1.7901,
+      "step": 972
+    },
+    {
+      "epoch": 1.0663013698630137,
+      "grad_norm": 2.166875123977661,
+      "learning_rate": 3.6088545897644195e-06,
+      "loss": 1.8969,
+      "step": 973
+    },
+    {
+      "epoch": 1.0673972602739725,
+      "grad_norm": 2.464716911315918,
+      "learning_rate": 3.6068237205523963e-06,
+      "loss": 1.8844,
+      "step": 974
+    },
+    {
+      "epoch": 1.0684931506849316,
+      "grad_norm": 2.337334156036377,
+      "learning_rate": 3.604792851340374e-06,
+      "loss": 1.8192,
+      "step": 975
+    },
+    {
+      "epoch": 1.0695890410958904,
+      "grad_norm": 2.117663621902466,
+      "learning_rate": 3.602761982128351e-06,
+      "loss": 1.8233,
+      "step": 976
+    },
+    {
+      "epoch": 1.0706849315068494,
+      "grad_norm": 2.3353586196899414,
+      "learning_rate": 3.6007311129163286e-06,
+      "loss": 1.9133,
+      "step": 977
+    },
+    {
+      "epoch": 1.0717808219178082,
+      "grad_norm": 2.444279909133911,
+      "learning_rate": 3.5987002437043054e-06,
+      "loss": 1.8803,
+      "step": 978
+    },
+    {
+      "epoch": 1.072876712328767,
+      "grad_norm": 2.3458383083343506,
+      "learning_rate": 3.596669374492283e-06,
+      "loss": 1.8108,
+      "step": 979
+    },
+    {
+      "epoch": 1.073972602739726,
+      "grad_norm": 2.207977771759033,
+      "learning_rate": 3.59463850528026e-06,
+      "loss": 1.8423,
+      "step": 980
+    },
+    {
+      "epoch": 1.073972602739726,
+      "eval_loss": 2.0578773021698,
+      "eval_runtime": 36.8123,
+      "eval_samples_per_second": 27.165,
+      "eval_steps_per_second": 1.141,
+      "step": 980
+    },
+    {
+      "epoch": 1.0750684931506849,
+      "grad_norm": 2.378127336502075,
+      "learning_rate": 3.5926076360682377e-06,
+      "loss": 1.9268,
+      "step": 981
+    },
+    {
+      "epoch": 1.076164383561644,
+      "grad_norm": 2.248105525970459,
+      "learning_rate": 3.5905767668562145e-06,
+      "loss": 1.8111,
+      "step": 982
+    },
+    {
+      "epoch": 1.0772602739726027,
+      "grad_norm": 2.275592088699341,
+      "learning_rate": 3.5885458976441922e-06,
+      "loss": 1.9099,
+      "step": 983
+    },
+    {
+      "epoch": 1.0783561643835617,
+      "grad_norm": 3.108377456665039,
+      "learning_rate": 3.586515028432169e-06,
+      "loss": 1.8728,
+      "step": 984
+    },
+    {
+      "epoch": 1.0794520547945206,
+      "grad_norm": 2.2702999114990234,
+      "learning_rate": 3.5844841592201464e-06,
+      "loss": 1.9007,
+      "step": 985
+    },
+    {
+      "epoch": 1.0805479452054794,
+      "grad_norm": 3.218647003173828,
+      "learning_rate": 3.5824532900081237e-06,
+      "loss": 1.9157,
+      "step": 986
+    },
+    {
+      "epoch": 1.0816438356164384,
+      "grad_norm": 2.3393027782440186,
+      "learning_rate": 3.580422420796101e-06,
+      "loss": 1.7936,
+      "step": 987
+    },
+    {
+      "epoch": 1.0827397260273972,
+      "grad_norm": 2.420541286468506,
+      "learning_rate": 3.578391551584078e-06,
+      "loss": 1.8608,
+      "step": 988
+    },
+    {
+      "epoch": 1.0838356164383562,
+      "grad_norm": 2.3354556560516357,
+      "learning_rate": 3.5763606823720555e-06,
+      "loss": 1.8684,
+      "step": 989
+    },
+    {
+      "epoch": 1.084931506849315,
+      "grad_norm": 2.200282573699951,
+      "learning_rate": 3.5743298131600328e-06,
+      "loss": 1.7814,
+      "step": 990
+    },
+    {
+      "epoch": 1.084931506849315,
+      "eval_loss": 2.0565249919891357,
+      "eval_runtime": 36.952,
+      "eval_samples_per_second": 27.062,
+      "eval_steps_per_second": 1.137,
+      "step": 990
+    },
+    {
+      "epoch": 1.086027397260274,
+      "grad_norm": 2.149033308029175,
+      "learning_rate": 3.57229894394801e-06,
+      "loss": 1.8708,
+      "step": 991
+    },
+    {
+      "epoch": 1.087123287671233,
+      "grad_norm": 2.4663679599761963,
+      "learning_rate": 3.5702680747359873e-06,
+      "loss": 1.8947,
+      "step": 992
+    },
+    {
+      "epoch": 1.0882191780821917,
+      "grad_norm": 2.3142261505126953,
+      "learning_rate": 3.5682372055239646e-06,
+      "loss": 1.8727,
+      "step": 993
+    },
+    {
+      "epoch": 1.0893150684931507,
+      "grad_norm": 2.2223033905029297,
+      "learning_rate": 3.566206336311942e-06,
+      "loss": 1.8203,
+      "step": 994
+    },
+    {
+      "epoch": 1.0904109589041096,
+      "grad_norm": 2.423855781555176,
+      "learning_rate": 3.564175467099919e-06,
+      "loss": 1.8687,
+      "step": 995
+    },
+    {
+      "epoch": 1.0915068493150686,
+      "grad_norm": 2.4022302627563477,
+      "learning_rate": 3.5621445978878964e-06,
+      "loss": 1.8215,
+      "step": 996
+    },
+    {
+      "epoch": 1.0926027397260274,
+      "grad_norm": 2.378236770629883,
+      "learning_rate": 3.5601137286758737e-06,
+      "loss": 1.7994,
+      "step": 997
+    },
+    {
+      "epoch": 1.0936986301369862,
+      "grad_norm": 2.4499096870422363,
+      "learning_rate": 3.5580828594638505e-06,
+      "loss": 1.8779,
+      "step": 998
+    },
+    {
+      "epoch": 1.0947945205479452,
+      "grad_norm": 2.3181755542755127,
+      "learning_rate": 3.5560519902518282e-06,
+      "loss": 1.8341,
+      "step": 999
+    },
+    {
+      "epoch": 1.095890410958904,
+      "grad_norm": 2.2406623363494873,
+      "learning_rate": 3.554021121039805e-06,
+      "loss": 1.874,
+      "step": 1000
+    },
+    {
+      "epoch": 1.095890410958904,
+      "eval_loss": 2.0523672103881836,
+      "eval_runtime": 37.0454,
+      "eval_samples_per_second": 26.994,
+      "eval_steps_per_second": 1.134,
+      "step": 1000
+    },
+    {
+      "epoch": 1.096986301369863,
+      "grad_norm": 2.254667043685913,
+      "learning_rate": 3.551990251827783e-06,
+      "loss": 1.9017,
+      "step": 1001
+    },
+    {
+      "epoch": 1.098082191780822,
+      "grad_norm": 2.2846896648406982,
+      "learning_rate": 3.5499593826157596e-06,
+      "loss": 1.8505,
+      "step": 1002
+    },
+    {
+      "epoch": 1.099178082191781,
+      "grad_norm": 2.183469772338867,
+      "learning_rate": 3.5479285134037373e-06,
+      "loss": 1.8266,
+      "step": 1003
+    },
+    {
+      "epoch": 1.1002739726027397,
+      "grad_norm": 2.151798963546753,
+      "learning_rate": 3.545897644191714e-06,
+      "loss": 1.8487,
+      "step": 1004
+    },
+    {
+      "epoch": 1.1013698630136985,
+      "grad_norm": 2.2039146423339844,
+      "learning_rate": 3.543866774979692e-06,
+      "loss": 1.8327,
+      "step": 1005
+    },
+    {
+      "epoch": 1.1024657534246576,
+      "grad_norm": 2.2091665267944336,
+      "learning_rate": 3.5418359057676687e-06,
+      "loss": 1.8214,
+      "step": 1006
+    },
+    {
+      "epoch": 1.1035616438356164,
+      "grad_norm": 2.2821404933929443,
+      "learning_rate": 3.5398050365556464e-06,
+      "loss": 1.8591,
+      "step": 1007
+    },
+    {
+      "epoch": 1.1046575342465754,
+      "grad_norm": 2.1726973056793213,
+      "learning_rate": 3.5377741673436233e-06,
+      "loss": 1.8265,
+      "step": 1008
+    },
+    {
+      "epoch": 1.1057534246575342,
+      "grad_norm": 2.2870233058929443,
+      "learning_rate": 3.535743298131601e-06,
+      "loss": 1.8783,
+      "step": 1009
+    },
+    {
+      "epoch": 1.106849315068493,
+      "grad_norm": 2.148080587387085,
+      "learning_rate": 3.533712428919578e-06,
+      "loss": 1.8272,
+      "step": 1010
+    },
+    {
+      "epoch": 1.106849315068493,
+      "eval_loss": 2.049504041671753,
+      "eval_runtime": 36.8717,
+      "eval_samples_per_second": 27.121,
+      "eval_steps_per_second": 1.139,
+      "step": 1010
+    },
+    {
+      "epoch": 1.107945205479452,
+      "grad_norm": 2.182035446166992,
+      "learning_rate": 3.5316815597075547e-06,
+      "loss": 1.8771,
+      "step": 1011
+    },
+    {
+      "epoch": 1.1090410958904109,
+      "grad_norm": 2.135085344314575,
+      "learning_rate": 3.5296506904955324e-06,
+      "loss": 1.9031,
+      "step": 1012
+    },
+    {
+      "epoch": 1.11013698630137,
+      "grad_norm": 2.1017255783081055,
+      "learning_rate": 3.5276198212835093e-06,
+      "loss": 1.8341,
+      "step": 1013
+    },
+    {
+      "epoch": 1.1112328767123287,
+      "grad_norm": 2.2773115634918213,
+      "learning_rate": 3.525588952071487e-06,
+      "loss": 1.8564,
+      "step": 1014
+    },
+    {
+      "epoch": 1.1123287671232878,
+      "grad_norm": 2.2773115634918213,
+      "learning_rate": 3.525588952071487e-06,
+      "loss": 1.8971,
+      "step": 1015
+    },
+    {
+      "epoch": 1.1134246575342466,
+      "grad_norm": 2.273375988006592,
+      "learning_rate": 3.523558082859464e-06,
+      "loss": 1.9074,
+      "step": 1016
+    },
+    {
+      "epoch": 1.1145205479452054,
+      "grad_norm": 2.2055182456970215,
+      "learning_rate": 3.5215272136474415e-06,
+      "loss": 1.8724,
+      "step": 1017
+    },
+    {
+      "epoch": 1.1156164383561644,
+      "grad_norm": 2.246126651763916,
+      "learning_rate": 3.5194963444354184e-06,
+      "loss": 1.8387,
+      "step": 1018
+    },
+    {
+      "epoch": 1.1167123287671232,
+      "grad_norm": 2.3076183795928955,
+      "learning_rate": 3.517465475223396e-06,
+      "loss": 1.8387,
+      "step": 1019
+    },
+    {
+      "epoch": 1.1178082191780823,
+      "grad_norm": 2.3365488052368164,
+      "learning_rate": 3.515434606011373e-06,
+      "loss": 1.8063,
+      "step": 1020
+    },
+    {
+      "epoch": 1.1178082191780823,
+      "eval_loss": 2.0528781414031982,
+      "eval_runtime": 36.8885,
+      "eval_samples_per_second": 27.109,
+      "eval_steps_per_second": 1.139,
+      "step": 1020
+    },
+    {
+      "epoch": 1.118904109589041,
+      "grad_norm": 2.319324493408203,
+      "learning_rate": 3.5134037367993506e-06,
+      "loss": 1.8843,
+      "step": 1021
+    },
+    {
+      "epoch": 1.12,
+      "grad_norm": 2.238720655441284,
+      "learning_rate": 3.5113728675873275e-06,
+      "loss": 1.9375,
+      "step": 1022
+    },
+    {
+      "epoch": 1.121095890410959,
+      "grad_norm": 2.2824833393096924,
+      "learning_rate": 3.5093419983753047e-06,
+      "loss": 1.8749,
+      "step": 1023
+    },
+    {
+      "epoch": 1.1221917808219177,
+      "grad_norm": 2.5120770931243896,
+      "learning_rate": 3.507311129163282e-06,
+      "loss": 1.8774,
+      "step": 1024
+    },
+    {
+      "epoch": 1.1232876712328768,
+      "grad_norm": 2.1762094497680664,
+      "learning_rate": 3.5052802599512593e-06,
+      "loss": 1.8018,
+      "step": 1025
+    },
+    {
+      "epoch": 1.1243835616438356,
+      "grad_norm": 2.2530343532562256,
+      "learning_rate": 3.5032493907392366e-06,
+      "loss": 1.8451,
+      "step": 1026
+    },
+    {
+      "epoch": 1.1254794520547946,
+      "grad_norm": 2.21649169921875,
+      "learning_rate": 3.501218521527214e-06,
+      "loss": 1.8613,
+      "step": 1027
+    },
+    {
+      "epoch": 1.1265753424657534,
+      "grad_norm": 2.306431293487549,
+      "learning_rate": 3.499187652315191e-06,
+      "loss": 1.7947,
+      "step": 1028
+    },
+    {
+      "epoch": 1.1276712328767124,
+      "grad_norm": 2.203005313873291,
+      "learning_rate": 3.4971567831031684e-06,
+      "loss": 1.8483,
+      "step": 1029
+    },
+    {
+      "epoch": 1.1287671232876713,
+      "grad_norm": 2.1645009517669678,
+      "learning_rate": 3.4951259138911457e-06,
+      "loss": 1.8915,
+      "step": 1030
+    },
+    {
+      "epoch": 1.1287671232876713,
+      "eval_loss": 2.053393602371216,
+      "eval_runtime": 36.855,
+      "eval_samples_per_second": 27.133,
+      "eval_steps_per_second": 1.14,
+      "step": 1030
+    },
+    {
+      "epoch": 1.12986301369863,
+      "grad_norm": 2.2863385677337646,
+      "learning_rate": 3.493095044679123e-06,
+      "loss": 1.8073,
+      "step": 1031
+    },
+    {
+      "epoch": 1.130958904109589,
+      "grad_norm": 2.1439812183380127,
+      "learning_rate": 3.4910641754671002e-06,
+      "loss": 1.9191,
+      "step": 1032
+    },
+    {
+      "epoch": 1.132054794520548,
+      "grad_norm": 2.224228858947754,
+      "learning_rate": 3.4890333062550775e-06,
+      "loss": 1.7478,
+      "step": 1033
+    },
+    {
+      "epoch": 1.133150684931507,
+      "grad_norm": 2.319059133529663,
+      "learning_rate": 3.4870024370430548e-06,
+      "loss": 1.8779,
+      "step": 1034
+    },
+    {
+      "epoch": 1.1342465753424658,
+      "grad_norm": 2.5003139972686768,
+      "learning_rate": 3.484971567831032e-06,
+      "loss": 1.776,
+      "step": 1035
+    },
+    {
+      "epoch": 1.1353424657534246,
+      "grad_norm": 2.2032580375671387,
+      "learning_rate": 3.482940698619009e-06,
+      "loss": 1.8221,
+      "step": 1036
+    },
+    {
+      "epoch": 1.1364383561643836,
+      "grad_norm": 2.5253560543060303,
+      "learning_rate": 3.4809098294069866e-06,
+      "loss": 1.8952,
+      "step": 1037
+    },
+    {
+      "epoch": 1.1375342465753424,
+      "grad_norm": 2.3215854167938232,
+      "learning_rate": 3.4788789601949634e-06,
+      "loss": 1.847,
+      "step": 1038
+    },
+    {
+      "epoch": 1.1386301369863014,
+      "grad_norm": 2.2543280124664307,
+      "learning_rate": 3.476848090982941e-06,
+      "loss": 1.8689,
+      "step": 1039
+    },
+    {
+      "epoch": 1.1397260273972603,
+      "grad_norm": 2.1592042446136475,
+      "learning_rate": 3.474817221770918e-06,
+      "loss": 1.8375,
+      "step": 1040
+    },
+    {
+      "epoch": 1.1397260273972603,
+      "eval_loss": 2.0524044036865234,
+      "eval_runtime": 36.8163,
+      "eval_samples_per_second": 27.162,
+      "eval_steps_per_second": 1.141,
+      "step": 1040
+    },
+    {
+      "epoch": 1.140821917808219,
+      "grad_norm": 2.2874810695648193,
+      "learning_rate": 3.4727863525588957e-06,
+      "loss": 1.9442,
+      "step": 1041
+    },
+    {
+      "epoch": 1.141917808219178,
+      "grad_norm": 2.288395404815674,
+      "learning_rate": 3.4707554833468725e-06,
+      "loss": 1.8881,
+      "step": 1042
+    },
+    {
+      "epoch": 1.143013698630137,
+      "grad_norm": 2.2264153957366943,
+      "learning_rate": 3.4687246141348502e-06,
+      "loss": 1.8518,
+      "step": 1043
+    },
+    {
+      "epoch": 1.144109589041096,
+      "grad_norm": 2.4413652420043945,
+      "learning_rate": 3.466693744922827e-06,
+      "loss": 1.8146,
+      "step": 1044
+    },
+    {
+      "epoch": 1.1452054794520548,
+      "grad_norm": 2.3597898483276367,
+      "learning_rate": 3.464662875710805e-06,
+      "loss": 1.8421,
+      "step": 1045
+    },
+    {
+      "epoch": 1.1463013698630138,
+      "grad_norm": 2.2489402294158936,
+      "learning_rate": 3.4626320064987817e-06,
+      "loss": 1.8643,
+      "step": 1046
+    },
+    {
+      "epoch": 1.1473972602739726,
+      "grad_norm": 2.3958940505981445,
+      "learning_rate": 3.4606011372867594e-06,
+      "loss": 1.8522,
+      "step": 1047
+    },
+    {
+      "epoch": 1.1484931506849314,
+      "grad_norm": 2.295349597930908,
+      "learning_rate": 3.458570268074736e-06,
+      "loss": 1.8677,
+      "step": 1048
+    },
+    {
+      "epoch": 1.1495890410958904,
+      "grad_norm": 2.2571280002593994,
+      "learning_rate": 3.4565393988627135e-06,
+      "loss": 1.8393,
+      "step": 1049
+    },
+    {
+      "epoch": 1.1506849315068493,
+      "grad_norm": 2.173983335494995,
+      "learning_rate": 3.4545085296506908e-06,
+      "loss": 1.8267,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1506849315068493,
+      "eval_loss": 2.0489604473114014,
+      "eval_runtime": 36.7678,
+      "eval_samples_per_second": 27.198,
+      "eval_steps_per_second": 1.142,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1517808219178083,
+      "grad_norm": 2.3395209312438965,
+      "learning_rate": 3.452477660438668e-06,
+      "loss": 1.8032,
+      "step": 1051
+    },
+    {
+      "epoch": 1.152876712328767,
+      "grad_norm": 2.1894757747650146,
+      "learning_rate": 3.4504467912266453e-06,
+      "loss": 1.8834,
+      "step": 1052
+    },
+    {
+      "epoch": 1.1539726027397261,
+      "grad_norm": 2.1978437900543213,
+      "learning_rate": 3.4484159220146226e-06,
+      "loss": 1.7897,
+      "step": 1053
+    },
+    {
+      "epoch": 1.155068493150685,
+      "grad_norm": 2.275792360305786,
+      "learning_rate": 3.4463850528026e-06,
+      "loss": 1.8566,
+      "step": 1054
+    },
+    {
+      "epoch": 1.1561643835616437,
+      "grad_norm": 2.2659800052642822,
+      "learning_rate": 3.444354183590577e-06,
+      "loss": 1.8307,
+      "step": 1055
+    },
+    {
+      "epoch": 1.1572602739726028,
+      "grad_norm": 2.232149600982666,
+      "learning_rate": 3.4423233143785544e-06,
+      "loss": 1.8701,
+      "step": 1056
+    },
+    {
+      "epoch": 1.1583561643835616,
+      "grad_norm": 2.3636281490325928,
+      "learning_rate": 3.4402924451665317e-06,
+      "loss": 1.8008,
+      "step": 1057
+    },
+    {
+      "epoch": 1.1594520547945206,
+      "grad_norm": 2.271533489227295,
+      "learning_rate": 3.438261575954509e-06,
+      "loss": 1.7751,
+      "step": 1058
+    },
+    {
+      "epoch": 1.1605479452054794,
+      "grad_norm": 2.3286566734313965,
+      "learning_rate": 3.4362307067424862e-06,
+      "loss": 1.865,
+      "step": 1059
+    },
+    {
+      "epoch": 1.1616438356164385,
+      "grad_norm": 2.46665096282959,
+      "learning_rate": 3.434199837530463e-06,
+      "loss": 1.9001,
+      "step": 1060
+    },
+    {
+      "epoch": 1.1616438356164385,
+      "eval_loss": 2.0477516651153564,
+      "eval_runtime": 36.7818,
+      "eval_samples_per_second": 27.187,
+      "eval_steps_per_second": 1.142,
+      "step": 1060
+    },
+    {
+      "epoch": 1.1627397260273973,
+      "grad_norm": 2.307744264602661,
+      "learning_rate": 3.432168968318441e-06,
+      "loss": 1.8637,
+      "step": 1061
+    },
+    {
+      "epoch": 1.163835616438356,
+      "grad_norm": 2.3935985565185547,
+      "learning_rate": 3.4301380991064176e-06,
+      "loss": 1.7867,
+      "step": 1062
+    },
+    {
+      "epoch": 1.1649315068493151,
+      "grad_norm": 2.290971040725708,
+      "learning_rate": 3.4281072298943953e-06,
+      "loss": 1.8708,
+      "step": 1063
+    },
+    {
+      "epoch": 1.166027397260274,
+      "grad_norm": 2.191629409790039,
+      "learning_rate": 3.426076360682372e-06,
+      "loss": 1.9031,
+      "step": 1064
+    },
+    {
+      "epoch": 1.167123287671233,
+      "grad_norm": 2.284426689147949,
+      "learning_rate": 3.42404549147035e-06,
+      "loss": 1.9026,
+      "step": 1065
+    },
+    {
+      "epoch": 1.1682191780821918,
+      "grad_norm": 2.2567849159240723,
+      "learning_rate": 3.4220146222583267e-06,
+      "loss": 1.8507,
+      "step": 1066
+    },
+    {
+      "epoch": 1.1693150684931506,
+      "grad_norm": 2.2093160152435303,
+      "learning_rate": 3.4199837530463044e-06,
+      "loss": 1.8001,
+      "step": 1067
+    },
+    {
+      "epoch": 1.1704109589041096,
+      "grad_norm": 2.092560291290283,
+      "learning_rate": 3.4179528838342813e-06,
+      "loss": 1.8393,
+      "step": 1068
+    },
+    {
+      "epoch": 1.1715068493150684,
+      "grad_norm": 2.313749074935913,
+      "learning_rate": 3.415922014622259e-06,
+      "loss": 1.9385,
+      "step": 1069
+    },
+    {
+      "epoch": 1.1726027397260275,
+      "grad_norm": 2.488673210144043,
+      "learning_rate": 3.413891145410236e-06,
+      "loss": 1.8924,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1726027397260275,
+      "eval_loss": 2.0482401847839355,
+      "eval_runtime": 36.7919,
+      "eval_samples_per_second": 27.18,
+      "eval_steps_per_second": 1.142,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1736986301369863,
+      "grad_norm": 2.0804412364959717,
+      "learning_rate": 3.4118602761982135e-06,
+      "loss": 1.881,
+      "step": 1071
+    },
+    {
+      "epoch": 1.174794520547945,
+      "grad_norm": 2.165426254272461,
+      "learning_rate": 3.4098294069861904e-06,
+      "loss": 1.9035,
+      "step": 1072
+    },
+    {
+      "epoch": 1.1758904109589041,
+      "grad_norm": 2.3034675121307373,
+      "learning_rate": 3.4077985377741673e-06,
+      "loss": 1.8554,
+      "step": 1073
+    },
+    {
+      "epoch": 1.176986301369863,
+      "grad_norm": 2.303159713745117,
+      "learning_rate": 3.405767668562145e-06,
+      "loss": 1.8255,
+      "step": 1074
+    },
+    {
+      "epoch": 1.178082191780822,
+      "grad_norm": 2.2402753829956055,
+      "learning_rate": 3.403736799350122e-06,
+      "loss": 1.9087,
+      "step": 1075
+    },
+    {
+      "epoch": 1.1791780821917808,
+      "grad_norm": 2.251638650894165,
+      "learning_rate": 3.4017059301380995e-06,
+      "loss": 1.8878,
+      "step": 1076
+    },
+    {
+      "epoch": 1.1802739726027398,
+      "grad_norm": 2.2661657333374023,
+      "learning_rate": 3.3996750609260764e-06,
+      "loss": 1.8244,
+      "step": 1077
+    },
+    {
+      "epoch": 1.1813698630136986,
+      "grad_norm": 2.3280904293060303,
+      "learning_rate": 3.397644191714054e-06,
+      "loss": 1.9328,
+      "step": 1078
+    },
+    {
+      "epoch": 1.1824657534246574,
+      "grad_norm": 2.315779447555542,
+      "learning_rate": 3.395613322502031e-06,
+      "loss": 1.8289,
+      "step": 1079
+    },
+    {
+      "epoch": 1.1835616438356165,
+      "grad_norm": 2.136967182159424,
+      "learning_rate": 3.3935824532900086e-06,
+      "loss": 1.8307,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1835616438356165,
+      "eval_loss": 2.0473902225494385,
+      "eval_runtime": 36.8094,
+      "eval_samples_per_second": 27.167,
+      "eval_steps_per_second": 1.141,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1846575342465753,
+      "grad_norm": 2.3701720237731934,
+      "learning_rate": 3.3915515840779855e-06,
+      "loss": 1.8123,
+      "step": 1081
+    },
+    {
+      "epoch": 1.1857534246575343,
+      "grad_norm": 2.2662320137023926,
+      "learning_rate": 3.389520714865963e-06,
+      "loss": 1.85,
+      "step": 1082
+    },
+    {
+      "epoch": 1.1868493150684931,
+      "grad_norm": 2.2151782512664795,
+      "learning_rate": 3.38748984565394e-06,
+      "loss": 1.845,
+      "step": 1083
+    },
+    {
+      "epoch": 1.1879452054794521,
+      "grad_norm": 2.2209115028381348,
+      "learning_rate": 3.3854589764419177e-06,
+      "loss": 1.8787,
+      "step": 1084
+    },
+    {
+      "epoch": 1.189041095890411,
+      "grad_norm": 2.4760477542877197,
+      "learning_rate": 3.3834281072298946e-06,
+      "loss": 1.818,
+      "step": 1085
+    },
+    {
+      "epoch": 1.1901369863013698,
+      "grad_norm": 2.309034585952759,
+      "learning_rate": 3.381397238017872e-06,
+      "loss": 1.8387,
+      "step": 1086
+    },
+    {
+      "epoch": 1.1912328767123288,
+      "grad_norm": 2.2174596786499023,
+      "learning_rate": 3.379366368805849e-06,
+      "loss": 1.8135,
+      "step": 1087
+    },
+    {
+      "epoch": 1.1923287671232876,
+      "grad_norm": 2.3043556213378906,
+      "learning_rate": 3.3773354995938264e-06,
+      "loss": 1.9156,
+      "step": 1088
+    },
+    {
+      "epoch": 1.1934246575342466,
+      "grad_norm": 2.2240524291992188,
+      "learning_rate": 3.3753046303818037e-06,
+      "loss": 1.8636,
+      "step": 1089
+    },
+    {
+      "epoch": 1.1945205479452055,
+      "grad_norm": 2.1108691692352295,
+      "learning_rate": 3.373273761169781e-06,
+      "loss": 1.8458,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1945205479452055,
+      "eval_loss": 2.04378604888916,
+      "eval_runtime": 36.9058,
+      "eval_samples_per_second": 27.096,
+      "eval_steps_per_second": 1.138,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1956164383561645,
+      "grad_norm": 2.1893398761749268,
+      "learning_rate": 3.3712428919577582e-06,
+      "loss": 1.8446,
+      "step": 1091
+    },
+    {
+      "epoch": 1.1967123287671233,
+      "grad_norm": 2.305298089981079,
+      "learning_rate": 3.3692120227457355e-06,
+      "loss": 1.8283,
+      "step": 1092
+    },
+    {
+      "epoch": 1.1978082191780821,
+      "grad_norm": 2.364433765411377,
+      "learning_rate": 3.3671811535337128e-06,
+      "loss": 1.8708,
+      "step": 1093
+    },
+    {
+      "epoch": 1.1989041095890411,
+      "grad_norm": 2.259892225265503,
+      "learning_rate": 3.36515028432169e-06,
+      "loss": 1.7966,
+      "step": 1094
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 2.2487263679504395,
+      "learning_rate": 3.3631194151096673e-06,
+      "loss": 1.8238,
+      "step": 1095
+    },
+    {
+      "epoch": 1.201095890410959,
+      "grad_norm": 2.1592557430267334,
+      "learning_rate": 3.3610885458976446e-06,
+      "loss": 1.8651,
+      "step": 1096
+    },
+    {
+      "epoch": 1.2021917808219178,
+      "grad_norm": 2.257737159729004,
+      "learning_rate": 3.359057676685622e-06,
+      "loss": 1.8926,
+      "step": 1097
+    },
+    {
+      "epoch": 1.2032876712328768,
+      "grad_norm": 2.0702707767486572,
+      "learning_rate": 3.357026807473599e-06,
+      "loss": 1.8833,
+      "step": 1098
+    },
+    {
+      "epoch": 1.2043835616438356,
+      "grad_norm": 2.1568455696105957,
+      "learning_rate": 3.354995938261576e-06,
+      "loss": 1.7732,
+      "step": 1099
+    },
+    {
+      "epoch": 1.2054794520547945,
+      "grad_norm": 2.294715404510498,
+      "learning_rate": 3.3529650690495537e-06,
+      "loss": 1.8205,
+      "step": 1100
+    },
+    {
+      "epoch": 1.2054794520547945,
+      "eval_loss": 2.044201612472534,
+      "eval_runtime": 36.9861,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 1100
+    },
+    {
+      "epoch": 1.2065753424657535,
+      "grad_norm": 2.3234570026397705,
+      "learning_rate": 3.3509341998375305e-06,
+      "loss": 1.7935,
+      "step": 1101
+    },
+    {
+      "epoch": 1.2076712328767123,
+      "grad_norm": 2.307224750518799,
+      "learning_rate": 3.3489033306255082e-06,
+      "loss": 1.8254,
+      "step": 1102
+    },
+    {
+      "epoch": 1.208767123287671,
+      "grad_norm": 2.3725457191467285,
+      "learning_rate": 3.346872461413485e-06,
+      "loss": 1.8493,
+      "step": 1103
+    },
+    {
+      "epoch": 1.2098630136986301,
+      "grad_norm": 2.2057435512542725,
+      "learning_rate": 3.344841592201463e-06,
+      "loss": 1.8241,
+      "step": 1104
+    },
+    {
+      "epoch": 1.210958904109589,
+      "grad_norm": 2.232947826385498,
+      "learning_rate": 3.3428107229894397e-06,
+      "loss": 1.8057,
+      "step": 1105
+    },
+    {
+      "epoch": 1.212054794520548,
+      "grad_norm": 2.4268925189971924,
+      "learning_rate": 3.3407798537774174e-06,
+      "loss": 1.8334,
+      "step": 1106
+    },
+    {
+      "epoch": 1.2131506849315068,
+      "grad_norm": 2.3134701251983643,
+      "learning_rate": 3.338748984565394e-06,
+      "loss": 1.8516,
+      "step": 1107
+    },
+    {
+      "epoch": 1.2142465753424658,
+      "grad_norm": 2.375873327255249,
+      "learning_rate": 3.336718115353372e-06,
+      "loss": 1.8942,
+      "step": 1108
+    },
+    {
+      "epoch": 1.2153424657534246,
+      "grad_norm": 2.622528076171875,
+      "learning_rate": 3.3346872461413488e-06,
+      "loss": 1.9027,
+      "step": 1109
+    },
+    {
+      "epoch": 1.2164383561643834,
+      "grad_norm": 2.290865898132324,
+      "learning_rate": 3.3326563769293256e-06,
+      "loss": 1.8258,
+      "step": 1110
+    },
+    {
+      "epoch": 1.2164383561643834,
+      "eval_loss": 2.0408973693847656,
+      "eval_runtime": 36.8596,
+      "eval_samples_per_second": 27.13,
+      "eval_steps_per_second": 1.139,
+      "step": 1110
+    },
+    {
+      "epoch": 1.2175342465753425,
+      "grad_norm": 2.1596908569335938,
+      "learning_rate": 3.3306255077173033e-06,
+      "loss": 1.8066,
+      "step": 1111
+    },
+    {
+      "epoch": 1.2186301369863013,
+      "grad_norm": 2.4652187824249268,
+      "learning_rate": 3.32859463850528e-06,
+      "loss": 1.8372,
+      "step": 1112
+    },
+    {
+      "epoch": 1.2197260273972603,
+      "grad_norm": 2.282440185546875,
+      "learning_rate": 3.326563769293258e-06,
+      "loss": 1.8535,
+      "step": 1113
+    },
+    {
+      "epoch": 1.2208219178082191,
+      "grad_norm": 2.2390308380126953,
+      "learning_rate": 3.3245329000812347e-06,
+      "loss": 1.7719,
+      "step": 1114
+    },
+    {
+      "epoch": 1.2219178082191782,
+      "grad_norm": 2.368565320968628,
+      "learning_rate": 3.3225020308692124e-06,
+      "loss": 1.8207,
+      "step": 1115
+    },
+    {
+      "epoch": 1.223013698630137,
+      "grad_norm": 2.463942766189575,
+      "learning_rate": 3.3204711616571893e-06,
+      "loss": 1.8402,
+      "step": 1116
+    },
+    {
+      "epoch": 1.2241095890410958,
+      "grad_norm": 2.266096830368042,
+      "learning_rate": 3.318440292445167e-06,
+      "loss": 1.7854,
+      "step": 1117
+    },
+    {
+      "epoch": 1.2252054794520548,
+      "grad_norm": 2.3463494777679443,
+      "learning_rate": 3.316409423233144e-06,
+      "loss": 1.8449,
+      "step": 1118
+    },
+    {
+      "epoch": 1.2263013698630136,
+      "grad_norm": 2.381727933883667,
+      "learning_rate": 3.3143785540211215e-06,
+      "loss": 1.8085,
+      "step": 1119
+    },
+    {
+      "epoch": 1.2273972602739727,
+      "grad_norm": 2.3251113891601562,
+      "learning_rate": 3.3123476848090984e-06,
+      "loss": 1.8922,
+      "step": 1120
+    },
+    {
+      "epoch": 1.2273972602739727,
+      "eval_loss": 2.0425429344177246,
+      "eval_runtime": 36.851,
+      "eval_samples_per_second": 27.136,
+      "eval_steps_per_second": 1.14,
+      "step": 1120
+    },
+    {
+      "epoch": 1.2284931506849315,
+      "grad_norm": 2.2314298152923584,
+      "learning_rate": 3.310316815597076e-06,
+      "loss": 1.8474,
+      "step": 1121
+    },
+    {
+      "epoch": 1.2295890410958905,
+      "grad_norm": 2.1847472190856934,
+      "learning_rate": 3.308285946385053e-06,
+      "loss": 1.8454,
+      "step": 1122
+    },
+    {
+      "epoch": 1.2306849315068493,
+      "grad_norm": 2.249422788619995,
+      "learning_rate": 3.30625507717303e-06,
+      "loss": 1.9257,
+      "step": 1123
+    },
+    {
+      "epoch": 1.2317808219178081,
+      "grad_norm": 2.223930835723877,
+      "learning_rate": 3.3042242079610075e-06,
+      "loss": 1.7816,
+      "step": 1124
+    },
+    {
+      "epoch": 1.2328767123287672,
+      "grad_norm": 2.3209826946258545,
+      "learning_rate": 3.3021933387489847e-06,
+      "loss": 1.8773,
+      "step": 1125
+    },
+    {
+      "epoch": 1.233972602739726,
+      "grad_norm": 2.2288055419921875,
+      "learning_rate": 3.300162469536962e-06,
+      "loss": 1.8931,
+      "step": 1126
+    },
+    {
+      "epoch": 1.235068493150685,
+      "grad_norm": 2.201449155807495,
+      "learning_rate": 3.2981316003249393e-06,
+      "loss": 1.8149,
+      "step": 1127
+    },
+    {
+      "epoch": 1.2361643835616438,
+      "grad_norm": 2.215824842453003,
+      "learning_rate": 3.2961007311129166e-06,
+      "loss": 1.8949,
+      "step": 1128
+    },
+    {
+      "epoch": 1.2372602739726029,
+      "grad_norm": 2.133671522140503,
+      "learning_rate": 3.294069861900894e-06,
+      "loss": 1.8669,
+      "step": 1129
+    },
+    {
+      "epoch": 1.2383561643835617,
+      "grad_norm": 2.2195792198181152,
+      "learning_rate": 3.292038992688871e-06,
+      "loss": 1.9245,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2383561643835617,
+      "eval_loss": 2.0442771911621094,
+      "eval_runtime": 36.8077,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2394520547945205,
+      "grad_norm": 2.310866117477417,
+      "learning_rate": 3.2900081234768484e-06,
+      "loss": 1.8523,
+      "step": 1131
+    },
+    {
+      "epoch": 1.2405479452054795,
+      "grad_norm": 2.318300247192383,
+      "learning_rate": 3.2879772542648257e-06,
+      "loss": 1.8665,
+      "step": 1132
+    },
+    {
+      "epoch": 1.2416438356164383,
+      "grad_norm": 2.082564115524292,
+      "learning_rate": 3.285946385052803e-06,
+      "loss": 1.8962,
+      "step": 1133
+    },
+    {
+      "epoch": 1.2427397260273974,
+      "grad_norm": 2.1982104778289795,
+      "learning_rate": 3.2839155158407802e-06,
+      "loss": 1.7722,
+      "step": 1134
+    },
+    {
+      "epoch": 1.2438356164383562,
+      "grad_norm": 2.330782890319824,
+      "learning_rate": 3.2818846466287575e-06,
+      "loss": 1.8787,
+      "step": 1135
+    },
+    {
+      "epoch": 1.244931506849315,
+      "grad_norm": 2.136747121810913,
+      "learning_rate": 3.2798537774167344e-06,
+      "loss": 1.7739,
+      "step": 1136
+    },
+    {
+      "epoch": 1.246027397260274,
+      "grad_norm": 2.205862283706665,
+      "learning_rate": 3.277822908204712e-06,
+      "loss": 1.9127,
+      "step": 1137
+    },
+    {
+      "epoch": 1.2471232876712328,
+      "grad_norm": 2.2060093879699707,
+      "learning_rate": 3.275792038992689e-06,
+      "loss": 1.8379,
+      "step": 1138
+    },
+    {
+      "epoch": 1.2482191780821918,
+      "grad_norm": 2.255772829055786,
+      "learning_rate": 3.2737611697806666e-06,
+      "loss": 1.7931,
+      "step": 1139
+    },
+    {
+      "epoch": 1.2493150684931507,
+      "grad_norm": 2.240095615386963,
+      "learning_rate": 3.2717303005686435e-06,
+      "loss": 1.8466,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2493150684931507,
+      "eval_loss": 2.0429375171661377,
+      "eval_runtime": 36.7983,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2504109589041095,
+      "grad_norm": 2.1491949558258057,
+      "learning_rate": 3.269699431356621e-06,
+      "loss": 1.8137,
+      "step": 1141
+    },
+    {
+      "epoch": 1.2515068493150685,
+      "grad_norm": 2.3854610919952393,
+      "learning_rate": 3.267668562144598e-06,
+      "loss": 1.8486,
+      "step": 1142
+    },
+    {
+      "epoch": 1.2526027397260273,
+      "grad_norm": 2.3498709201812744,
+      "learning_rate": 3.2656376929325757e-06,
+      "loss": 1.837,
+      "step": 1143
+    },
+    {
+      "epoch": 1.2536986301369863,
+      "grad_norm": 2.1651649475097656,
+      "learning_rate": 3.2636068237205526e-06,
+      "loss": 1.8,
+      "step": 1144
+    },
+    {
+      "epoch": 1.2547945205479452,
+      "grad_norm": 2.2500200271606445,
+      "learning_rate": 3.2615759545085303e-06,
+      "loss": 1.8691,
+      "step": 1145
+    },
+    {
+      "epoch": 1.2558904109589042,
+      "grad_norm": 2.2178192138671875,
+      "learning_rate": 3.259545085296507e-06,
+      "loss": 1.8345,
+      "step": 1146
+    },
+    {
+      "epoch": 1.256986301369863,
+      "grad_norm": 2.238114595413208,
+      "learning_rate": 3.257514216084484e-06,
+      "loss": 1.8935,
+      "step": 1147
+    },
+    {
+      "epoch": 1.2580821917808218,
+      "grad_norm": 2.37042498588562,
+      "learning_rate": 3.2554833468724617e-06,
+      "loss": 1.7872,
+      "step": 1148
+    },
+    {
+      "epoch": 1.2591780821917808,
+      "grad_norm": 2.1831908226013184,
+      "learning_rate": 3.2534524776604385e-06,
+      "loss": 1.8913,
+      "step": 1149
+    },
+    {
+      "epoch": 1.2602739726027397,
+      "grad_norm": 2.251382827758789,
+      "learning_rate": 3.2514216084484162e-06,
+      "loss": 1.8611,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2602739726027397,
+      "eval_loss": 2.041306257247925,
+      "eval_runtime": 36.8199,
+      "eval_samples_per_second": 27.159,
+      "eval_steps_per_second": 1.141,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2613698630136987,
+      "grad_norm": 2.2690420150756836,
+      "learning_rate": 3.249390739236393e-06,
+      "loss": 1.8283,
+      "step": 1151
+    },
+    {
+      "epoch": 1.2624657534246575,
+      "grad_norm": 2.255833387374878,
+      "learning_rate": 3.2473598700243708e-06,
+      "loss": 1.8563,
+      "step": 1152
+    },
+    {
+      "epoch": 1.2635616438356165,
+      "grad_norm": 2.1362855434417725,
+      "learning_rate": 3.2453290008123476e-06,
+      "loss": 1.809,
+      "step": 1153
+    },
+    {
+      "epoch": 1.2646575342465753,
+      "grad_norm": 2.1227219104766846,
+      "learning_rate": 3.2432981316003253e-06,
+      "loss": 1.848,
+      "step": 1154
+    },
+    {
+      "epoch": 1.2657534246575342,
+      "grad_norm": 2.3639719486236572,
+      "learning_rate": 3.241267262388302e-06,
+      "loss": 1.8472,
+      "step": 1155
+    },
+    {
+      "epoch": 1.2668493150684932,
+      "grad_norm": 2.334404468536377,
+      "learning_rate": 3.23923639317628e-06,
+      "loss": 1.8073,
+      "step": 1156
+    },
+    {
+      "epoch": 1.267945205479452,
+      "grad_norm": 2.231656312942505,
+      "learning_rate": 3.2372055239642567e-06,
+      "loss": 1.8446,
+      "step": 1157
+    },
+    {
+      "epoch": 1.269041095890411,
+      "grad_norm": 2.371375560760498,
+      "learning_rate": 3.2351746547522344e-06,
+      "loss": 1.8002,
+      "step": 1158
+    },
+    {
+      "epoch": 1.2701369863013698,
+      "grad_norm": 2.5212879180908203,
+      "learning_rate": 3.2331437855402113e-06,
+      "loss": 1.863,
+      "step": 1159
+    },
+    {
+      "epoch": 1.2712328767123289,
+      "grad_norm": 2.3285911083221436,
+      "learning_rate": 3.2311129163281885e-06,
+      "loss": 1.9101,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2712328767123289,
+      "eval_loss": 2.038487195968628,
+      "eval_runtime": 36.8002,
+      "eval_samples_per_second": 27.174,
+      "eval_steps_per_second": 1.141,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2723287671232877,
+      "grad_norm": 2.1316046714782715,
+      "learning_rate": 3.229082047116166e-06,
+      "loss": 1.7952,
+      "step": 1161
+    },
+    {
+      "epoch": 1.2734246575342465,
+      "grad_norm": 2.4078125953674316,
+      "learning_rate": 3.227051177904143e-06,
+      "loss": 1.869,
+      "step": 1162
+    },
+    {
+      "epoch": 1.2745205479452055,
+      "grad_norm": 2.363672971725464,
+      "learning_rate": 3.2250203086921204e-06,
+      "loss": 1.8478,
+      "step": 1163
+    },
+    {
+      "epoch": 1.2756164383561643,
+      "grad_norm": 2.240558385848999,
+      "learning_rate": 3.2229894394800977e-06,
+      "loss": 1.8483,
+      "step": 1164
+    },
+    {
+      "epoch": 1.2767123287671232,
+      "grad_norm": 2.9142000675201416,
+      "learning_rate": 3.220958570268075e-06,
+      "loss": 1.8789,
+      "step": 1165
+    },
+    {
+      "epoch": 1.2778082191780822,
+      "grad_norm": 2.1930558681488037,
+      "learning_rate": 3.218927701056052e-06,
+      "loss": 1.7659,
+      "step": 1166
+    },
+    {
+      "epoch": 1.2789041095890412,
+      "grad_norm": 2.312471866607666,
+      "learning_rate": 3.2168968318440295e-06,
+      "loss": 1.8374,
+      "step": 1167
+    },
+    {
+      "epoch": 1.28,
+      "grad_norm": 2.6222236156463623,
+      "learning_rate": 3.2148659626320068e-06,
+      "loss": 1.7539,
+      "step": 1168
+    },
+    {
+      "epoch": 1.2810958904109588,
+      "grad_norm": 2.3013322353363037,
+      "learning_rate": 3.212835093419984e-06,
+      "loss": 1.9011,
+      "step": 1169
+    },
+    {
+      "epoch": 1.2821917808219179,
+      "grad_norm": 2.215240716934204,
+      "learning_rate": 3.2108042242079613e-06,
+      "loss": 1.8032,
+      "step": 1170
+    },
+    {
+      "epoch": 1.2821917808219179,
+      "eval_loss": 2.0365872383117676,
+      "eval_runtime": 36.8265,
+      "eval_samples_per_second": 27.154,
+      "eval_steps_per_second": 1.14,
+      "step": 1170
+    },
+    {
+      "epoch": 1.2832876712328767,
+      "grad_norm": 2.3620481491088867,
+      "learning_rate": 3.2087733549959386e-06,
+      "loss": 1.8258,
+      "step": 1171
+    },
+    {
+      "epoch": 1.2843835616438355,
+      "grad_norm": 2.2650208473205566,
+      "learning_rate": 3.206742485783916e-06,
+      "loss": 1.8888,
+      "step": 1172
+    },
+    {
+      "epoch": 1.2854794520547945,
+      "grad_norm": 2.2564077377319336,
+      "learning_rate": 3.2047116165718927e-06,
+      "loss": 1.7874,
+      "step": 1173
+    },
+    {
+      "epoch": 1.2865753424657536,
+      "grad_norm": 2.2855281829833984,
+      "learning_rate": 3.2026807473598704e-06,
+      "loss": 1.8003,
+      "step": 1174
+    },
+    {
+      "epoch": 1.2876712328767124,
+      "grad_norm": 2.353893756866455,
+      "learning_rate": 3.2006498781478473e-06,
+      "loss": 1.8466,
+      "step": 1175
+    },
+    {
+      "epoch": 1.2887671232876712,
+      "grad_norm": 2.1893045902252197,
+      "learning_rate": 3.198619008935825e-06,
+      "loss": 1.8652,
+      "step": 1176
+    },
+    {
+      "epoch": 1.2898630136986302,
+      "grad_norm": 2.6629793643951416,
+      "learning_rate": 3.196588139723802e-06,
+      "loss": 1.797,
+      "step": 1177
+    },
+    {
+      "epoch": 1.290958904109589,
+      "grad_norm": 2.2427361011505127,
+      "learning_rate": 3.1945572705117795e-06,
+      "loss": 1.8294,
+      "step": 1178
+    },
+    {
+      "epoch": 1.2920547945205478,
+      "grad_norm": 2.321878671646118,
+      "learning_rate": 3.1925264012997564e-06,
+      "loss": 1.8494,
+      "step": 1179
+    },
+    {
+      "epoch": 1.2931506849315069,
+      "grad_norm": 2.6928281784057617,
+      "learning_rate": 3.190495532087734e-06,
+      "loss": 1.8574,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2931506849315069,
+      "eval_loss": 2.0426652431488037,
+      "eval_runtime": 36.8482,
+      "eval_samples_per_second": 27.138,
+      "eval_steps_per_second": 1.14,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2942465753424657,
+      "grad_norm": 2.382164239883423,
+      "learning_rate": 3.188464662875711e-06,
+      "loss": 1.8518,
+      "step": 1181
+    },
+    {
+      "epoch": 1.2953424657534247,
+      "grad_norm": 2.4109256267547607,
+      "learning_rate": 3.1864337936636886e-06,
+      "loss": 1.9083,
+      "step": 1182
+    },
+    {
+      "epoch": 1.2964383561643835,
+      "grad_norm": 2.468062400817871,
+      "learning_rate": 3.1844029244516655e-06,
+      "loss": 1.8247,
+      "step": 1183
+    },
+    {
+      "epoch": 1.2975342465753426,
+      "grad_norm": 2.459378242492676,
+      "learning_rate": 3.182372055239643e-06,
+      "loss": 1.791,
+      "step": 1184
+    },
+    {
+      "epoch": 1.2986301369863014,
+      "grad_norm": 2.3331072330474854,
+      "learning_rate": 3.18034118602762e-06,
+      "loss": 1.7869,
+      "step": 1185
+    },
+    {
+      "epoch": 1.2997260273972602,
+      "grad_norm": 2.339428424835205,
+      "learning_rate": 3.1783103168155973e-06,
+      "loss": 1.8729,
+      "step": 1186
+    },
+    {
+      "epoch": 1.3008219178082192,
+      "grad_norm": 2.3033406734466553,
+      "learning_rate": 3.1762794476035746e-06,
+      "loss": 1.8455,
+      "step": 1187
+    },
+    {
+      "epoch": 1.301917808219178,
+      "grad_norm": 2.2480714321136475,
+      "learning_rate": 3.174248578391552e-06,
+      "loss": 1.8233,
+      "step": 1188
+    },
+    {
+      "epoch": 1.303013698630137,
+      "grad_norm": 2.2953624725341797,
+      "learning_rate": 3.172217709179529e-06,
+      "loss": 1.8261,
+      "step": 1189
+    },
+    {
+      "epoch": 1.3041095890410959,
+      "grad_norm": 2.494307518005371,
+      "learning_rate": 3.1701868399675064e-06,
+      "loss": 1.8153,
+      "step": 1190
+    },
+    {
+      "epoch": 1.3041095890410959,
+      "eval_loss": 2.037719964981079,
+      "eval_runtime": 36.9788,
+      "eval_samples_per_second": 27.043,
+      "eval_steps_per_second": 1.136,
+      "step": 1190
+    },
+    {
+      "epoch": 1.305205479452055,
+      "grad_norm": 2.314159631729126,
+      "learning_rate": 3.1681559707554837e-06,
+      "loss": 1.8406,
+      "step": 1191
+    },
+    {
+      "epoch": 1.3063013698630137,
+      "grad_norm": 2.4262588024139404,
+      "learning_rate": 3.166125101543461e-06,
+      "loss": 1.7803,
+      "step": 1192
+    },
+    {
+      "epoch": 1.3073972602739725,
+      "grad_norm": 2.3800835609436035,
+      "learning_rate": 3.1640942323314382e-06,
+      "loss": 1.8521,
+      "step": 1193
+    },
+    {
+      "epoch": 1.3084931506849315,
+      "grad_norm": 2.234551429748535,
+      "learning_rate": 3.1620633631194155e-06,
+      "loss": 1.845,
+      "step": 1194
+    },
+    {
+      "epoch": 1.3095890410958904,
+      "grad_norm": 2.167635440826416,
+      "learning_rate": 3.1600324939073928e-06,
+      "loss": 1.789,
+      "step": 1195
+    },
+    {
+      "epoch": 1.3106849315068494,
+      "grad_norm": 2.320247173309326,
+      "learning_rate": 3.15800162469537e-06,
+      "loss": 1.8311,
+      "step": 1196
+    },
+    {
+      "epoch": 1.3117808219178082,
+      "grad_norm": 2.2715561389923096,
+      "learning_rate": 3.155970755483347e-06,
+      "loss": 1.8308,
+      "step": 1197
+    },
+    {
+      "epoch": 1.3128767123287672,
+      "grad_norm": 2.2535455226898193,
+      "learning_rate": 3.1539398862713246e-06,
+      "loss": 1.8358,
+      "step": 1198
+    },
+    {
+      "epoch": 1.313972602739726,
+      "grad_norm": 2.2399163246154785,
+      "learning_rate": 3.1519090170593015e-06,
+      "loss": 1.8513,
+      "step": 1199
+    },
+    {
+      "epoch": 1.3150684931506849,
+      "grad_norm": 2.2984132766723633,
+      "learning_rate": 3.149878147847279e-06,
+      "loss": 1.8377,
+      "step": 1200
+    },
+    {
+      "epoch": 1.3150684931506849,
+      "eval_loss": 2.0334064960479736,
+      "eval_runtime": 37.0346,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1200
+    },
+    {
+      "epoch": 1.316164383561644,
+      "grad_norm": 2.1296892166137695,
+      "learning_rate": 3.147847278635256e-06,
+      "loss": 1.9023,
+      "step": 1201
+    },
+    {
+      "epoch": 1.3172602739726027,
+      "grad_norm": 2.1034350395202637,
+      "learning_rate": 3.1458164094232337e-06,
+      "loss": 1.917,
+      "step": 1202
+    },
+    {
+      "epoch": 1.3183561643835615,
+      "grad_norm": 2.258863687515259,
+      "learning_rate": 3.1437855402112106e-06,
+      "loss": 1.8093,
+      "step": 1203
+    },
+    {
+      "epoch": 1.3194520547945205,
+      "grad_norm": 2.20119309425354,
+      "learning_rate": 3.1417546709991883e-06,
+      "loss": 1.7796,
+      "step": 1204
+    },
+    {
+      "epoch": 1.3205479452054796,
+      "grad_norm": 2.3005869388580322,
+      "learning_rate": 3.139723801787165e-06,
+      "loss": 1.8496,
+      "step": 1205
+    },
+    {
+      "epoch": 1.3216438356164384,
+      "grad_norm": 2.2009031772613525,
+      "learning_rate": 3.137692932575143e-06,
+      "loss": 1.8244,
+      "step": 1206
+    },
+    {
+      "epoch": 1.3227397260273972,
+      "grad_norm": 2.3467633724212646,
+      "learning_rate": 3.1356620633631197e-06,
+      "loss": 1.8803,
+      "step": 1207
+    },
+    {
+      "epoch": 1.3238356164383562,
+      "grad_norm": 2.318711519241333,
+      "learning_rate": 3.1336311941510974e-06,
+      "loss": 1.7999,
+      "step": 1208
+    },
+    {
+      "epoch": 1.324931506849315,
+      "grad_norm": 2.182602882385254,
+      "learning_rate": 3.1316003249390742e-06,
+      "loss": 1.836,
+      "step": 1209
+    },
+    {
+      "epoch": 1.3260273972602739,
+      "grad_norm": 2.217298984527588,
+      "learning_rate": 3.129569455727051e-06,
+      "loss": 1.8776,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3260273972602739,
+      "eval_loss": 2.033573627471924,
+      "eval_runtime": 36.8029,
+      "eval_samples_per_second": 27.172,
+      "eval_steps_per_second": 1.141,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3271232876712329,
+      "grad_norm": 2.2814879417419434,
+      "learning_rate": 3.1275385865150288e-06,
+      "loss": 1.8739,
+      "step": 1211
+    },
+    {
+      "epoch": 1.3282191780821917,
+      "grad_norm": 2.1692471504211426,
+      "learning_rate": 3.1255077173030056e-06,
+      "loss": 1.8539,
+      "step": 1212
+    },
+    {
+      "epoch": 1.3293150684931507,
+      "grad_norm": 2.2472000122070312,
+      "learning_rate": 3.1234768480909833e-06,
+      "loss": 1.8527,
+      "step": 1213
+    },
+    {
+      "epoch": 1.3304109589041095,
+      "grad_norm": 2.1969852447509766,
+      "learning_rate": 3.12144597887896e-06,
+      "loss": 1.8429,
+      "step": 1214
+    },
+    {
+      "epoch": 1.3315068493150686,
+      "grad_norm": 2.141629219055176,
+      "learning_rate": 3.119415109666938e-06,
+      "loss": 1.868,
+      "step": 1215
+    },
+    {
+      "epoch": 1.3326027397260274,
+      "grad_norm": 2.282562494277954,
+      "learning_rate": 3.1173842404549147e-06,
+      "loss": 1.8913,
+      "step": 1216
+    },
+    {
+      "epoch": 1.3336986301369862,
+      "grad_norm": 2.2375850677490234,
+      "learning_rate": 3.1153533712428924e-06,
+      "loss": 1.8462,
+      "step": 1217
+    },
+    {
+      "epoch": 1.3347945205479452,
+      "grad_norm": 2.3529837131500244,
+      "learning_rate": 3.1133225020308693e-06,
+      "loss": 1.8445,
+      "step": 1218
+    },
+    {
+      "epoch": 1.335890410958904,
+      "grad_norm": 2.385857343673706,
+      "learning_rate": 3.111291632818847e-06,
+      "loss": 1.89,
+      "step": 1219
+    },
+    {
+      "epoch": 1.336986301369863,
+      "grad_norm": 2.301253080368042,
+      "learning_rate": 3.109260763606824e-06,
+      "loss": 1.8882,
+      "step": 1220
+    },
+    {
+      "epoch": 1.336986301369863,
+      "eval_loss": 2.040025234222412,
+      "eval_runtime": 36.839,
+      "eval_samples_per_second": 27.145,
+      "eval_steps_per_second": 1.14,
+      "step": 1220
+    },
+    {
+      "epoch": 1.3380821917808219,
+      "grad_norm": 2.3772335052490234,
+      "learning_rate": 3.1072298943948015e-06,
+      "loss": 1.8545,
+      "step": 1221
+    },
+    {
+      "epoch": 1.339178082191781,
+      "grad_norm": 2.387830972671509,
+      "learning_rate": 3.1051990251827784e-06,
+      "loss": 1.8653,
+      "step": 1222
+    },
+    {
+      "epoch": 1.3402739726027397,
+      "grad_norm": 2.4285190105438232,
+      "learning_rate": 3.1031681559707557e-06,
+      "loss": 1.8268,
+      "step": 1223
+    },
+    {
+      "epoch": 1.3413698630136985,
+      "grad_norm": 2.1892075538635254,
+      "learning_rate": 3.101137286758733e-06,
+      "loss": 1.8818,
+      "step": 1224
+    },
+    {
+      "epoch": 1.3424657534246576,
+      "grad_norm": 2.2510695457458496,
+      "learning_rate": 3.09910641754671e-06,
+      "loss": 1.879,
+      "step": 1225
+    },
+    {
+      "epoch": 1.3435616438356164,
+      "grad_norm": 2.505671977996826,
+      "learning_rate": 3.0970755483346875e-06,
+      "loss": 1.872,
+      "step": 1226
+    },
+    {
+      "epoch": 1.3446575342465754,
+      "grad_norm": 2.5315067768096924,
+      "learning_rate": 3.0950446791226648e-06,
+      "loss": 1.8651,
+      "step": 1227
+    },
+    {
+      "epoch": 1.3457534246575342,
+      "grad_norm": 2.223600387573242,
+      "learning_rate": 3.093013809910642e-06,
+      "loss": 1.9167,
+      "step": 1228
+    },
+    {
+      "epoch": 1.3468493150684933,
+      "grad_norm": 2.357802391052246,
+      "learning_rate": 3.0909829406986193e-06,
+      "loss": 1.8593,
+      "step": 1229
+    },
+    {
+      "epoch": 1.347945205479452,
+      "grad_norm": 2.3102364540100098,
+      "learning_rate": 3.0889520714865966e-06,
+      "loss": 1.8994,
+      "step": 1230
+    },
+    {
+      "epoch": 1.347945205479452,
+      "eval_loss": 2.0332930088043213,
+      "eval_runtime": 36.989,
+      "eval_samples_per_second": 27.035,
+      "eval_steps_per_second": 1.135,
+      "step": 1230
+    },
+    {
+      "epoch": 1.3490410958904109,
+      "grad_norm": 2.085080623626709,
+      "learning_rate": 3.086921202274574e-06,
+      "loss": 1.8741,
+      "step": 1231
+    },
+    {
+      "epoch": 1.35013698630137,
+      "grad_norm": 2.267305612564087,
+      "learning_rate": 3.084890333062551e-06,
+      "loss": 1.8426,
+      "step": 1232
+    },
+    {
+      "epoch": 1.3512328767123287,
+      "grad_norm": 2.181033134460449,
+      "learning_rate": 3.0828594638505284e-06,
+      "loss": 1.8199,
+      "step": 1233
+    },
+    {
+      "epoch": 1.3523287671232875,
+      "grad_norm": 2.288386821746826,
+      "learning_rate": 3.0808285946385057e-06,
+      "loss": 1.8469,
+      "step": 1234
+    },
+    {
+      "epoch": 1.3534246575342466,
+      "grad_norm": 2.1516127586364746,
+      "learning_rate": 3.078797725426483e-06,
+      "loss": 1.8056,
+      "step": 1235
+    },
+    {
+      "epoch": 1.3545205479452056,
+      "grad_norm": 2.2699198722839355,
+      "learning_rate": 3.07676685621446e-06,
+      "loss": 1.7535,
+      "step": 1236
+    },
+    {
+      "epoch": 1.3556164383561644,
+      "grad_norm": 2.193251132965088,
+      "learning_rate": 3.0747359870024375e-06,
+      "loss": 1.8407,
+      "step": 1237
+    },
+    {
+      "epoch": 1.3567123287671232,
+      "grad_norm": 2.2217774391174316,
+      "learning_rate": 3.0727051177904144e-06,
+      "loss": 1.8062,
+      "step": 1238
+    },
+    {
+      "epoch": 1.3578082191780823,
+      "grad_norm": 2.2376203536987305,
+      "learning_rate": 3.070674248578392e-06,
+      "loss": 1.829,
+      "step": 1239
+    },
+    {
+      "epoch": 1.358904109589041,
+      "grad_norm": 2.387559413909912,
+      "learning_rate": 3.068643379366369e-06,
+      "loss": 1.8459,
+      "step": 1240
+    },
+    {
+      "epoch": 1.358904109589041,
+      "eval_loss": 2.029298782348633,
+      "eval_runtime": 37.0491,
+      "eval_samples_per_second": 26.991,
+      "eval_steps_per_second": 1.134,
+      "step": 1240
+    },
+    {
+      "epoch": 1.3599999999999999,
+      "grad_norm": 2.1940178871154785,
+      "learning_rate": 3.0666125101543466e-06,
+      "loss": 1.831,
+      "step": 1241
+    },
+    {
+      "epoch": 1.361095890410959,
+      "grad_norm": 2.145514488220215,
+      "learning_rate": 3.0645816409423235e-06,
+      "loss": 1.7919,
+      "step": 1242
+    },
+    {
+      "epoch": 1.362191780821918,
+      "grad_norm": 2.319164276123047,
+      "learning_rate": 3.062550771730301e-06,
+      "loss": 1.8472,
+      "step": 1243
+    },
+    {
+      "epoch": 1.3632876712328768,
+      "grad_norm": 2.218986749649048,
+      "learning_rate": 3.060519902518278e-06,
+      "loss": 1.8184,
+      "step": 1244
+    },
+    {
+      "epoch": 1.3643835616438356,
+      "grad_norm": 2.3755338191986084,
+      "learning_rate": 3.0584890333062557e-06,
+      "loss": 1.8166,
+      "step": 1245
+    },
+    {
+      "epoch": 1.3654794520547946,
+      "grad_norm": 2.3186213970184326,
+      "learning_rate": 3.0564581640942326e-06,
+      "loss": 1.7599,
+      "step": 1246
+    },
+    {
+      "epoch": 1.3665753424657534,
+      "grad_norm": 2.193958282470703,
+      "learning_rate": 3.0544272948822094e-06,
+      "loss": 1.8404,
+      "step": 1247
+    },
+    {
+      "epoch": 1.3676712328767122,
+      "grad_norm": 2.40419602394104,
+      "learning_rate": 3.052396425670187e-06,
+      "loss": 1.821,
+      "step": 1248
+    },
+    {
+      "epoch": 1.3687671232876713,
+      "grad_norm": 2.445542573928833,
+      "learning_rate": 3.050365556458164e-06,
+      "loss": 1.8518,
+      "step": 1249
+    },
+    {
+      "epoch": 1.36986301369863,
+      "grad_norm": 2.2776567935943604,
+      "learning_rate": 3.0483346872461417e-06,
+      "loss": 1.8121,
+      "step": 1250
+    },
+    {
+      "epoch": 1.36986301369863,
+      "eval_loss": 2.030043125152588,
+      "eval_runtime": 37.0721,
+      "eval_samples_per_second": 26.974,
+      "eval_steps_per_second": 1.133,
+      "step": 1250
+    },
+    {
+      "epoch": 1.370958904109589,
+      "grad_norm": 2.420448064804077,
+      "learning_rate": 3.0463038180341185e-06,
+      "loss": 1.8168,
+      "step": 1251
+    },
+    {
+      "epoch": 1.372054794520548,
+      "grad_norm": 2.252065658569336,
+      "learning_rate": 3.0442729488220962e-06,
+      "loss": 1.8757,
+      "step": 1252
+    },
+    {
+      "epoch": 1.373150684931507,
+      "grad_norm": 2.4772560596466064,
+      "learning_rate": 3.042242079610073e-06,
+      "loss": 1.8711,
+      "step": 1253
+    },
+    {
+      "epoch": 1.3742465753424657,
+      "grad_norm": 2.3986129760742188,
+      "learning_rate": 3.0402112103980508e-06,
+      "loss": 1.8636,
+      "step": 1254
+    },
+    {
+      "epoch": 1.3753424657534246,
+      "grad_norm": 2.257274866104126,
+      "learning_rate": 3.0381803411860276e-06,
+      "loss": 1.7858,
+      "step": 1255
+    },
+    {
+      "epoch": 1.3764383561643836,
+      "grad_norm": 2.2373499870300293,
+      "learning_rate": 3.0361494719740053e-06,
+      "loss": 1.817,
+      "step": 1256
+    },
+    {
+      "epoch": 1.3775342465753424,
+      "grad_norm": 2.292388916015625,
+      "learning_rate": 3.034118602761982e-06,
+      "loss": 1.827,
+      "step": 1257
+    },
+    {
+      "epoch": 1.3786301369863014,
+      "grad_norm": 2.4618642330169678,
+      "learning_rate": 3.03208773354996e-06,
+      "loss": 1.8218,
+      "step": 1258
+    },
+    {
+      "epoch": 1.3797260273972602,
+      "grad_norm": 2.271958112716675,
+      "learning_rate": 3.0300568643379367e-06,
+      "loss": 1.8092,
+      "step": 1259
+    },
+    {
+      "epoch": 1.3808219178082193,
+      "grad_norm": 2.3990538120269775,
+      "learning_rate": 3.028025995125914e-06,
+      "loss": 1.8633,
+      "step": 1260
+    },
+    {
+      "epoch": 1.3808219178082193,
+      "eval_loss": 2.028458595275879,
+      "eval_runtime": 37.0659,
+      "eval_samples_per_second": 26.979,
+      "eval_steps_per_second": 1.133,
+      "step": 1260
+    },
+    {
+      "epoch": 1.381917808219178,
+      "grad_norm": 2.208803653717041,
+      "learning_rate": 3.0259951259138913e-06,
+      "loss": 1.8463,
+      "step": 1261
+    },
+    {
+      "epoch": 1.383013698630137,
+      "grad_norm": 2.3269028663635254,
+      "learning_rate": 3.0239642567018686e-06,
+      "loss": 1.7784,
+      "step": 1262
+    },
+    {
+      "epoch": 1.384109589041096,
+      "grad_norm": 2.367910623550415,
+      "learning_rate": 3.021933387489846e-06,
+      "loss": 1.8281,
+      "step": 1263
+    },
+    {
+      "epoch": 1.3852054794520547,
+      "grad_norm": 2.273686647415161,
+      "learning_rate": 3.019902518277823e-06,
+      "loss": 1.7832,
+      "step": 1264
+    },
+    {
+      "epoch": 1.3863013698630138,
+      "grad_norm": 2.421964168548584,
+      "learning_rate": 3.0178716490658004e-06,
+      "loss": 1.8407,
+      "step": 1265
+    },
+    {
+      "epoch": 1.3873972602739726,
+      "grad_norm": 2.2627696990966797,
+      "learning_rate": 3.0158407798537777e-06,
+      "loss": 1.8718,
+      "step": 1266
+    },
+    {
+      "epoch": 1.3884931506849316,
+      "grad_norm": 2.1981618404388428,
+      "learning_rate": 3.013809910641755e-06,
+      "loss": 1.7763,
+      "step": 1267
+    },
+    {
+      "epoch": 1.3895890410958904,
+      "grad_norm": 2.245253562927246,
+      "learning_rate": 3.0117790414297322e-06,
+      "loss": 1.7915,
+      "step": 1268
+    },
+    {
+      "epoch": 1.3906849315068492,
+      "grad_norm": 2.2292444705963135,
+      "learning_rate": 3.0097481722177095e-06,
+      "loss": 1.7909,
+      "step": 1269
+    },
+    {
+      "epoch": 1.3917808219178083,
+      "grad_norm": 2.2290430068969727,
+      "learning_rate": 3.0077173030056868e-06,
+      "loss": 1.8364,
+      "step": 1270
+    },
+    {
+      "epoch": 1.3917808219178083,
+      "eval_loss": 2.0290327072143555,
+      "eval_runtime": 37.0666,
+      "eval_samples_per_second": 26.978,
+      "eval_steps_per_second": 1.133,
+      "step": 1270
+    },
+    {
+      "epoch": 1.392876712328767,
+      "grad_norm": 2.220881700515747,
+      "learning_rate": 3.005686433793664e-06,
+      "loss": 1.7968,
+      "step": 1271
+    },
+    {
+      "epoch": 1.393972602739726,
+      "grad_norm": 2.375332832336426,
+      "learning_rate": 3.0036555645816413e-06,
+      "loss": 1.8502,
+      "step": 1272
+    },
+    {
+      "epoch": 1.395068493150685,
+      "grad_norm": 2.3065083026885986,
+      "learning_rate": 3.001624695369618e-06,
+      "loss": 1.8116,
+      "step": 1273
+    },
+    {
+      "epoch": 1.396164383561644,
+      "grad_norm": 2.7262511253356934,
+      "learning_rate": 2.999593826157596e-06,
+      "loss": 1.8641,
+      "step": 1274
+    },
+    {
+      "epoch": 1.3972602739726028,
+      "grad_norm": 2.3381216526031494,
+      "learning_rate": 2.9975629569455727e-06,
+      "loss": 1.8128,
+      "step": 1275
+    },
+    {
+      "epoch": 1.3983561643835616,
+      "grad_norm": 2.1994192600250244,
+      "learning_rate": 2.9955320877335504e-06,
+      "loss": 1.8563,
+      "step": 1276
+    },
+    {
+      "epoch": 1.3994520547945206,
+      "grad_norm": 2.3810958862304688,
+      "learning_rate": 2.9935012185215273e-06,
+      "loss": 1.8196,
+      "step": 1277
+    },
+    {
+      "epoch": 1.4005479452054794,
+      "grad_norm": 2.4343955516815186,
+      "learning_rate": 2.991470349309505e-06,
+      "loss": 1.8564,
+      "step": 1278
+    },
+    {
+      "epoch": 1.4016438356164382,
+      "grad_norm": 2.1929988861083984,
+      "learning_rate": 2.989439480097482e-06,
+      "loss": 1.8119,
+      "step": 1279
+    },
+    {
+      "epoch": 1.4027397260273973,
+      "grad_norm": 2.2305309772491455,
+      "learning_rate": 2.9874086108854595e-06,
+      "loss": 1.8289,
+      "step": 1280
+    },
+    {
+      "epoch": 1.4027397260273973,
+      "eval_loss": 2.032022476196289,
+      "eval_runtime": 37.084,
+      "eval_samples_per_second": 26.966,
+      "eval_steps_per_second": 1.133,
+      "step": 1280
+    },
+    {
+      "epoch": 1.403835616438356,
+      "grad_norm": 2.6255314350128174,
+      "learning_rate": 2.9853777416734364e-06,
+      "loss": 1.8176,
+      "step": 1281
+    },
+    {
+      "epoch": 1.4049315068493151,
+      "grad_norm": 2.238891363143921,
+      "learning_rate": 2.983346872461414e-06,
+      "loss": 1.8679,
+      "step": 1282
+    },
+    {
+      "epoch": 1.406027397260274,
+      "grad_norm": 2.9972007274627686,
+      "learning_rate": 2.981316003249391e-06,
+      "loss": 1.8839,
+      "step": 1283
+    },
+    {
+      "epoch": 1.407123287671233,
+      "grad_norm": 2.3255932331085205,
+      "learning_rate": 2.9792851340373678e-06,
+      "loss": 1.8451,
+      "step": 1284
+    },
+    {
+      "epoch": 1.4082191780821918,
+      "grad_norm": 2.290395498275757,
+      "learning_rate": 2.9772542648253455e-06,
+      "loss": 1.8244,
+      "step": 1285
+    },
+    {
+      "epoch": 1.4093150684931506,
+      "grad_norm": 2.288728713989258,
+      "learning_rate": 2.9752233956133223e-06,
+      "loss": 1.8628,
+      "step": 1286
+    },
+    {
+      "epoch": 1.4104109589041096,
+      "grad_norm": 2.177381992340088,
+      "learning_rate": 2.9731925264013e-06,
+      "loss": 1.8111,
+      "step": 1287
+    },
+    {
+      "epoch": 1.4115068493150684,
+      "grad_norm": 2.214956760406494,
+      "learning_rate": 2.971161657189277e-06,
+      "loss": 1.8377,
+      "step": 1288
+    },
+    {
+      "epoch": 1.4126027397260275,
+      "grad_norm": 2.3859663009643555,
+      "learning_rate": 2.9691307879772546e-06,
+      "loss": 1.8154,
+      "step": 1289
+    },
+    {
+      "epoch": 1.4136986301369863,
+      "grad_norm": 2.0873706340789795,
+      "learning_rate": 2.9670999187652314e-06,
+      "loss": 1.8098,
+      "step": 1290
+    },
+    {
+      "epoch": 1.4136986301369863,
+      "eval_loss": 2.029996633529663,
+      "eval_runtime": 37.0979,
+      "eval_samples_per_second": 26.956,
+      "eval_steps_per_second": 1.132,
+      "step": 1290
+    },
+    {
+      "epoch": 1.4147945205479453,
+      "grad_norm": 2.2632675170898438,
+      "learning_rate": 2.965069049553209e-06,
+      "loss": 1.7974,
+      "step": 1291
+    },
+    {
+      "epoch": 1.415890410958904,
+      "grad_norm": 2.208515167236328,
+      "learning_rate": 2.963038180341186e-06,
+      "loss": 1.8856,
+      "step": 1292
+    },
+    {
+      "epoch": 1.416986301369863,
+      "grad_norm": 2.176525115966797,
+      "learning_rate": 2.9610073111291637e-06,
+      "loss": 1.7783,
+      "step": 1293
+    },
+    {
+      "epoch": 1.418082191780822,
+      "grad_norm": 2.226682186126709,
+      "learning_rate": 2.9589764419171405e-06,
+      "loss": 1.865,
+      "step": 1294
+    },
+    {
+      "epoch": 1.4191780821917808,
+      "grad_norm": 2.3544883728027344,
+      "learning_rate": 2.9569455727051182e-06,
+      "loss": 1.8808,
+      "step": 1295
+    },
+    {
+      "epoch": 1.4202739726027398,
+      "grad_norm": 2.1729962825775146,
+      "learning_rate": 2.954914703493095e-06,
+      "loss": 1.7927,
+      "step": 1296
+    },
+    {
+      "epoch": 1.4213698630136986,
+      "grad_norm": 2.168168067932129,
+      "learning_rate": 2.9528838342810724e-06,
+      "loss": 1.8253,
+      "step": 1297
+    },
+    {
+      "epoch": 1.4224657534246576,
+      "grad_norm": 2.2479910850524902,
+      "learning_rate": 2.9508529650690496e-06,
+      "loss": 1.7976,
+      "step": 1298
+    },
+    {
+      "epoch": 1.4235616438356165,
+      "grad_norm": 2.1733837127685547,
+      "learning_rate": 2.948822095857027e-06,
+      "loss": 1.8015,
+      "step": 1299
+    },
+    {
+      "epoch": 1.4246575342465753,
+      "grad_norm": 2.2553391456604004,
+      "learning_rate": 2.946791226645004e-06,
+      "loss": 1.8545,
+      "step": 1300
+    },
+    {
+      "epoch": 1.4246575342465753,
+      "eval_loss": 2.0245048999786377,
+      "eval_runtime": 37.0943,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 1300
+    },
+    {
+      "epoch": 1.4257534246575343,
+      "grad_norm": 2.2817769050598145,
+      "learning_rate": 2.9447603574329815e-06,
+      "loss": 1.8017,
+      "step": 1301
+    },
+    {
+      "epoch": 1.426849315068493,
+      "grad_norm": 2.170865297317505,
+      "learning_rate": 2.9427294882209587e-06,
+      "loss": 1.7738,
+      "step": 1302
+    },
+    {
+      "epoch": 1.427945205479452,
+      "grad_norm": 2.296919107437134,
+      "learning_rate": 2.940698619008936e-06,
+      "loss": 1.8098,
+      "step": 1303
+    },
+    {
+      "epoch": 1.429041095890411,
+      "grad_norm": 2.476266622543335,
+      "learning_rate": 2.9386677497969133e-06,
+      "loss": 1.8229,
+      "step": 1304
+    },
+    {
+      "epoch": 1.43013698630137,
+      "grad_norm": 2.400918483734131,
+      "learning_rate": 2.9366368805848906e-06,
+      "loss": 1.8057,
+      "step": 1305
+    },
+    {
+      "epoch": 1.4312328767123288,
+      "grad_norm": 2.393892765045166,
+      "learning_rate": 2.934606011372868e-06,
+      "loss": 1.8568,
+      "step": 1306
+    },
+    {
+      "epoch": 1.4323287671232876,
+      "grad_norm": 2.2551307678222656,
+      "learning_rate": 2.932575142160845e-06,
+      "loss": 1.8493,
+      "step": 1307
+    },
+    {
+      "epoch": 1.4334246575342466,
+      "grad_norm": 2.3091611862182617,
+      "learning_rate": 2.9305442729488224e-06,
+      "loss": 1.8297,
+      "step": 1308
+    },
+    {
+      "epoch": 1.4345205479452054,
+      "grad_norm": 2.231614828109741,
+      "learning_rate": 2.9285134037367997e-06,
+      "loss": 1.8345,
+      "step": 1309
+    },
+    {
+      "epoch": 1.4356164383561643,
+      "grad_norm": 2.24942946434021,
+      "learning_rate": 2.9264825345247765e-06,
+      "loss": 1.8284,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4356164383561643,
+      "eval_loss": 2.024843215942383,
+      "eval_runtime": 36.7795,
+      "eval_samples_per_second": 27.189,
+      "eval_steps_per_second": 1.142,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4367123287671233,
+      "grad_norm": 2.2798643112182617,
+      "learning_rate": 2.9244516653127542e-06,
+      "loss": 1.8554,
+      "step": 1311
+    },
+    {
+      "epoch": 1.4378082191780823,
+      "grad_norm": 2.342576265335083,
+      "learning_rate": 2.922420796100731e-06,
+      "loss": 1.8296,
+      "step": 1312
+    },
+    {
+      "epoch": 1.4389041095890411,
+      "grad_norm": 2.268016815185547,
+      "learning_rate": 2.9203899268887088e-06,
+      "loss": 1.7467,
+      "step": 1313
+    },
+    {
+      "epoch": 1.44,
+      "grad_norm": 2.2520217895507812,
+      "learning_rate": 2.9183590576766856e-06,
+      "loss": 1.8748,
+      "step": 1314
+    },
+    {
+      "epoch": 1.441095890410959,
+      "grad_norm": 2.2091598510742188,
+      "learning_rate": 2.9163281884646633e-06,
+      "loss": 1.7715,
+      "step": 1315
+    },
+    {
+      "epoch": 1.4421917808219178,
+      "grad_norm": 2.197575092315674,
+      "learning_rate": 2.91429731925264e-06,
+      "loss": 1.8265,
+      "step": 1316
+    },
+    {
+      "epoch": 1.4432876712328766,
+      "grad_norm": 2.261122465133667,
+      "learning_rate": 2.912266450040618e-06,
+      "loss": 1.8758,
+      "step": 1317
+    },
+    {
+      "epoch": 1.4443835616438356,
+      "grad_norm": 2.3291778564453125,
+      "learning_rate": 2.9102355808285947e-06,
+      "loss": 1.8213,
+      "step": 1318
+    },
+    {
+      "epoch": 1.4454794520547944,
+      "grad_norm": 2.2182250022888184,
+      "learning_rate": 2.9082047116165724e-06,
+      "loss": 1.822,
+      "step": 1319
+    },
+    {
+      "epoch": 1.4465753424657535,
+      "grad_norm": 2.295978546142578,
+      "learning_rate": 2.9061738424045493e-06,
+      "loss": 1.8008,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4465753424657535,
+      "eval_loss": 2.0230674743652344,
+      "eval_runtime": 37.0013,
+      "eval_samples_per_second": 27.026,
+      "eval_steps_per_second": 1.135,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4476712328767123,
+      "grad_norm": 2.3187007904052734,
+      "learning_rate": 2.904142973192527e-06,
+      "loss": 1.7923,
+      "step": 1321
+    },
+    {
+      "epoch": 1.4487671232876713,
+      "grad_norm": 2.242851972579956,
+      "learning_rate": 2.902112103980504e-06,
+      "loss": 1.8874,
+      "step": 1322
+    },
+    {
+      "epoch": 1.4498630136986301,
+      "grad_norm": 2.2560081481933594,
+      "learning_rate": 2.900081234768481e-06,
+      "loss": 1.8284,
+      "step": 1323
+    },
+    {
+      "epoch": 1.450958904109589,
+      "grad_norm": 2.3017799854278564,
+      "learning_rate": 2.8980503655564584e-06,
+      "loss": 1.8263,
+      "step": 1324
+    },
+    {
+      "epoch": 1.452054794520548,
+      "grad_norm": 2.3016278743743896,
+      "learning_rate": 2.8960194963444357e-06,
+      "loss": 1.8174,
+      "step": 1325
+    },
+    {
+      "epoch": 1.4531506849315068,
+      "grad_norm": 2.1965372562408447,
+      "learning_rate": 2.893988627132413e-06,
+      "loss": 1.7812,
+      "step": 1326
+    },
+    {
+      "epoch": 1.4542465753424658,
+      "grad_norm": 2.302969455718994,
+      "learning_rate": 2.8919577579203902e-06,
+      "loss": 1.8115,
+      "step": 1327
+    },
+    {
+      "epoch": 1.4553424657534246,
+      "grad_norm": 2.338066339492798,
+      "learning_rate": 2.8899268887083675e-06,
+      "loss": 1.7535,
+      "step": 1328
+    },
+    {
+      "epoch": 1.4564383561643837,
+      "grad_norm": 2.446503162384033,
+      "learning_rate": 2.8878960194963448e-06,
+      "loss": 1.8052,
+      "step": 1329
+    },
+    {
+      "epoch": 1.4575342465753425,
+      "grad_norm": 2.2319273948669434,
+      "learning_rate": 2.885865150284322e-06,
+      "loss": 1.8226,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4575342465753425,
+      "eval_loss": 2.022364854812622,
+      "eval_runtime": 37.0901,
+      "eval_samples_per_second": 26.961,
+      "eval_steps_per_second": 1.132,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4586301369863013,
+      "grad_norm": 2.391080856323242,
+      "learning_rate": 2.8838342810722993e-06,
+      "loss": 1.8529,
+      "step": 1331
+    },
+    {
+      "epoch": 1.4597260273972603,
+      "grad_norm": 2.274463415145874,
+      "learning_rate": 2.8818034118602766e-06,
+      "loss": 1.8043,
+      "step": 1332
+    },
+    {
+      "epoch": 1.4608219178082191,
+      "grad_norm": 2.4040963649749756,
+      "learning_rate": 2.879772542648254e-06,
+      "loss": 1.7842,
+      "step": 1333
+    },
+    {
+      "epoch": 1.4619178082191782,
+      "grad_norm": 2.268629312515259,
+      "learning_rate": 2.8777416734362307e-06,
+      "loss": 1.859,
+      "step": 1334
+    },
+    {
+      "epoch": 1.463013698630137,
+      "grad_norm": 2.4727842807769775,
+      "learning_rate": 2.8757108042242084e-06,
+      "loss": 1.7615,
+      "step": 1335
+    },
+    {
+      "epoch": 1.464109589041096,
+      "grad_norm": 2.179358959197998,
+      "learning_rate": 2.8736799350121853e-06,
+      "loss": 1.9085,
+      "step": 1336
+    },
+    {
+      "epoch": 1.4652054794520548,
+      "grad_norm": 2.220917224884033,
+      "learning_rate": 2.871649065800163e-06,
+      "loss": 1.8618,
+      "step": 1337
+    },
+    {
+      "epoch": 1.4663013698630136,
+      "grad_norm": 2.3807291984558105,
+      "learning_rate": 2.86961819658814e-06,
+      "loss": 1.88,
+      "step": 1338
+    },
+    {
+      "epoch": 1.4673972602739727,
+      "grad_norm": 2.272507667541504,
+      "learning_rate": 2.8675873273761175e-06,
+      "loss": 1.8045,
+      "step": 1339
+    },
+    {
+      "epoch": 1.4684931506849315,
+      "grad_norm": 2.1832377910614014,
+      "learning_rate": 2.8655564581640944e-06,
+      "loss": 1.8273,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4684931506849315,
+      "eval_loss": 2.0218358039855957,
+      "eval_runtime": 37.1223,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4695890410958903,
+      "grad_norm": 2.2655251026153564,
+      "learning_rate": 2.863525588952072e-06,
+      "loss": 1.8162,
+      "step": 1341
+    },
+    {
+      "epoch": 1.4706849315068493,
+      "grad_norm": 2.191852331161499,
+      "learning_rate": 2.861494719740049e-06,
+      "loss": 1.8778,
+      "step": 1342
+    },
+    {
+      "epoch": 1.4717808219178083,
+      "grad_norm": 2.1869187355041504,
+      "learning_rate": 2.8594638505280266e-06,
+      "loss": 1.7695,
+      "step": 1343
+    },
+    {
+      "epoch": 1.4728767123287672,
+      "grad_norm": 2.248108148574829,
+      "learning_rate": 2.8574329813160035e-06,
+      "loss": 1.7759,
+      "step": 1344
+    },
+    {
+      "epoch": 1.473972602739726,
+      "grad_norm": 2.2599704265594482,
+      "learning_rate": 2.855402112103981e-06,
+      "loss": 1.8214,
+      "step": 1345
+    },
+    {
+      "epoch": 1.475068493150685,
+      "grad_norm": 2.2920355796813965,
+      "learning_rate": 2.853371242891958e-06,
+      "loss": 1.8061,
+      "step": 1346
+    },
+    {
+      "epoch": 1.4761643835616438,
+      "grad_norm": 2.26715087890625,
+      "learning_rate": 2.851340373679935e-06,
+      "loss": 1.8244,
+      "step": 1347
+    },
+    {
+      "epoch": 1.4772602739726026,
+      "grad_norm": 2.349169969558716,
+      "learning_rate": 2.8493095044679126e-06,
+      "loss": 1.8401,
+      "step": 1348
+    },
+    {
+      "epoch": 1.4783561643835617,
+      "grad_norm": 2.3571698665618896,
+      "learning_rate": 2.8472786352558894e-06,
+      "loss": 1.8913,
+      "step": 1349
+    },
+    {
+      "epoch": 1.4794520547945205,
+      "grad_norm": 2.334057331085205,
+      "learning_rate": 2.845247766043867e-06,
+      "loss": 1.8434,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4794520547945205,
+      "eval_loss": 2.0251200199127197,
+      "eval_runtime": 37.1298,
+      "eval_samples_per_second": 26.933,
+      "eval_steps_per_second": 1.131,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4805479452054795,
+      "grad_norm": 2.363516330718994,
+      "learning_rate": 2.843216896831844e-06,
+      "loss": 1.7846,
+      "step": 1351
+    },
+    {
+      "epoch": 1.4816438356164383,
+      "grad_norm": 2.372403860092163,
+      "learning_rate": 2.8411860276198217e-06,
+      "loss": 1.7791,
+      "step": 1352
+    },
+    {
+      "epoch": 1.4827397260273973,
+      "grad_norm": 2.2666590213775635,
+      "learning_rate": 2.8391551584077985e-06,
+      "loss": 1.7969,
+      "step": 1353
+    },
+    {
+      "epoch": 1.4838356164383562,
+      "grad_norm": 2.3651468753814697,
+      "learning_rate": 2.8371242891957762e-06,
+      "loss": 1.8565,
+      "step": 1354
+    },
+    {
+      "epoch": 1.484931506849315,
+      "grad_norm": 2.3559513092041016,
+      "learning_rate": 2.835093419983753e-06,
+      "loss": 1.8765,
+      "step": 1355
+    },
+    {
+      "epoch": 1.486027397260274,
+      "grad_norm": 2.232041358947754,
+      "learning_rate": 2.833062550771731e-06,
+      "loss": 1.805,
+      "step": 1356
+    },
+    {
+      "epoch": 1.4871232876712328,
+      "grad_norm": 2.2423202991485596,
+      "learning_rate": 2.8310316815597076e-06,
+      "loss": 1.8137,
+      "step": 1357
+    },
+    {
+      "epoch": 1.4882191780821918,
+      "grad_norm": 2.3454222679138184,
+      "learning_rate": 2.8290008123476853e-06,
+      "loss": 1.7839,
+      "step": 1358
+    },
+    {
+      "epoch": 1.4893150684931507,
+      "grad_norm": 2.351266384124756,
+      "learning_rate": 2.826969943135662e-06,
+      "loss": 1.8208,
+      "step": 1359
+    },
+    {
+      "epoch": 1.4904109589041097,
+      "grad_norm": 2.3025388717651367,
+      "learning_rate": 2.8249390739236395e-06,
+      "loss": 1.8059,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4904109589041097,
+      "eval_loss": 2.0235962867736816,
+      "eval_runtime": 37.1359,
+      "eval_samples_per_second": 26.928,
+      "eval_steps_per_second": 1.131,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4915068493150685,
+      "grad_norm": 2.2277750968933105,
+      "learning_rate": 2.8229082047116167e-06,
+      "loss": 1.8238,
+      "step": 1361
+    },
+    {
+      "epoch": 1.4926027397260273,
+      "grad_norm": 2.318286180496216,
+      "learning_rate": 2.820877335499594e-06,
+      "loss": 1.7867,
+      "step": 1362
+    },
+    {
+      "epoch": 1.4936986301369863,
+      "grad_norm": 2.296259880065918,
+      "learning_rate": 2.8188464662875713e-06,
+      "loss": 1.8068,
+      "step": 1363
+    },
+    {
+      "epoch": 1.4947945205479451,
+      "grad_norm": 2.2617714405059814,
+      "learning_rate": 2.8168155970755486e-06,
+      "loss": 1.7607,
+      "step": 1364
+    },
+    {
+      "epoch": 1.4958904109589042,
+      "grad_norm": 2.318294048309326,
+      "learning_rate": 2.814784727863526e-06,
+      "loss": 1.7872,
+      "step": 1365
+    },
+    {
+      "epoch": 1.496986301369863,
+      "grad_norm": 2.260286808013916,
+      "learning_rate": 2.812753858651503e-06,
+      "loss": 1.8013,
+      "step": 1366
+    },
+    {
+      "epoch": 1.498082191780822,
+      "grad_norm": 2.222074508666992,
+      "learning_rate": 2.8107229894394804e-06,
+      "loss": 1.8338,
+      "step": 1367
+    },
+    {
+      "epoch": 1.4991780821917808,
+      "grad_norm": 2.6403002738952637,
+      "learning_rate": 2.8086921202274577e-06,
+      "loss": 1.8126,
+      "step": 1368
+    },
+    {
+      "epoch": 1.5002739726027396,
+      "grad_norm": 2.2439680099487305,
+      "learning_rate": 2.806661251015435e-06,
+      "loss": 1.7669,
+      "step": 1369
+    },
+    {
+      "epoch": 1.5013698630136987,
+      "grad_norm": 2.3914687633514404,
+      "learning_rate": 2.8046303818034122e-06,
+      "loss": 1.8421,
+      "step": 1370
+    },
+    {
+      "epoch": 1.5013698630136987,
+      "eval_loss": 2.0222702026367188,
+      "eval_runtime": 37.1598,
+      "eval_samples_per_second": 26.911,
+      "eval_steps_per_second": 1.13,
+      "step": 1370
+    },
+    {
+      "epoch": 1.5024657534246575,
+      "grad_norm": 2.3510093688964844,
+      "learning_rate": 2.802599512591389e-06,
+      "loss": 1.8902,
+      "step": 1371
+    },
+    {
+      "epoch": 1.5035616438356163,
+      "grad_norm": 2.4427363872528076,
+      "learning_rate": 2.8005686433793668e-06,
+      "loss": 1.9376,
+      "step": 1372
+    },
+    {
+      "epoch": 1.5046575342465753,
+      "grad_norm": 2.299356460571289,
+      "learning_rate": 2.7985377741673436e-06,
+      "loss": 1.812,
+      "step": 1373
+    },
+    {
+      "epoch": 1.5057534246575344,
+      "grad_norm": 2.3738906383514404,
+      "learning_rate": 2.7965069049553213e-06,
+      "loss": 1.7666,
+      "step": 1374
+    },
+    {
+      "epoch": 1.5068493150684932,
+      "grad_norm": 2.3620493412017822,
+      "learning_rate": 2.794476035743298e-06,
+      "loss": 1.824,
+      "step": 1375
+    },
+    {
+      "epoch": 1.507945205479452,
+      "grad_norm": 2.539048671722412,
+      "learning_rate": 2.792445166531276e-06,
+      "loss": 1.8712,
+      "step": 1376
+    },
+    {
+      "epoch": 1.509041095890411,
+      "grad_norm": 2.2931456565856934,
+      "learning_rate": 2.7904142973192527e-06,
+      "loss": 1.8414,
+      "step": 1377
+    },
+    {
+      "epoch": 1.5101369863013698,
+      "grad_norm": 2.4368224143981934,
+      "learning_rate": 2.7883834281072304e-06,
+      "loss": 1.8419,
+      "step": 1378
+    },
+    {
+      "epoch": 1.5112328767123286,
+      "grad_norm": 2.298302173614502,
+      "learning_rate": 2.7863525588952073e-06,
+      "loss": 1.7846,
+      "step": 1379
+    },
+    {
+      "epoch": 1.5123287671232877,
+      "grad_norm": 2.3726806640625,
+      "learning_rate": 2.784321689683185e-06,
+      "loss": 1.7733,
+      "step": 1380
+    },
+    {
+      "epoch": 1.5123287671232877,
+      "eval_loss": 2.0201807022094727,
+      "eval_runtime": 37.1012,
+      "eval_samples_per_second": 26.953,
+      "eval_steps_per_second": 1.132,
+      "step": 1380
+    },
+    {
+      "epoch": 1.5134246575342467,
+      "grad_norm": 2.6515002250671387,
+      "learning_rate": 2.782290820471162e-06,
+      "loss": 1.7764,
+      "step": 1381
+    },
+    {
+      "epoch": 1.5145205479452055,
+      "grad_norm": 2.273272752761841,
+      "learning_rate": 2.7802599512591395e-06,
+      "loss": 1.7747,
+      "step": 1382
+    },
+    {
+      "epoch": 1.5156164383561643,
+      "grad_norm": 2.570009708404541,
+      "learning_rate": 2.7782290820471164e-06,
+      "loss": 1.8612,
+      "step": 1383
+    },
+    {
+      "epoch": 1.5167123287671234,
+      "grad_norm": 2.567878007888794,
+      "learning_rate": 2.7761982128350932e-06,
+      "loss": 1.8402,
+      "step": 1384
+    },
+    {
+      "epoch": 1.5178082191780822,
+      "grad_norm": 2.3768470287323,
+      "learning_rate": 2.774167343623071e-06,
+      "loss": 1.8473,
+      "step": 1385
+    },
+    {
+      "epoch": 1.518904109589041,
+      "grad_norm": 2.1383397579193115,
+      "learning_rate": 2.772136474411048e-06,
+      "loss": 1.8593,
+      "step": 1386
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 2.4437034130096436,
+      "learning_rate": 2.7701056051990255e-06,
+      "loss": 1.792,
+      "step": 1387
+    },
+    {
+      "epoch": 1.521095890410959,
+      "grad_norm": 2.433933734893799,
+      "learning_rate": 2.7680747359870023e-06,
+      "loss": 1.7312,
+      "step": 1388
+    },
+    {
+      "epoch": 1.5221917808219179,
+      "grad_norm": 2.340994119644165,
+      "learning_rate": 2.76604386677498e-06,
+      "loss": 1.8326,
+      "step": 1389
+    },
+    {
+      "epoch": 1.5232876712328767,
+      "grad_norm": 2.4730165004730225,
+      "learning_rate": 2.764012997562957e-06,
+      "loss": 1.8129,
+      "step": 1390
+    },
+    {
+      "epoch": 1.5232876712328767,
+      "eval_loss": 2.0158324241638184,
+      "eval_runtime": 37.0997,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 1390
+    },
+    {
+      "epoch": 1.5243835616438357,
+      "grad_norm": 2.2898151874542236,
+      "learning_rate": 2.7619821283509346e-06,
+      "loss": 1.8726,
+      "step": 1391
+    },
+    {
+      "epoch": 1.5254794520547945,
+      "grad_norm": 2.255439519882202,
+      "learning_rate": 2.7599512591389114e-06,
+      "loss": 1.7587,
+      "step": 1392
+    },
+    {
+      "epoch": 1.5265753424657533,
+      "grad_norm": 2.4269862174987793,
+      "learning_rate": 2.757920389926889e-06,
+      "loss": 1.8498,
+      "step": 1393
+    },
+    {
+      "epoch": 1.5276712328767124,
+      "grad_norm": 2.3411412239074707,
+      "learning_rate": 2.755889520714866e-06,
+      "loss": 1.8103,
+      "step": 1394
+    },
+    {
+      "epoch": 1.5287671232876714,
+      "grad_norm": 2.2859277725219727,
+      "learning_rate": 2.7538586515028437e-06,
+      "loss": 1.7909,
+      "step": 1395
+    },
+    {
+      "epoch": 1.52986301369863,
+      "grad_norm": 2.2675201892852783,
+      "learning_rate": 2.7518277822908205e-06,
+      "loss": 1.7672,
+      "step": 1396
+    },
+    {
+      "epoch": 1.530958904109589,
+      "grad_norm": 2.240610361099243,
+      "learning_rate": 2.749796913078798e-06,
+      "loss": 1.8162,
+      "step": 1397
+    },
+    {
+      "epoch": 1.532054794520548,
+      "grad_norm": 2.272045373916626,
+      "learning_rate": 2.747766043866775e-06,
+      "loss": 1.8593,
+      "step": 1398
+    },
+    {
+      "epoch": 1.5331506849315069,
+      "grad_norm": 2.1736183166503906,
+      "learning_rate": 2.7457351746547524e-06,
+      "loss": 1.8154,
+      "step": 1399
+    },
+    {
+      "epoch": 1.5342465753424657,
+      "grad_norm": 2.197085380554199,
+      "learning_rate": 2.7437043054427297e-06,
+      "loss": 1.8758,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5342465753424657,
+      "eval_loss": 2.017298698425293,
+      "eval_runtime": 37.1666,
+      "eval_samples_per_second": 26.906,
+      "eval_steps_per_second": 1.13,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5353424657534247,
+      "grad_norm": 2.2274084091186523,
+      "learning_rate": 2.741673436230707e-06,
+      "loss": 1.9241,
+      "step": 1401
+    },
+    {
+      "epoch": 1.5364383561643835,
+      "grad_norm": 2.3320305347442627,
+      "learning_rate": 2.739642567018684e-06,
+      "loss": 1.8209,
+      "step": 1402
+    },
+    {
+      "epoch": 1.5375342465753423,
+      "grad_norm": 2.370651960372925,
+      "learning_rate": 2.7376116978066615e-06,
+      "loss": 1.7897,
+      "step": 1403
+    },
+    {
+      "epoch": 1.5386301369863014,
+      "grad_norm": 2.387906312942505,
+      "learning_rate": 2.7355808285946388e-06,
+      "loss": 1.8385,
+      "step": 1404
+    },
+    {
+      "epoch": 1.5397260273972604,
+      "grad_norm": 2.340228796005249,
+      "learning_rate": 2.733549959382616e-06,
+      "loss": 1.8279,
+      "step": 1405
+    },
+    {
+      "epoch": 1.5408219178082192,
+      "grad_norm": 2.338785409927368,
+      "learning_rate": 2.7315190901705933e-06,
+      "loss": 1.7887,
+      "step": 1406
+    },
+    {
+      "epoch": 1.541917808219178,
+      "grad_norm": 2.350057601928711,
+      "learning_rate": 2.7294882209585706e-06,
+      "loss": 1.8772,
+      "step": 1407
+    },
+    {
+      "epoch": 1.543013698630137,
+      "grad_norm": 2.1129322052001953,
+      "learning_rate": 2.727457351746548e-06,
+      "loss": 1.8393,
+      "step": 1408
+    },
+    {
+      "epoch": 1.5441095890410959,
+      "grad_norm": 2.282553195953369,
+      "learning_rate": 2.725426482534525e-06,
+      "loss": 1.7893,
+      "step": 1409
+    },
+    {
+      "epoch": 1.5452054794520547,
+      "grad_norm": 2.3291871547698975,
+      "learning_rate": 2.723395613322502e-06,
+      "loss": 1.7921,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5452054794520547,
+      "eval_loss": 2.0166046619415283,
+      "eval_runtime": 36.7906,
+      "eval_samples_per_second": 27.181,
+      "eval_steps_per_second": 1.142,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5463013698630137,
+      "grad_norm": 2.3050696849823,
+      "learning_rate": 2.7213647441104797e-06,
+      "loss": 1.8292,
+      "step": 1411
+    },
+    {
+      "epoch": 1.5473972602739727,
+      "grad_norm": 2.2762458324432373,
+      "learning_rate": 2.7193338748984565e-06,
+      "loss": 1.811,
+      "step": 1412
+    },
+    {
+      "epoch": 1.5484931506849315,
+      "grad_norm": 2.5544228553771973,
+      "learning_rate": 2.7173030056864342e-06,
+      "loss": 1.8203,
+      "step": 1413
+    },
+    {
+      "epoch": 1.5495890410958904,
+      "grad_norm": 2.6392276287078857,
+      "learning_rate": 2.715272136474411e-06,
+      "loss": 1.9724,
+      "step": 1414
+    },
+    {
+      "epoch": 1.5506849315068494,
+      "grad_norm": 2.445366382598877,
+      "learning_rate": 2.713241267262389e-06,
+      "loss": 1.7589,
+      "step": 1415
+    },
+    {
+      "epoch": 1.5517808219178082,
+      "grad_norm": 2.5305662155151367,
+      "learning_rate": 2.7112103980503656e-06,
+      "loss": 1.8311,
+      "step": 1416
+    },
+    {
+      "epoch": 1.552876712328767,
+      "grad_norm": 2.405482769012451,
+      "learning_rate": 2.7091795288383433e-06,
+      "loss": 1.7893,
+      "step": 1417
+    },
+    {
+      "epoch": 1.553972602739726,
+      "grad_norm": 2.258361577987671,
+      "learning_rate": 2.70714865962632e-06,
+      "loss": 1.8489,
+      "step": 1418
+    },
+    {
+      "epoch": 1.555068493150685,
+      "grad_norm": 2.1798577308654785,
+      "learning_rate": 2.705117790414298e-06,
+      "loss": 1.7983,
+      "step": 1419
+    },
+    {
+      "epoch": 1.5561643835616439,
+      "grad_norm": 2.1985363960266113,
+      "learning_rate": 2.7030869212022747e-06,
+      "loss": 1.7934,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5561643835616439,
+      "eval_loss": 2.0219907760620117,
+      "eval_runtime": 36.9416,
+      "eval_samples_per_second": 27.07,
+      "eval_steps_per_second": 1.137,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5572602739726027,
+      "grad_norm": 2.273252248764038,
+      "learning_rate": 2.7010560519902516e-06,
+      "loss": 1.8571,
+      "step": 1421
+    },
+    {
+      "epoch": 1.5583561643835617,
+      "grad_norm": 2.22920298576355,
+      "learning_rate": 2.6990251827782293e-06,
+      "loss": 1.7785,
+      "step": 1422
+    },
+    {
+      "epoch": 1.5594520547945205,
+      "grad_norm": 2.1784098148345947,
+      "learning_rate": 2.696994313566206e-06,
+      "loss": 1.8053,
+      "step": 1423
+    },
+    {
+      "epoch": 1.5605479452054793,
+      "grad_norm": 2.223404884338379,
+      "learning_rate": 2.694963444354184e-06,
+      "loss": 1.7965,
+      "step": 1424
+    },
+    {
+      "epoch": 1.5616438356164384,
+      "grad_norm": 2.2004787921905518,
+      "learning_rate": 2.6929325751421607e-06,
+      "loss": 1.8851,
+      "step": 1425
+    },
+    {
+      "epoch": 1.5627397260273974,
+      "grad_norm": 2.202467441558838,
+      "learning_rate": 2.6909017059301384e-06,
+      "loss": 1.7884,
+      "step": 1426
+    },
+    {
+      "epoch": 1.563835616438356,
+      "grad_norm": 2.337430238723755,
+      "learning_rate": 2.6888708367181153e-06,
+      "loss": 1.8543,
+      "step": 1427
+    },
+    {
+      "epoch": 1.564931506849315,
+      "grad_norm": 2.322848081588745,
+      "learning_rate": 2.686839967506093e-06,
+      "loss": 1.7607,
+      "step": 1428
+    },
+    {
+      "epoch": 1.566027397260274,
+      "grad_norm": 2.277765989303589,
+      "learning_rate": 2.68480909829407e-06,
+      "loss": 1.8354,
+      "step": 1429
+    },
+    {
+      "epoch": 1.5671232876712329,
+      "grad_norm": 2.3229448795318604,
+      "learning_rate": 2.6827782290820475e-06,
+      "loss": 1.7277,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5671232876712329,
+      "eval_loss": 2.0220985412597656,
+      "eval_runtime": 37.0054,
+      "eval_samples_per_second": 27.023,
+      "eval_steps_per_second": 1.135,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5682191780821917,
+      "grad_norm": 2.49545955657959,
+      "learning_rate": 2.6807473598700244e-06,
+      "loss": 1.7351,
+      "step": 1431
+    },
+    {
+      "epoch": 1.5693150684931507,
+      "grad_norm": 2.530553102493286,
+      "learning_rate": 2.678716490658002e-06,
+      "loss": 1.8614,
+      "step": 1432
+    },
+    {
+      "epoch": 1.5704109589041098,
+      "grad_norm": 2.3352842330932617,
+      "learning_rate": 2.676685621445979e-06,
+      "loss": 1.8218,
+      "step": 1433
+    },
+    {
+      "epoch": 1.5715068493150683,
+      "grad_norm": 2.3533058166503906,
+      "learning_rate": 2.674654752233956e-06,
+      "loss": 1.8282,
+      "step": 1434
+    },
+    {
+      "epoch": 1.5726027397260274,
+      "grad_norm": 2.5509746074676514,
+      "learning_rate": 2.6726238830219335e-06,
+      "loss": 1.8333,
+      "step": 1435
+    },
+    {
+      "epoch": 1.5736986301369864,
+      "grad_norm": 2.185175895690918,
+      "learning_rate": 2.6705930138099107e-06,
+      "loss": 1.8299,
+      "step": 1436
+    },
+    {
+      "epoch": 1.5747945205479452,
+      "grad_norm": 2.2338390350341797,
+      "learning_rate": 2.668562144597888e-06,
+      "loss": 1.857,
+      "step": 1437
+    },
+    {
+      "epoch": 1.575890410958904,
+      "grad_norm": 2.4094536304473877,
+      "learning_rate": 2.6665312753858653e-06,
+      "loss": 1.7333,
+      "step": 1438
+    },
+    {
+      "epoch": 1.576986301369863,
+      "grad_norm": 2.4650957584381104,
+      "learning_rate": 2.6645004061738426e-06,
+      "loss": 1.8237,
+      "step": 1439
+    },
+    {
+      "epoch": 1.5780821917808219,
+      "grad_norm": 2.427415132522583,
+      "learning_rate": 2.66246953696182e-06,
+      "loss": 1.8206,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5780821917808219,
+      "eval_loss": 2.0222699642181396,
+      "eval_runtime": 37.0252,
+      "eval_samples_per_second": 27.009,
+      "eval_steps_per_second": 1.134,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5791780821917807,
+      "grad_norm": 2.3878679275512695,
+      "learning_rate": 2.660438667749797e-06,
+      "loss": 1.775,
+      "step": 1441
+    },
+    {
+      "epoch": 1.5802739726027397,
+      "grad_norm": 2.317819595336914,
+      "learning_rate": 2.6584077985377744e-06,
+      "loss": 1.7479,
+      "step": 1442
+    },
+    {
+      "epoch": 1.5813698630136988,
+      "grad_norm": 2.34951114654541,
+      "learning_rate": 2.6563769293257517e-06,
+      "loss": 1.8025,
+      "step": 1443
+    },
+    {
+      "epoch": 1.5824657534246576,
+      "grad_norm": 2.2849795818328857,
+      "learning_rate": 2.654346060113729e-06,
+      "loss": 1.8152,
+      "step": 1444
+    },
+    {
+      "epoch": 1.5835616438356164,
+      "grad_norm": 2.211489677429199,
+      "learning_rate": 2.6523151909017062e-06,
+      "loss": 1.7666,
+      "step": 1445
+    },
+    {
+      "epoch": 1.5846575342465754,
+      "grad_norm": 2.1876230239868164,
+      "learning_rate": 2.6502843216896835e-06,
+      "loss": 1.7933,
+      "step": 1446
+    },
+    {
+      "epoch": 1.5857534246575342,
+      "grad_norm": 2.367628335952759,
+      "learning_rate": 2.6482534524776603e-06,
+      "loss": 1.7756,
+      "step": 1447
+    },
+    {
+      "epoch": 1.586849315068493,
+      "grad_norm": 2.1198348999023438,
+      "learning_rate": 2.646222583265638e-06,
+      "loss": 1.8908,
+      "step": 1448
+    },
+    {
+      "epoch": 1.587945205479452,
+      "grad_norm": 2.192788600921631,
+      "learning_rate": 2.644191714053615e-06,
+      "loss": 1.8432,
+      "step": 1449
+    },
+    {
+      "epoch": 1.589041095890411,
+      "grad_norm": 2.4480655193328857,
+      "learning_rate": 2.6421608448415926e-06,
+      "loss": 1.8223,
+      "step": 1450
+    },
+    {
+      "epoch": 1.589041095890411,
+      "eval_loss": 2.019976854324341,
+      "eval_runtime": 37.0117,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1450
+    },
+    {
+      "epoch": 1.59013698630137,
+      "grad_norm": 2.573078155517578,
+      "learning_rate": 2.6401299756295694e-06,
+      "loss": 1.7881,
+      "step": 1451
+    },
+    {
+      "epoch": 1.5912328767123287,
+      "grad_norm": 2.2820355892181396,
+      "learning_rate": 2.638099106417547e-06,
+      "loss": 1.7703,
+      "step": 1452
+    },
+    {
+      "epoch": 1.5923287671232877,
+      "grad_norm": 2.526477575302124,
+      "learning_rate": 2.636068237205524e-06,
+      "loss": 1.8281,
+      "step": 1453
+    },
+    {
+      "epoch": 1.5934246575342466,
+      "grad_norm": 2.873647689819336,
+      "learning_rate": 2.6340373679935017e-06,
+      "loss": 1.7771,
+      "step": 1454
+    },
+    {
+      "epoch": 1.5945205479452054,
+      "grad_norm": 2.3782413005828857,
+      "learning_rate": 2.6320064987814785e-06,
+      "loss": 1.8864,
+      "step": 1455
+    },
+    {
+      "epoch": 1.5956164383561644,
+      "grad_norm": 2.3153343200683594,
+      "learning_rate": 2.6299756295694562e-06,
+      "loss": 1.734,
+      "step": 1456
+    },
+    {
+      "epoch": 1.5967123287671234,
+      "grad_norm": 2.631880283355713,
+      "learning_rate": 2.627944760357433e-06,
+      "loss": 1.8295,
+      "step": 1457
+    },
+    {
+      "epoch": 1.5978082191780822,
+      "grad_norm": 2.5827910900115967,
+      "learning_rate": 2.6259138911454104e-06,
+      "loss": 1.8738,
+      "step": 1458
+    },
+    {
+      "epoch": 1.598904109589041,
+      "grad_norm": 2.3519935607910156,
+      "learning_rate": 2.6238830219333877e-06,
+      "loss": 1.8832,
+      "step": 1459
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 2.5236215591430664,
+      "learning_rate": 2.621852152721365e-06,
+      "loss": 1.8595,
+      "step": 1460
+    },
+    {
+      "epoch": 1.6,
+      "eval_loss": 2.0170512199401855,
+      "eval_runtime": 37.0302,
+      "eval_samples_per_second": 27.005,
+      "eval_steps_per_second": 1.134,
+      "step": 1460
+    },
+    {
+      "epoch": 1.601095890410959,
+      "grad_norm": 2.725083112716675,
+      "learning_rate": 2.619821283509342e-06,
+      "loss": 1.7476,
+      "step": 1461
+    },
+    {
+      "epoch": 1.6021917808219177,
+      "grad_norm": 2.212644100189209,
+      "learning_rate": 2.6177904142973195e-06,
+      "loss": 1.8475,
+      "step": 1462
+    },
+    {
+      "epoch": 1.6032876712328767,
+      "grad_norm": 2.346808910369873,
+      "learning_rate": 2.6157595450852968e-06,
+      "loss": 1.7762,
+      "step": 1463
+    },
+    {
+      "epoch": 1.6043835616438358,
+      "grad_norm": 2.702331304550171,
+      "learning_rate": 2.613728675873274e-06,
+      "loss": 1.8296,
+      "step": 1464
+    },
+    {
+      "epoch": 1.6054794520547944,
+      "grad_norm": 2.4175899028778076,
+      "learning_rate": 2.6116978066612513e-06,
+      "loss": 1.8476,
+      "step": 1465
+    },
+    {
+      "epoch": 1.6065753424657534,
+      "grad_norm": 2.325457811355591,
+      "learning_rate": 2.6096669374492286e-06,
+      "loss": 1.8706,
+      "step": 1466
+    },
+    {
+      "epoch": 1.6076712328767124,
+      "grad_norm": 2.383842945098877,
+      "learning_rate": 2.607636068237206e-06,
+      "loss": 1.8246,
+      "step": 1467
+    },
+    {
+      "epoch": 1.6087671232876712,
+      "grad_norm": 2.289515256881714,
+      "learning_rate": 2.605605199025183e-06,
+      "loss": 1.7728,
+      "step": 1468
+    },
+    {
+      "epoch": 1.60986301369863,
+      "grad_norm": 2.5116689205169678,
+      "learning_rate": 2.6035743298131604e-06,
+      "loss": 1.8359,
+      "step": 1469
+    },
+    {
+      "epoch": 1.610958904109589,
+      "grad_norm": 2.279695749282837,
+      "learning_rate": 2.6015434606011377e-06,
+      "loss": 1.8155,
+      "step": 1470
+    },
+    {
+      "epoch": 1.610958904109589,
+      "eval_loss": 2.0142691135406494,
+      "eval_runtime": 37.0119,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1470
+    },
+    {
+      "epoch": 1.612054794520548,
+      "grad_norm": 2.4397690296173096,
+      "learning_rate": 2.5995125913891145e-06,
+      "loss": 1.8424,
+      "step": 1471
+    },
+    {
+      "epoch": 1.6131506849315067,
+      "grad_norm": 2.433577299118042,
+      "learning_rate": 2.5974817221770922e-06,
+      "loss": 1.8196,
+      "step": 1472
+    },
+    {
+      "epoch": 1.6142465753424657,
+      "grad_norm": 2.365459442138672,
+      "learning_rate": 2.595450852965069e-06,
+      "loss": 1.8124,
+      "step": 1473
+    },
+    {
+      "epoch": 1.6153424657534248,
+      "grad_norm": 2.253133535385132,
+      "learning_rate": 2.593419983753047e-06,
+      "loss": 1.8443,
+      "step": 1474
+    },
+    {
+      "epoch": 1.6164383561643836,
+      "grad_norm": 2.227625608444214,
+      "learning_rate": 2.5913891145410236e-06,
+      "loss": 1.8524,
+      "step": 1475
+    },
+    {
+      "epoch": 1.6175342465753424,
+      "grad_norm": 2.403702974319458,
+      "learning_rate": 2.5893582453290013e-06,
+      "loss": 1.7983,
+      "step": 1476
+    },
+    {
+      "epoch": 1.6186301369863014,
+      "grad_norm": 2.24507999420166,
+      "learning_rate": 2.587327376116978e-06,
+      "loss": 1.8563,
+      "step": 1477
+    },
+    {
+      "epoch": 1.6197260273972602,
+      "grad_norm": 2.284700393676758,
+      "learning_rate": 2.585296506904956e-06,
+      "loss": 1.7958,
+      "step": 1478
+    },
+    {
+      "epoch": 1.620821917808219,
+      "grad_norm": 2.288215398788452,
+      "learning_rate": 2.5832656376929327e-06,
+      "loss": 1.8021,
+      "step": 1479
+    },
+    {
+      "epoch": 1.621917808219178,
+      "grad_norm": 2.2312939167022705,
+      "learning_rate": 2.5812347684809104e-06,
+      "loss": 1.7856,
+      "step": 1480
+    },
+    {
+      "epoch": 1.621917808219178,
+      "eval_loss": 2.01674222946167,
+      "eval_runtime": 36.9943,
+      "eval_samples_per_second": 27.031,
+      "eval_steps_per_second": 1.135,
+      "step": 1480
+    },
+    {
+      "epoch": 1.6230136986301371,
+      "grad_norm": 2.319363832473755,
+      "learning_rate": 2.5792038992688873e-06,
+      "loss": 1.7263,
+      "step": 1481
+    },
+    {
+      "epoch": 1.624109589041096,
+      "grad_norm": 2.4342033863067627,
+      "learning_rate": 2.577173030056865e-06,
+      "loss": 1.897,
+      "step": 1482
+    },
+    {
+      "epoch": 1.6252054794520547,
+      "grad_norm": 2.3052282333374023,
+      "learning_rate": 2.575142160844842e-06,
+      "loss": 1.7165,
+      "step": 1483
+    },
+    {
+      "epoch": 1.6263013698630138,
+      "grad_norm": 2.2476744651794434,
+      "learning_rate": 2.5731112916328187e-06,
+      "loss": 1.8347,
+      "step": 1484
+    },
+    {
+      "epoch": 1.6273972602739726,
+      "grad_norm": 2.3128387928009033,
+      "learning_rate": 2.5710804224207964e-06,
+      "loss": 1.8403,
+      "step": 1485
+    },
+    {
+      "epoch": 1.6284931506849314,
+      "grad_norm": 2.3483974933624268,
+      "learning_rate": 2.5690495532087733e-06,
+      "loss": 1.781,
+      "step": 1486
+    },
+    {
+      "epoch": 1.6295890410958904,
+      "grad_norm": 2.328298568725586,
+      "learning_rate": 2.567018683996751e-06,
+      "loss": 1.7356,
+      "step": 1487
+    },
+    {
+      "epoch": 1.6306849315068495,
+      "grad_norm": 2.2756922245025635,
+      "learning_rate": 2.564987814784728e-06,
+      "loss": 1.8449,
+      "step": 1488
+    },
+    {
+      "epoch": 1.6317808219178083,
+      "grad_norm": 2.363874912261963,
+      "learning_rate": 2.5629569455727055e-06,
+      "loss": 1.7767,
+      "step": 1489
+    },
+    {
+      "epoch": 1.632876712328767,
+      "grad_norm": 2.2872767448425293,
+      "learning_rate": 2.5609260763606824e-06,
+      "loss": 1.8052,
+      "step": 1490
+    },
+    {
+      "epoch": 1.632876712328767,
+      "eval_loss": 2.0134243965148926,
+      "eval_runtime": 37.029,
+      "eval_samples_per_second": 27.006,
+      "eval_steps_per_second": 1.134,
+      "step": 1490
+    },
+    {
+      "epoch": 1.633972602739726,
+      "grad_norm": 2.354851007461548,
+      "learning_rate": 2.55889520714866e-06,
+      "loss": 1.7953,
+      "step": 1491
+    },
+    {
+      "epoch": 1.635068493150685,
+      "grad_norm": 2.4440531730651855,
+      "learning_rate": 2.556864337936637e-06,
+      "loss": 1.9063,
+      "step": 1492
+    },
+    {
+      "epoch": 1.6361643835616437,
+      "grad_norm": 2.296835422515869,
+      "learning_rate": 2.5548334687246146e-06,
+      "loss": 1.7961,
+      "step": 1493
+    },
+    {
+      "epoch": 1.6372602739726028,
+      "grad_norm": 2.7421159744262695,
+      "learning_rate": 2.5528025995125915e-06,
+      "loss": 1.809,
+      "step": 1494
+    },
+    {
+      "epoch": 1.6383561643835618,
+      "grad_norm": 2.819125175476074,
+      "learning_rate": 2.550771730300569e-06,
+      "loss": 1.853,
+      "step": 1495
+    },
+    {
+      "epoch": 1.6394520547945204,
+      "grad_norm": 2.2536680698394775,
+      "learning_rate": 2.548740861088546e-06,
+      "loss": 1.8297,
+      "step": 1496
+    },
+    {
+      "epoch": 1.6405479452054794,
+      "grad_norm": 2.7130825519561768,
+      "learning_rate": 2.5467099918765233e-06,
+      "loss": 1.843,
+      "step": 1497
+    },
+    {
+      "epoch": 1.6416438356164385,
+      "grad_norm": 3.0625038146972656,
+      "learning_rate": 2.5446791226645006e-06,
+      "loss": 1.7462,
+      "step": 1498
+    },
+    {
+      "epoch": 1.6427397260273973,
+      "grad_norm": 2.3095271587371826,
+      "learning_rate": 2.542648253452478e-06,
+      "loss": 1.7363,
+      "step": 1499
+    },
+    {
+      "epoch": 1.643835616438356,
+      "grad_norm": 2.536547899246216,
+      "learning_rate": 2.540617384240455e-06,
+      "loss": 1.8422,
+      "step": 1500
+    },
+    {
+      "epoch": 1.643835616438356,
+      "eval_loss": 2.0126097202301025,
+      "eval_runtime": 37.0469,
+      "eval_samples_per_second": 26.993,
+      "eval_steps_per_second": 1.134,
+      "step": 1500
+    },
+    {
+      "epoch": 1.644931506849315,
+      "grad_norm": 2.7454309463500977,
+      "learning_rate": 2.5385865150284324e-06,
+      "loss": 1.8183,
+      "step": 1501
+    },
+    {
+      "epoch": 1.6460273972602741,
+      "grad_norm": 2.3443901538848877,
+      "learning_rate": 2.5365556458164097e-06,
+      "loss": 1.7827,
+      "step": 1502
+    },
+    {
+      "epoch": 1.6471232876712327,
+      "grad_norm": 2.2412657737731934,
+      "learning_rate": 2.534524776604387e-06,
+      "loss": 1.797,
+      "step": 1503
+    },
+    {
+      "epoch": 1.6482191780821918,
+      "grad_norm": 2.6193699836730957,
+      "learning_rate": 2.5324939073923642e-06,
+      "loss": 1.8103,
+      "step": 1504
+    },
+    {
+      "epoch": 1.6493150684931508,
+      "grad_norm": 2.363919258117676,
+      "learning_rate": 2.5304630381803415e-06,
+      "loss": 1.6987,
+      "step": 1505
+    },
+    {
+      "epoch": 1.6504109589041096,
+      "grad_norm": 2.303164482116699,
+      "learning_rate": 2.5284321689683188e-06,
+      "loss": 1.8005,
+      "step": 1506
+    },
+    {
+      "epoch": 1.6515068493150684,
+      "grad_norm": 2.444174289703369,
+      "learning_rate": 2.526401299756296e-06,
+      "loss": 1.8009,
+      "step": 1507
+    },
+    {
+      "epoch": 1.6526027397260274,
+      "grad_norm": 2.418454647064209,
+      "learning_rate": 2.524370430544273e-06,
+      "loss": 1.782,
+      "step": 1508
+    },
+    {
+      "epoch": 1.6536986301369863,
+      "grad_norm": 2.48840069770813,
+      "learning_rate": 2.5223395613322506e-06,
+      "loss": 1.8143,
+      "step": 1509
+    },
+    {
+      "epoch": 1.654794520547945,
+      "grad_norm": 2.2814347743988037,
+      "learning_rate": 2.5203086921202274e-06,
+      "loss": 1.7687,
+      "step": 1510
+    },
+    {
+      "epoch": 1.654794520547945,
+      "eval_loss": 2.0148253440856934,
+      "eval_runtime": 36.7833,
+      "eval_samples_per_second": 27.186,
+      "eval_steps_per_second": 1.142,
+      "step": 1510
+    },
+    {
+      "epoch": 1.655890410958904,
+      "grad_norm": 2.228933572769165,
+      "learning_rate": 2.518277822908205e-06,
+      "loss": 1.8416,
+      "step": 1511
+    },
+    {
+      "epoch": 1.6569863013698631,
+      "grad_norm": 2.3696742057800293,
+      "learning_rate": 2.516246953696182e-06,
+      "loss": 1.7868,
+      "step": 1512
+    },
+    {
+      "epoch": 1.658082191780822,
+      "grad_norm": 2.216360092163086,
+      "learning_rate": 2.5142160844841597e-06,
+      "loss": 1.8312,
+      "step": 1513
+    },
+    {
+      "epoch": 1.6591780821917808,
+      "grad_norm": 2.454345226287842,
+      "learning_rate": 2.5121852152721365e-06,
+      "loss": 1.8309,
+      "step": 1514
+    },
+    {
+      "epoch": 1.6602739726027398,
+      "grad_norm": 2.2106776237487793,
+      "learning_rate": 2.5101543460601142e-06,
+      "loss": 1.8155,
+      "step": 1515
+    },
+    {
+      "epoch": 1.6613698630136986,
+      "grad_norm": 2.289006233215332,
+      "learning_rate": 2.508123476848091e-06,
+      "loss": 1.8162,
+      "step": 1516
+    },
+    {
+      "epoch": 1.6624657534246574,
+      "grad_norm": 2.297246217727661,
+      "learning_rate": 2.506092607636069e-06,
+      "loss": 1.7559,
+      "step": 1517
+    },
+    {
+      "epoch": 1.6635616438356164,
+      "grad_norm": 2.3909482955932617,
+      "learning_rate": 2.5040617384240457e-06,
+      "loss": 1.8389,
+      "step": 1518
+    },
+    {
+      "epoch": 1.6646575342465755,
+      "grad_norm": 2.398171901702881,
+      "learning_rate": 2.5020308692120234e-06,
+      "loss": 1.7785,
+      "step": 1519
+    },
+    {
+      "epoch": 1.6657534246575343,
+      "grad_norm": 2.377707004547119,
+      "learning_rate": 2.5e-06,
+      "loss": 1.768,
+      "step": 1520
+    },
+    {
+      "epoch": 1.6657534246575343,
+      "eval_loss": 2.012815475463867,
+      "eval_runtime": 36.9603,
+      "eval_samples_per_second": 27.056,
+      "eval_steps_per_second": 1.136,
+      "step": 1520
+    },
+    {
+      "epoch": 1.666849315068493,
+      "grad_norm": 2.336160898208618,
+      "learning_rate": 2.4979691307879775e-06,
+      "loss": 1.8653,
+      "step": 1521
+    },
+    {
+      "epoch": 1.6679452054794521,
+      "grad_norm": 2.288424253463745,
+      "learning_rate": 2.4959382615759548e-06,
+      "loss": 1.7798,
+      "step": 1522
+    },
+    {
+      "epoch": 1.669041095890411,
+      "grad_norm": 2.2599472999572754,
+      "learning_rate": 2.493907392363932e-06,
+      "loss": 1.7736,
+      "step": 1523
+    },
+    {
+      "epoch": 1.6701369863013698,
+      "grad_norm": 2.242276430130005,
+      "learning_rate": 2.4918765231519093e-06,
+      "loss": 1.8753,
+      "step": 1524
+    },
+    {
+      "epoch": 1.6712328767123288,
+      "grad_norm": 2.1897733211517334,
+      "learning_rate": 2.4898456539398866e-06,
+      "loss": 1.8353,
+      "step": 1525
+    },
+    {
+      "epoch": 1.6723287671232878,
+      "grad_norm": 2.254009246826172,
+      "learning_rate": 2.487814784727864e-06,
+      "loss": 1.7825,
+      "step": 1526
+    },
+    {
+      "epoch": 1.6734246575342466,
+      "grad_norm": 2.3704216480255127,
+      "learning_rate": 2.4857839155158407e-06,
+      "loss": 1.8742,
+      "step": 1527
+    },
+    {
+      "epoch": 1.6745205479452054,
+      "grad_norm": 2.1590118408203125,
+      "learning_rate": 2.483753046303818e-06,
+      "loss": 1.7974,
+      "step": 1528
+    },
+    {
+      "epoch": 1.6756164383561645,
+      "grad_norm": 2.2389416694641113,
+      "learning_rate": 2.4817221770917953e-06,
+      "loss": 1.8007,
+      "step": 1529
+    },
+    {
+      "epoch": 1.6767123287671233,
+      "grad_norm": 2.324131488800049,
+      "learning_rate": 2.4796913078797725e-06,
+      "loss": 1.8079,
+      "step": 1530
+    },
+    {
+      "epoch": 1.6767123287671233,
+      "eval_loss": 2.013923168182373,
+      "eval_runtime": 37.0828,
+      "eval_samples_per_second": 26.967,
+      "eval_steps_per_second": 1.133,
+      "step": 1530
+    },
+    {
+      "epoch": 1.677808219178082,
+      "grad_norm": 2.406456708908081,
+      "learning_rate": 2.47766043866775e-06,
+      "loss": 1.8356,
+      "step": 1531
+    },
+    {
+      "epoch": 1.6789041095890411,
+      "grad_norm": 2.289919137954712,
+      "learning_rate": 2.475629569455727e-06,
+      "loss": 1.8239,
+      "step": 1532
+    },
+    {
+      "epoch": 1.6800000000000002,
+      "grad_norm": 2.2617733478546143,
+      "learning_rate": 2.4735987002437044e-06,
+      "loss": 1.8025,
+      "step": 1533
+    },
+    {
+      "epoch": 1.6810958904109587,
+      "grad_norm": 2.133652925491333,
+      "learning_rate": 2.4715678310316816e-06,
+      "loss": 1.7743,
+      "step": 1534
+    },
+    {
+      "epoch": 1.6821917808219178,
+      "grad_norm": 2.215322256088257,
+      "learning_rate": 2.469536961819659e-06,
+      "loss": 1.7979,
+      "step": 1535
+    },
+    {
+      "epoch": 1.6832876712328768,
+      "grad_norm": 2.336021900177002,
+      "learning_rate": 2.467506092607636e-06,
+      "loss": 1.8027,
+      "step": 1536
+    },
+    {
+      "epoch": 1.6843835616438356,
+      "grad_norm": 2.3567142486572266,
+      "learning_rate": 2.4654752233956135e-06,
+      "loss": 1.7877,
+      "step": 1537
+    },
+    {
+      "epoch": 1.6854794520547944,
+      "grad_norm": 2.2642927169799805,
+      "learning_rate": 2.4634443541835907e-06,
+      "loss": 1.7695,
+      "step": 1538
+    },
+    {
+      "epoch": 1.6865753424657535,
+      "grad_norm": 2.353303909301758,
+      "learning_rate": 2.461413484971568e-06,
+      "loss": 1.867,
+      "step": 1539
+    },
+    {
+      "epoch": 1.6876712328767123,
+      "grad_norm": 2.252074956893921,
+      "learning_rate": 2.4593826157595453e-06,
+      "loss": 1.8009,
+      "step": 1540
+    },
+    {
+      "epoch": 1.6876712328767123,
+      "eval_loss": 2.0113608837127686,
+      "eval_runtime": 37.0875,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 1540
+    },
+    {
+      "epoch": 1.688767123287671,
+      "grad_norm": 2.2303407192230225,
+      "learning_rate": 2.4573517465475226e-06,
+      "loss": 1.8881,
+      "step": 1541
+    },
+    {
+      "epoch": 1.6898630136986301,
+      "grad_norm": 2.3359081745147705,
+      "learning_rate": 2.4553208773355e-06,
+      "loss": 1.7388,
+      "step": 1542
+    },
+    {
+      "epoch": 1.6909589041095892,
+      "grad_norm": 2.3783583641052246,
+      "learning_rate": 2.453290008123477e-06,
+      "loss": 1.7993,
+      "step": 1543
+    },
+    {
+      "epoch": 1.692054794520548,
+      "grad_norm": 2.159264326095581,
+      "learning_rate": 2.4512591389114544e-06,
+      "loss": 1.81,
+      "step": 1544
+    },
+    {
+      "epoch": 1.6931506849315068,
+      "grad_norm": 2.2232203483581543,
+      "learning_rate": 2.4492282696994317e-06,
+      "loss": 1.8095,
+      "step": 1545
+    },
+    {
+      "epoch": 1.6942465753424658,
+      "grad_norm": 2.1755056381225586,
+      "learning_rate": 2.447197400487409e-06,
+      "loss": 1.8149,
+      "step": 1546
+    },
+    {
+      "epoch": 1.6953424657534246,
+      "grad_norm": 2.325871467590332,
+      "learning_rate": 2.4451665312753862e-06,
+      "loss": 1.8154,
+      "step": 1547
+    },
+    {
+      "epoch": 1.6964383561643834,
+      "grad_norm": 2.286848545074463,
+      "learning_rate": 2.4431356620633635e-06,
+      "loss": 1.8409,
+      "step": 1548
+    },
+    {
+      "epoch": 1.6975342465753425,
+      "grad_norm": 2.359510898590088,
+      "learning_rate": 2.4411047928513408e-06,
+      "loss": 1.843,
+      "step": 1549
+    },
+    {
+      "epoch": 1.6986301369863015,
+      "grad_norm": 2.474428176879883,
+      "learning_rate": 2.439073923639318e-06,
+      "loss": 1.8089,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6986301369863015,
+      "eval_loss": 2.0077731609344482,
+      "eval_runtime": 37.1092,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6997260273972603,
+      "grad_norm": 2.3662912845611572,
+      "learning_rate": 2.4370430544272953e-06,
+      "loss": 1.8176,
+      "step": 1551
+    },
+    {
+      "epoch": 1.7008219178082191,
+      "grad_norm": 2.448948860168457,
+      "learning_rate": 2.435012185215272e-06,
+      "loss": 1.7776,
+      "step": 1552
+    },
+    {
+      "epoch": 1.7019178082191782,
+      "grad_norm": 2.333739757537842,
+      "learning_rate": 2.4329813160032495e-06,
+      "loss": 1.7552,
+      "step": 1553
+    },
+    {
+      "epoch": 1.703013698630137,
+      "grad_norm": 2.170365571975708,
+      "learning_rate": 2.4309504467912267e-06,
+      "loss": 1.7759,
+      "step": 1554
+    },
+    {
+      "epoch": 1.7041095890410958,
+      "grad_norm": 2.2343010902404785,
+      "learning_rate": 2.428919577579204e-06,
+      "loss": 1.8073,
+      "step": 1555
+    },
+    {
+      "epoch": 1.7052054794520548,
+      "grad_norm": 2.2245287895202637,
+      "learning_rate": 2.4268887083671813e-06,
+      "loss": 1.7826,
+      "step": 1556
+    },
+    {
+      "epoch": 1.7063013698630138,
+      "grad_norm": 2.339942216873169,
+      "learning_rate": 2.4248578391551586e-06,
+      "loss": 1.8766,
+      "step": 1557
+    },
+    {
+      "epoch": 1.7073972602739727,
+      "grad_norm": 2.1585116386413574,
+      "learning_rate": 2.422826969943136e-06,
+      "loss": 1.7605,
+      "step": 1558
+    },
+    {
+      "epoch": 1.7084931506849315,
+      "grad_norm": 2.2108142375946045,
+      "learning_rate": 2.420796100731113e-06,
+      "loss": 1.7855,
+      "step": 1559
+    },
+    {
+      "epoch": 1.7095890410958905,
+      "grad_norm": 2.192457675933838,
+      "learning_rate": 2.4187652315190904e-06,
+      "loss": 1.7489,
+      "step": 1560
+    },
+    {
+      "epoch": 1.7095890410958905,
+      "eval_loss": 2.0085604190826416,
+      "eval_runtime": 37.1105,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1560
+    },
+    {
+      "epoch": 1.7106849315068493,
+      "grad_norm": 2.291106939315796,
+      "learning_rate": 2.4167343623070677e-06,
+      "loss": 1.7981,
+      "step": 1561
+    },
+    {
+      "epoch": 1.7117808219178081,
+      "grad_norm": 2.509237766265869,
+      "learning_rate": 2.414703493095045e-06,
+      "loss": 1.8146,
+      "step": 1562
+    },
+    {
+      "epoch": 1.7128767123287671,
+      "grad_norm": 2.497162342071533,
+      "learning_rate": 2.4126726238830222e-06,
+      "loss": 1.807,
+      "step": 1563
+    },
+    {
+      "epoch": 1.7139726027397262,
+      "grad_norm": 2.256833076477051,
+      "learning_rate": 2.410641754670999e-06,
+      "loss": 1.8304,
+      "step": 1564
+    },
+    {
+      "epoch": 1.7150684931506848,
+      "grad_norm": 2.2523772716522217,
+      "learning_rate": 2.4086108854589763e-06,
+      "loss": 1.7794,
+      "step": 1565
+    },
+    {
+      "epoch": 1.7161643835616438,
+      "grad_norm": 2.289182424545288,
+      "learning_rate": 2.4065800162469536e-06,
+      "loss": 1.8425,
+      "step": 1566
+    },
+    {
+      "epoch": 1.7172602739726028,
+      "grad_norm": 2.734487771987915,
+      "learning_rate": 2.404549147034931e-06,
+      "loss": 1.7791,
+      "step": 1567
+    },
+    {
+      "epoch": 1.7183561643835616,
+      "grad_norm": 2.3307888507843018,
+      "learning_rate": 2.402518277822908e-06,
+      "loss": 1.8064,
+      "step": 1568
+    },
+    {
+      "epoch": 1.7194520547945205,
+      "grad_norm": 2.5016298294067383,
+      "learning_rate": 2.4004874086108854e-06,
+      "loss": 1.8544,
+      "step": 1569
+    },
+    {
+      "epoch": 1.7205479452054795,
+      "grad_norm": 2.142246961593628,
+      "learning_rate": 2.3984565393988627e-06,
+      "loss": 1.7653,
+      "step": 1570
+    },
+    {
+      "epoch": 1.7205479452054795,
+      "eval_loss": 2.0068864822387695,
+      "eval_runtime": 37.1078,
+      "eval_samples_per_second": 26.949,
+      "eval_steps_per_second": 1.132,
+      "step": 1570
+    },
+    {
+      "epoch": 1.7216438356164385,
+      "grad_norm": 2.5692505836486816,
+      "learning_rate": 2.39642567018684e-06,
+      "loss": 1.8444,
+      "step": 1571
+    },
+    {
+      "epoch": 1.7227397260273971,
+      "grad_norm": 2.2596213817596436,
+      "learning_rate": 2.3943948009748173e-06,
+      "loss": 1.7694,
+      "step": 1572
+    },
+    {
+      "epoch": 1.7238356164383561,
+      "grad_norm": 2.230984926223755,
+      "learning_rate": 2.3923639317627945e-06,
+      "loss": 1.786,
+      "step": 1573
+    },
+    {
+      "epoch": 1.7249315068493152,
+      "grad_norm": 2.366178274154663,
+      "learning_rate": 2.390333062550772e-06,
+      "loss": 1.8138,
+      "step": 1574
+    },
+    {
+      "epoch": 1.726027397260274,
+      "grad_norm": 2.23783016204834,
+      "learning_rate": 2.388302193338749e-06,
+      "loss": 1.7615,
+      "step": 1575
+    },
+    {
+      "epoch": 1.7271232876712328,
+      "grad_norm": 2.3766677379608154,
+      "learning_rate": 2.3862713241267264e-06,
+      "loss": 1.6811,
+      "step": 1576
+    },
+    {
+      "epoch": 1.7282191780821918,
+      "grad_norm": 2.3434040546417236,
+      "learning_rate": 2.3842404549147037e-06,
+      "loss": 1.8806,
+      "step": 1577
+    },
+    {
+      "epoch": 1.7293150684931506,
+      "grad_norm": 2.2444276809692383,
+      "learning_rate": 2.382209585702681e-06,
+      "loss": 1.7968,
+      "step": 1578
+    },
+    {
+      "epoch": 1.7304109589041095,
+      "grad_norm": 2.2524919509887695,
+      "learning_rate": 2.380178716490658e-06,
+      "loss": 1.7602,
+      "step": 1579
+    },
+    {
+      "epoch": 1.7315068493150685,
+      "grad_norm": 2.3310587406158447,
+      "learning_rate": 2.3781478472786355e-06,
+      "loss": 1.8414,
+      "step": 1580
+    },
+    {
+      "epoch": 1.7315068493150685,
+      "eval_loss": 1.9981482028961182,
+      "eval_runtime": 37.0827,
+      "eval_samples_per_second": 26.967,
+      "eval_steps_per_second": 1.133,
+      "step": 1580
+    },
+    {
+      "epoch": 1.7326027397260275,
+      "grad_norm": 2.4081802368164062,
+      "learning_rate": 2.3761169780666128e-06,
+      "loss": 1.8002,
+      "step": 1581
+    },
+    {
+      "epoch": 1.7336986301369863,
+      "grad_norm": 2.293506622314453,
+      "learning_rate": 2.37408610885459e-06,
+      "loss": 1.8365,
+      "step": 1582
+    },
+    {
+      "epoch": 1.7347945205479451,
+      "grad_norm": 2.2015790939331055,
+      "learning_rate": 2.3720552396425673e-06,
+      "loss": 1.8571,
+      "step": 1583
+    },
+    {
+      "epoch": 1.7358904109589042,
+      "grad_norm": 2.308809518814087,
+      "learning_rate": 2.3700243704305446e-06,
+      "loss": 1.7821,
+      "step": 1584
+    },
+    {
+      "epoch": 1.736986301369863,
+      "grad_norm": 2.376842975616455,
+      "learning_rate": 2.367993501218522e-06,
+      "loss": 1.8701,
+      "step": 1585
+    },
+    {
+      "epoch": 1.7380821917808218,
+      "grad_norm": 2.4102022647857666,
+      "learning_rate": 2.365962632006499e-06,
+      "loss": 1.8055,
+      "step": 1586
+    },
+    {
+      "epoch": 1.7391780821917808,
+      "grad_norm": 2.2734079360961914,
+      "learning_rate": 2.3639317627944764e-06,
+      "loss": 1.8117,
+      "step": 1587
+    },
+    {
+      "epoch": 1.7402739726027399,
+      "grad_norm": 2.3426599502563477,
+      "learning_rate": 2.3619008935824537e-06,
+      "loss": 1.8337,
+      "step": 1588
+    },
+    {
+      "epoch": 1.7413698630136987,
+      "grad_norm": 2.340383768081665,
+      "learning_rate": 2.3598700243704305e-06,
+      "loss": 1.8247,
+      "step": 1589
+    },
+    {
+      "epoch": 1.7424657534246575,
+      "grad_norm": 2.2561380863189697,
+      "learning_rate": 2.357839155158408e-06,
+      "loss": 1.8249,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7424657534246575,
+      "eval_loss": 1.9979513883590698,
+      "eval_runtime": 37.0948,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7435616438356165,
+      "grad_norm": 2.2459654808044434,
+      "learning_rate": 2.355808285946385e-06,
+      "loss": 1.8273,
+      "step": 1591
+    },
+    {
+      "epoch": 1.7446575342465753,
+      "grad_norm": 2.2017974853515625,
+      "learning_rate": 2.3537774167343624e-06,
+      "loss": 1.828,
+      "step": 1592
+    },
+    {
+      "epoch": 1.7457534246575341,
+      "grad_norm": 2.286304235458374,
+      "learning_rate": 2.3517465475223396e-06,
+      "loss": 1.8497,
+      "step": 1593
+    },
+    {
+      "epoch": 1.7468493150684932,
+      "grad_norm": 2.296494722366333,
+      "learning_rate": 2.349715678310317e-06,
+      "loss": 1.8155,
+      "step": 1594
+    },
+    {
+      "epoch": 1.7479452054794522,
+      "grad_norm": 2.2989203929901123,
+      "learning_rate": 2.347684809098294e-06,
+      "loss": 1.8214,
+      "step": 1595
+    },
+    {
+      "epoch": 1.749041095890411,
+      "grad_norm": 2.3034965991973877,
+      "learning_rate": 2.3456539398862715e-06,
+      "loss": 1.7792,
+      "step": 1596
+    },
+    {
+      "epoch": 1.7501369863013698,
+      "grad_norm": 2.3823719024658203,
+      "learning_rate": 2.3436230706742487e-06,
+      "loss": 1.8326,
+      "step": 1597
+    },
+    {
+      "epoch": 1.7512328767123289,
+      "grad_norm": 2.231095552444458,
+      "learning_rate": 2.341592201462226e-06,
+      "loss": 1.7713,
+      "step": 1598
+    },
+    {
+      "epoch": 1.7523287671232877,
+      "grad_norm": 2.287524938583374,
+      "learning_rate": 2.3395613322502033e-06,
+      "loss": 1.8143,
+      "step": 1599
+    },
+    {
+      "epoch": 1.7534246575342465,
+      "grad_norm": 2.267552614212036,
+      "learning_rate": 2.3375304630381806e-06,
+      "loss": 1.7883,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7534246575342465,
+      "eval_loss": 1.998821496963501,
+      "eval_runtime": 37.1222,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7545205479452055,
+      "grad_norm": 2.2986714839935303,
+      "learning_rate": 2.335499593826158e-06,
+      "loss": 1.7615,
+      "step": 1601
+    },
+    {
+      "epoch": 1.7556164383561645,
+      "grad_norm": 2.2913246154785156,
+      "learning_rate": 2.333468724614135e-06,
+      "loss": 1.7584,
+      "step": 1602
+    },
+    {
+      "epoch": 1.7567123287671231,
+      "grad_norm": 2.2486486434936523,
+      "learning_rate": 2.3314378554021124e-06,
+      "loss": 1.8089,
+      "step": 1603
+    },
+    {
+      "epoch": 1.7578082191780822,
+      "grad_norm": 2.319581985473633,
+      "learning_rate": 2.3294069861900897e-06,
+      "loss": 1.8364,
+      "step": 1604
+    },
+    {
+      "epoch": 1.7589041095890412,
+      "grad_norm": 2.2775609493255615,
+      "learning_rate": 2.327376116978067e-06,
+      "loss": 1.7661,
+      "step": 1605
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 2.3445472717285156,
+      "learning_rate": 2.3253452477660442e-06,
+      "loss": 1.7583,
+      "step": 1606
+    },
+    {
+      "epoch": 1.7610958904109588,
+      "grad_norm": 2.420865058898926,
+      "learning_rate": 2.3233143785540215e-06,
+      "loss": 1.8411,
+      "step": 1607
+    },
+    {
+      "epoch": 1.7621917808219179,
+      "grad_norm": 2.294821262359619,
+      "learning_rate": 2.3212835093419988e-06,
+      "loss": 1.7812,
+      "step": 1608
+    },
+    {
+      "epoch": 1.7632876712328767,
+      "grad_norm": 2.332784414291382,
+      "learning_rate": 2.319252640129976e-06,
+      "loss": 1.8252,
+      "step": 1609
+    },
+    {
+      "epoch": 1.7643835616438355,
+      "grad_norm": 2.433619499206543,
+      "learning_rate": 2.3172217709179533e-06,
+      "loss": 1.7709,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7643835616438355,
+      "eval_loss": 2.0010287761688232,
+      "eval_runtime": 36.7858,
+      "eval_samples_per_second": 27.184,
+      "eval_steps_per_second": 1.142,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7654794520547945,
+      "grad_norm": 2.532618522644043,
+      "learning_rate": 2.3151909017059306e-06,
+      "loss": 1.8677,
+      "step": 1611
+    },
+    {
+      "epoch": 1.7665753424657535,
+      "grad_norm": 2.354578733444214,
+      "learning_rate": 2.313160032493908e-06,
+      "loss": 1.8412,
+      "step": 1612
+    },
+    {
+      "epoch": 1.7676712328767124,
+      "grad_norm": 2.382774591445923,
+      "learning_rate": 2.311129163281885e-06,
+      "loss": 1.7543,
+      "step": 1613
+    },
+    {
+      "epoch": 1.7687671232876712,
+      "grad_norm": 2.3443398475646973,
+      "learning_rate": 2.309098294069862e-06,
+      "loss": 1.8267,
+      "step": 1614
+    },
+    {
+      "epoch": 1.7698630136986302,
+      "grad_norm": 2.3804502487182617,
+      "learning_rate": 2.3070674248578393e-06,
+      "loss": 1.7582,
+      "step": 1615
+    },
+    {
+      "epoch": 1.770958904109589,
+      "grad_norm": 2.3163580894470215,
+      "learning_rate": 2.3050365556458166e-06,
+      "loss": 1.7539,
+      "step": 1616
+    },
+    {
+      "epoch": 1.7720547945205478,
+      "grad_norm": 2.605738639831543,
+      "learning_rate": 2.303005686433794e-06,
+      "loss": 1.8124,
+      "step": 1617
+    },
+    {
+      "epoch": 1.7731506849315068,
+      "grad_norm": 2.3859915733337402,
+      "learning_rate": 2.300974817221771e-06,
+      "loss": 1.8271,
+      "step": 1618
+    },
+    {
+      "epoch": 1.7742465753424659,
+      "grad_norm": 2.3715479373931885,
+      "learning_rate": 2.2989439480097484e-06,
+      "loss": 1.7823,
+      "step": 1619
+    },
+    {
+      "epoch": 1.7753424657534247,
+      "grad_norm": 2.441100597381592,
+      "learning_rate": 2.2969130787977257e-06,
+      "loss": 1.7461,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7753424657534247,
+      "eval_loss": 2.002152442932129,
+      "eval_runtime": 37.0246,
+      "eval_samples_per_second": 27.009,
+      "eval_steps_per_second": 1.134,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7764383561643835,
+      "grad_norm": 2.4154021739959717,
+      "learning_rate": 2.294882209585703e-06,
+      "loss": 1.8012,
+      "step": 1621
+    },
+    {
+      "epoch": 1.7775342465753425,
+      "grad_norm": 2.333400249481201,
+      "learning_rate": 2.2928513403736802e-06,
+      "loss": 1.8222,
+      "step": 1622
+    },
+    {
+      "epoch": 1.7786301369863013,
+      "grad_norm": 2.3183937072753906,
+      "learning_rate": 2.2908204711616575e-06,
+      "loss": 1.7561,
+      "step": 1623
+    },
+    {
+      "epoch": 1.7797260273972602,
+      "grad_norm": 2.302823066711426,
+      "learning_rate": 2.2887896019496348e-06,
+      "loss": 1.7975,
+      "step": 1624
+    },
+    {
+      "epoch": 1.7808219178082192,
+      "grad_norm": 2.3702917098999023,
+      "learning_rate": 2.286758732737612e-06,
+      "loss": 1.8478,
+      "step": 1625
+    },
+    {
+      "epoch": 1.7819178082191782,
+      "grad_norm": 2.4997549057006836,
+      "learning_rate": 2.284727863525589e-06,
+      "loss": 1.7855,
+      "step": 1626
+    },
+    {
+      "epoch": 1.783013698630137,
+      "grad_norm": 2.296191930770874,
+      "learning_rate": 2.282696994313566e-06,
+      "loss": 1.7922,
+      "step": 1627
+    },
+    {
+      "epoch": 1.7841095890410958,
+      "grad_norm": 2.2567873001098633,
+      "learning_rate": 2.2806661251015434e-06,
+      "loss": 1.8455,
+      "step": 1628
+    },
+    {
+      "epoch": 1.7852054794520549,
+      "grad_norm": 2.393336534500122,
+      "learning_rate": 2.2786352558895207e-06,
+      "loss": 1.8141,
+      "step": 1629
+    },
+    {
+      "epoch": 1.7863013698630137,
+      "grad_norm": 2.610466480255127,
+      "learning_rate": 2.276604386677498e-06,
+      "loss": 1.7875,
+      "step": 1630
+    },
+    {
+      "epoch": 1.7863013698630137,
+      "eval_loss": 1.9985637664794922,
+      "eval_runtime": 37.1001,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 1630
+    },
+    {
+      "epoch": 1.7873972602739725,
+      "grad_norm": 2.2163283824920654,
+      "learning_rate": 2.2745735174654753e-06,
+      "loss": 1.7479,
+      "step": 1631
+    },
+    {
+      "epoch": 1.7884931506849315,
+      "grad_norm": 2.2641165256500244,
+      "learning_rate": 2.2725426482534526e-06,
+      "loss": 1.7749,
+      "step": 1632
+    },
+    {
+      "epoch": 1.7895890410958906,
+      "grad_norm": 2.2786946296691895,
+      "learning_rate": 2.27051177904143e-06,
+      "loss": 1.8403,
+      "step": 1633
+    },
+    {
+      "epoch": 1.7906849315068492,
+      "grad_norm": 2.660356044769287,
+      "learning_rate": 2.268480909829407e-06,
+      "loss": 1.7994,
+      "step": 1634
+    },
+    {
+      "epoch": 1.7917808219178082,
+      "grad_norm": 2.322181224822998,
+      "learning_rate": 2.2664500406173844e-06,
+      "loss": 1.7917,
+      "step": 1635
+    },
+    {
+      "epoch": 1.7928767123287672,
+      "grad_norm": 2.3020682334899902,
+      "learning_rate": 2.2644191714053617e-06,
+      "loss": 1.7983,
+      "step": 1636
+    },
+    {
+      "epoch": 1.793972602739726,
+      "grad_norm": 2.5280368328094482,
+      "learning_rate": 2.262388302193339e-06,
+      "loss": 1.8378,
+      "step": 1637
+    },
+    {
+      "epoch": 1.7950684931506848,
+      "grad_norm": 2.3975653648376465,
+      "learning_rate": 2.260357432981316e-06,
+      "loss": 1.8081,
+      "step": 1638
+    },
+    {
+      "epoch": 1.7961643835616439,
+      "grad_norm": 2.419806957244873,
+      "learning_rate": 2.2583265637692935e-06,
+      "loss": 1.793,
+      "step": 1639
+    },
+    {
+      "epoch": 1.7972602739726027,
+      "grad_norm": 2.4601471424102783,
+      "learning_rate": 2.2562956945572708e-06,
+      "loss": 1.8116,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7972602739726027,
+      "eval_loss": 1.9951634407043457,
+      "eval_runtime": 37.1095,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7983561643835615,
+      "grad_norm": 2.17830753326416,
+      "learning_rate": 2.254264825345248e-06,
+      "loss": 1.7845,
+      "step": 1641
+    },
+    {
+      "epoch": 1.7994520547945205,
+      "grad_norm": 2.3332247734069824,
+      "learning_rate": 2.2522339561332253e-06,
+      "loss": 1.7945,
+      "step": 1642
+    },
+    {
+      "epoch": 1.8005479452054796,
+      "grad_norm": 2.3841960430145264,
+      "learning_rate": 2.2502030869212026e-06,
+      "loss": 1.7994,
+      "step": 1643
+    },
+    {
+      "epoch": 1.8016438356164384,
+      "grad_norm": 2.3669073581695557,
+      "learning_rate": 2.24817221770918e-06,
+      "loss": 1.8053,
+      "step": 1644
+    },
+    {
+      "epoch": 1.8027397260273972,
+      "grad_norm": 2.406858205795288,
+      "learning_rate": 2.246141348497157e-06,
+      "loss": 1.8647,
+      "step": 1645
+    },
+    {
+      "epoch": 1.8038356164383562,
+      "grad_norm": 2.2854161262512207,
+      "learning_rate": 2.2441104792851344e-06,
+      "loss": 1.8312,
+      "step": 1646
+    },
+    {
+      "epoch": 1.804931506849315,
+      "grad_norm": 2.252976655960083,
+      "learning_rate": 2.2420796100731117e-06,
+      "loss": 1.7834,
+      "step": 1647
+    },
+    {
+      "epoch": 1.8060273972602738,
+      "grad_norm": 2.4131624698638916,
+      "learning_rate": 2.240048740861089e-06,
+      "loss": 1.8204,
+      "step": 1648
+    },
+    {
+      "epoch": 1.8071232876712329,
+      "grad_norm": 2.4727940559387207,
+      "learning_rate": 2.2380178716490662e-06,
+      "loss": 1.8326,
+      "step": 1649
+    },
+    {
+      "epoch": 1.808219178082192,
+      "grad_norm": 2.4432709217071533,
+      "learning_rate": 2.2359870024370435e-06,
+      "loss": 1.6911,
+      "step": 1650
+    },
+    {
+      "epoch": 1.808219178082192,
+      "eval_loss": 1.996634840965271,
+      "eval_runtime": 37.072,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 1650
+    },
+    {
+      "epoch": 1.8093150684931507,
+      "grad_norm": 2.3810362815856934,
+      "learning_rate": 2.2339561332250204e-06,
+      "loss": 1.8256,
+      "step": 1651
+    },
+    {
+      "epoch": 1.8104109589041095,
+      "grad_norm": 2.4489142894744873,
+      "learning_rate": 2.2319252640129976e-06,
+      "loss": 1.8368,
+      "step": 1652
+    },
+    {
+      "epoch": 1.8115068493150686,
+      "grad_norm": 2.458686113357544,
+      "learning_rate": 2.229894394800975e-06,
+      "loss": 1.8176,
+      "step": 1653
+    },
+    {
+      "epoch": 1.8126027397260274,
+      "grad_norm": 2.3977842330932617,
+      "learning_rate": 2.227863525588952e-06,
+      "loss": 1.7516,
+      "step": 1654
+    },
+    {
+      "epoch": 1.8136986301369862,
+      "grad_norm": 2.3777663707733154,
+      "learning_rate": 2.2258326563769295e-06,
+      "loss": 1.732,
+      "step": 1655
+    },
+    {
+      "epoch": 1.8147945205479452,
+      "grad_norm": 2.2874817848205566,
+      "learning_rate": 2.2238017871649067e-06,
+      "loss": 1.749,
+      "step": 1656
+    },
+    {
+      "epoch": 1.8158904109589042,
+      "grad_norm": 2.288114547729492,
+      "learning_rate": 2.221770917952884e-06,
+      "loss": 1.7357,
+      "step": 1657
+    },
+    {
+      "epoch": 1.816986301369863,
+      "grad_norm": 2.3532187938690186,
+      "learning_rate": 2.2197400487408613e-06,
+      "loss": 1.7953,
+      "step": 1658
+    },
+    {
+      "epoch": 1.8180821917808219,
+      "grad_norm": 2.2829396724700928,
+      "learning_rate": 2.2177091795288386e-06,
+      "loss": 1.8358,
+      "step": 1659
+    },
+    {
+      "epoch": 1.819178082191781,
+      "grad_norm": 2.3282570838928223,
+      "learning_rate": 2.215678310316816e-06,
+      "loss": 1.7448,
+      "step": 1660
+    },
+    {
+      "epoch": 1.819178082191781,
+      "eval_loss": 1.999374270439148,
+      "eval_runtime": 37.0739,
+      "eval_samples_per_second": 26.973,
+      "eval_steps_per_second": 1.133,
+      "step": 1660
+    },
+    {
+      "epoch": 1.8202739726027397,
+      "grad_norm": 2.474106788635254,
+      "learning_rate": 2.213647441104793e-06,
+      "loss": 1.7979,
+      "step": 1661
+    },
+    {
+      "epoch": 1.8213698630136985,
+      "grad_norm": 2.159498453140259,
+      "learning_rate": 2.2116165718927704e-06,
+      "loss": 1.8075,
+      "step": 1662
+    },
+    {
+      "epoch": 1.8224657534246576,
+      "grad_norm": 2.2503573894500732,
+      "learning_rate": 2.2095857026807473e-06,
+      "loss": 1.7742,
+      "step": 1663
+    },
+    {
+      "epoch": 1.8235616438356166,
+      "grad_norm": 2.4465060234069824,
+      "learning_rate": 2.2075548334687245e-06,
+      "loss": 1.8067,
+      "step": 1664
+    },
+    {
+      "epoch": 1.8246575342465754,
+      "grad_norm": 2.462158679962158,
+      "learning_rate": 2.205523964256702e-06,
+      "loss": 1.7456,
+      "step": 1665
+    },
+    {
+      "epoch": 1.8257534246575342,
+      "grad_norm": 2.354051113128662,
+      "learning_rate": 2.203493095044679e-06,
+      "loss": 1.7688,
+      "step": 1666
+    },
+    {
+      "epoch": 1.8268493150684932,
+      "grad_norm": 2.3165128231048584,
+      "learning_rate": 2.2014622258326564e-06,
+      "loss": 1.7986,
+      "step": 1667
+    },
+    {
+      "epoch": 1.827945205479452,
+      "grad_norm": 2.3131113052368164,
+      "learning_rate": 2.1994313566206336e-06,
+      "loss": 1.8348,
+      "step": 1668
+    },
+    {
+      "epoch": 1.8290410958904109,
+      "grad_norm": 2.303086996078491,
+      "learning_rate": 2.197400487408611e-06,
+      "loss": 1.8413,
+      "step": 1669
+    },
+    {
+      "epoch": 1.83013698630137,
+      "grad_norm": 2.392995834350586,
+      "learning_rate": 2.195369618196588e-06,
+      "loss": 1.7663,
+      "step": 1670
+    },
+    {
+      "epoch": 1.83013698630137,
+      "eval_loss": 1.9936721324920654,
+      "eval_runtime": 37.0717,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 1670
+    },
+    {
+      "epoch": 1.831232876712329,
+      "grad_norm": 2.323596477508545,
+      "learning_rate": 2.1933387489845655e-06,
+      "loss": 1.7463,
+      "step": 1671
+    },
+    {
+      "epoch": 1.8323287671232875,
+      "grad_norm": 2.3553171157836914,
+      "learning_rate": 2.1913078797725427e-06,
+      "loss": 1.7844,
+      "step": 1672
+    },
+    {
+      "epoch": 1.8334246575342465,
+      "grad_norm": 2.3724591732025146,
+      "learning_rate": 2.18927701056052e-06,
+      "loss": 1.7955,
+      "step": 1673
+    },
+    {
+      "epoch": 1.8345205479452056,
+      "grad_norm": 2.4300284385681152,
+      "learning_rate": 2.1872461413484973e-06,
+      "loss": 1.8697,
+      "step": 1674
+    },
+    {
+      "epoch": 1.8356164383561644,
+      "grad_norm": 2.4300284385681152,
+      "learning_rate": 2.1872461413484973e-06,
+      "loss": 1.8065,
+      "step": 1675
+    },
+    {
+      "epoch": 1.8367123287671232,
+      "grad_norm": 2.523627519607544,
+      "learning_rate": 2.1852152721364746e-06,
+      "loss": 1.801,
+      "step": 1676
+    },
+    {
+      "epoch": 1.8378082191780822,
+      "grad_norm": 2.2884891033172607,
+      "learning_rate": 2.183184402924452e-06,
+      "loss": 1.7512,
+      "step": 1677
+    },
+    {
+      "epoch": 1.838904109589041,
+      "grad_norm": 2.3114986419677734,
+      "learning_rate": 2.181153533712429e-06,
+      "loss": 1.8245,
+      "step": 1678
+    },
+    {
+      "epoch": 1.8399999999999999,
+      "grad_norm": 2.326855421066284,
+      "learning_rate": 2.1791226645004064e-06,
+      "loss": 1.7921,
+      "step": 1679
+    },
+    {
+      "epoch": 1.841095890410959,
+      "grad_norm": 2.3448753356933594,
+      "learning_rate": 2.1770917952883837e-06,
+      "loss": 1.7709,
+      "step": 1680
+    },
+    {
+      "epoch": 1.841095890410959,
+      "eval_loss": 1.996195673942566,
+      "eval_runtime": 37.0481,
+      "eval_samples_per_second": 26.992,
+      "eval_steps_per_second": 1.134,
+      "step": 1680
+    },
+    {
+      "epoch": 1.842191780821918,
+      "grad_norm": 2.223470687866211,
+      "learning_rate": 2.175060926076361e-06,
+      "loss": 1.8468,
+      "step": 1681
+    },
+    {
+      "epoch": 1.8432876712328767,
+      "grad_norm": 2.29726505279541,
+      "learning_rate": 2.1730300568643382e-06,
+      "loss": 1.8753,
+      "step": 1682
+    },
+    {
+      "epoch": 1.8443835616438355,
+      "grad_norm": 2.3869571685791016,
+      "learning_rate": 2.1709991876523155e-06,
+      "loss": 1.7835,
+      "step": 1683
+    },
+    {
+      "epoch": 1.8454794520547946,
+      "grad_norm": 2.192608594894409,
+      "learning_rate": 2.1689683184402928e-06,
+      "loss": 1.7556,
+      "step": 1684
+    },
+    {
+      "epoch": 1.8465753424657534,
+      "grad_norm": 2.355975866317749,
+      "learning_rate": 2.16693744922827e-06,
+      "loss": 1.8323,
+      "step": 1685
+    },
+    {
+      "epoch": 1.8476712328767122,
+      "grad_norm": 2.243938446044922,
+      "learning_rate": 2.1649065800162473e-06,
+      "loss": 1.7942,
+      "step": 1686
+    },
+    {
+      "epoch": 1.8487671232876712,
+      "grad_norm": 2.337069511413574,
+      "learning_rate": 2.1628757108042246e-06,
+      "loss": 1.7709,
+      "step": 1687
+    },
+    {
+      "epoch": 1.8498630136986303,
+      "grad_norm": 2.257504463195801,
+      "learning_rate": 2.160844841592202e-06,
+      "loss": 1.7496,
+      "step": 1688
+    },
+    {
+      "epoch": 1.850958904109589,
+      "grad_norm": 2.287336826324463,
+      "learning_rate": 2.1588139723801787e-06,
+      "loss": 1.7454,
+      "step": 1689
+    },
+    {
+      "epoch": 1.8520547945205479,
+      "grad_norm": 2.2864668369293213,
+      "learning_rate": 2.156783103168156e-06,
+      "loss": 1.7807,
+      "step": 1690
+    },
+    {
+      "epoch": 1.8520547945205479,
+      "eval_loss": 1.9940879344940186,
+      "eval_runtime": 37.0261,
+      "eval_samples_per_second": 27.008,
+      "eval_steps_per_second": 1.134,
+      "step": 1690
+    },
+    {
+      "epoch": 1.853150684931507,
+      "grad_norm": 2.1782145500183105,
+      "learning_rate": 2.1547522339561333e-06,
+      "loss": 1.7567,
+      "step": 1691
+    },
+    {
+      "epoch": 1.8542465753424657,
+      "grad_norm": 2.252023458480835,
+      "learning_rate": 2.1527213647441106e-06,
+      "loss": 1.811,
+      "step": 1692
+    },
+    {
+      "epoch": 1.8553424657534245,
+      "grad_norm": 2.2532849311828613,
+      "learning_rate": 2.150690495532088e-06,
+      "loss": 1.8491,
+      "step": 1693
+    },
+    {
+      "epoch": 1.8564383561643836,
+      "grad_norm": 2.35648250579834,
+      "learning_rate": 2.148659626320065e-06,
+      "loss": 1.8125,
+      "step": 1694
+    },
+    {
+      "epoch": 1.8575342465753426,
+      "grad_norm": 2.3490521907806396,
+      "learning_rate": 2.1466287571080424e-06,
+      "loss": 1.8035,
+      "step": 1695
+    },
+    {
+      "epoch": 1.8586301369863014,
+      "grad_norm": 2.2851879596710205,
+      "learning_rate": 2.1445978878960197e-06,
+      "loss": 1.7755,
+      "step": 1696
+    },
+    {
+      "epoch": 1.8597260273972602,
+      "grad_norm": 2.166754961013794,
+      "learning_rate": 2.142567018683997e-06,
+      "loss": 1.7861,
+      "step": 1697
+    },
+    {
+      "epoch": 1.8608219178082193,
+      "grad_norm": 2.2356410026550293,
+      "learning_rate": 2.140536149471974e-06,
+      "loss": 1.8412,
+      "step": 1698
+    },
+    {
+      "epoch": 1.861917808219178,
+      "grad_norm": 2.215618133544922,
+      "learning_rate": 2.1385052802599515e-06,
+      "loss": 1.7623,
+      "step": 1699
+    },
+    {
+      "epoch": 1.8630136986301369,
+      "grad_norm": 2.283298969268799,
+      "learning_rate": 2.1364744110479288e-06,
+      "loss": 1.7692,
+      "step": 1700
+    },
+    {
+      "epoch": 1.8630136986301369,
+      "eval_loss": 1.9961501359939575,
+      "eval_runtime": 37.0338,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1700
+    },
+    {
+      "epoch": 1.864109589041096,
+      "grad_norm": 2.292818546295166,
+      "learning_rate": 2.134443541835906e-06,
+      "loss": 1.8251,
+      "step": 1701
+    },
+    {
+      "epoch": 1.865205479452055,
+      "grad_norm": 2.235895872116089,
+      "learning_rate": 2.132412672623883e-06,
+      "loss": 1.7726,
+      "step": 1702
+    },
+    {
+      "epoch": 1.8663013698630135,
+      "grad_norm": 2.239413022994995,
+      "learning_rate": 2.13038180341186e-06,
+      "loss": 1.7509,
+      "step": 1703
+    },
+    {
+      "epoch": 1.8673972602739726,
+      "grad_norm": 2.4125585556030273,
+      "learning_rate": 2.1283509341998374e-06,
+      "loss": 1.7973,
+      "step": 1704
+    },
+    {
+      "epoch": 1.8684931506849316,
+      "grad_norm": 2.3774733543395996,
+      "learning_rate": 2.1263200649878147e-06,
+      "loss": 1.8526,
+      "step": 1705
+    },
+    {
+      "epoch": 1.8695890410958904,
+      "grad_norm": 2.344918966293335,
+      "learning_rate": 2.124289195775792e-06,
+      "loss": 1.7708,
+      "step": 1706
+    },
+    {
+      "epoch": 1.8706849315068492,
+      "grad_norm": 2.1933865547180176,
+      "learning_rate": 2.1222583265637693e-06,
+      "loss": 1.7935,
+      "step": 1707
+    },
+    {
+      "epoch": 1.8717808219178083,
+      "grad_norm": 2.4939582347869873,
+      "learning_rate": 2.1202274573517465e-06,
+      "loss": 1.706,
+      "step": 1708
+    },
+    {
+      "epoch": 1.872876712328767,
+      "grad_norm": 2.407532215118408,
+      "learning_rate": 2.118196588139724e-06,
+      "loss": 1.7522,
+      "step": 1709
+    },
+    {
+      "epoch": 1.8739726027397259,
+      "grad_norm": 2.297900915145874,
+      "learning_rate": 2.116165718927701e-06,
+      "loss": 1.7558,
+      "step": 1710
+    },
+    {
+      "epoch": 1.8739726027397259,
+      "eval_loss": 1.9946404695510864,
+      "eval_runtime": 36.7569,
+      "eval_samples_per_second": 27.206,
+      "eval_steps_per_second": 1.143,
+      "step": 1710
+    },
+    {
+      "epoch": 1.875068493150685,
+      "grad_norm": 2.2298660278320312,
+      "learning_rate": 2.1141348497156784e-06,
+      "loss": 1.7836,
+      "step": 1711
+    },
+    {
+      "epoch": 1.876164383561644,
+      "grad_norm": 2.308060884475708,
+      "learning_rate": 2.1121039805036556e-06,
+      "loss": 1.7616,
+      "step": 1712
+    },
+    {
+      "epoch": 1.8772602739726028,
+      "grad_norm": 2.288381338119507,
+      "learning_rate": 2.110073111291633e-06,
+      "loss": 1.7502,
+      "step": 1713
+    },
+    {
+      "epoch": 1.8783561643835616,
+      "grad_norm": 2.2191097736358643,
+      "learning_rate": 2.10804224207961e-06,
+      "loss": 1.7946,
+      "step": 1714
+    },
+    {
+      "epoch": 1.8794520547945206,
+      "grad_norm": 2.211777925491333,
+      "learning_rate": 2.1060113728675875e-06,
+      "loss": 1.8238,
+      "step": 1715
+    },
+    {
+      "epoch": 1.8805479452054794,
+      "grad_norm": 2.3946406841278076,
+      "learning_rate": 2.1039805036555647e-06,
+      "loss": 1.8146,
+      "step": 1716
+    },
+    {
+      "epoch": 1.8816438356164382,
+      "grad_norm": 2.2949914932250977,
+      "learning_rate": 2.101949634443542e-06,
+      "loss": 1.6877,
+      "step": 1717
+    },
+    {
+      "epoch": 1.8827397260273973,
+      "grad_norm": 2.2905256748199463,
+      "learning_rate": 2.0999187652315193e-06,
+      "loss": 1.7868,
+      "step": 1718
+    },
+    {
+      "epoch": 1.8838356164383563,
+      "grad_norm": 2.289280891418457,
+      "learning_rate": 2.0978878960194966e-06,
+      "loss": 1.7387,
+      "step": 1719
+    },
+    {
+      "epoch": 1.884931506849315,
+      "grad_norm": 2.35497784614563,
+      "learning_rate": 2.095857026807474e-06,
+      "loss": 1.8002,
+      "step": 1720
+    },
+    {
+      "epoch": 1.884931506849315,
+      "eval_loss": 1.9928913116455078,
+      "eval_runtime": 36.9289,
+      "eval_samples_per_second": 27.079,
+      "eval_steps_per_second": 1.137,
+      "step": 1720
+    },
+    {
+      "epoch": 1.886027397260274,
+      "grad_norm": 2.4404468536376953,
+      "learning_rate": 2.093826157595451e-06,
+      "loss": 1.7981,
+      "step": 1721
+    },
+    {
+      "epoch": 1.887123287671233,
+      "grad_norm": 2.315718173980713,
+      "learning_rate": 2.0917952883834284e-06,
+      "loss": 1.7581,
+      "step": 1722
+    },
+    {
+      "epoch": 1.8882191780821918,
+      "grad_norm": 2.2984843254089355,
+      "learning_rate": 2.0897644191714057e-06,
+      "loss": 1.8153,
+      "step": 1723
+    },
+    {
+      "epoch": 1.8893150684931506,
+      "grad_norm": 2.3332557678222656,
+      "learning_rate": 2.087733549959383e-06,
+      "loss": 1.831,
+      "step": 1724
+    },
+    {
+      "epoch": 1.8904109589041096,
+      "grad_norm": 2.3394649028778076,
+      "learning_rate": 2.0857026807473602e-06,
+      "loss": 1.7486,
+      "step": 1725
+    },
+    {
+      "epoch": 1.8915068493150686,
+      "grad_norm": 2.2985246181488037,
+      "learning_rate": 2.0836718115353375e-06,
+      "loss": 1.721,
+      "step": 1726
+    },
+    {
+      "epoch": 1.8926027397260274,
+      "grad_norm": 2.318504810333252,
+      "learning_rate": 2.0816409423233144e-06,
+      "loss": 1.8469,
+      "step": 1727
+    },
+    {
+      "epoch": 1.8936986301369862,
+      "grad_norm": 2.4938831329345703,
+      "learning_rate": 2.0796100731112916e-06,
+      "loss": 1.7525,
+      "step": 1728
+    },
+    {
+      "epoch": 1.8947945205479453,
+      "grad_norm": 2.3391780853271484,
+      "learning_rate": 2.077579203899269e-06,
+      "loss": 1.7733,
+      "step": 1729
+    },
+    {
+      "epoch": 1.895890410958904,
+      "grad_norm": 2.437540054321289,
+      "learning_rate": 2.075548334687246e-06,
+      "loss": 1.7344,
+      "step": 1730
+    },
+    {
+      "epoch": 1.895890410958904,
+      "eval_loss": 1.990493655204773,
+      "eval_runtime": 37.0347,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1730
+    },
+    {
+      "epoch": 1.896986301369863,
+      "grad_norm": 2.413628578186035,
+      "learning_rate": 2.0735174654752235e-06,
+      "loss": 1.7978,
+      "step": 1731
+    },
+    {
+      "epoch": 1.898082191780822,
+      "grad_norm": 2.3673973083496094,
+      "learning_rate": 2.0714865962632007e-06,
+      "loss": 1.7414,
+      "step": 1732
+    },
+    {
+      "epoch": 1.899178082191781,
+      "grad_norm": 2.3961968421936035,
+      "learning_rate": 2.069455727051178e-06,
+      "loss": 1.7854,
+      "step": 1733
+    },
+    {
+      "epoch": 1.9002739726027398,
+      "grad_norm": 2.5583736896514893,
+      "learning_rate": 2.0674248578391553e-06,
+      "loss": 1.7629,
+      "step": 1734
+    },
+    {
+      "epoch": 1.9013698630136986,
+      "grad_norm": 2.256418228149414,
+      "learning_rate": 2.0653939886271326e-06,
+      "loss": 1.7845,
+      "step": 1735
+    },
+    {
+      "epoch": 1.9024657534246576,
+      "grad_norm": 2.4235012531280518,
+      "learning_rate": 2.06336311941511e-06,
+      "loss": 1.8151,
+      "step": 1736
+    },
+    {
+      "epoch": 1.9035616438356164,
+      "grad_norm": 2.2945358753204346,
+      "learning_rate": 2.061332250203087e-06,
+      "loss": 1.7665,
+      "step": 1737
+    },
+    {
+      "epoch": 1.9046575342465752,
+      "grad_norm": 2.371809959411621,
+      "learning_rate": 2.0593013809910644e-06,
+      "loss": 1.7355,
+      "step": 1738
+    },
+    {
+      "epoch": 1.9057534246575343,
+      "grad_norm": 2.26947021484375,
+      "learning_rate": 2.0572705117790417e-06,
+      "loss": 1.7702,
+      "step": 1739
+    },
+    {
+      "epoch": 1.9068493150684933,
+      "grad_norm": 2.314237356185913,
+      "learning_rate": 2.055239642567019e-06,
+      "loss": 1.8084,
+      "step": 1740
+    },
+    {
+      "epoch": 1.9068493150684933,
+      "eval_loss": 1.9961000680923462,
+      "eval_runtime": 37.0775,
+      "eval_samples_per_second": 26.971,
+      "eval_steps_per_second": 1.133,
+      "step": 1740
+    },
+    {
+      "epoch": 1.907945205479452,
+      "grad_norm": 2.248680591583252,
+      "learning_rate": 2.0532087733549962e-06,
+      "loss": 1.7898,
+      "step": 1741
+    },
+    {
+      "epoch": 1.909041095890411,
+      "grad_norm": 2.336378812789917,
+      "learning_rate": 2.0511779041429735e-06,
+      "loss": 1.795,
+      "step": 1742
+    },
+    {
+      "epoch": 1.91013698630137,
+      "grad_norm": 2.2420554161071777,
+      "learning_rate": 2.0491470349309508e-06,
+      "loss": 1.76,
+      "step": 1743
+    },
+    {
+      "epoch": 1.9112328767123288,
+      "grad_norm": 2.3305511474609375,
+      "learning_rate": 2.047116165718928e-06,
+      "loss": 1.7959,
+      "step": 1744
+    },
+    {
+      "epoch": 1.9123287671232876,
+      "grad_norm": 2.2677786350250244,
+      "learning_rate": 2.0450852965069053e-06,
+      "loss": 1.7922,
+      "step": 1745
+    },
+    {
+      "epoch": 1.9134246575342466,
+      "grad_norm": 2.224148988723755,
+      "learning_rate": 2.0430544272948826e-06,
+      "loss": 1.841,
+      "step": 1746
+    },
+    {
+      "epoch": 1.9145205479452054,
+      "grad_norm": 2.172091245651245,
+      "learning_rate": 2.04102355808286e-06,
+      "loss": 1.784,
+      "step": 1747
+    },
+    {
+      "epoch": 1.9156164383561642,
+      "grad_norm": 2.319962501525879,
+      "learning_rate": 2.038992688870837e-06,
+      "loss": 1.7238,
+      "step": 1748
+    },
+    {
+      "epoch": 1.9167123287671233,
+      "grad_norm": 2.299760103225708,
+      "learning_rate": 2.0369618196588144e-06,
+      "loss": 1.7501,
+      "step": 1749
+    },
+    {
+      "epoch": 1.9178082191780823,
+      "grad_norm": 2.218407154083252,
+      "learning_rate": 2.0349309504467917e-06,
+      "loss": 1.7628,
+      "step": 1750
+    },
+    {
+      "epoch": 1.9178082191780823,
+      "eval_loss": 1.9934331178665161,
+      "eval_runtime": 37.0986,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 1750
+    },
+    {
+      "epoch": 1.9189041095890411,
+      "grad_norm": 2.4121744632720947,
+      "learning_rate": 2.032900081234769e-06,
+      "loss": 1.8128,
+      "step": 1751
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 2.253225326538086,
+      "learning_rate": 2.030869212022746e-06,
+      "loss": 1.8161,
+      "step": 1752
+    },
+    {
+      "epoch": 1.921095890410959,
+      "grad_norm": 2.28702449798584,
+      "learning_rate": 2.028838342810723e-06,
+      "loss": 1.7769,
+      "step": 1753
+    },
+    {
+      "epoch": 1.9221917808219178,
+      "grad_norm": 2.3004465103149414,
+      "learning_rate": 2.0268074735987004e-06,
+      "loss": 1.7631,
+      "step": 1754
+    },
+    {
+      "epoch": 1.9232876712328766,
+      "grad_norm": 2.2733960151672363,
+      "learning_rate": 2.0247766043866777e-06,
+      "loss": 1.75,
+      "step": 1755
+    },
+    {
+      "epoch": 1.9243835616438356,
+      "grad_norm": 2.274019241333008,
+      "learning_rate": 2.022745735174655e-06,
+      "loss": 1.8317,
+      "step": 1756
+    },
+    {
+      "epoch": 1.9254794520547946,
+      "grad_norm": 2.3094100952148438,
+      "learning_rate": 2.020714865962632e-06,
+      "loss": 1.7841,
+      "step": 1757
+    },
+    {
+      "epoch": 1.9265753424657535,
+      "grad_norm": 2.269321918487549,
+      "learning_rate": 2.0186839967506095e-06,
+      "loss": 1.8052,
+      "step": 1758
+    },
+    {
+      "epoch": 1.9276712328767123,
+      "grad_norm": 2.5324459075927734,
+      "learning_rate": 2.0166531275385868e-06,
+      "loss": 1.7916,
+      "step": 1759
+    },
+    {
+      "epoch": 1.9287671232876713,
+      "grad_norm": 2.302569627761841,
+      "learning_rate": 2.014622258326564e-06,
+      "loss": 1.7499,
+      "step": 1760
+    },
+    {
+      "epoch": 1.9287671232876713,
+      "eval_loss": 1.9879337549209595,
+      "eval_runtime": 37.1186,
+      "eval_samples_per_second": 26.941,
+      "eval_steps_per_second": 1.132,
+      "step": 1760
+    },
+    {
+      "epoch": 1.9298630136986301,
+      "grad_norm": 2.29470157623291,
+      "learning_rate": 2.0125913891145413e-06,
+      "loss": 1.7922,
+      "step": 1761
+    },
+    {
+      "epoch": 1.930958904109589,
+      "grad_norm": 2.4307754039764404,
+      "learning_rate": 2.0105605199025186e-06,
+      "loss": 1.7802,
+      "step": 1762
+    },
+    {
+      "epoch": 1.932054794520548,
+      "grad_norm": 2.179588794708252,
+      "learning_rate": 2.008529650690496e-06,
+      "loss": 1.7391,
+      "step": 1763
+    },
+    {
+      "epoch": 1.933150684931507,
+      "grad_norm": 2.5824451446533203,
+      "learning_rate": 2.0064987814784727e-06,
+      "loss": 1.8106,
+      "step": 1764
+    },
+    {
+      "epoch": 1.9342465753424658,
+      "grad_norm": 2.32308292388916,
+      "learning_rate": 2.00446791226645e-06,
+      "loss": 1.7689,
+      "step": 1765
+    },
+    {
+      "epoch": 1.9353424657534246,
+      "grad_norm": 2.262533664703369,
+      "learning_rate": 2.0024370430544273e-06,
+      "loss": 1.7604,
+      "step": 1766
+    },
+    {
+      "epoch": 1.9364383561643836,
+      "grad_norm": 2.325253963470459,
+      "learning_rate": 2.0004061738424045e-06,
+      "loss": 1.7695,
+      "step": 1767
+    },
+    {
+      "epoch": 1.9375342465753425,
+      "grad_norm": 2.2855470180511475,
+      "learning_rate": 1.998375304630382e-06,
+      "loss": 1.8233,
+      "step": 1768
+    },
+    {
+      "epoch": 1.9386301369863013,
+      "grad_norm": 2.3688673973083496,
+      "learning_rate": 1.996344435418359e-06,
+      "loss": 1.7799,
+      "step": 1769
+    },
+    {
+      "epoch": 1.9397260273972603,
+      "grad_norm": 2.4682323932647705,
+      "learning_rate": 1.9943135662063364e-06,
+      "loss": 1.8036,
+      "step": 1770
+    },
+    {
+      "epoch": 1.9397260273972603,
+      "eval_loss": 1.9906389713287354,
+      "eval_runtime": 37.0959,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 1770
+    },
+    {
+      "epoch": 1.9408219178082193,
+      "grad_norm": 2.369166374206543,
+      "learning_rate": 1.9922826969943136e-06,
+      "loss": 1.7878,
+      "step": 1771
+    },
+    {
+      "epoch": 1.941917808219178,
+      "grad_norm": 2.4669666290283203,
+      "learning_rate": 1.990251827782291e-06,
+      "loss": 1.8228,
+      "step": 1772
+    },
+    {
+      "epoch": 1.943013698630137,
+      "grad_norm": 2.3083250522613525,
+      "learning_rate": 1.988220958570268e-06,
+      "loss": 1.7997,
+      "step": 1773
+    },
+    {
+      "epoch": 1.944109589041096,
+      "grad_norm": 2.4884281158447266,
+      "learning_rate": 1.9861900893582455e-06,
+      "loss": 1.8105,
+      "step": 1774
+    },
+    {
+      "epoch": 1.9452054794520548,
+      "grad_norm": 2.3504858016967773,
+      "learning_rate": 1.9841592201462227e-06,
+      "loss": 1.8159,
+      "step": 1775
+    },
+    {
+      "epoch": 1.9463013698630136,
+      "grad_norm": 2.618431329727173,
+      "learning_rate": 1.9821283509342e-06,
+      "loss": 1.7352,
+      "step": 1776
+    },
+    {
+      "epoch": 1.9473972602739726,
+      "grad_norm": 2.3452653884887695,
+      "learning_rate": 1.9800974817221773e-06,
+      "loss": 1.7451,
+      "step": 1777
+    },
+    {
+      "epoch": 1.9484931506849315,
+      "grad_norm": 2.5966849327087402,
+      "learning_rate": 1.9780666125101546e-06,
+      "loss": 1.7965,
+      "step": 1778
+    },
+    {
+      "epoch": 1.9495890410958903,
+      "grad_norm": 2.3017661571502686,
+      "learning_rate": 1.976035743298132e-06,
+      "loss": 1.7629,
+      "step": 1779
+    },
+    {
+      "epoch": 1.9506849315068493,
+      "grad_norm": 2.3766098022460938,
+      "learning_rate": 1.974004874086109e-06,
+      "loss": 1.8403,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9506849315068493,
+      "eval_loss": 1.9936374425888062,
+      "eval_runtime": 37.1158,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9517808219178083,
+      "grad_norm": 2.2327704429626465,
+      "learning_rate": 1.9719740048740864e-06,
+      "loss": 1.7096,
+      "step": 1781
+    },
+    {
+      "epoch": 1.9528767123287671,
+      "grad_norm": 2.458707809448242,
+      "learning_rate": 1.9699431356620637e-06,
+      "loss": 1.7636,
+      "step": 1782
+    },
+    {
+      "epoch": 1.953972602739726,
+      "grad_norm": 2.4143803119659424,
+      "learning_rate": 1.967912266450041e-06,
+      "loss": 1.757,
+      "step": 1783
+    },
+    {
+      "epoch": 1.955068493150685,
+      "grad_norm": 2.465148448944092,
+      "learning_rate": 1.9658813972380182e-06,
+      "loss": 1.7526,
+      "step": 1784
+    },
+    {
+      "epoch": 1.9561643835616438,
+      "grad_norm": 2.356210231781006,
+      "learning_rate": 1.9638505280259955e-06,
+      "loss": 1.8705,
+      "step": 1785
+    },
+    {
+      "epoch": 1.9572602739726026,
+      "grad_norm": 2.7456002235412598,
+      "learning_rate": 1.9618196588139728e-06,
+      "loss": 1.7574,
+      "step": 1786
+    },
+    {
+      "epoch": 1.9583561643835616,
+      "grad_norm": 2.3724262714385986,
+      "learning_rate": 1.95978878960195e-06,
+      "loss": 1.7854,
+      "step": 1787
+    },
+    {
+      "epoch": 1.9594520547945207,
+      "grad_norm": 2.275827646255493,
+      "learning_rate": 1.9577579203899273e-06,
+      "loss": 1.8447,
+      "step": 1788
+    },
+    {
+      "epoch": 1.9605479452054795,
+      "grad_norm": 2.369992256164551,
+      "learning_rate": 1.955727051177904e-06,
+      "loss": 1.8299,
+      "step": 1789
+    },
+    {
+      "epoch": 1.9616438356164383,
+      "grad_norm": 2.462372064590454,
+      "learning_rate": 1.9536961819658815e-06,
+      "loss": 1.7596,
+      "step": 1790
+    },
+    {
+      "epoch": 1.9616438356164383,
+      "eval_loss": 1.9947243928909302,
+      "eval_runtime": 37.111,
+      "eval_samples_per_second": 26.946,
+      "eval_steps_per_second": 1.132,
+      "step": 1790
+    },
+    {
+      "epoch": 1.9627397260273973,
+      "grad_norm": 2.3040034770965576,
+      "learning_rate": 1.9516653127538587e-06,
+      "loss": 1.7391,
+      "step": 1791
+    },
+    {
+      "epoch": 1.9638356164383561,
+      "grad_norm": 2.449328899383545,
+      "learning_rate": 1.949634443541836e-06,
+      "loss": 1.7428,
+      "step": 1792
+    },
+    {
+      "epoch": 1.964931506849315,
+      "grad_norm": 2.2679781913757324,
+      "learning_rate": 1.9476035743298133e-06,
+      "loss": 1.738,
+      "step": 1793
+    },
+    {
+      "epoch": 1.966027397260274,
+      "grad_norm": 2.3444032669067383,
+      "learning_rate": 1.9455727051177906e-06,
+      "loss": 1.8089,
+      "step": 1794
+    },
+    {
+      "epoch": 1.967123287671233,
+      "grad_norm": 2.3005306720733643,
+      "learning_rate": 1.943541835905768e-06,
+      "loss": 1.7718,
+      "step": 1795
+    },
+    {
+      "epoch": 1.9682191780821918,
+      "grad_norm": 2.6022629737854004,
+      "learning_rate": 1.941510966693745e-06,
+      "loss": 1.8325,
+      "step": 1796
+    },
+    {
+      "epoch": 1.9693150684931506,
+      "grad_norm": 2.407585382461548,
+      "learning_rate": 1.9394800974817224e-06,
+      "loss": 1.737,
+      "step": 1797
+    },
+    {
+      "epoch": 1.9704109589041097,
+      "grad_norm": 2.349339723587036,
+      "learning_rate": 1.9374492282696997e-06,
+      "loss": 1.81,
+      "step": 1798
+    },
+    {
+      "epoch": 1.9715068493150685,
+      "grad_norm": 2.296192169189453,
+      "learning_rate": 1.935418359057677e-06,
+      "loss": 1.8424,
+      "step": 1799
+    },
+    {
+      "epoch": 1.9726027397260273,
+      "grad_norm": 2.175875663757324,
+      "learning_rate": 1.9333874898456542e-06,
+      "loss": 1.7977,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9726027397260273,
+      "eval_loss": 1.9929841756820679,
+      "eval_runtime": 37.1159,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9736986301369863,
+      "grad_norm": 2.37428617477417,
+      "learning_rate": 1.931356620633631e-06,
+      "loss": 1.8248,
+      "step": 1801
+    },
+    {
+      "epoch": 1.9747945205479454,
+      "grad_norm": 2.297621965408325,
+      "learning_rate": 1.9293257514216083e-06,
+      "loss": 1.7215,
+      "step": 1802
+    },
+    {
+      "epoch": 1.9758904109589042,
+      "grad_norm": 2.217275857925415,
+      "learning_rate": 1.9272948822095856e-06,
+      "loss": 1.7828,
+      "step": 1803
+    },
+    {
+      "epoch": 1.976986301369863,
+      "grad_norm": 2.250854015350342,
+      "learning_rate": 1.925264012997563e-06,
+      "loss": 1.7245,
+      "step": 1804
+    },
+    {
+      "epoch": 1.978082191780822,
+      "grad_norm": 2.2556822299957275,
+      "learning_rate": 1.92323314378554e-06,
+      "loss": 1.7715,
+      "step": 1805
+    },
+    {
+      "epoch": 1.9791780821917808,
+      "grad_norm": 2.316389322280884,
+      "learning_rate": 1.9212022745735174e-06,
+      "loss": 1.8063,
+      "step": 1806
+    },
+    {
+      "epoch": 1.9802739726027396,
+      "grad_norm": 2.4358153343200684,
+      "learning_rate": 1.9191714053614947e-06,
+      "loss": 1.747,
+      "step": 1807
+    },
+    {
+      "epoch": 1.9813698630136987,
+      "grad_norm": 2.437631368637085,
+      "learning_rate": 1.917140536149472e-06,
+      "loss": 1.7485,
+      "step": 1808
+    },
+    {
+      "epoch": 1.9824657534246577,
+      "grad_norm": 2.3554601669311523,
+      "learning_rate": 1.9151096669374493e-06,
+      "loss": 1.7195,
+      "step": 1809
+    },
+    {
+      "epoch": 1.9835616438356163,
+      "grad_norm": 2.575559616088867,
+      "learning_rate": 1.9130787977254266e-06,
+      "loss": 1.8368,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9835616438356163,
+      "eval_loss": 1.9903514385223389,
+      "eval_runtime": 36.8082,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9846575342465753,
+      "grad_norm": 2.3074066638946533,
+      "learning_rate": 1.911047928513404e-06,
+      "loss": 1.7952,
+      "step": 1811
+    },
+    {
+      "epoch": 1.9857534246575343,
+      "grad_norm": 2.3162307739257812,
+      "learning_rate": 1.909017059301381e-06,
+      "loss": 1.7455,
+      "step": 1812
+    },
+    {
+      "epoch": 1.9868493150684932,
+      "grad_norm": 2.3085129261016846,
+      "learning_rate": 1.9069861900893586e-06,
+      "loss": 1.7513,
+      "step": 1813
+    },
+    {
+      "epoch": 1.987945205479452,
+      "grad_norm": 2.3454785346984863,
+      "learning_rate": 1.9049553208773354e-06,
+      "loss": 1.803,
+      "step": 1814
+    },
+    {
+      "epoch": 1.989041095890411,
+      "grad_norm": 2.4522876739501953,
+      "learning_rate": 1.9029244516653127e-06,
+      "loss": 1.7567,
+      "step": 1815
+    },
+    {
+      "epoch": 1.9901369863013698,
+      "grad_norm": 2.3876044750213623,
+      "learning_rate": 1.90089358245329e-06,
+      "loss": 1.7126,
+      "step": 1816
+    },
+    {
+      "epoch": 1.9912328767123286,
+      "grad_norm": 2.4061570167541504,
+      "learning_rate": 1.8988627132412673e-06,
+      "loss": 1.7618,
+      "step": 1817
+    },
+    {
+      "epoch": 1.9923287671232877,
+      "grad_norm": 2.307445764541626,
+      "learning_rate": 1.8968318440292445e-06,
+      "loss": 1.8014,
+      "step": 1818
+    },
+    {
+      "epoch": 1.9934246575342467,
+      "grad_norm": 2.4414806365966797,
+      "learning_rate": 1.8948009748172218e-06,
+      "loss": 1.7202,
+      "step": 1819
+    },
+    {
+      "epoch": 1.9945205479452055,
+      "grad_norm": 2.3933329582214355,
+      "learning_rate": 1.892770105605199e-06,
+      "loss": 1.8133,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9945205479452055,
+      "eval_loss": 1.991702675819397,
+      "eval_runtime": 37.0127,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9956164383561643,
+      "grad_norm": 2.237699508666992,
+      "learning_rate": 1.8907392363931764e-06,
+      "loss": 1.8026,
+      "step": 1821
+    },
+    {
+      "epoch": 1.9967123287671233,
+      "grad_norm": 2.290004014968872,
+      "learning_rate": 1.8887083671811536e-06,
+      "loss": 1.7689,
+      "step": 1822
+    },
+    {
+      "epoch": 1.9978082191780822,
+      "grad_norm": 2.2264676094055176,
+      "learning_rate": 1.886677497969131e-06,
+      "loss": 1.7862,
+      "step": 1823
+    },
+    {
+      "epoch": 1.998904109589041,
+      "grad_norm": 2.2764949798583984,
+      "learning_rate": 1.8846466287571082e-06,
+      "loss": 1.6763,
+      "step": 1824
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 2.3024826049804688,
+      "learning_rate": 1.8826157595450855e-06,
+      "loss": 1.8076,
+      "step": 1825
+    },
+    {
+      "epoch": 2.001095890410959,
+      "grad_norm": 2.362882614135742,
+      "learning_rate": 1.8805848903330625e-06,
+      "loss": 1.7009,
+      "step": 1826
+    },
+    {
+      "epoch": 2.0021917808219176,
+      "grad_norm": 2.432370185852051,
+      "learning_rate": 1.8785540211210398e-06,
+      "loss": 1.8712,
+      "step": 1827
+    },
+    {
+      "epoch": 2.0032876712328767,
+      "grad_norm": 2.3584744930267334,
+      "learning_rate": 1.876523151909017e-06,
+      "loss": 1.7869,
+      "step": 1828
+    },
+    {
+      "epoch": 2.0043835616438357,
+      "grad_norm": 2.534205198287964,
+      "learning_rate": 1.8744922826969944e-06,
+      "loss": 1.7376,
+      "step": 1829
+    },
+    {
+      "epoch": 2.0054794520547947,
+      "grad_norm": 2.58241605758667,
+      "learning_rate": 1.8724614134849716e-06,
+      "loss": 1.7204,
+      "step": 1830
+    },
+    {
+      "epoch": 2.0054794520547947,
+      "eval_loss": 1.9912636280059814,
+      "eval_runtime": 37.0969,
+      "eval_samples_per_second": 26.956,
+      "eval_steps_per_second": 1.132,
+      "step": 1830
+    },
+    {
+      "epoch": 2.0065753424657533,
+      "grad_norm": 2.3742504119873047,
+      "learning_rate": 1.870430544272949e-06,
+      "loss": 1.7537,
+      "step": 1831
+    },
+    {
+      "epoch": 2.0076712328767123,
+      "grad_norm": 2.6560916900634766,
+      "learning_rate": 1.8683996750609262e-06,
+      "loss": 1.8524,
+      "step": 1832
+    },
+    {
+      "epoch": 2.0087671232876714,
+      "grad_norm": 2.523041248321533,
+      "learning_rate": 1.8663688058489035e-06,
+      "loss": 1.7405,
+      "step": 1833
+    },
+    {
+      "epoch": 2.00986301369863,
+      "grad_norm": 2.5444304943084717,
+      "learning_rate": 1.8643379366368807e-06,
+      "loss": 1.7588,
+      "step": 1834
+    },
+    {
+      "epoch": 2.010958904109589,
+      "grad_norm": 2.708984613418579,
+      "learning_rate": 1.862307067424858e-06,
+      "loss": 1.7686,
+      "step": 1835
+    },
+    {
+      "epoch": 2.012054794520548,
+      "grad_norm": 2.726691961288452,
+      "learning_rate": 1.8602761982128353e-06,
+      "loss": 1.7733,
+      "step": 1836
+    },
+    {
+      "epoch": 2.013150684931507,
+      "grad_norm": 2.5085904598236084,
+      "learning_rate": 1.8582453290008126e-06,
+      "loss": 1.81,
+      "step": 1837
+    },
+    {
+      "epoch": 2.0142465753424657,
+      "grad_norm": 2.4699466228485107,
+      "learning_rate": 1.8562144597887898e-06,
+      "loss": 1.7369,
+      "step": 1838
+    },
+    {
+      "epoch": 2.0153424657534247,
+      "grad_norm": 2.531027317047119,
+      "learning_rate": 1.854183590576767e-06,
+      "loss": 1.7611,
+      "step": 1839
+    },
+    {
+      "epoch": 2.0164383561643837,
+      "grad_norm": 2.5061161518096924,
+      "learning_rate": 1.8521527213647442e-06,
+      "loss": 1.811,
+      "step": 1840
+    },
+    {
+      "epoch": 2.0164383561643837,
+      "eval_loss": 1.9893144369125366,
+      "eval_runtime": 37.1262,
+      "eval_samples_per_second": 26.935,
+      "eval_steps_per_second": 1.131,
+      "step": 1840
+    },
+    {
+      "epoch": 2.0175342465753423,
+      "grad_norm": 2.5739994049072266,
+      "learning_rate": 1.8501218521527215e-06,
+      "loss": 1.7153,
+      "step": 1841
+    },
+    {
+      "epoch": 2.0186301369863013,
+      "grad_norm": 2.5673015117645264,
+      "learning_rate": 1.8480909829406987e-06,
+      "loss": 1.6879,
+      "step": 1842
+    },
+    {
+      "epoch": 2.0197260273972604,
+      "grad_norm": 2.717430353164673,
+      "learning_rate": 1.846060113728676e-06,
+      "loss": 1.723,
+      "step": 1843
+    },
+    {
+      "epoch": 2.020821917808219,
+      "grad_norm": 2.4905285835266113,
+      "learning_rate": 1.8440292445166533e-06,
+      "loss": 1.7651,
+      "step": 1844
+    },
+    {
+      "epoch": 2.021917808219178,
+      "grad_norm": 2.6798791885375977,
+      "learning_rate": 1.8419983753046306e-06,
+      "loss": 1.7308,
+      "step": 1845
+    },
+    {
+      "epoch": 2.023013698630137,
+      "grad_norm": 2.591614007949829,
+      "learning_rate": 1.8399675060926078e-06,
+      "loss": 1.786,
+      "step": 1846
+    },
+    {
+      "epoch": 2.024109589041096,
+      "grad_norm": 2.5405824184417725,
+      "learning_rate": 1.8379366368805851e-06,
+      "loss": 1.7699,
+      "step": 1847
+    },
+    {
+      "epoch": 2.0252054794520546,
+      "grad_norm": 2.4189934730529785,
+      "learning_rate": 1.8359057676685624e-06,
+      "loss": 1.732,
+      "step": 1848
+    },
+    {
+      "epoch": 2.0263013698630137,
+      "grad_norm": 2.5857982635498047,
+      "learning_rate": 1.8338748984565397e-06,
+      "loss": 1.7123,
+      "step": 1849
+    },
+    {
+      "epoch": 2.0273972602739727,
+      "grad_norm": 2.4710893630981445,
+      "learning_rate": 1.831844029244517e-06,
+      "loss": 1.7419,
+      "step": 1850
+    },
+    {
+      "epoch": 2.0273972602739727,
+      "eval_loss": 1.9874999523162842,
+      "eval_runtime": 37.1588,
+      "eval_samples_per_second": 26.912,
+      "eval_steps_per_second": 1.13,
+      "step": 1850
+    },
+    {
+      "epoch": 2.0284931506849313,
+      "grad_norm": 2.3414645195007324,
+      "learning_rate": 1.829813160032494e-06,
+      "loss": 1.7102,
+      "step": 1851
+    },
+    {
+      "epoch": 2.0295890410958903,
+      "grad_norm": 2.4148900508880615,
+      "learning_rate": 1.8277822908204713e-06,
+      "loss": 1.7287,
+      "step": 1852
+    },
+    {
+      "epoch": 2.0306849315068494,
+      "grad_norm": 2.5501346588134766,
+      "learning_rate": 1.8257514216084486e-06,
+      "loss": 1.7633,
+      "step": 1853
+    },
+    {
+      "epoch": 2.0317808219178084,
+      "grad_norm": 2.52215576171875,
+      "learning_rate": 1.8237205523964258e-06,
+      "loss": 1.7385,
+      "step": 1854
+    },
+    {
+      "epoch": 2.032876712328767,
+      "grad_norm": 2.433335065841675,
+      "learning_rate": 1.8216896831844031e-06,
+      "loss": 1.7566,
+      "step": 1855
+    },
+    {
+      "epoch": 2.033972602739726,
+      "grad_norm": 2.441232442855835,
+      "learning_rate": 1.8196588139723804e-06,
+      "loss": 1.8027,
+      "step": 1856
+    },
+    {
+      "epoch": 2.035068493150685,
+      "grad_norm": 2.4601082801818848,
+      "learning_rate": 1.8176279447603577e-06,
+      "loss": 1.6859,
+      "step": 1857
+    },
+    {
+      "epoch": 2.0361643835616436,
+      "grad_norm": 2.5008108615875244,
+      "learning_rate": 1.815597075548335e-06,
+      "loss": 1.7294,
+      "step": 1858
+    },
+    {
+      "epoch": 2.0372602739726027,
+      "grad_norm": 2.433915615081787,
+      "learning_rate": 1.8135662063363122e-06,
+      "loss": 1.755,
+      "step": 1859
+    },
+    {
+      "epoch": 2.0383561643835617,
+      "grad_norm": 2.4938697814941406,
+      "learning_rate": 1.8115353371242895e-06,
+      "loss": 1.7904,
+      "step": 1860
+    },
+    {
+      "epoch": 2.0383561643835617,
+      "eval_loss": 1.9939266443252563,
+      "eval_runtime": 37.1539,
+      "eval_samples_per_second": 26.915,
+      "eval_steps_per_second": 1.13,
+      "step": 1860
+    },
+    {
+      "epoch": 2.0394520547945207,
+      "grad_norm": 2.382324695587158,
+      "learning_rate": 1.8095044679122668e-06,
+      "loss": 1.6856,
+      "step": 1861
+    },
+    {
+      "epoch": 2.0405479452054793,
+      "grad_norm": 2.5250437259674072,
+      "learning_rate": 1.807473598700244e-06,
+      "loss": 1.7434,
+      "step": 1862
+    },
+    {
+      "epoch": 2.0416438356164384,
+      "grad_norm": 2.63969087600708,
+      "learning_rate": 1.805442729488221e-06,
+      "loss": 1.7343,
+      "step": 1863
+    },
+    {
+      "epoch": 2.0427397260273974,
+      "grad_norm": 2.4852068424224854,
+      "learning_rate": 1.8034118602761982e-06,
+      "loss": 1.7494,
+      "step": 1864
+    },
+    {
+      "epoch": 2.043835616438356,
+      "grad_norm": 2.393967390060425,
+      "learning_rate": 1.8013809910641754e-06,
+      "loss": 1.7503,
+      "step": 1865
+    },
+    {
+      "epoch": 2.044931506849315,
+      "grad_norm": 2.4447648525238037,
+      "learning_rate": 1.7993501218521527e-06,
+      "loss": 1.7539,
+      "step": 1866
+    },
+    {
+      "epoch": 2.046027397260274,
+      "grad_norm": 2.483530044555664,
+      "learning_rate": 1.79731925264013e-06,
+      "loss": 1.7001,
+      "step": 1867
+    },
+    {
+      "epoch": 2.047123287671233,
+      "grad_norm": 2.56669282913208,
+      "learning_rate": 1.7952883834281073e-06,
+      "loss": 1.7828,
+      "step": 1868
+    },
+    {
+      "epoch": 2.0482191780821917,
+      "grad_norm": 2.4894964694976807,
+      "learning_rate": 1.7932575142160846e-06,
+      "loss": 1.7208,
+      "step": 1869
+    },
+    {
+      "epoch": 2.0493150684931507,
+      "grad_norm": 2.4713854789733887,
+      "learning_rate": 1.7912266450040618e-06,
+      "loss": 1.7687,
+      "step": 1870
+    },
+    {
+      "epoch": 2.0493150684931507,
+      "eval_loss": 1.9938212633132935,
+      "eval_runtime": 37.1298,
+      "eval_samples_per_second": 26.933,
+      "eval_steps_per_second": 1.131,
+      "step": 1870
+    },
+    {
+      "epoch": 2.0504109589041097,
+      "grad_norm": 2.6204898357391357,
+      "learning_rate": 1.789195775792039e-06,
+      "loss": 1.7718,
+      "step": 1871
+    },
+    {
+      "epoch": 2.0515068493150683,
+      "grad_norm": 2.5194199085235596,
+      "learning_rate": 1.7871649065800164e-06,
+      "loss": 1.6653,
+      "step": 1872
+    },
+    {
+      "epoch": 2.0526027397260274,
+      "grad_norm": 2.5077965259552,
+      "learning_rate": 1.7851340373679937e-06,
+      "loss": 1.8098,
+      "step": 1873
+    },
+    {
+      "epoch": 2.0536986301369864,
+      "grad_norm": 2.5841965675354004,
+      "learning_rate": 1.783103168155971e-06,
+      "loss": 1.7126,
+      "step": 1874
+    },
+    {
+      "epoch": 2.0547945205479454,
+      "grad_norm": 2.617910623550415,
+      "learning_rate": 1.7810722989439482e-06,
+      "loss": 1.7294,
+      "step": 1875
+    },
+    {
+      "epoch": 2.055890410958904,
+      "grad_norm": 2.519253969192505,
+      "learning_rate": 1.7790414297319253e-06,
+      "loss": 1.7497,
+      "step": 1876
+    },
+    {
+      "epoch": 2.056986301369863,
+      "grad_norm": 2.5518412590026855,
+      "learning_rate": 1.7770105605199025e-06,
+      "loss": 1.7188,
+      "step": 1877
+    },
+    {
+      "epoch": 2.058082191780822,
+      "grad_norm": 2.5372824668884277,
+      "learning_rate": 1.7749796913078798e-06,
+      "loss": 1.6998,
+      "step": 1878
+    },
+    {
+      "epoch": 2.0591780821917807,
+      "grad_norm": 2.6649036407470703,
+      "learning_rate": 1.772948822095857e-06,
+      "loss": 1.7621,
+      "step": 1879
+    },
+    {
+      "epoch": 2.0602739726027397,
+      "grad_norm": 2.5216212272644043,
+      "learning_rate": 1.7709179528838344e-06,
+      "loss": 1.7963,
+      "step": 1880
+    },
+    {
+      "epoch": 2.0602739726027397,
+      "eval_loss": 1.993401050567627,
+      "eval_runtime": 37.0954,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 1880
+    },
+    {
+      "epoch": 2.0613698630136987,
+      "grad_norm": 2.5226478576660156,
+      "learning_rate": 1.7688870836718116e-06,
+      "loss": 1.7841,
+      "step": 1881
+    },
+    {
+      "epoch": 2.0624657534246573,
+      "grad_norm": 2.584319829940796,
+      "learning_rate": 1.766856214459789e-06,
+      "loss": 1.7449,
+      "step": 1882
+    },
+    {
+      "epoch": 2.0635616438356164,
+      "grad_norm": 2.5305304527282715,
+      "learning_rate": 1.7648253452477662e-06,
+      "loss": 1.7373,
+      "step": 1883
+    },
+    {
+      "epoch": 2.0646575342465754,
+      "grad_norm": 2.48797607421875,
+      "learning_rate": 1.7627944760357435e-06,
+      "loss": 1.722,
+      "step": 1884
+    },
+    {
+      "epoch": 2.0657534246575344,
+      "grad_norm": 2.4461028575897217,
+      "learning_rate": 1.7607636068237208e-06,
+      "loss": 1.7983,
+      "step": 1885
+    },
+    {
+      "epoch": 2.066849315068493,
+      "grad_norm": 2.624058961868286,
+      "learning_rate": 1.758732737611698e-06,
+      "loss": 1.7709,
+      "step": 1886
+    },
+    {
+      "epoch": 2.067945205479452,
+      "grad_norm": 2.848062515258789,
+      "learning_rate": 1.7567018683996753e-06,
+      "loss": 1.7513,
+      "step": 1887
+    },
+    {
+      "epoch": 2.069041095890411,
+      "grad_norm": 2.5219593048095703,
+      "learning_rate": 1.7546709991876524e-06,
+      "loss": 1.7387,
+      "step": 1888
+    },
+    {
+      "epoch": 2.0701369863013697,
+      "grad_norm": 2.5296876430511475,
+      "learning_rate": 1.7526401299756296e-06,
+      "loss": 1.7378,
+      "step": 1889
+    },
+    {
+      "epoch": 2.0712328767123287,
+      "grad_norm": 2.5452613830566406,
+      "learning_rate": 1.750609260763607e-06,
+      "loss": 1.7402,
+      "step": 1890
+    },
+    {
+      "epoch": 2.0712328767123287,
+      "eval_loss": 1.9914054870605469,
+      "eval_runtime": 37.0548,
+      "eval_samples_per_second": 26.987,
+      "eval_steps_per_second": 1.133,
+      "step": 1890
+    },
+    {
+      "epoch": 2.0723287671232877,
+      "grad_norm": 2.653596878051758,
+      "learning_rate": 1.7485783915515842e-06,
+      "loss": 1.7369,
+      "step": 1891
+    },
+    {
+      "epoch": 2.0734246575342468,
+      "grad_norm": 2.615581750869751,
+      "learning_rate": 1.7465475223395615e-06,
+      "loss": 1.7317,
+      "step": 1892
+    },
+    {
+      "epoch": 2.0745205479452054,
+      "grad_norm": 2.610304832458496,
+      "learning_rate": 1.7445166531275387e-06,
+      "loss": 1.7121,
+      "step": 1893
+    },
+    {
+      "epoch": 2.0756164383561644,
+      "grad_norm": 2.6463494300842285,
+      "learning_rate": 1.742485783915516e-06,
+      "loss": 1.8523,
+      "step": 1894
+    },
+    {
+      "epoch": 2.0767123287671234,
+      "grad_norm": 2.4814131259918213,
+      "learning_rate": 1.7404549147034933e-06,
+      "loss": 1.8239,
+      "step": 1895
+    },
+    {
+      "epoch": 2.077808219178082,
+      "grad_norm": 2.7416090965270996,
+      "learning_rate": 1.7384240454914706e-06,
+      "loss": 1.669,
+      "step": 1896
+    },
+    {
+      "epoch": 2.078904109589041,
+      "grad_norm": 2.9458305835723877,
+      "learning_rate": 1.7363931762794478e-06,
+      "loss": 1.7504,
+      "step": 1897
+    },
+    {
+      "epoch": 2.08,
+      "grad_norm": 2.5118932723999023,
+      "learning_rate": 1.7343623070674251e-06,
+      "loss": 1.7854,
+      "step": 1898
+    },
+    {
+      "epoch": 2.081095890410959,
+      "grad_norm": 2.479421615600586,
+      "learning_rate": 1.7323314378554024e-06,
+      "loss": 1.7325,
+      "step": 1899
+    },
+    {
+      "epoch": 2.0821917808219177,
+      "grad_norm": 2.7400424480438232,
+      "learning_rate": 1.7303005686433797e-06,
+      "loss": 1.7434,
+      "step": 1900
+    },
+    {
+      "epoch": 2.0821917808219177,
+      "eval_loss": 1.9887213706970215,
+      "eval_runtime": 37.0484,
+      "eval_samples_per_second": 26.992,
+      "eval_steps_per_second": 1.134,
+      "step": 1900
+    },
+    {
+      "epoch": 2.0832876712328767,
+      "grad_norm": 2.587714433670044,
+      "learning_rate": 1.7282696994313567e-06,
+      "loss": 1.7335,
+      "step": 1901
+    },
+    {
+      "epoch": 2.0843835616438358,
+      "grad_norm": 2.526818037033081,
+      "learning_rate": 1.726238830219334e-06,
+      "loss": 1.7538,
+      "step": 1902
+    },
+    {
+      "epoch": 2.0854794520547943,
+      "grad_norm": 2.595221519470215,
+      "learning_rate": 1.7242079610073113e-06,
+      "loss": 1.7331,
+      "step": 1903
+    },
+    {
+      "epoch": 2.0865753424657534,
+      "grad_norm": 2.5018110275268555,
+      "learning_rate": 1.7221770917952886e-06,
+      "loss": 1.8069,
+      "step": 1904
+    },
+    {
+      "epoch": 2.0876712328767124,
+      "grad_norm": 2.4378256797790527,
+      "learning_rate": 1.7201462225832658e-06,
+      "loss": 1.757,
+      "step": 1905
+    },
+    {
+      "epoch": 2.0887671232876714,
+      "grad_norm": 2.651167392730713,
+      "learning_rate": 1.7181153533712431e-06,
+      "loss": 1.7696,
+      "step": 1906
+    },
+    {
+      "epoch": 2.08986301369863,
+      "grad_norm": 2.550239086151123,
+      "learning_rate": 1.7160844841592204e-06,
+      "loss": 1.6551,
+      "step": 1907
+    },
+    {
+      "epoch": 2.090958904109589,
+      "grad_norm": 2.4192144870758057,
+      "learning_rate": 1.7140536149471977e-06,
+      "loss": 1.7031,
+      "step": 1908
+    },
+    {
+      "epoch": 2.092054794520548,
+      "grad_norm": 2.5327398777008057,
+      "learning_rate": 1.712022745735175e-06,
+      "loss": 1.734,
+      "step": 1909
+    },
+    {
+      "epoch": 2.0931506849315067,
+      "grad_norm": 2.5962865352630615,
+      "learning_rate": 1.7099918765231522e-06,
+      "loss": 1.732,
+      "step": 1910
+    },
+    {
+      "epoch": 2.0931506849315067,
+      "eval_loss": 1.9915482997894287,
+      "eval_runtime": 36.7856,
+      "eval_samples_per_second": 27.185,
+      "eval_steps_per_second": 1.142,
+      "step": 1910
+    },
+    {
+      "epoch": 2.0942465753424657,
+      "grad_norm": 2.561943531036377,
+      "learning_rate": 1.7079610073111295e-06,
+      "loss": 1.7556,
+      "step": 1911
+    },
+    {
+      "epoch": 2.0953424657534248,
+      "grad_norm": 2.529829263687134,
+      "learning_rate": 1.7059301380991068e-06,
+      "loss": 1.6839,
+      "step": 1912
+    },
+    {
+      "epoch": 2.0964383561643833,
+      "grad_norm": 2.636559009552002,
+      "learning_rate": 1.7038992688870836e-06,
+      "loss": 1.7393,
+      "step": 1913
+    },
+    {
+      "epoch": 2.0975342465753424,
+      "grad_norm": 2.558166742324829,
+      "learning_rate": 1.701868399675061e-06,
+      "loss": 1.7346,
+      "step": 1914
+    },
+    {
+      "epoch": 2.0986301369863014,
+      "grad_norm": 2.5193750858306885,
+      "learning_rate": 1.6998375304630382e-06,
+      "loss": 1.7666,
+      "step": 1915
+    },
+    {
+      "epoch": 2.0997260273972604,
+      "grad_norm": 2.498244047164917,
+      "learning_rate": 1.6978066612510155e-06,
+      "loss": 1.827,
+      "step": 1916
+    },
+    {
+      "epoch": 2.100821917808219,
+      "grad_norm": 2.799603223800659,
+      "learning_rate": 1.6957757920389927e-06,
+      "loss": 1.788,
+      "step": 1917
+    },
+    {
+      "epoch": 2.101917808219178,
+      "grad_norm": 2.733391761779785,
+      "learning_rate": 1.69374492282697e-06,
+      "loss": 1.7487,
+      "step": 1918
+    },
+    {
+      "epoch": 2.103013698630137,
+      "grad_norm": 2.5135183334350586,
+      "learning_rate": 1.6917140536149473e-06,
+      "loss": 1.7114,
+      "step": 1919
+    },
+    {
+      "epoch": 2.1041095890410957,
+      "grad_norm": 2.5275800228118896,
+      "learning_rate": 1.6896831844029246e-06,
+      "loss": 1.797,
+      "step": 1920
+    },
+    {
+      "epoch": 2.1041095890410957,
+      "eval_loss": 1.9926362037658691,
+      "eval_runtime": 36.9326,
+      "eval_samples_per_second": 27.076,
+      "eval_steps_per_second": 1.137,
+      "step": 1920
+    },
+    {
+      "epoch": 2.1052054794520547,
+      "grad_norm": 2.682774782180786,
+      "learning_rate": 1.6876523151909018e-06,
+      "loss": 1.7581,
+      "step": 1921
+    },
+    {
+      "epoch": 2.1063013698630138,
+      "grad_norm": 2.625620126724243,
+      "learning_rate": 1.6856214459788791e-06,
+      "loss": 1.7606,
+      "step": 1922
+    },
+    {
+      "epoch": 2.107397260273973,
+      "grad_norm": 2.5005173683166504,
+      "learning_rate": 1.6835905767668564e-06,
+      "loss": 1.7139,
+      "step": 1923
+    },
+    {
+      "epoch": 2.1084931506849314,
+      "grad_norm": 2.6731526851654053,
+      "learning_rate": 1.6815597075548337e-06,
+      "loss": 1.7174,
+      "step": 1924
+    },
+    {
+      "epoch": 2.1095890410958904,
+      "grad_norm": 2.5407724380493164,
+      "learning_rate": 1.679528838342811e-06,
+      "loss": 1.8049,
+      "step": 1925
+    },
+    {
+      "epoch": 2.1106849315068494,
+      "grad_norm": 2.502427101135254,
+      "learning_rate": 1.677497969130788e-06,
+      "loss": 1.7068,
+      "step": 1926
+    },
+    {
+      "epoch": 2.111780821917808,
+      "grad_norm": 2.5063414573669434,
+      "learning_rate": 1.6754670999187653e-06,
+      "loss": 1.7174,
+      "step": 1927
+    },
+    {
+      "epoch": 2.112876712328767,
+      "grad_norm": 2.524588108062744,
+      "learning_rate": 1.6734362307067426e-06,
+      "loss": 1.7442,
+      "step": 1928
+    },
+    {
+      "epoch": 2.113972602739726,
+      "grad_norm": 2.531909465789795,
+      "learning_rate": 1.6714053614947198e-06,
+      "loss": 1.692,
+      "step": 1929
+    },
+    {
+      "epoch": 2.115068493150685,
+      "grad_norm": 2.5104269981384277,
+      "learning_rate": 1.669374492282697e-06,
+      "loss": 1.7039,
+      "step": 1930
+    },
+    {
+      "epoch": 2.115068493150685,
+      "eval_loss": 1.9927221536636353,
+      "eval_runtime": 37.004,
+      "eval_samples_per_second": 27.024,
+      "eval_steps_per_second": 1.135,
+      "step": 1930
+    },
+    {
+      "epoch": 2.1161643835616437,
+      "grad_norm": 2.5443670749664307,
+      "learning_rate": 1.6673436230706744e-06,
+      "loss": 1.7896,
+      "step": 1931
+    },
+    {
+      "epoch": 2.1172602739726027,
+      "grad_norm": 2.6495745182037354,
+      "learning_rate": 1.6653127538586517e-06,
+      "loss": 1.7735,
+      "step": 1932
+    },
+    {
+      "epoch": 2.118356164383562,
+      "grad_norm": 2.7233121395111084,
+      "learning_rate": 1.663281884646629e-06,
+      "loss": 1.7186,
+      "step": 1933
+    },
+    {
+      "epoch": 2.1194520547945204,
+      "grad_norm": 2.4411816596984863,
+      "learning_rate": 1.6612510154346062e-06,
+      "loss": 1.7378,
+      "step": 1934
+    },
+    {
+      "epoch": 2.1205479452054794,
+      "grad_norm": 2.6478421688079834,
+      "learning_rate": 1.6592201462225835e-06,
+      "loss": 1.7989,
+      "step": 1935
+    },
+    {
+      "epoch": 2.1216438356164384,
+      "grad_norm": 2.56093168258667,
+      "learning_rate": 1.6571892770105608e-06,
+      "loss": 1.7663,
+      "step": 1936
+    },
+    {
+      "epoch": 2.1227397260273975,
+      "grad_norm": 2.5598859786987305,
+      "learning_rate": 1.655158407798538e-06,
+      "loss": 1.7956,
+      "step": 1937
+    },
+    {
+      "epoch": 2.123835616438356,
+      "grad_norm": 2.4958112239837646,
+      "learning_rate": 1.653127538586515e-06,
+      "loss": 1.7482,
+      "step": 1938
+    },
+    {
+      "epoch": 2.124931506849315,
+      "grad_norm": 2.5133280754089355,
+      "learning_rate": 1.6510966693744924e-06,
+      "loss": 1.7811,
+      "step": 1939
+    },
+    {
+      "epoch": 2.126027397260274,
+      "grad_norm": 2.492892026901245,
+      "learning_rate": 1.6490658001624696e-06,
+      "loss": 1.7883,
+      "step": 1940
+    },
+    {
+      "epoch": 2.126027397260274,
+      "eval_loss": 1.9917834997177124,
+      "eval_runtime": 37.0222,
+      "eval_samples_per_second": 27.011,
+      "eval_steps_per_second": 1.134,
+      "step": 1940
+    },
+    {
+      "epoch": 2.1271232876712327,
+      "grad_norm": 2.560483455657959,
+      "learning_rate": 1.647034930950447e-06,
+      "loss": 1.7871,
+      "step": 1941
+    },
+    {
+      "epoch": 2.1282191780821917,
+      "grad_norm": 2.549100160598755,
+      "learning_rate": 1.6450040617384242e-06,
+      "loss": 1.7097,
+      "step": 1942
+    },
+    {
+      "epoch": 2.1293150684931508,
+      "grad_norm": 2.4832775592803955,
+      "learning_rate": 1.6429731925264015e-06,
+      "loss": 1.736,
+      "step": 1943
+    },
+    {
+      "epoch": 2.1304109589041094,
+      "grad_norm": 2.60006046295166,
+      "learning_rate": 1.6409423233143788e-06,
+      "loss": 1.765,
+      "step": 1944
+    },
+    {
+      "epoch": 2.1315068493150684,
+      "grad_norm": 2.4875144958496094,
+      "learning_rate": 1.638911454102356e-06,
+      "loss": 1.7317,
+      "step": 1945
+    },
+    {
+      "epoch": 2.1326027397260274,
+      "grad_norm": 2.5562970638275146,
+      "learning_rate": 1.6368805848903333e-06,
+      "loss": 1.6982,
+      "step": 1946
+    },
+    {
+      "epoch": 2.1336986301369865,
+      "grad_norm": 2.4928972721099854,
+      "learning_rate": 1.6348497156783106e-06,
+      "loss": 1.7821,
+      "step": 1947
+    },
+    {
+      "epoch": 2.134794520547945,
+      "grad_norm": 2.6159043312072754,
+      "learning_rate": 1.6328188464662879e-06,
+      "loss": 1.7223,
+      "step": 1948
+    },
+    {
+      "epoch": 2.135890410958904,
+      "grad_norm": 2.541520833969116,
+      "learning_rate": 1.6307879772542651e-06,
+      "loss": 1.7321,
+      "step": 1949
+    },
+    {
+      "epoch": 2.136986301369863,
+      "grad_norm": 2.5165696144104004,
+      "learning_rate": 1.628757108042242e-06,
+      "loss": 1.7571,
+      "step": 1950
+    },
+    {
+      "epoch": 2.136986301369863,
+      "eval_loss": 1.9868476390838623,
+      "eval_runtime": 37.0367,
+      "eval_samples_per_second": 27.0,
+      "eval_steps_per_second": 1.134,
+      "step": 1950
+    },
+    {
+      "epoch": 2.138082191780822,
+      "grad_norm": 2.5660274028778076,
+      "learning_rate": 1.6267262388302193e-06,
+      "loss": 1.7855,
+      "step": 1951
+    },
+    {
+      "epoch": 2.1391780821917807,
+      "grad_norm": 2.4606971740722656,
+      "learning_rate": 1.6246953696181965e-06,
+      "loss": 1.7324,
+      "step": 1952
+    },
+    {
+      "epoch": 2.1402739726027398,
+      "grad_norm": 2.6528608798980713,
+      "learning_rate": 1.6226645004061738e-06,
+      "loss": 1.7715,
+      "step": 1953
+    },
+    {
+      "epoch": 2.141369863013699,
+      "grad_norm": 2.7232985496520996,
+      "learning_rate": 1.620633631194151e-06,
+      "loss": 1.837,
+      "step": 1954
+    },
+    {
+      "epoch": 2.1424657534246574,
+      "grad_norm": 2.6264185905456543,
+      "learning_rate": 1.6186027619821284e-06,
+      "loss": 1.7678,
+      "step": 1955
+    },
+    {
+      "epoch": 2.1435616438356164,
+      "grad_norm": 2.6248252391815186,
+      "learning_rate": 1.6165718927701056e-06,
+      "loss": 1.7218,
+      "step": 1956
+    },
+    {
+      "epoch": 2.1446575342465755,
+      "grad_norm": 3.00539493560791,
+      "learning_rate": 1.614541023558083e-06,
+      "loss": 1.7678,
+      "step": 1957
+    },
+    {
+      "epoch": 2.145753424657534,
+      "grad_norm": 2.5368528366088867,
+      "learning_rate": 1.6125101543460602e-06,
+      "loss": 1.7737,
+      "step": 1958
+    },
+    {
+      "epoch": 2.146849315068493,
+      "grad_norm": 2.5805916786193848,
+      "learning_rate": 1.6104792851340375e-06,
+      "loss": 1.7394,
+      "step": 1959
+    },
+    {
+      "epoch": 2.147945205479452,
+      "grad_norm": 2.615856647491455,
+      "learning_rate": 1.6084484159220147e-06,
+      "loss": 1.766,
+      "step": 1960
+    },
+    {
+      "epoch": 2.147945205479452,
+      "eval_loss": 1.987140417098999,
+      "eval_runtime": 37.0593,
+      "eval_samples_per_second": 26.984,
+      "eval_steps_per_second": 1.133,
+      "step": 1960
+    },
+    {
+      "epoch": 2.149041095890411,
+      "grad_norm": 2.6550803184509277,
+      "learning_rate": 1.606417546709992e-06,
+      "loss": 1.7342,
+      "step": 1961
+    },
+    {
+      "epoch": 2.1501369863013697,
+      "grad_norm": 2.5999467372894287,
+      "learning_rate": 1.6043866774979693e-06,
+      "loss": 1.7175,
+      "step": 1962
+    },
+    {
+      "epoch": 2.1512328767123288,
+      "grad_norm": 2.561642646789551,
+      "learning_rate": 1.6023558082859464e-06,
+      "loss": 1.7409,
+      "step": 1963
+    },
+    {
+      "epoch": 2.152328767123288,
+      "grad_norm": 2.6062211990356445,
+      "learning_rate": 1.6003249390739236e-06,
+      "loss": 1.7052,
+      "step": 1964
+    },
+    {
+      "epoch": 2.1534246575342464,
+      "grad_norm": 2.5122230052948,
+      "learning_rate": 1.598294069861901e-06,
+      "loss": 1.7444,
+      "step": 1965
+    },
+    {
+      "epoch": 2.1545205479452054,
+      "grad_norm": 2.5838866233825684,
+      "learning_rate": 1.5962632006498782e-06,
+      "loss": 1.716,
+      "step": 1966
+    },
+    {
+      "epoch": 2.1556164383561645,
+      "grad_norm": 2.5454747676849365,
+      "learning_rate": 1.5942323314378555e-06,
+      "loss": 1.7902,
+      "step": 1967
+    },
+    {
+      "epoch": 2.1567123287671235,
+      "grad_norm": 2.57637619972229,
+      "learning_rate": 1.5922014622258327e-06,
+      "loss": 1.6963,
+      "step": 1968
+    },
+    {
+      "epoch": 2.157808219178082,
+      "grad_norm": 2.5542941093444824,
+      "learning_rate": 1.59017059301381e-06,
+      "loss": 1.7151,
+      "step": 1969
+    },
+    {
+      "epoch": 2.158904109589041,
+      "grad_norm": 2.4912102222442627,
+      "learning_rate": 1.5881397238017873e-06,
+      "loss": 1.7624,
+      "step": 1970
+    },
+    {
+      "epoch": 2.158904109589041,
+      "eval_loss": 1.9855093955993652,
+      "eval_runtime": 37.0796,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 1970
+    },
+    {
+      "epoch": 2.16,
+      "grad_norm": 2.4793057441711426,
+      "learning_rate": 1.5861088545897646e-06,
+      "loss": 1.7617,
+      "step": 1971
+    },
+    {
+      "epoch": 2.1610958904109587,
+      "grad_norm": 2.5231704711914062,
+      "learning_rate": 1.5840779853777418e-06,
+      "loss": 1.6845,
+      "step": 1972
+    },
+    {
+      "epoch": 2.1621917808219178,
+      "grad_norm": 2.5126073360443115,
+      "learning_rate": 1.5820471161657191e-06,
+      "loss": 1.7468,
+      "step": 1973
+    },
+    {
+      "epoch": 2.163287671232877,
+      "grad_norm": 2.5558865070343018,
+      "learning_rate": 1.5800162469536964e-06,
+      "loss": 1.7253,
+      "step": 1974
+    },
+    {
+      "epoch": 2.1643835616438354,
+      "grad_norm": 2.808987855911255,
+      "learning_rate": 1.5779853777416735e-06,
+      "loss": 1.7844,
+      "step": 1975
+    },
+    {
+      "epoch": 2.1654794520547944,
+      "grad_norm": 2.765275001525879,
+      "learning_rate": 1.5759545085296507e-06,
+      "loss": 1.7658,
+      "step": 1976
+    },
+    {
+      "epoch": 2.1665753424657535,
+      "grad_norm": 2.5721709728240967,
+      "learning_rate": 1.573923639317628e-06,
+      "loss": 1.74,
+      "step": 1977
+    },
+    {
+      "epoch": 2.1676712328767125,
+      "grad_norm": 2.4276933670043945,
+      "learning_rate": 1.5718927701056053e-06,
+      "loss": 1.7613,
+      "step": 1978
+    },
+    {
+      "epoch": 2.168767123287671,
+      "grad_norm": 2.5665223598480225,
+      "learning_rate": 1.5698619008935826e-06,
+      "loss": 1.7514,
+      "step": 1979
+    },
+    {
+      "epoch": 2.16986301369863,
+      "grad_norm": 2.5588295459747314,
+      "learning_rate": 1.5678310316815598e-06,
+      "loss": 1.7749,
+      "step": 1980
+    },
+    {
+      "epoch": 2.16986301369863,
+      "eval_loss": 1.9858319759368896,
+      "eval_runtime": 37.0705,
+      "eval_samples_per_second": 26.976,
+      "eval_steps_per_second": 1.133,
+      "step": 1980
+    },
+    {
+      "epoch": 2.170958904109589,
+      "grad_norm": 2.5413172245025635,
+      "learning_rate": 1.5658001624695371e-06,
+      "loss": 1.706,
+      "step": 1981
+    },
+    {
+      "epoch": 2.172054794520548,
+      "grad_norm": 2.5246310234069824,
+      "learning_rate": 1.5637692932575144e-06,
+      "loss": 1.7825,
+      "step": 1982
+    },
+    {
+      "epoch": 2.1731506849315068,
+      "grad_norm": 2.385380506515503,
+      "learning_rate": 1.5617384240454917e-06,
+      "loss": 1.7268,
+      "step": 1983
+    },
+    {
+      "epoch": 2.174246575342466,
+      "grad_norm": 2.4967970848083496,
+      "learning_rate": 1.559707554833469e-06,
+      "loss": 1.7424,
+      "step": 1984
+    },
+    {
+      "epoch": 2.175342465753425,
+      "grad_norm": 2.563683271408081,
+      "learning_rate": 1.5576766856214462e-06,
+      "loss": 1.7904,
+      "step": 1985
+    },
+    {
+      "epoch": 2.1764383561643834,
+      "grad_norm": 2.518751859664917,
+      "learning_rate": 1.5556458164094235e-06,
+      "loss": 1.6616,
+      "step": 1986
+    },
+    {
+      "epoch": 2.1775342465753424,
+      "grad_norm": 2.5789380073547363,
+      "learning_rate": 1.5536149471974008e-06,
+      "loss": 1.7363,
+      "step": 1987
+    },
+    {
+      "epoch": 2.1786301369863015,
+      "grad_norm": 2.567406415939331,
+      "learning_rate": 1.5515840779853778e-06,
+      "loss": 1.6835,
+      "step": 1988
+    },
+    {
+      "epoch": 2.17972602739726,
+      "grad_norm": 2.509427309036255,
+      "learning_rate": 1.549553208773355e-06,
+      "loss": 1.7493,
+      "step": 1989
+    },
+    {
+      "epoch": 2.180821917808219,
+      "grad_norm": 2.6355340480804443,
+      "learning_rate": 1.5475223395613324e-06,
+      "loss": 1.7292,
+      "step": 1990
+    },
+    {
+      "epoch": 2.180821917808219,
+      "eval_loss": 1.9893935918807983,
+      "eval_runtime": 37.0928,
+      "eval_samples_per_second": 26.959,
+      "eval_steps_per_second": 1.132,
+      "step": 1990
+    },
+    {
+      "epoch": 2.181917808219178,
+      "grad_norm": 2.564587116241455,
+      "learning_rate": 1.5454914703493097e-06,
+      "loss": 1.7248,
+      "step": 1991
+    },
+    {
+      "epoch": 2.183013698630137,
+      "grad_norm": 2.6708462238311768,
+      "learning_rate": 1.543460601137287e-06,
+      "loss": 1.8123,
+      "step": 1992
+    },
+    {
+      "epoch": 2.1841095890410958,
+      "grad_norm": 2.6396312713623047,
+      "learning_rate": 1.5414297319252642e-06,
+      "loss": 1.7742,
+      "step": 1993
+    },
+    {
+      "epoch": 2.185205479452055,
+      "grad_norm": 2.553544044494629,
+      "learning_rate": 1.5393988627132415e-06,
+      "loss": 1.7705,
+      "step": 1994
+    },
+    {
+      "epoch": 2.186301369863014,
+      "grad_norm": 2.554997682571411,
+      "learning_rate": 1.5373679935012188e-06,
+      "loss": 1.737,
+      "step": 1995
+    },
+    {
+      "epoch": 2.1873972602739724,
+      "grad_norm": 2.5650475025177,
+      "learning_rate": 1.535337124289196e-06,
+      "loss": 1.6865,
+      "step": 1996
+    },
+    {
+      "epoch": 2.1884931506849314,
+      "grad_norm": 2.8019111156463623,
+      "learning_rate": 1.5333062550771733e-06,
+      "loss": 1.7428,
+      "step": 1997
+    },
+    {
+      "epoch": 2.1895890410958905,
+      "grad_norm": 2.5704092979431152,
+      "learning_rate": 1.5312753858651506e-06,
+      "loss": 1.8291,
+      "step": 1998
+    },
+    {
+      "epoch": 2.1906849315068495,
+      "grad_norm": 2.5332226753234863,
+      "learning_rate": 1.5292445166531279e-06,
+      "loss": 1.7274,
+      "step": 1999
+    },
+    {
+      "epoch": 2.191780821917808,
+      "grad_norm": 2.555288553237915,
+      "learning_rate": 1.5272136474411047e-06,
+      "loss": 1.7497,
+      "step": 2000
+    },
+    {
+      "epoch": 2.191780821917808,
+      "eval_loss": 1.9881768226623535,
+      "eval_runtime": 37.0996,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 2000
+    },
+    {
+      "epoch": 2.192876712328767,
+      "grad_norm": 2.6068379878997803,
+      "learning_rate": 1.525182778229082e-06,
+      "loss": 1.7955,
+      "step": 2001
+    },
+    {
+      "epoch": 2.193972602739726,
+      "grad_norm": 2.479121208190918,
+      "learning_rate": 1.5231519090170593e-06,
+      "loss": 1.7342,
+      "step": 2002
+    },
+    {
+      "epoch": 2.1950684931506848,
+      "grad_norm": 2.495670795440674,
+      "learning_rate": 1.5211210398050365e-06,
+      "loss": 1.7125,
+      "step": 2003
+    },
+    {
+      "epoch": 2.196164383561644,
+      "grad_norm": 2.4988651275634766,
+      "learning_rate": 1.5190901705930138e-06,
+      "loss": 1.7137,
+      "step": 2004
+    },
+    {
+      "epoch": 2.197260273972603,
+      "grad_norm": 2.5191574096679688,
+      "learning_rate": 1.517059301380991e-06,
+      "loss": 1.7672,
+      "step": 2005
+    },
+    {
+      "epoch": 2.198356164383562,
+      "grad_norm": 2.5469892024993896,
+      "learning_rate": 1.5150284321689684e-06,
+      "loss": 1.7127,
+      "step": 2006
+    },
+    {
+      "epoch": 2.1994520547945204,
+      "grad_norm": 2.490896701812744,
+      "learning_rate": 1.5129975629569456e-06,
+      "loss": 1.7397,
+      "step": 2007
+    },
+    {
+      "epoch": 2.2005479452054795,
+      "grad_norm": 2.5142805576324463,
+      "learning_rate": 1.510966693744923e-06,
+      "loss": 1.7431,
+      "step": 2008
+    },
+    {
+      "epoch": 2.2016438356164385,
+      "grad_norm": 2.5397887229919434,
+      "learning_rate": 1.5089358245329002e-06,
+      "loss": 1.7439,
+      "step": 2009
+    },
+    {
+      "epoch": 2.202739726027397,
+      "grad_norm": 2.5063931941986084,
+      "learning_rate": 1.5069049553208775e-06,
+      "loss": 1.7259,
+      "step": 2010
+    },
+    {
+      "epoch": 2.202739726027397,
+      "eval_loss": 1.9853808879852295,
+      "eval_runtime": 36.8066,
+      "eval_samples_per_second": 27.169,
+      "eval_steps_per_second": 1.141,
+      "step": 2010
+    },
+    {
+      "epoch": 2.203835616438356,
+      "grad_norm": 2.6017589569091797,
+      "learning_rate": 1.5048740861088547e-06,
+      "loss": 1.7714,
+      "step": 2011
+    },
+    {
+      "epoch": 2.204931506849315,
+      "grad_norm": 2.5295040607452393,
+      "learning_rate": 1.502843216896832e-06,
+      "loss": 1.7561,
+      "step": 2012
+    },
+    {
+      "epoch": 2.206027397260274,
+      "grad_norm": 2.5187482833862305,
+      "learning_rate": 1.500812347684809e-06,
+      "loss": 1.6761,
+      "step": 2013
+    },
+    {
+      "epoch": 2.207123287671233,
+      "grad_norm": 2.5163650512695312,
+      "learning_rate": 1.4987814784727864e-06,
+      "loss": 1.7469,
+      "step": 2014
+    },
+    {
+      "epoch": 2.208219178082192,
+      "grad_norm": 2.731238603591919,
+      "learning_rate": 1.4967506092607636e-06,
+      "loss": 1.7527,
+      "step": 2015
+    },
+    {
+      "epoch": 2.209315068493151,
+      "grad_norm": 2.5298311710357666,
+      "learning_rate": 1.494719740048741e-06,
+      "loss": 1.6912,
+      "step": 2016
+    },
+    {
+      "epoch": 2.2104109589041094,
+      "grad_norm": 2.6385974884033203,
+      "learning_rate": 1.4926888708367182e-06,
+      "loss": 1.7664,
+      "step": 2017
+    },
+    {
+      "epoch": 2.2115068493150685,
+      "grad_norm": 2.6150033473968506,
+      "learning_rate": 1.4906580016246955e-06,
+      "loss": 1.7708,
+      "step": 2018
+    },
+    {
+      "epoch": 2.2126027397260275,
+      "grad_norm": 2.589329957962036,
+      "learning_rate": 1.4886271324126727e-06,
+      "loss": 1.7875,
+      "step": 2019
+    },
+    {
+      "epoch": 2.213698630136986,
+      "grad_norm": 2.6769933700561523,
+      "learning_rate": 1.48659626320065e-06,
+      "loss": 1.8136,
+      "step": 2020
+    },
+    {
+      "epoch": 2.213698630136986,
+      "eval_loss": 1.9827086925506592,
+      "eval_runtime": 37.0026,
+      "eval_samples_per_second": 27.025,
+      "eval_steps_per_second": 1.135,
+      "step": 2020
+    },
+    {
+      "epoch": 2.214794520547945,
+      "grad_norm": 2.7220497131347656,
+      "learning_rate": 1.4845653939886273e-06,
+      "loss": 1.6627,
+      "step": 2021
+    },
+    {
+      "epoch": 2.215890410958904,
+      "grad_norm": 2.6115968227386475,
+      "learning_rate": 1.4825345247766046e-06,
+      "loss": 1.6676,
+      "step": 2022
+    },
+    {
+      "epoch": 2.216986301369863,
+      "grad_norm": 2.4612581729888916,
+      "learning_rate": 1.4805036555645818e-06,
+      "loss": 1.6644,
+      "step": 2023
+    },
+    {
+      "epoch": 2.2180821917808218,
+      "grad_norm": 2.5708696842193604,
+      "learning_rate": 1.4784727863525591e-06,
+      "loss": 1.7357,
+      "step": 2024
+    },
+    {
+      "epoch": 2.219178082191781,
+      "grad_norm": 2.578158140182495,
+      "learning_rate": 1.4764419171405362e-06,
+      "loss": 1.7011,
+      "step": 2025
+    },
+    {
+      "epoch": 2.22027397260274,
+      "grad_norm": 2.547666072845459,
+      "learning_rate": 1.4744110479285135e-06,
+      "loss": 1.7787,
+      "step": 2026
+    },
+    {
+      "epoch": 2.2213698630136984,
+      "grad_norm": 2.617551326751709,
+      "learning_rate": 1.4723801787164907e-06,
+      "loss": 1.784,
+      "step": 2027
+    },
+    {
+      "epoch": 2.2224657534246575,
+      "grad_norm": 2.4385478496551514,
+      "learning_rate": 1.470349309504468e-06,
+      "loss": 1.7764,
+      "step": 2028
+    },
+    {
+      "epoch": 2.2235616438356165,
+      "grad_norm": 2.6088693141937256,
+      "learning_rate": 1.4683184402924453e-06,
+      "loss": 1.7477,
+      "step": 2029
+    },
+    {
+      "epoch": 2.2246575342465755,
+      "grad_norm": 2.6165895462036133,
+      "learning_rate": 1.4662875710804226e-06,
+      "loss": 1.7531,
+      "step": 2030
+    },
+    {
+      "epoch": 2.2246575342465755,
+      "eval_loss": 1.9826008081436157,
+      "eval_runtime": 37.085,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 2030
+    },
+    {
+      "epoch": 2.225753424657534,
+      "grad_norm": 2.6285006999969482,
+      "learning_rate": 1.4642567018683998e-06,
+      "loss": 1.8403,
+      "step": 2031
+    },
+    {
+      "epoch": 2.226849315068493,
+      "grad_norm": 2.549326181411743,
+      "learning_rate": 1.4622258326563771e-06,
+      "loss": 1.7304,
+      "step": 2032
+    },
+    {
+      "epoch": 2.227945205479452,
+      "grad_norm": 2.6198782920837402,
+      "learning_rate": 1.4601949634443544e-06,
+      "loss": 1.7147,
+      "step": 2033
+    },
+    {
+      "epoch": 2.2290410958904108,
+      "grad_norm": 2.536947727203369,
+      "learning_rate": 1.4581640942323317e-06,
+      "loss": 1.7639,
+      "step": 2034
+    },
+    {
+      "epoch": 2.23013698630137,
+      "grad_norm": 2.6757264137268066,
+      "learning_rate": 1.456133225020309e-06,
+      "loss": 1.8434,
+      "step": 2035
+    },
+    {
+      "epoch": 2.231232876712329,
+      "grad_norm": 2.598095655441284,
+      "learning_rate": 1.4541023558082862e-06,
+      "loss": 1.7593,
+      "step": 2036
+    },
+    {
+      "epoch": 2.232328767123288,
+      "grad_norm": 2.6104440689086914,
+      "learning_rate": 1.4520714865962635e-06,
+      "loss": 1.7473,
+      "step": 2037
+    },
+    {
+      "epoch": 2.2334246575342465,
+      "grad_norm": 2.6350162029266357,
+      "learning_rate": 1.4500406173842406e-06,
+      "loss": 1.777,
+      "step": 2038
+    },
+    {
+      "epoch": 2.2345205479452055,
+      "grad_norm": 2.593874216079712,
+      "learning_rate": 1.4480097481722178e-06,
+      "loss": 1.7506,
+      "step": 2039
+    },
+    {
+      "epoch": 2.2356164383561645,
+      "grad_norm": 2.6991093158721924,
+      "learning_rate": 1.4459788789601951e-06,
+      "loss": 1.7375,
+      "step": 2040
+    },
+    {
+      "epoch": 2.2356164383561645,
+      "eval_loss": 1.9815349578857422,
+      "eval_runtime": 37.1126,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 2040
+    },
+    {
+      "epoch": 2.236712328767123,
+      "grad_norm": 2.6339058876037598,
+      "learning_rate": 1.4439480097481724e-06,
+      "loss": 1.7045,
+      "step": 2041
+    },
+    {
+      "epoch": 2.237808219178082,
+      "grad_norm": 2.691631555557251,
+      "learning_rate": 1.4419171405361497e-06,
+      "loss": 1.7063,
+      "step": 2042
+    },
+    {
+      "epoch": 2.238904109589041,
+      "grad_norm": 3.021667003631592,
+      "learning_rate": 1.439886271324127e-06,
+      "loss": 1.7607,
+      "step": 2043
+    },
+    {
+      "epoch": 2.24,
+      "grad_norm": 2.690802812576294,
+      "learning_rate": 1.4378554021121042e-06,
+      "loss": 1.7594,
+      "step": 2044
+    },
+    {
+      "epoch": 2.241095890410959,
+      "grad_norm": 2.5371177196502686,
+      "learning_rate": 1.4358245329000815e-06,
+      "loss": 1.7177,
+      "step": 2045
+    },
+    {
+      "epoch": 2.242191780821918,
+      "grad_norm": 2.5586729049682617,
+      "learning_rate": 1.4337936636880588e-06,
+      "loss": 1.7864,
+      "step": 2046
+    },
+    {
+      "epoch": 2.243287671232877,
+      "grad_norm": 2.6621484756469727,
+      "learning_rate": 1.431762794476036e-06,
+      "loss": 1.7616,
+      "step": 2047
+    },
+    {
+      "epoch": 2.2443835616438355,
+      "grad_norm": 2.787508010864258,
+      "learning_rate": 1.4297319252640133e-06,
+      "loss": 1.7056,
+      "step": 2048
+    },
+    {
+      "epoch": 2.2454794520547945,
+      "grad_norm": 2.6896064281463623,
+      "learning_rate": 1.4277010560519906e-06,
+      "loss": 1.7413,
+      "step": 2049
+    },
+    {
+      "epoch": 2.2465753424657535,
+      "grad_norm": 2.540524959564209,
+      "learning_rate": 1.4256701868399674e-06,
+      "loss": 1.7096,
+      "step": 2050
+    },
+    {
+      "epoch": 2.2465753424657535,
+      "eval_loss": 1.9832353591918945,
+      "eval_runtime": 37.1202,
+      "eval_samples_per_second": 26.94,
+      "eval_steps_per_second": 1.131,
+      "step": 2050
+    },
+    {
+      "epoch": 2.247671232876712,
+      "grad_norm": 2.582171678543091,
+      "learning_rate": 1.4236393176279447e-06,
+      "loss": 1.8262,
+      "step": 2051
+    },
+    {
+      "epoch": 2.248767123287671,
+      "grad_norm": 2.659044027328491,
+      "learning_rate": 1.421608448415922e-06,
+      "loss": 1.7458,
+      "step": 2052
+    },
+    {
+      "epoch": 2.24986301369863,
+      "grad_norm": 2.7265775203704834,
+      "learning_rate": 1.4195775792038993e-06,
+      "loss": 1.8341,
+      "step": 2053
+    },
+    {
+      "epoch": 2.250958904109589,
+      "grad_norm": 2.922330379486084,
+      "learning_rate": 1.4175467099918765e-06,
+      "loss": 1.7103,
+      "step": 2054
+    },
+    {
+      "epoch": 2.252054794520548,
+      "grad_norm": 2.4665236473083496,
+      "learning_rate": 1.4155158407798538e-06,
+      "loss": 1.7461,
+      "step": 2055
+    },
+    {
+      "epoch": 2.253150684931507,
+      "grad_norm": 2.639538049697876,
+      "learning_rate": 1.413484971567831e-06,
+      "loss": 1.7683,
+      "step": 2056
+    },
+    {
+      "epoch": 2.254246575342466,
+      "grad_norm": 2.558363676071167,
+      "learning_rate": 1.4114541023558084e-06,
+      "loss": 1.6821,
+      "step": 2057
+    },
+    {
+      "epoch": 2.255342465753425,
+      "grad_norm": 2.6308321952819824,
+      "learning_rate": 1.4094232331437856e-06,
+      "loss": 1.7226,
+      "step": 2058
+    },
+    {
+      "epoch": 2.2564383561643835,
+      "grad_norm": 2.6325299739837646,
+      "learning_rate": 1.407392363931763e-06,
+      "loss": 1.783,
+      "step": 2059
+    },
+    {
+      "epoch": 2.2575342465753425,
+      "grad_norm": 2.6024723052978516,
+      "learning_rate": 1.4053614947197402e-06,
+      "loss": 1.8402,
+      "step": 2060
+    },
+    {
+      "epoch": 2.2575342465753425,
+      "eval_loss": 1.983332633972168,
+      "eval_runtime": 37.1133,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 2060
+    },
+    {
+      "epoch": 2.2586301369863016,
+      "grad_norm": 2.548522710800171,
+      "learning_rate": 1.4033306255077175e-06,
+      "loss": 1.7625,
+      "step": 2061
+    },
+    {
+      "epoch": 2.25972602739726,
+      "grad_norm": 2.5953831672668457,
+      "learning_rate": 1.4012997562956945e-06,
+      "loss": 1.6632,
+      "step": 2062
+    },
+    {
+      "epoch": 2.260821917808219,
+      "grad_norm": 2.5880215167999268,
+      "learning_rate": 1.3992688870836718e-06,
+      "loss": 1.7957,
+      "step": 2063
+    },
+    {
+      "epoch": 2.261917808219178,
+      "grad_norm": 2.439326047897339,
+      "learning_rate": 1.397238017871649e-06,
+      "loss": 1.76,
+      "step": 2064
+    },
+    {
+      "epoch": 2.263013698630137,
+      "grad_norm": 2.5185437202453613,
+      "learning_rate": 1.3952071486596264e-06,
+      "loss": 1.7385,
+      "step": 2065
+    },
+    {
+      "epoch": 2.264109589041096,
+      "grad_norm": 2.5943732261657715,
+      "learning_rate": 1.3931762794476036e-06,
+      "loss": 1.7066,
+      "step": 2066
+    },
+    {
+      "epoch": 2.265205479452055,
+      "grad_norm": 2.6375551223754883,
+      "learning_rate": 1.391145410235581e-06,
+      "loss": 1.7884,
+      "step": 2067
+    },
+    {
+      "epoch": 2.266301369863014,
+      "grad_norm": 2.4962515830993652,
+      "learning_rate": 1.3891145410235582e-06,
+      "loss": 1.7295,
+      "step": 2068
+    },
+    {
+      "epoch": 2.2673972602739725,
+      "grad_norm": 2.5873053073883057,
+      "learning_rate": 1.3870836718115355e-06,
+      "loss": 1.6697,
+      "step": 2069
+    },
+    {
+      "epoch": 2.2684931506849315,
+      "grad_norm": 2.573211908340454,
+      "learning_rate": 1.3850528025995127e-06,
+      "loss": 1.7198,
+      "step": 2070
+    },
+    {
+      "epoch": 2.2684931506849315,
+      "eval_loss": 1.9826557636260986,
+      "eval_runtime": 37.1271,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 2070
+    },
+    {
+      "epoch": 2.2695890410958905,
+      "grad_norm": 2.68015718460083,
+      "learning_rate": 1.38302193338749e-06,
+      "loss": 1.7294,
+      "step": 2071
+    },
+    {
+      "epoch": 2.270684931506849,
+      "grad_norm": 2.4882888793945312,
+      "learning_rate": 1.3809910641754673e-06,
+      "loss": 1.6592,
+      "step": 2072
+    },
+    {
+      "epoch": 2.271780821917808,
+      "grad_norm": 2.587048053741455,
+      "learning_rate": 1.3789601949634446e-06,
+      "loss": 1.7605,
+      "step": 2073
+    },
+    {
+      "epoch": 2.272876712328767,
+      "grad_norm": 2.6452062129974365,
+      "learning_rate": 1.3769293257514218e-06,
+      "loss": 1.6664,
+      "step": 2074
+    },
+    {
+      "epoch": 2.2739726027397262,
+      "grad_norm": 2.5663232803344727,
+      "learning_rate": 1.374898456539399e-06,
+      "loss": 1.771,
+      "step": 2075
+    },
+    {
+      "epoch": 2.275068493150685,
+      "grad_norm": 2.652989149093628,
+      "learning_rate": 1.3728675873273762e-06,
+      "loss": 1.7096,
+      "step": 2076
+    },
+    {
+      "epoch": 2.276164383561644,
+      "grad_norm": 2.538435220718384,
+      "learning_rate": 1.3708367181153535e-06,
+      "loss": 1.798,
+      "step": 2077
+    },
+    {
+      "epoch": 2.277260273972603,
+      "grad_norm": 2.4685096740722656,
+      "learning_rate": 1.3688058489033307e-06,
+      "loss": 1.7904,
+      "step": 2078
+    },
+    {
+      "epoch": 2.2783561643835615,
+      "grad_norm": 2.566723108291626,
+      "learning_rate": 1.366774979691308e-06,
+      "loss": 1.7569,
+      "step": 2079
+    },
+    {
+      "epoch": 2.2794520547945205,
+      "grad_norm": 2.560960531234741,
+      "learning_rate": 1.3647441104792853e-06,
+      "loss": 1.7734,
+      "step": 2080
+    },
+    {
+      "epoch": 2.2794520547945205,
+      "eval_loss": 1.9809880256652832,
+      "eval_runtime": 37.1137,
+      "eval_samples_per_second": 26.944,
+      "eval_steps_per_second": 1.132,
+      "step": 2080
+    },
+    {
+      "epoch": 2.2805479452054795,
+      "grad_norm": 2.576608419418335,
+      "learning_rate": 1.3627132412672626e-06,
+      "loss": 1.7431,
+      "step": 2081
+    },
+    {
+      "epoch": 2.281643835616438,
+      "grad_norm": 2.648301601409912,
+      "learning_rate": 1.3606823720552398e-06,
+      "loss": 1.6959,
+      "step": 2082
+    },
+    {
+      "epoch": 2.282739726027397,
+      "grad_norm": 2.516493082046509,
+      "learning_rate": 1.3586515028432171e-06,
+      "loss": 1.7519,
+      "step": 2083
+    },
+    {
+      "epoch": 2.283835616438356,
+      "grad_norm": 2.474616050720215,
+      "learning_rate": 1.3566206336311944e-06,
+      "loss": 1.6796,
+      "step": 2084
+    },
+    {
+      "epoch": 2.2849315068493152,
+      "grad_norm": 2.482044219970703,
+      "learning_rate": 1.3545897644191717e-06,
+      "loss": 1.7443,
+      "step": 2085
+    },
+    {
+      "epoch": 2.286027397260274,
+      "grad_norm": 2.6574456691741943,
+      "learning_rate": 1.352558895207149e-06,
+      "loss": 1.7145,
+      "step": 2086
+    },
+    {
+      "epoch": 2.287123287671233,
+      "grad_norm": 2.5350115299224854,
+      "learning_rate": 1.3505280259951258e-06,
+      "loss": 1.765,
+      "step": 2087
+    },
+    {
+      "epoch": 2.288219178082192,
+      "grad_norm": 2.6488730907440186,
+      "learning_rate": 1.348497156783103e-06,
+      "loss": 1.7407,
+      "step": 2088
+    },
+    {
+      "epoch": 2.289315068493151,
+      "grad_norm": 2.6020500659942627,
+      "learning_rate": 1.3464662875710804e-06,
+      "loss": 1.779,
+      "step": 2089
+    },
+    {
+      "epoch": 2.2904109589041095,
+      "grad_norm": 2.633013963699341,
+      "learning_rate": 1.3444354183590576e-06,
+      "loss": 1.6824,
+      "step": 2090
+    },
+    {
+      "epoch": 2.2904109589041095,
+      "eval_loss": 1.982380986213684,
+      "eval_runtime": 37.1166,
+      "eval_samples_per_second": 26.942,
+      "eval_steps_per_second": 1.132,
+      "step": 2090
+    },
+    {
+      "epoch": 2.2915068493150685,
+      "grad_norm": 2.6399149894714355,
+      "learning_rate": 1.342404549147035e-06,
+      "loss": 1.7634,
+      "step": 2091
+    },
+    {
+      "epoch": 2.2926027397260276,
+      "grad_norm": 2.580634832382202,
+      "learning_rate": 1.3403736799350122e-06,
+      "loss": 1.7099,
+      "step": 2092
+    },
+    {
+      "epoch": 2.293698630136986,
+      "grad_norm": 2.523400068283081,
+      "learning_rate": 1.3383428107229895e-06,
+      "loss": 1.7705,
+      "step": 2093
+    },
+    {
+      "epoch": 2.294794520547945,
+      "grad_norm": 2.5114550590515137,
+      "learning_rate": 1.3363119415109667e-06,
+      "loss": 1.7961,
+      "step": 2094
+    },
+    {
+      "epoch": 2.2958904109589042,
+      "grad_norm": 2.457369804382324,
+      "learning_rate": 1.334281072298944e-06,
+      "loss": 1.7507,
+      "step": 2095
+    },
+    {
+      "epoch": 2.296986301369863,
+      "grad_norm": 2.5671467781066895,
+      "learning_rate": 1.3322502030869213e-06,
+      "loss": 1.7607,
+      "step": 2096
+    },
+    {
+      "epoch": 2.298082191780822,
+      "grad_norm": 2.459540367126465,
+      "learning_rate": 1.3302193338748986e-06,
+      "loss": 1.7193,
+      "step": 2097
+    },
+    {
+      "epoch": 2.299178082191781,
+      "grad_norm": 2.4934308528900146,
+      "learning_rate": 1.3281884646628758e-06,
+      "loss": 1.7775,
+      "step": 2098
+    },
+    {
+      "epoch": 2.30027397260274,
+      "grad_norm": 2.520641565322876,
+      "learning_rate": 1.3261575954508531e-06,
+      "loss": 1.7847,
+      "step": 2099
+    },
+    {
+      "epoch": 2.3013698630136985,
+      "grad_norm": 2.4607160091400146,
+      "learning_rate": 1.3241267262388302e-06,
+      "loss": 1.7812,
+      "step": 2100
+    },
+    {
+      "epoch": 2.3013698630136985,
+      "eval_loss": 1.9836152791976929,
+      "eval_runtime": 37.1239,
+      "eval_samples_per_second": 26.937,
+      "eval_steps_per_second": 1.131,
+      "step": 2100
+    },
+    {
+      "epoch": 2.3024657534246575,
+      "grad_norm": 2.577761650085449,
+      "learning_rate": 1.3220958570268074e-06,
+      "loss": 1.7452,
+      "step": 2101
+    },
+    {
+      "epoch": 2.3035616438356166,
+      "grad_norm": 2.5331356525421143,
+      "learning_rate": 1.3200649878147847e-06,
+      "loss": 1.6989,
+      "step": 2102
+    },
+    {
+      "epoch": 2.304657534246575,
+      "grad_norm": 2.632581949234009,
+      "learning_rate": 1.318034118602762e-06,
+      "loss": 1.7368,
+      "step": 2103
+    },
+    {
+      "epoch": 2.305753424657534,
+      "grad_norm": 2.55940318107605,
+      "learning_rate": 1.3160032493907393e-06,
+      "loss": 1.7806,
+      "step": 2104
+    },
+    {
+      "epoch": 2.3068493150684932,
+      "grad_norm": 2.5680153369903564,
+      "learning_rate": 1.3139723801787166e-06,
+      "loss": 1.7482,
+      "step": 2105
+    },
+    {
+      "epoch": 2.3079452054794523,
+      "grad_norm": 2.7915756702423096,
+      "learning_rate": 1.3119415109666938e-06,
+      "loss": 1.832,
+      "step": 2106
+    },
+    {
+      "epoch": 2.309041095890411,
+      "grad_norm": 2.518031597137451,
+      "learning_rate": 1.309910641754671e-06,
+      "loss": 1.7401,
+      "step": 2107
+    },
+    {
+      "epoch": 2.31013698630137,
+      "grad_norm": 2.7410221099853516,
+      "learning_rate": 1.3078797725426484e-06,
+      "loss": 1.6992,
+      "step": 2108
+    },
+    {
+      "epoch": 2.311232876712329,
+      "grad_norm": 2.611825466156006,
+      "learning_rate": 1.3058489033306257e-06,
+      "loss": 1.7085,
+      "step": 2109
+    },
+    {
+      "epoch": 2.3123287671232875,
+      "grad_norm": 2.5833330154418945,
+      "learning_rate": 1.303818034118603e-06,
+      "loss": 1.7129,
+      "step": 2110
+    },
+    {
+      "epoch": 2.3123287671232875,
+      "eval_loss": 1.9824191331863403,
+      "eval_runtime": 36.773,
+      "eval_samples_per_second": 27.194,
+      "eval_steps_per_second": 1.142,
+      "step": 2110
+    },
+    {
+      "epoch": 2.3134246575342465,
+      "grad_norm": 2.587327003479004,
+      "learning_rate": 1.3017871649065802e-06,
+      "loss": 1.7439,
+      "step": 2111
+    },
+    {
+      "epoch": 2.3145205479452056,
+      "grad_norm": 2.574925422668457,
+      "learning_rate": 1.2997562956945573e-06,
+      "loss": 1.7173,
+      "step": 2112
+    },
+    {
+      "epoch": 2.315616438356164,
+      "grad_norm": 2.6404287815093994,
+      "learning_rate": 1.2977254264825345e-06,
+      "loss": 1.7052,
+      "step": 2113
+    },
+    {
+      "epoch": 2.316712328767123,
+      "grad_norm": 2.634584903717041,
+      "learning_rate": 1.2956945572705118e-06,
+      "loss": 1.741,
+      "step": 2114
+    },
+    {
+      "epoch": 2.317808219178082,
+      "grad_norm": 2.66159987449646,
+      "learning_rate": 1.293663688058489e-06,
+      "loss": 1.76,
+      "step": 2115
+    },
+    {
+      "epoch": 2.3189041095890413,
+      "grad_norm": 2.6337332725524902,
+      "learning_rate": 1.2916328188464664e-06,
+      "loss": 1.7035,
+      "step": 2116
+    },
+    {
+      "epoch": 2.32,
+      "grad_norm": 2.528524398803711,
+      "learning_rate": 1.2896019496344436e-06,
+      "loss": 1.6439,
+      "step": 2117
+    },
+    {
+      "epoch": 2.321095890410959,
+      "grad_norm": 2.5938384532928467,
+      "learning_rate": 1.287571080422421e-06,
+      "loss": 1.7134,
+      "step": 2118
+    },
+    {
+      "epoch": 2.322191780821918,
+      "grad_norm": 2.606271505355835,
+      "learning_rate": 1.2855402112103982e-06,
+      "loss": 1.7758,
+      "step": 2119
+    },
+    {
+      "epoch": 2.323287671232877,
+      "grad_norm": 2.530071258544922,
+      "learning_rate": 1.2835093419983755e-06,
+      "loss": 1.6927,
+      "step": 2120
+    },
+    {
+      "epoch": 2.323287671232877,
+      "eval_loss": 1.9856306314468384,
+      "eval_runtime": 36.9674,
+      "eval_samples_per_second": 27.051,
+      "eval_steps_per_second": 1.136,
+      "step": 2120
+    },
+    {
+      "epoch": 2.3243835616438355,
+      "grad_norm": 2.8125321865081787,
+      "learning_rate": 1.2814784727863528e-06,
+      "loss": 1.7225,
+      "step": 2121
+    },
+    {
+      "epoch": 2.3254794520547946,
+      "grad_norm": 2.57012939453125,
+      "learning_rate": 1.27944760357433e-06,
+      "loss": 1.7177,
+      "step": 2122
+    },
+    {
+      "epoch": 2.3265753424657536,
+      "grad_norm": 2.6595780849456787,
+      "learning_rate": 1.2774167343623073e-06,
+      "loss": 1.7807,
+      "step": 2123
+    },
+    {
+      "epoch": 2.327671232876712,
+      "grad_norm": 2.557115077972412,
+      "learning_rate": 1.2753858651502846e-06,
+      "loss": 1.7762,
+      "step": 2124
+    },
+    {
+      "epoch": 2.328767123287671,
+      "grad_norm": 2.763218879699707,
+      "learning_rate": 1.2733549959382616e-06,
+      "loss": 1.7638,
+      "step": 2125
+    },
+    {
+      "epoch": 2.3298630136986302,
+      "grad_norm": 2.7284505367279053,
+      "learning_rate": 1.271324126726239e-06,
+      "loss": 1.7387,
+      "step": 2126
+    },
+    {
+      "epoch": 2.330958904109589,
+      "grad_norm": 2.5893704891204834,
+      "learning_rate": 1.2692932575142162e-06,
+      "loss": 1.7322,
+      "step": 2127
+    },
+    {
+      "epoch": 2.332054794520548,
+      "grad_norm": 2.625582456588745,
+      "learning_rate": 1.2672623883021935e-06,
+      "loss": 1.6865,
+      "step": 2128
+    },
+    {
+      "epoch": 2.333150684931507,
+      "grad_norm": 2.5458641052246094,
+      "learning_rate": 1.2652315190901707e-06,
+      "loss": 1.6817,
+      "step": 2129
+    },
+    {
+      "epoch": 2.334246575342466,
+      "grad_norm": 2.5718271732330322,
+      "learning_rate": 1.263200649878148e-06,
+      "loss": 1.8076,
+      "step": 2130
+    },
+    {
+      "epoch": 2.334246575342466,
+      "eval_loss": 1.9806523323059082,
+      "eval_runtime": 36.9843,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 2130
+    },
+    {
+      "epoch": 2.3353424657534245,
+      "grad_norm": 2.4864919185638428,
+      "learning_rate": 1.2611697806661253e-06,
+      "loss": 1.7308,
+      "step": 2131
+    },
+    {
+      "epoch": 2.3364383561643836,
+      "grad_norm": 2.6007416248321533,
+      "learning_rate": 1.2591389114541026e-06,
+      "loss": 1.7523,
+      "step": 2132
+    },
+    {
+      "epoch": 2.3375342465753426,
+      "grad_norm": 2.600177049636841,
+      "learning_rate": 1.2571080422420798e-06,
+      "loss": 1.7422,
+      "step": 2133
+    },
+    {
+      "epoch": 2.338630136986301,
+      "grad_norm": 2.489025831222534,
+      "learning_rate": 1.2550771730300571e-06,
+      "loss": 1.7184,
+      "step": 2134
+    },
+    {
+      "epoch": 2.33972602739726,
+      "grad_norm": 2.5677363872528076,
+      "learning_rate": 1.2530463038180344e-06,
+      "loss": 1.7922,
+      "step": 2135
+    },
+    {
+      "epoch": 2.3408219178082192,
+      "grad_norm": 2.616976737976074,
+      "learning_rate": 1.2510154346060117e-06,
+      "loss": 1.7354,
+      "step": 2136
+    },
+    {
+      "epoch": 2.3419178082191783,
+      "grad_norm": 2.646456241607666,
+      "learning_rate": 1.2489845653939887e-06,
+      "loss": 1.7377,
+      "step": 2137
+    },
+    {
+      "epoch": 2.343013698630137,
+      "grad_norm": 2.550020933151245,
+      "learning_rate": 1.246953696181966e-06,
+      "loss": 1.7751,
+      "step": 2138
+    },
+    {
+      "epoch": 2.344109589041096,
+      "grad_norm": 2.634279251098633,
+      "learning_rate": 1.2449228269699433e-06,
+      "loss": 1.8157,
+      "step": 2139
+    },
+    {
+      "epoch": 2.345205479452055,
+      "grad_norm": 2.5621821880340576,
+      "learning_rate": 1.2428919577579204e-06,
+      "loss": 1.7383,
+      "step": 2140
+    },
+    {
+      "epoch": 2.345205479452055,
+      "eval_loss": 1.9809924364089966,
+      "eval_runtime": 37.004,
+      "eval_samples_per_second": 27.024,
+      "eval_steps_per_second": 1.135,
+      "step": 2140
+    },
+    {
+      "epoch": 2.3463013698630135,
+      "grad_norm": 2.638831615447998,
+      "learning_rate": 1.2408610885458976e-06,
+      "loss": 1.6988,
+      "step": 2141
+    },
+    {
+      "epoch": 2.3473972602739726,
+      "grad_norm": 2.634612798690796,
+      "learning_rate": 1.238830219333875e-06,
+      "loss": 1.7331,
+      "step": 2142
+    },
+    {
+      "epoch": 2.3484931506849316,
+      "grad_norm": 2.552955389022827,
+      "learning_rate": 1.2367993501218522e-06,
+      "loss": 1.7148,
+      "step": 2143
+    },
+    {
+      "epoch": 2.34958904109589,
+      "grad_norm": 2.6475753784179688,
+      "learning_rate": 1.2347684809098295e-06,
+      "loss": 1.7815,
+      "step": 2144
+    },
+    {
+      "epoch": 2.350684931506849,
+      "grad_norm": 2.7666079998016357,
+      "learning_rate": 1.2327376116978067e-06,
+      "loss": 1.7634,
+      "step": 2145
+    },
+    {
+      "epoch": 2.3517808219178082,
+      "grad_norm": 2.6810429096221924,
+      "learning_rate": 1.230706742485784e-06,
+      "loss": 1.6769,
+      "step": 2146
+    },
+    {
+      "epoch": 2.3528767123287673,
+      "grad_norm": 2.671382188796997,
+      "learning_rate": 1.2286758732737613e-06,
+      "loss": 1.6884,
+      "step": 2147
+    },
+    {
+      "epoch": 2.353972602739726,
+      "grad_norm": 2.614684820175171,
+      "learning_rate": 1.2266450040617386e-06,
+      "loss": 1.8226,
+      "step": 2148
+    },
+    {
+      "epoch": 2.355068493150685,
+      "grad_norm": 2.47189998626709,
+      "learning_rate": 1.2246141348497158e-06,
+      "loss": 1.819,
+      "step": 2149
+    },
+    {
+      "epoch": 2.356164383561644,
+      "grad_norm": 2.4641449451446533,
+      "learning_rate": 1.2225832656376931e-06,
+      "loss": 1.7494,
+      "step": 2150
+    },
+    {
+      "epoch": 2.356164383561644,
+      "eval_loss": 1.9806426763534546,
+      "eval_runtime": 37.0153,
+      "eval_samples_per_second": 27.016,
+      "eval_steps_per_second": 1.135,
+      "step": 2150
+    },
+    {
+      "epoch": 2.357260273972603,
+      "grad_norm": 2.7310373783111572,
+      "learning_rate": 1.2205523964256704e-06,
+      "loss": 1.7432,
+      "step": 2151
+    },
+    {
+      "epoch": 2.3583561643835615,
+      "grad_norm": 2.5554895401000977,
+      "learning_rate": 1.2185215272136477e-06,
+      "loss": 1.7213,
+      "step": 2152
+    },
+    {
+      "epoch": 2.3594520547945206,
+      "grad_norm": 2.5401952266693115,
+      "learning_rate": 1.2164906580016247e-06,
+      "loss": 1.7469,
+      "step": 2153
+    },
+    {
+      "epoch": 2.3605479452054796,
+      "grad_norm": 2.5213980674743652,
+      "learning_rate": 1.214459788789602e-06,
+      "loss": 1.7459,
+      "step": 2154
+    },
+    {
+      "epoch": 2.361643835616438,
+      "grad_norm": 2.5585155487060547,
+      "learning_rate": 1.2124289195775793e-06,
+      "loss": 1.6587,
+      "step": 2155
+    },
+    {
+      "epoch": 2.3627397260273972,
+      "grad_norm": 2.6329963207244873,
+      "learning_rate": 1.2103980503655566e-06,
+      "loss": 1.6925,
+      "step": 2156
+    },
+    {
+      "epoch": 2.3638356164383563,
+      "grad_norm": 2.678351640701294,
+      "learning_rate": 1.2083671811535338e-06,
+      "loss": 1.7322,
+      "step": 2157
+    },
+    {
+      "epoch": 2.364931506849315,
+      "grad_norm": 2.595313549041748,
+      "learning_rate": 1.2063363119415111e-06,
+      "loss": 1.7142,
+      "step": 2158
+    },
+    {
+      "epoch": 2.366027397260274,
+      "grad_norm": 2.5650861263275146,
+      "learning_rate": 1.2043054427294882e-06,
+      "loss": 1.7024,
+      "step": 2159
+    },
+    {
+      "epoch": 2.367123287671233,
+      "grad_norm": 2.495866537094116,
+      "learning_rate": 1.2022745735174654e-06,
+      "loss": 1.7205,
+      "step": 2160
+    },
+    {
+      "epoch": 2.367123287671233,
+      "eval_loss": 1.976922631263733,
+      "eval_runtime": 37.012,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 2160
+    },
+    {
+      "epoch": 2.368219178082192,
+      "grad_norm": 2.570307493209839,
+      "learning_rate": 1.2002437043054427e-06,
+      "loss": 1.7601,
+      "step": 2161
+    },
+    {
+      "epoch": 2.3693150684931505,
+      "grad_norm": 2.444373846054077,
+      "learning_rate": 1.19821283509342e-06,
+      "loss": 1.7857,
+      "step": 2162
+    },
+    {
+      "epoch": 2.3704109589041096,
+      "grad_norm": 2.525505781173706,
+      "learning_rate": 1.1961819658813973e-06,
+      "loss": 1.7875,
+      "step": 2163
+    },
+    {
+      "epoch": 2.3715068493150686,
+      "grad_norm": 2.622260093688965,
+      "learning_rate": 1.1941510966693746e-06,
+      "loss": 1.7596,
+      "step": 2164
+    },
+    {
+      "epoch": 2.3726027397260276,
+      "grad_norm": 2.542076349258423,
+      "learning_rate": 1.1921202274573518e-06,
+      "loss": 1.7248,
+      "step": 2165
+    },
+    {
+      "epoch": 2.3736986301369862,
+      "grad_norm": 2.510263681411743,
+      "learning_rate": 1.190089358245329e-06,
+      "loss": 1.7168,
+      "step": 2166
+    },
+    {
+      "epoch": 2.3747945205479453,
+      "grad_norm": 2.605023145675659,
+      "learning_rate": 1.1880584890333064e-06,
+      "loss": 1.798,
+      "step": 2167
+    },
+    {
+      "epoch": 2.3758904109589043,
+      "grad_norm": 2.593261480331421,
+      "learning_rate": 1.1860276198212837e-06,
+      "loss": 1.7247,
+      "step": 2168
+    },
+    {
+      "epoch": 2.376986301369863,
+      "grad_norm": 2.6010820865631104,
+      "learning_rate": 1.183996750609261e-06,
+      "loss": 1.7607,
+      "step": 2169
+    },
+    {
+      "epoch": 2.378082191780822,
+      "grad_norm": 2.62886118888855,
+      "learning_rate": 1.1819658813972382e-06,
+      "loss": 1.7641,
+      "step": 2170
+    },
+    {
+      "epoch": 2.378082191780822,
+      "eval_loss": 1.9790476560592651,
+      "eval_runtime": 37.0413,
+      "eval_samples_per_second": 26.997,
+      "eval_steps_per_second": 1.134,
+      "step": 2170
+    },
+    {
+      "epoch": 2.379178082191781,
+      "grad_norm": 2.5506582260131836,
+      "learning_rate": 1.1799350121852153e-06,
+      "loss": 1.7407,
+      "step": 2171
+    },
+    {
+      "epoch": 2.3802739726027395,
+      "grad_norm": 2.547445774078369,
+      "learning_rate": 1.1779041429731925e-06,
+      "loss": 1.7808,
+      "step": 2172
+    },
+    {
+      "epoch": 2.3813698630136986,
+      "grad_norm": 2.5443003177642822,
+      "learning_rate": 1.1758732737611698e-06,
+      "loss": 1.7842,
+      "step": 2173
+    },
+    {
+      "epoch": 2.3824657534246576,
+      "grad_norm": 2.5733375549316406,
+      "learning_rate": 1.173842404549147e-06,
+      "loss": 1.6902,
+      "step": 2174
+    },
+    {
+      "epoch": 2.383561643835616,
+      "grad_norm": 2.63702392578125,
+      "learning_rate": 1.1718115353371244e-06,
+      "loss": 1.8049,
+      "step": 2175
+    },
+    {
+      "epoch": 2.3846575342465752,
+      "grad_norm": 2.5577163696289062,
+      "learning_rate": 1.1697806661251016e-06,
+      "loss": 1.7162,
+      "step": 2176
+    },
+    {
+      "epoch": 2.3857534246575343,
+      "grad_norm": 2.4259274005889893,
+      "learning_rate": 1.167749796913079e-06,
+      "loss": 1.7883,
+      "step": 2177
+    },
+    {
+      "epoch": 2.3868493150684933,
+      "grad_norm": 2.537519931793213,
+      "learning_rate": 1.1657189277010562e-06,
+      "loss": 1.7882,
+      "step": 2178
+    },
+    {
+      "epoch": 2.387945205479452,
+      "grad_norm": 2.5480241775512695,
+      "learning_rate": 1.1636880584890335e-06,
+      "loss": 1.7245,
+      "step": 2179
+    },
+    {
+      "epoch": 2.389041095890411,
+      "grad_norm": 2.630337953567505,
+      "learning_rate": 1.1616571892770108e-06,
+      "loss": 1.7165,
+      "step": 2180
+    },
+    {
+      "epoch": 2.389041095890411,
+      "eval_loss": 1.9772392511367798,
+      "eval_runtime": 37.1283,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 2180
+    },
+    {
+      "epoch": 2.39013698630137,
+      "grad_norm": 2.5741171836853027,
+      "learning_rate": 1.159626320064988e-06,
+      "loss": 1.7806,
+      "step": 2181
+    },
+    {
+      "epoch": 2.391232876712329,
+      "grad_norm": 2.6166305541992188,
+      "learning_rate": 1.1575954508529653e-06,
+      "loss": 1.744,
+      "step": 2182
+    },
+    {
+      "epoch": 2.3923287671232876,
+      "grad_norm": 2.6965837478637695,
+      "learning_rate": 1.1555645816409426e-06,
+      "loss": 1.7736,
+      "step": 2183
+    },
+    {
+      "epoch": 2.3934246575342466,
+      "grad_norm": 2.586827516555786,
+      "learning_rate": 1.1535337124289196e-06,
+      "loss": 1.7455,
+      "step": 2184
+    },
+    {
+      "epoch": 2.3945205479452056,
+      "grad_norm": 2.587235689163208,
+      "learning_rate": 1.151502843216897e-06,
+      "loss": 1.6999,
+      "step": 2185
+    },
+    {
+      "epoch": 2.3956164383561642,
+      "grad_norm": 2.656740665435791,
+      "learning_rate": 1.1494719740048742e-06,
+      "loss": 1.6518,
+      "step": 2186
+    },
+    {
+      "epoch": 2.3967123287671233,
+      "grad_norm": 2.5353455543518066,
+      "learning_rate": 1.1474411047928515e-06,
+      "loss": 1.7389,
+      "step": 2187
+    },
+    {
+      "epoch": 2.3978082191780823,
+      "grad_norm": 2.622314453125,
+      "learning_rate": 1.1454102355808287e-06,
+      "loss": 1.7309,
+      "step": 2188
+    },
+    {
+      "epoch": 2.398904109589041,
+      "grad_norm": 2.6286797523498535,
+      "learning_rate": 1.143379366368806e-06,
+      "loss": 1.7094,
+      "step": 2189
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 2.5203921794891357,
+      "learning_rate": 1.141348497156783e-06,
+      "loss": 1.6905,
+      "step": 2190
+    },
+    {
+      "epoch": 2.4,
+      "eval_loss": 1.9822566509246826,
+      "eval_runtime": 37.0581,
+      "eval_samples_per_second": 26.985,
+      "eval_steps_per_second": 1.133,
+      "step": 2190
+    },
+    {
+      "epoch": 2.401095890410959,
+      "grad_norm": 2.6562373638153076,
+      "learning_rate": 1.1393176279447604e-06,
+      "loss": 1.8293,
+      "step": 2191
+    },
+    {
+      "epoch": 2.402191780821918,
+      "grad_norm": 2.5581510066986084,
+      "learning_rate": 1.1372867587327376e-06,
+      "loss": 1.7936,
+      "step": 2192
+    },
+    {
+      "epoch": 2.4032876712328766,
+      "grad_norm": 2.534675121307373,
+      "learning_rate": 1.135255889520715e-06,
+      "loss": 1.6648,
+      "step": 2193
+    },
+    {
+      "epoch": 2.4043835616438356,
+      "grad_norm": 2.7236168384552,
+      "learning_rate": 1.1332250203086922e-06,
+      "loss": 1.7654,
+      "step": 2194
+    },
+    {
+      "epoch": 2.4054794520547946,
+      "grad_norm": 2.5706138610839844,
+      "learning_rate": 1.1311941510966695e-06,
+      "loss": 1.7427,
+      "step": 2195
+    },
+    {
+      "epoch": 2.4065753424657537,
+      "grad_norm": 2.7030224800109863,
+      "learning_rate": 1.1291632818846467e-06,
+      "loss": 1.7216,
+      "step": 2196
+    },
+    {
+      "epoch": 2.4076712328767123,
+      "grad_norm": 2.4936161041259766,
+      "learning_rate": 1.127132412672624e-06,
+      "loss": 1.7033,
+      "step": 2197
+    },
+    {
+      "epoch": 2.4087671232876713,
+      "grad_norm": 2.5384793281555176,
+      "learning_rate": 1.1251015434606013e-06,
+      "loss": 1.7023,
+      "step": 2198
+    },
+    {
+      "epoch": 2.4098630136986303,
+      "grad_norm": 2.5271613597869873,
+      "learning_rate": 1.1230706742485786e-06,
+      "loss": 1.738,
+      "step": 2199
+    },
+    {
+      "epoch": 2.410958904109589,
+      "grad_norm": 2.557610034942627,
+      "learning_rate": 1.1210398050365558e-06,
+      "loss": 1.7927,
+      "step": 2200
+    },
+    {
+      "epoch": 2.410958904109589,
+      "eval_loss": 1.9795563220977783,
+      "eval_runtime": 37.0832,
+      "eval_samples_per_second": 26.966,
+      "eval_steps_per_second": 1.133,
+      "step": 2200
+    },
+    {
+      "epoch": 2.412054794520548,
+      "grad_norm": 2.5497145652770996,
+      "learning_rate": 1.1190089358245331e-06,
+      "loss": 1.7361,
+      "step": 2201
+    },
+    {
+      "epoch": 2.413150684931507,
+      "grad_norm": 2.622793436050415,
+      "learning_rate": 1.1169780666125102e-06,
+      "loss": 1.6624,
+      "step": 2202
+    },
+    {
+      "epoch": 2.4142465753424656,
+      "grad_norm": 2.694563865661621,
+      "learning_rate": 1.1149471974004875e-06,
+      "loss": 1.7755,
+      "step": 2203
+    },
+    {
+      "epoch": 2.4153424657534246,
+      "grad_norm": 2.68461012840271,
+      "learning_rate": 1.1129163281884647e-06,
+      "loss": 1.7132,
+      "step": 2204
+    },
+    {
+      "epoch": 2.4164383561643836,
+      "grad_norm": 2.65791654586792,
+      "learning_rate": 1.110885458976442e-06,
+      "loss": 1.7728,
+      "step": 2205
+    },
+    {
+      "epoch": 2.417534246575342,
+      "grad_norm": 2.46262788772583,
+      "learning_rate": 1.1088545897644193e-06,
+      "loss": 1.6973,
+      "step": 2206
+    },
+    {
+      "epoch": 2.4186301369863012,
+      "grad_norm": 2.5602169036865234,
+      "learning_rate": 1.1068237205523966e-06,
+      "loss": 1.7651,
+      "step": 2207
+    },
+    {
+      "epoch": 2.4197260273972603,
+      "grad_norm": 2.5492701530456543,
+      "learning_rate": 1.1047928513403736e-06,
+      "loss": 1.7674,
+      "step": 2208
+    },
+    {
+      "epoch": 2.4208219178082193,
+      "grad_norm": 2.798940420150757,
+      "learning_rate": 1.102761982128351e-06,
+      "loss": 1.7719,
+      "step": 2209
+    },
+    {
+      "epoch": 2.421917808219178,
+      "grad_norm": 2.6989545822143555,
+      "learning_rate": 1.1007311129163282e-06,
+      "loss": 1.6821,
+      "step": 2210
+    },
+    {
+      "epoch": 2.421917808219178,
+      "eval_loss": 1.9763859510421753,
+      "eval_runtime": 36.7821,
+      "eval_samples_per_second": 27.187,
+      "eval_steps_per_second": 1.142,
+      "step": 2210
+    },
+    {
+      "epoch": 2.423013698630137,
+      "grad_norm": 2.752720594406128,
+      "learning_rate": 1.0987002437043055e-06,
+      "loss": 1.7432,
+      "step": 2211
+    },
+    {
+      "epoch": 2.424109589041096,
+      "grad_norm": 2.6162850856781006,
+      "learning_rate": 1.0966693744922827e-06,
+      "loss": 1.718,
+      "step": 2212
+    },
+    {
+      "epoch": 2.425205479452055,
+      "grad_norm": 2.938995838165283,
+      "learning_rate": 1.09463850528026e-06,
+      "loss": 1.6761,
+      "step": 2213
+    },
+    {
+      "epoch": 2.4263013698630136,
+      "grad_norm": 2.7038822174072266,
+      "learning_rate": 1.0926076360682373e-06,
+      "loss": 1.7267,
+      "step": 2214
+    },
+    {
+      "epoch": 2.4273972602739726,
+      "grad_norm": 2.6025028228759766,
+      "learning_rate": 1.0905767668562146e-06,
+      "loss": 1.7021,
+      "step": 2215
+    },
+    {
+      "epoch": 2.4284931506849317,
+      "grad_norm": 2.5922977924346924,
+      "learning_rate": 1.0885458976441918e-06,
+      "loss": 1.7221,
+      "step": 2216
+    },
+    {
+      "epoch": 2.4295890410958902,
+      "grad_norm": 2.634716033935547,
+      "learning_rate": 1.0865150284321691e-06,
+      "loss": 1.8047,
+      "step": 2217
+    },
+    {
+      "epoch": 2.4306849315068493,
+      "grad_norm": 2.655155658721924,
+      "learning_rate": 1.0844841592201464e-06,
+      "loss": 1.7426,
+      "step": 2218
+    },
+    {
+      "epoch": 2.4317808219178083,
+      "grad_norm": 2.5151286125183105,
+      "learning_rate": 1.0824532900081237e-06,
+      "loss": 1.7627,
+      "step": 2219
+    },
+    {
+      "epoch": 2.432876712328767,
+      "grad_norm": 2.696597099304199,
+      "learning_rate": 1.080422420796101e-06,
+      "loss": 1.7325,
+      "step": 2220
+    },
+    {
+      "epoch": 2.432876712328767,
+      "eval_loss": 1.9818732738494873,
+      "eval_runtime": 36.9854,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 2220
+    },
+    {
+      "epoch": 2.433972602739726,
+      "grad_norm": 2.7444307804107666,
+      "learning_rate": 1.078391551584078e-06,
+      "loss": 1.743,
+      "step": 2221
+    },
+    {
+      "epoch": 2.435068493150685,
+      "grad_norm": 2.503706455230713,
+      "learning_rate": 1.0763606823720553e-06,
+      "loss": 1.7699,
+      "step": 2222
+    },
+    {
+      "epoch": 2.436164383561644,
+      "grad_norm": 2.546041965484619,
+      "learning_rate": 1.0743298131600326e-06,
+      "loss": 1.8077,
+      "step": 2223
+    },
+    {
+      "epoch": 2.4372602739726026,
+      "grad_norm": 2.5515291690826416,
+      "learning_rate": 1.0722989439480098e-06,
+      "loss": 1.7637,
+      "step": 2224
+    },
+    {
+      "epoch": 2.4383561643835616,
+      "grad_norm": 2.564180850982666,
+      "learning_rate": 1.070268074735987e-06,
+      "loss": 1.7299,
+      "step": 2225
+    },
+    {
+      "epoch": 2.4394520547945207,
+      "grad_norm": 2.532708168029785,
+      "learning_rate": 1.0682372055239644e-06,
+      "loss": 1.713,
+      "step": 2226
+    },
+    {
+      "epoch": 2.4405479452054797,
+      "grad_norm": 2.672358512878418,
+      "learning_rate": 1.0662063363119414e-06,
+      "loss": 1.7103,
+      "step": 2227
+    },
+    {
+      "epoch": 2.4416438356164383,
+      "grad_norm": 2.7836480140686035,
+      "learning_rate": 1.0641754670999187e-06,
+      "loss": 1.7129,
+      "step": 2228
+    },
+    {
+      "epoch": 2.4427397260273973,
+      "grad_norm": 2.6253440380096436,
+      "learning_rate": 1.062144597887896e-06,
+      "loss": 1.7389,
+      "step": 2229
+    },
+    {
+      "epoch": 2.4438356164383563,
+      "grad_norm": 2.5863685607910156,
+      "learning_rate": 1.0601137286758733e-06,
+      "loss": 1.7293,
+      "step": 2230
+    },
+    {
+      "epoch": 2.4438356164383563,
+      "eval_loss": 1.9771078824996948,
+      "eval_runtime": 37.0603,
+      "eval_samples_per_second": 26.983,
+      "eval_steps_per_second": 1.133,
+      "step": 2230
+    },
+    {
+      "epoch": 2.444931506849315,
+      "grad_norm": 2.732042074203491,
+      "learning_rate": 1.0580828594638505e-06,
+      "loss": 1.8012,
+      "step": 2231
+    },
+    {
+      "epoch": 2.446027397260274,
+      "grad_norm": 2.5275564193725586,
+      "learning_rate": 1.0560519902518278e-06,
+      "loss": 1.6482,
+      "step": 2232
+    },
+    {
+      "epoch": 2.447123287671233,
+      "grad_norm": 2.7054405212402344,
+      "learning_rate": 1.054021121039805e-06,
+      "loss": 1.7799,
+      "step": 2233
+    },
+    {
+      "epoch": 2.4482191780821916,
+      "grad_norm": 2.587719678878784,
+      "learning_rate": 1.0519902518277824e-06,
+      "loss": 1.6797,
+      "step": 2234
+    },
+    {
+      "epoch": 2.4493150684931506,
+      "grad_norm": 2.600907564163208,
+      "learning_rate": 1.0499593826157596e-06,
+      "loss": 1.6727,
+      "step": 2235
+    },
+    {
+      "epoch": 2.4504109589041096,
+      "grad_norm": 2.6509485244750977,
+      "learning_rate": 1.047928513403737e-06,
+      "loss": 1.7535,
+      "step": 2236
+    },
+    {
+      "epoch": 2.4515068493150687,
+      "grad_norm": 2.6827805042266846,
+      "learning_rate": 1.0458976441917142e-06,
+      "loss": 1.7285,
+      "step": 2237
+    },
+    {
+      "epoch": 2.4526027397260273,
+      "grad_norm": 2.7095749378204346,
+      "learning_rate": 1.0438667749796915e-06,
+      "loss": 1.6546,
+      "step": 2238
+    },
+    {
+      "epoch": 2.4536986301369863,
+      "grad_norm": 2.7757186889648438,
+      "learning_rate": 1.0418359057676688e-06,
+      "loss": 1.6659,
+      "step": 2239
+    },
+    {
+      "epoch": 2.4547945205479453,
+      "grad_norm": 2.7696826457977295,
+      "learning_rate": 1.0398050365556458e-06,
+      "loss": 1.7108,
+      "step": 2240
+    },
+    {
+      "epoch": 2.4547945205479453,
+      "eval_loss": 1.980337142944336,
+      "eval_runtime": 37.0883,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 2240
+    },
+    {
+      "epoch": 2.455890410958904,
+      "grad_norm": 2.96648907661438,
+      "learning_rate": 1.037774167343623e-06,
+      "loss": 1.77,
+      "step": 2241
+    },
+    {
+      "epoch": 2.456986301369863,
+      "grad_norm": 2.561461925506592,
+      "learning_rate": 1.0357432981316004e-06,
+      "loss": 1.6962,
+      "step": 2242
+    },
+    {
+      "epoch": 2.458082191780822,
+      "grad_norm": 2.5870988368988037,
+      "learning_rate": 1.0337124289195776e-06,
+      "loss": 1.7385,
+      "step": 2243
+    },
+    {
+      "epoch": 2.459178082191781,
+      "grad_norm": 2.607558488845825,
+      "learning_rate": 1.031681559707555e-06,
+      "loss": 1.7771,
+      "step": 2244
+    },
+    {
+      "epoch": 2.4602739726027396,
+      "grad_norm": 2.6373603343963623,
+      "learning_rate": 1.0296506904955322e-06,
+      "loss": 1.7845,
+      "step": 2245
+    },
+    {
+      "epoch": 2.4613698630136986,
+      "grad_norm": 2.5508599281311035,
+      "learning_rate": 1.0276198212835095e-06,
+      "loss": 1.7384,
+      "step": 2246
+    },
+    {
+      "epoch": 2.4624657534246577,
+      "grad_norm": 2.730696201324463,
+      "learning_rate": 1.0255889520714867e-06,
+      "loss": 1.7458,
+      "step": 2247
+    },
+    {
+      "epoch": 2.4635616438356163,
+      "grad_norm": 2.4905753135681152,
+      "learning_rate": 1.023558082859464e-06,
+      "loss": 1.7195,
+      "step": 2248
+    },
+    {
+      "epoch": 2.4646575342465753,
+      "grad_norm": 2.664184331893921,
+      "learning_rate": 1.0215272136474413e-06,
+      "loss": 1.6986,
+      "step": 2249
+    },
+    {
+      "epoch": 2.4657534246575343,
+      "grad_norm": 2.683645009994507,
+      "learning_rate": 1.0194963444354186e-06,
+      "loss": 1.7981,
+      "step": 2250
+    },
+    {
+      "epoch": 2.4657534246575343,
+      "eval_loss": 1.9768283367156982,
+      "eval_runtime": 37.1086,
+      "eval_samples_per_second": 26.948,
+      "eval_steps_per_second": 1.132,
+      "step": 2250
+    },
+    {
+      "epoch": 2.466849315068493,
+      "grad_norm": 2.633633852005005,
+      "learning_rate": 1.0174654752233958e-06,
+      "loss": 1.6807,
+      "step": 2251
+    },
+    {
+      "epoch": 2.467945205479452,
+      "grad_norm": 2.655794382095337,
+      "learning_rate": 1.015434606011373e-06,
+      "loss": 1.7705,
+      "step": 2252
+    },
+    {
+      "epoch": 2.469041095890411,
+      "grad_norm": 2.6457812786102295,
+      "learning_rate": 1.0134037367993502e-06,
+      "loss": 1.7201,
+      "step": 2253
+    },
+    {
+      "epoch": 2.47013698630137,
+      "grad_norm": 2.623656749725342,
+      "learning_rate": 1.0113728675873275e-06,
+      "loss": 1.7402,
+      "step": 2254
+    },
+    {
+      "epoch": 2.4712328767123286,
+      "grad_norm": 2.65552020072937,
+      "learning_rate": 1.0093419983753047e-06,
+      "loss": 1.7372,
+      "step": 2255
+    },
+    {
+      "epoch": 2.4723287671232876,
+      "grad_norm": 2.707531213760376,
+      "learning_rate": 1.007311129163282e-06,
+      "loss": 1.7299,
+      "step": 2256
+    },
+    {
+      "epoch": 2.4734246575342467,
+      "grad_norm": 2.504805564880371,
+      "learning_rate": 1.0052802599512593e-06,
+      "loss": 1.6858,
+      "step": 2257
+    },
+    {
+      "epoch": 2.4745205479452057,
+      "grad_norm": 2.584251642227173,
+      "learning_rate": 1.0032493907392364e-06,
+      "loss": 1.6867,
+      "step": 2258
+    },
+    {
+      "epoch": 2.4756164383561643,
+      "grad_norm": 2.5356390476226807,
+      "learning_rate": 1.0012185215272136e-06,
+      "loss": 1.7442,
+      "step": 2259
+    },
+    {
+      "epoch": 2.4767123287671233,
+      "grad_norm": 2.5929319858551025,
+      "learning_rate": 9.99187652315191e-07,
+      "loss": 1.6777,
+      "step": 2260
+    },
+    {
+      "epoch": 2.4767123287671233,
+      "eval_loss": 1.9772632122039795,
+      "eval_runtime": 37.0951,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 2260
+    },
+    {
+      "epoch": 2.4778082191780824,
+      "grad_norm": 2.6396026611328125,
+      "learning_rate": 9.971567831031682e-07,
+      "loss": 1.7638,
+      "step": 2261
+    },
+    {
+      "epoch": 2.478904109589041,
+      "grad_norm": 2.5710740089416504,
+      "learning_rate": 9.951259138911455e-07,
+      "loss": 1.7566,
+      "step": 2262
+    },
+    {
+      "epoch": 2.48,
+      "grad_norm": 2.5611767768859863,
+      "learning_rate": 9.930950446791227e-07,
+      "loss": 1.6639,
+      "step": 2263
+    },
+    {
+      "epoch": 2.481095890410959,
+      "grad_norm": 2.5390372276306152,
+      "learning_rate": 9.910641754671e-07,
+      "loss": 1.7887,
+      "step": 2264
+    },
+    {
+      "epoch": 2.4821917808219176,
+      "grad_norm": 2.6124560832977295,
+      "learning_rate": 9.890333062550773e-07,
+      "loss": 1.8221,
+      "step": 2265
+    },
+    {
+      "epoch": 2.4832876712328766,
+      "grad_norm": 2.6506896018981934,
+      "learning_rate": 9.870024370430546e-07,
+      "loss": 1.725,
+      "step": 2266
+    },
+    {
+      "epoch": 2.4843835616438357,
+      "grad_norm": 2.617889642715454,
+      "learning_rate": 9.849715678310318e-07,
+      "loss": 1.7566,
+      "step": 2267
+    },
+    {
+      "epoch": 2.4854794520547947,
+      "grad_norm": 2.571110486984253,
+      "learning_rate": 9.829406986190091e-07,
+      "loss": 1.7327,
+      "step": 2268
+    },
+    {
+      "epoch": 2.4865753424657533,
+      "grad_norm": 2.657665252685547,
+      "learning_rate": 9.809098294069864e-07,
+      "loss": 1.7471,
+      "step": 2269
+    },
+    {
+      "epoch": 2.4876712328767123,
+      "grad_norm": 2.5402324199676514,
+      "learning_rate": 9.788789601949637e-07,
+      "loss": 1.6597,
+      "step": 2270
+    },
+    {
+      "epoch": 2.4876712328767123,
+      "eval_loss": 1.975923776626587,
+      "eval_runtime": 37.0909,
+      "eval_samples_per_second": 26.961,
+      "eval_steps_per_second": 1.132,
+      "step": 2270
+    },
+    {
+      "epoch": 2.4887671232876714,
+      "grad_norm": 2.619765281677246,
+      "learning_rate": 9.768480909829407e-07,
+      "loss": 1.7434,
+      "step": 2271
+    },
+    {
+      "epoch": 2.48986301369863,
+      "grad_norm": 2.664238214492798,
+      "learning_rate": 9.74817221770918e-07,
+      "loss": 1.7937,
+      "step": 2272
+    },
+    {
+      "epoch": 2.490958904109589,
+      "grad_norm": 2.6743860244750977,
+      "learning_rate": 9.727863525588953e-07,
+      "loss": 1.7203,
+      "step": 2273
+    },
+    {
+      "epoch": 2.492054794520548,
+      "grad_norm": 2.695051670074463,
+      "learning_rate": 9.707554833468726e-07,
+      "loss": 1.7473,
+      "step": 2274
+    },
+    {
+      "epoch": 2.493150684931507,
+      "grad_norm": 2.6115036010742188,
+      "learning_rate": 9.687246141348498e-07,
+      "loss": 1.7358,
+      "step": 2275
+    },
+    {
+      "epoch": 2.4942465753424656,
+      "grad_norm": 2.7951300144195557,
+      "learning_rate": 9.666937449228271e-07,
+      "loss": 1.723,
+      "step": 2276
+    },
+    {
+      "epoch": 2.4953424657534247,
+      "grad_norm": 2.6242685317993164,
+      "learning_rate": 9.646628757108042e-07,
+      "loss": 1.7389,
+      "step": 2277
+    },
+    {
+      "epoch": 2.4964383561643837,
+      "grad_norm": 2.567150831222534,
+      "learning_rate": 9.626320064987814e-07,
+      "loss": 1.6775,
+      "step": 2278
+    },
+    {
+      "epoch": 2.4975342465753423,
+      "grad_norm": 2.6157355308532715,
+      "learning_rate": 9.606011372867587e-07,
+      "loss": 1.7624,
+      "step": 2279
+    },
+    {
+      "epoch": 2.4986301369863013,
+      "grad_norm": 2.5663843154907227,
+      "learning_rate": 9.58570268074736e-07,
+      "loss": 1.6955,
+      "step": 2280
+    },
+    {
+      "epoch": 2.4986301369863013,
+      "eval_loss": 1.9744558334350586,
+      "eval_runtime": 37.1094,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 2280
+    },
+    {
+      "epoch": 2.4997260273972604,
+      "grad_norm": 2.667611598968506,
+      "learning_rate": 9.565393988627133e-07,
+      "loss": 1.7261,
+      "step": 2281
+    },
+    {
+      "epoch": 2.500821917808219,
+      "grad_norm": 2.5702202320098877,
+      "learning_rate": 9.545085296506906e-07,
+      "loss": 1.7252,
+      "step": 2282
+    },
+    {
+      "epoch": 2.501917808219178,
+      "grad_norm": 2.5271430015563965,
+      "learning_rate": 9.524776604386677e-07,
+      "loss": 1.7407,
+      "step": 2283
+    },
+    {
+      "epoch": 2.503013698630137,
+      "grad_norm": 2.5804171562194824,
+      "learning_rate": 9.50446791226645e-07,
+      "loss": 1.739,
+      "step": 2284
+    },
+    {
+      "epoch": 2.504109589041096,
+      "grad_norm": 2.571345090866089,
+      "learning_rate": 9.484159220146223e-07,
+      "loss": 1.668,
+      "step": 2285
+    },
+    {
+      "epoch": 2.5052054794520546,
+      "grad_norm": 2.5697009563446045,
+      "learning_rate": 9.463850528025995e-07,
+      "loss": 1.7677,
+      "step": 2286
+    },
+    {
+      "epoch": 2.5063013698630137,
+      "grad_norm": 2.5846893787384033,
+      "learning_rate": 9.443541835905768e-07,
+      "loss": 1.7589,
+      "step": 2287
+    },
+    {
+      "epoch": 2.5073972602739727,
+      "grad_norm": 2.803380250930786,
+      "learning_rate": 9.423233143785541e-07,
+      "loss": 1.8104,
+      "step": 2288
+    },
+    {
+      "epoch": 2.5084931506849317,
+      "grad_norm": 2.6451611518859863,
+      "learning_rate": 9.402924451665313e-07,
+      "loss": 1.8061,
+      "step": 2289
+    },
+    {
+      "epoch": 2.5095890410958903,
+      "grad_norm": 2.6164097785949707,
+      "learning_rate": 9.382615759545085e-07,
+      "loss": 1.6899,
+      "step": 2290
+    },
+    {
+      "epoch": 2.5095890410958903,
+      "eval_loss": 1.9739093780517578,
+      "eval_runtime": 37.1222,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 2290
+    },
+    {
+      "epoch": 2.5106849315068493,
+      "grad_norm": 2.5874385833740234,
+      "learning_rate": 9.362307067424858e-07,
+      "loss": 1.7692,
+      "step": 2291
+    },
+    {
+      "epoch": 2.5117808219178084,
+      "grad_norm": 2.5705063343048096,
+      "learning_rate": 9.341998375304631e-07,
+      "loss": 1.7303,
+      "step": 2292
+    },
+    {
+      "epoch": 2.512876712328767,
+      "grad_norm": 2.6130380630493164,
+      "learning_rate": 9.321689683184404e-07,
+      "loss": 1.7892,
+      "step": 2293
+    },
+    {
+      "epoch": 2.513972602739726,
+      "grad_norm": 2.5991110801696777,
+      "learning_rate": 9.301380991064176e-07,
+      "loss": 1.6877,
+      "step": 2294
+    },
+    {
+      "epoch": 2.515068493150685,
+      "grad_norm": 2.6092689037323,
+      "learning_rate": 9.281072298943949e-07,
+      "loss": 1.6729,
+      "step": 2295
+    },
+    {
+      "epoch": 2.5161643835616436,
+      "grad_norm": 2.5563883781433105,
+      "learning_rate": 9.260763606823721e-07,
+      "loss": 1.7181,
+      "step": 2296
+    },
+    {
+      "epoch": 2.5172602739726027,
+      "grad_norm": 2.623948574066162,
+      "learning_rate": 9.240454914703494e-07,
+      "loss": 1.7528,
+      "step": 2297
+    },
+    {
+      "epoch": 2.5183561643835617,
+      "grad_norm": 2.661874294281006,
+      "learning_rate": 9.220146222583266e-07,
+      "loss": 1.7582,
+      "step": 2298
+    },
+    {
+      "epoch": 2.5194520547945203,
+      "grad_norm": 2.5936238765716553,
+      "learning_rate": 9.199837530463039e-07,
+      "loss": 1.6987,
+      "step": 2299
+    },
+    {
+      "epoch": 2.5205479452054793,
+      "grad_norm": 2.5134992599487305,
+      "learning_rate": 9.179528838342812e-07,
+      "loss": 1.7265,
+      "step": 2300
+    },
+    {
+      "epoch": 2.5205479452054793,
+      "eval_loss": 1.9737082719802856,
+      "eval_runtime": 37.1105,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 2300
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2736,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.169811641532416e+17,
+  "train_batch_size": 3,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/training_args.bin b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..b7adbe5a2dbf27f05befec6cc65fbc93e90eefd0
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aaa882279512a8c278376c0c601206933f3c5662ddcd3c13c53bf32ae3680d14
+size 6011
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2300/zero_to_fp32.py
@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/config.json
@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "meta-llama/Llama-3.2-3B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "torch_dtype": "float16",
+  "transformers_version": "4.45.1",
+  "use_cache": true,
+  "vocab_size": 128256
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/generation_config.json
@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": 128001,
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.45.1"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/latest b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/latest
new file mode 100644
index 0000000000000000000000000000000000000000..361f8a1a19370f3a8801a82732f02b0e9e639e14
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/latest
@@ -0,0 +1 @@
+global_step2400
\ No newline at end of file
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/model-00002-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..657d9f764560e2c0f93370710f2c7dfac0a4c45d
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/model-00002-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:583bc3f5e1793930f7e6cc39ddf7112b5737346d56bff6c521e99c11fab535f4
+size 2247734920
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/model.safetensors.index.json
new file mode 100644
index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/model.safetensors.index.json
@@ -0,0 +1,262 @@
+{
+  "metadata": {
+    "total_size": 7213504512
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00002-of-00002.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/rng_state_0.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..7cebe108f7c295f538afca780f8bff98aa6232f2
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c7105b68cc7706a616e5184c049404c7551d0ad422b5b1a292aa19facaa3d0f
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/rng_state_1.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..618e221236c3debbaad7ada8aebf8a3335d91f6e
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a994ddb08d50eccc51680080b7c8b587407a3900fb66b09b933d12402a0da4b
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/rng_state_2.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..f378850538621142c5cd15d9d6f50bc9a0f23b02
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0fd0c5d9fbbbbb845dde3ee7d26db4ac79f6c6261f597d751fb6b57476387a4b
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/scheduler.pt b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..e0d6b99496210d40206b6eefd90c34950f2f39b0
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e9cc4799e7e0cae3fc9b7650fd2c5e15389f6a18bff6987ca144ed98209d731
+size 627
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/tokenizer.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb
+size 17210374
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128256": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|end_of_text|>",
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "[PAD]",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..96add4ccbfbb3220812dec785da1b835dbfbf9a6
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/trainer_state.json
@@ -0,0 +1,18753 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.6301369863013697,
+  "eval_steps": 10,
+  "global_step": 2400,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001095890410958904,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0636,
+      "step": 1
+    },
+    {
+      "epoch": 0.002191780821917808,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0501,
+      "step": 2
+    },
+    {
+      "epoch": 0.0032876712328767125,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0238,
+      "step": 3
+    },
+    {
+      "epoch": 0.004383561643835616,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0346,
+      "step": 4
+    },
+    {
+      "epoch": 0.005479452054794521,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0108,
+      "step": 5
+    },
+    {
+      "epoch": 0.006575342465753425,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.131,
+      "step": 6
+    },
+    {
+      "epoch": 0.007671232876712329,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0488,
+      "step": 7
+    },
+    {
+      "epoch": 0.008767123287671232,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0905,
+      "step": 8
+    },
+    {
+      "epoch": 0.009863013698630137,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0493,
+      "step": 9
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.1038,
+      "step": 10
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "eval_loss": 3.100078821182251,
+      "eval_runtime": 37.5838,
+      "eval_samples_per_second": 26.607,
+      "eval_steps_per_second": 1.118,
+      "step": 10
+    },
+    {
+      "epoch": 0.012054794520547946,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0804,
+      "step": 11
+    },
+    {
+      "epoch": 0.01315068493150685,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0053,
+      "step": 12
+    },
+    {
+      "epoch": 0.014246575342465753,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0607,
+      "step": 13
+    },
+    {
+      "epoch": 0.015342465753424657,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0539,
+      "step": 14
+    },
+    {
+      "epoch": 0.01643835616438356,
+      "grad_norm": 5.420984745025635,
+      "learning_rate": 1.8248175182481753e-08,
+      "loss": 3.0941,
+      "step": 15
+    },
+    {
+      "epoch": 0.017534246575342465,
+      "grad_norm": 5.76840877532959,
+      "learning_rate": 3.6496350364963505e-08,
+      "loss": 3.036,
+      "step": 16
+    },
+    {
+      "epoch": 0.01863013698630137,
+      "grad_norm": 5.905357360839844,
+      "learning_rate": 5.4744525547445264e-08,
+      "loss": 3.0491,
+      "step": 17
+    },
+    {
+      "epoch": 0.019726027397260273,
+      "grad_norm": 5.981553554534912,
+      "learning_rate": 7.299270072992701e-08,
+      "loss": 3.0775,
+      "step": 18
+    },
+    {
+      "epoch": 0.020821917808219178,
+      "grad_norm": 6.240172863006592,
+      "learning_rate": 9.124087591240876e-08,
+      "loss": 3.0345,
+      "step": 19
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "grad_norm": 5.59375,
+      "learning_rate": 1.0948905109489053e-07,
+      "loss": 3.1063,
+      "step": 20
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "eval_loss": 3.098820686340332,
+      "eval_runtime": 37.2695,
+      "eval_samples_per_second": 26.832,
+      "eval_steps_per_second": 1.127,
+      "step": 20
+    },
+    {
+      "epoch": 0.023013698630136987,
+      "grad_norm": 5.819795608520508,
+      "learning_rate": 1.2773722627737227e-07,
+      "loss": 3.0808,
+      "step": 21
+    },
+    {
+      "epoch": 0.02410958904109589,
+      "grad_norm": 5.7714762687683105,
+      "learning_rate": 1.4598540145985402e-07,
+      "loss": 3.0055,
+      "step": 22
+    },
+    {
+      "epoch": 0.025205479452054796,
+      "grad_norm": 5.51552152633667,
+      "learning_rate": 1.642335766423358e-07,
+      "loss": 3.0396,
+      "step": 23
+    },
+    {
+      "epoch": 0.0263013698630137,
+      "grad_norm": 5.594012260437012,
+      "learning_rate": 1.8248175182481753e-07,
+      "loss": 3.0824,
+      "step": 24
+    },
+    {
+      "epoch": 0.0273972602739726,
+      "grad_norm": 5.1974029541015625,
+      "learning_rate": 2.0072992700729928e-07,
+      "loss": 3.08,
+      "step": 25
+    },
+    {
+      "epoch": 0.028493150684931506,
+      "grad_norm": 4.733968257904053,
+      "learning_rate": 2.1897810218978106e-07,
+      "loss": 3.0597,
+      "step": 26
+    },
+    {
+      "epoch": 0.02958904109589041,
+      "grad_norm": 4.717319011688232,
+      "learning_rate": 2.3722627737226278e-07,
+      "loss": 3.0696,
+      "step": 27
+    },
+    {
+      "epoch": 0.030684931506849315,
+      "grad_norm": 4.6303534507751465,
+      "learning_rate": 2.5547445255474454e-07,
+      "loss": 3.0273,
+      "step": 28
+    },
+    {
+      "epoch": 0.03178082191780822,
+      "grad_norm": 4.413999557495117,
+      "learning_rate": 2.737226277372263e-07,
+      "loss": 3.0202,
+      "step": 29
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "grad_norm": 4.516043186187744,
+      "learning_rate": 2.9197080291970804e-07,
+      "loss": 2.9926,
+      "step": 30
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "eval_loss": 3.0561797618865967,
+      "eval_runtime": 37.0351,
+      "eval_samples_per_second": 27.001,
+      "eval_steps_per_second": 1.134,
+      "step": 30
+    },
+    {
+      "epoch": 0.03397260273972603,
+      "grad_norm": 3.7179207801818848,
+      "learning_rate": 3.102189781021898e-07,
+      "loss": 3.0086,
+      "step": 31
+    },
+    {
+      "epoch": 0.03506849315068493,
+      "grad_norm": 3.7840425968170166,
+      "learning_rate": 3.284671532846716e-07,
+      "loss": 2.9986,
+      "step": 32
+    },
+    {
+      "epoch": 0.03616438356164384,
+      "grad_norm": 3.843116521835327,
+      "learning_rate": 3.467153284671533e-07,
+      "loss": 2.9677,
+      "step": 33
+    },
+    {
+      "epoch": 0.03726027397260274,
+      "grad_norm": 3.8528714179992676,
+      "learning_rate": 3.6496350364963505e-07,
+      "loss": 2.9732,
+      "step": 34
+    },
+    {
+      "epoch": 0.038356164383561646,
+      "grad_norm": 3.76963210105896,
+      "learning_rate": 3.8321167883211686e-07,
+      "loss": 2.9855,
+      "step": 35
+    },
+    {
+      "epoch": 0.03945205479452055,
+      "grad_norm": 3.6240015029907227,
+      "learning_rate": 4.0145985401459856e-07,
+      "loss": 3.0117,
+      "step": 36
+    },
+    {
+      "epoch": 0.040547945205479455,
+      "grad_norm": 3.1380624771118164,
+      "learning_rate": 4.197080291970803e-07,
+      "loss": 2.9481,
+      "step": 37
+    },
+    {
+      "epoch": 0.041643835616438356,
+      "grad_norm": 3.565329074859619,
+      "learning_rate": 4.379562043795621e-07,
+      "loss": 2.9175,
+      "step": 38
+    },
+    {
+      "epoch": 0.042739726027397264,
+      "grad_norm": 3.8062448501586914,
+      "learning_rate": 4.562043795620438e-07,
+      "loss": 3.0074,
+      "step": 39
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "grad_norm": 3.412046432495117,
+      "learning_rate": 4.7445255474452557e-07,
+      "loss": 2.9119,
+      "step": 40
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "eval_loss": 2.9874370098114014,
+      "eval_runtime": 37.0514,
+      "eval_samples_per_second": 26.99,
+      "eval_steps_per_second": 1.134,
+      "step": 40
+    },
+    {
+      "epoch": 0.044931506849315066,
+      "grad_norm": 3.227437973022461,
+      "learning_rate": 4.927007299270073e-07,
+      "loss": 2.9436,
+      "step": 41
+    },
+    {
+      "epoch": 0.046027397260273974,
+      "grad_norm": 2.9633443355560303,
+      "learning_rate": 5.109489051094891e-07,
+      "loss": 2.967,
+      "step": 42
+    },
+    {
+      "epoch": 0.047123287671232875,
+      "grad_norm": 2.942678928375244,
+      "learning_rate": 5.291970802919708e-07,
+      "loss": 2.9375,
+      "step": 43
+    },
+    {
+      "epoch": 0.04821917808219178,
+      "grad_norm": 2.9806888103485107,
+      "learning_rate": 5.474452554744526e-07,
+      "loss": 2.9474,
+      "step": 44
+    },
+    {
+      "epoch": 0.049315068493150684,
+      "grad_norm": 3.000063419342041,
+      "learning_rate": 5.656934306569343e-07,
+      "loss": 2.9289,
+      "step": 45
+    },
+    {
+      "epoch": 0.05041095890410959,
+      "grad_norm": 3.1481194496154785,
+      "learning_rate": 5.839416058394161e-07,
+      "loss": 2.8869,
+      "step": 46
+    },
+    {
+      "epoch": 0.05150684931506849,
+      "grad_norm": 2.949202060699463,
+      "learning_rate": 6.021897810218979e-07,
+      "loss": 2.8783,
+      "step": 47
+    },
+    {
+      "epoch": 0.0526027397260274,
+      "grad_norm": 2.706526279449463,
+      "learning_rate": 6.204379562043796e-07,
+      "loss": 2.9024,
+      "step": 48
+    },
+    {
+      "epoch": 0.0536986301369863,
+      "grad_norm": 2.7317655086517334,
+      "learning_rate": 6.386861313868614e-07,
+      "loss": 2.8325,
+      "step": 49
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "grad_norm": 2.25644588470459,
+      "learning_rate": 6.569343065693432e-07,
+      "loss": 2.9409,
+      "step": 50
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "eval_loss": 2.909320116043091,
+      "eval_runtime": 37.0714,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 50
+    },
+    {
+      "epoch": 0.05589041095890411,
+      "grad_norm": 2.8789613246917725,
+      "learning_rate": 6.751824817518248e-07,
+      "loss": 2.8459,
+      "step": 51
+    },
+    {
+      "epoch": 0.05698630136986301,
+      "grad_norm": 2.686873435974121,
+      "learning_rate": 6.934306569343066e-07,
+      "loss": 2.8653,
+      "step": 52
+    },
+    {
+      "epoch": 0.05808219178082192,
+      "grad_norm": 2.54382586479187,
+      "learning_rate": 7.116788321167883e-07,
+      "loss": 2.7973,
+      "step": 53
+    },
+    {
+      "epoch": 0.05917808219178082,
+      "grad_norm": 2.1871724128723145,
+      "learning_rate": 7.299270072992701e-07,
+      "loss": 2.7982,
+      "step": 54
+    },
+    {
+      "epoch": 0.06027397260273973,
+      "grad_norm": 2.145303249359131,
+      "learning_rate": 7.48175182481752e-07,
+      "loss": 2.8195,
+      "step": 55
+    },
+    {
+      "epoch": 0.06136986301369863,
+      "grad_norm": 2.1635069847106934,
+      "learning_rate": 7.664233576642337e-07,
+      "loss": 2.825,
+      "step": 56
+    },
+    {
+      "epoch": 0.06246575342465754,
+      "grad_norm": 2.3530752658843994,
+      "learning_rate": 7.846715328467154e-07,
+      "loss": 2.8407,
+      "step": 57
+    },
+    {
+      "epoch": 0.06356164383561644,
+      "grad_norm": 2.2455942630767822,
+      "learning_rate": 8.029197080291971e-07,
+      "loss": 2.8182,
+      "step": 58
+    },
+    {
+      "epoch": 0.06465753424657535,
+      "grad_norm": 1.8985673189163208,
+      "learning_rate": 8.211678832116789e-07,
+      "loss": 2.772,
+      "step": 59
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "grad_norm": 1.9699499607086182,
+      "learning_rate": 8.394160583941606e-07,
+      "loss": 2.7367,
+      "step": 60
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "eval_loss": 2.8451859951019287,
+      "eval_runtime": 37.0793,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 60
+    },
+    {
+      "epoch": 0.06684931506849315,
+      "grad_norm": 2.159547805786133,
+      "learning_rate": 8.576642335766425e-07,
+      "loss": 2.7635,
+      "step": 61
+    },
+    {
+      "epoch": 0.06794520547945206,
+      "grad_norm": 2.12705659866333,
+      "learning_rate": 8.759124087591242e-07,
+      "loss": 2.7624,
+      "step": 62
+    },
+    {
+      "epoch": 0.06904109589041096,
+      "grad_norm": 1.8393241167068481,
+      "learning_rate": 8.94160583941606e-07,
+      "loss": 2.7763,
+      "step": 63
+    },
+    {
+      "epoch": 0.07013698630136986,
+      "grad_norm": 1.683522343635559,
+      "learning_rate": 9.124087591240876e-07,
+      "loss": 2.7762,
+      "step": 64
+    },
+    {
+      "epoch": 0.07123287671232877,
+      "grad_norm": 1.9260376691818237,
+      "learning_rate": 9.306569343065694e-07,
+      "loss": 2.7418,
+      "step": 65
+    },
+    {
+      "epoch": 0.07232876712328767,
+      "grad_norm": 2.09322190284729,
+      "learning_rate": 9.489051094890511e-07,
+      "loss": 2.7625,
+      "step": 66
+    },
+    {
+      "epoch": 0.07342465753424658,
+      "grad_norm": 1.9183429479599,
+      "learning_rate": 9.67153284671533e-07,
+      "loss": 2.7453,
+      "step": 67
+    },
+    {
+      "epoch": 0.07452054794520548,
+      "grad_norm": 1.9792778491973877,
+      "learning_rate": 9.854014598540146e-07,
+      "loss": 2.7858,
+      "step": 68
+    },
+    {
+      "epoch": 0.07561643835616438,
+      "grad_norm": 2.1397311687469482,
+      "learning_rate": 1.0036496350364965e-06,
+      "loss": 2.7242,
+      "step": 69
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "grad_norm": 1.8556463718414307,
+      "learning_rate": 1.0218978102189781e-06,
+      "loss": 2.6853,
+      "step": 70
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "eval_loss": 2.7787046432495117,
+      "eval_runtime": 37.1017,
+      "eval_samples_per_second": 26.953,
+      "eval_steps_per_second": 1.132,
+      "step": 70
+    },
+    {
+      "epoch": 0.07780821917808219,
+      "grad_norm": 1.9776800870895386,
+      "learning_rate": 1.04014598540146e-06,
+      "loss": 2.749,
+      "step": 71
+    },
+    {
+      "epoch": 0.0789041095890411,
+      "grad_norm": 2.3972630500793457,
+      "learning_rate": 1.0583941605839416e-06,
+      "loss": 2.6457,
+      "step": 72
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 2.090210199356079,
+      "learning_rate": 1.0766423357664233e-06,
+      "loss": 2.6401,
+      "step": 73
+    },
+    {
+      "epoch": 0.08109589041095891,
+      "grad_norm": 2.4117701053619385,
+      "learning_rate": 1.0948905109489052e-06,
+      "loss": 2.6649,
+      "step": 74
+    },
+    {
+      "epoch": 0.0821917808219178,
+      "grad_norm": 2.4845173358917236,
+      "learning_rate": 1.113138686131387e-06,
+      "loss": 2.6855,
+      "step": 75
+    },
+    {
+      "epoch": 0.08328767123287671,
+      "grad_norm": 1.7779242992401123,
+      "learning_rate": 1.1313868613138687e-06,
+      "loss": 2.6926,
+      "step": 76
+    },
+    {
+      "epoch": 0.08438356164383562,
+      "grad_norm": 2.3163535594940186,
+      "learning_rate": 1.1496350364963505e-06,
+      "loss": 2.6302,
+      "step": 77
+    },
+    {
+      "epoch": 0.08547945205479453,
+      "grad_norm": 2.2296528816223145,
+      "learning_rate": 1.1678832116788322e-06,
+      "loss": 2.6054,
+      "step": 78
+    },
+    {
+      "epoch": 0.08657534246575342,
+      "grad_norm": 2.253431797027588,
+      "learning_rate": 1.1861313868613138e-06,
+      "loss": 2.6104,
+      "step": 79
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "grad_norm": 2.490609645843506,
+      "learning_rate": 1.2043795620437959e-06,
+      "loss": 2.6078,
+      "step": 80
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "eval_loss": 2.706103563308716,
+      "eval_runtime": 37.0892,
+      "eval_samples_per_second": 26.962,
+      "eval_steps_per_second": 1.132,
+      "step": 80
+    },
+    {
+      "epoch": 0.08876712328767124,
+      "grad_norm": 2.3048248291015625,
+      "learning_rate": 1.2226277372262775e-06,
+      "loss": 2.6096,
+      "step": 81
+    },
+    {
+      "epoch": 0.08986301369863013,
+      "grad_norm": 2.440666913986206,
+      "learning_rate": 1.2408759124087592e-06,
+      "loss": 2.6412,
+      "step": 82
+    },
+    {
+      "epoch": 0.09095890410958904,
+      "grad_norm": 1.9456374645233154,
+      "learning_rate": 1.259124087591241e-06,
+      "loss": 2.591,
+      "step": 83
+    },
+    {
+      "epoch": 0.09205479452054795,
+      "grad_norm": 2.598172664642334,
+      "learning_rate": 1.2773722627737229e-06,
+      "loss": 2.6093,
+      "step": 84
+    },
+    {
+      "epoch": 0.09315068493150686,
+      "grad_norm": 2.2587032318115234,
+      "learning_rate": 1.2956204379562045e-06,
+      "loss": 2.5949,
+      "step": 85
+    },
+    {
+      "epoch": 0.09424657534246575,
+      "grad_norm": 2.2927448749542236,
+      "learning_rate": 1.3138686131386864e-06,
+      "loss": 2.5464,
+      "step": 86
+    },
+    {
+      "epoch": 0.09534246575342466,
+      "grad_norm": 2.5940468311309814,
+      "learning_rate": 1.332116788321168e-06,
+      "loss": 2.5613,
+      "step": 87
+    },
+    {
+      "epoch": 0.09643835616438357,
+      "grad_norm": 2.073207378387451,
+      "learning_rate": 1.3503649635036497e-06,
+      "loss": 2.4955,
+      "step": 88
+    },
+    {
+      "epoch": 0.09753424657534246,
+      "grad_norm": 3.211766242980957,
+      "learning_rate": 1.3686131386861315e-06,
+      "loss": 2.5315,
+      "step": 89
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "grad_norm": 2.2590038776397705,
+      "learning_rate": 1.3868613138686132e-06,
+      "loss": 2.4641,
+      "step": 90
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "eval_loss": 2.625087261199951,
+      "eval_runtime": 37.0852,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 90
+    },
+    {
+      "epoch": 0.09972602739726028,
+      "grad_norm": 2.3959836959838867,
+      "learning_rate": 1.405109489051095e-06,
+      "loss": 2.5537,
+      "step": 91
+    },
+    {
+      "epoch": 0.10082191780821918,
+      "grad_norm": 2.596761465072632,
+      "learning_rate": 1.4233576642335767e-06,
+      "loss": 2.5135,
+      "step": 92
+    },
+    {
+      "epoch": 0.10191780821917808,
+      "grad_norm": 2.475759267807007,
+      "learning_rate": 1.4416058394160583e-06,
+      "loss": 2.5436,
+      "step": 93
+    },
+    {
+      "epoch": 0.10301369863013699,
+      "grad_norm": 3.0613298416137695,
+      "learning_rate": 1.4598540145985402e-06,
+      "loss": 2.5066,
+      "step": 94
+    },
+    {
+      "epoch": 0.10410958904109589,
+      "grad_norm": 2.406097650527954,
+      "learning_rate": 1.478102189781022e-06,
+      "loss": 2.4646,
+      "step": 95
+    },
+    {
+      "epoch": 0.1052054794520548,
+      "grad_norm": 2.5872504711151123,
+      "learning_rate": 1.496350364963504e-06,
+      "loss": 2.5351,
+      "step": 96
+    },
+    {
+      "epoch": 0.1063013698630137,
+      "grad_norm": 2.4294471740722656,
+      "learning_rate": 1.5145985401459856e-06,
+      "loss": 2.4953,
+      "step": 97
+    },
+    {
+      "epoch": 0.1073972602739726,
+      "grad_norm": 2.674884080886841,
+      "learning_rate": 1.5328467153284674e-06,
+      "loss": 2.5133,
+      "step": 98
+    },
+    {
+      "epoch": 0.10849315068493151,
+      "grad_norm": 2.5096852779388428,
+      "learning_rate": 1.551094890510949e-06,
+      "loss": 2.4693,
+      "step": 99
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "grad_norm": 2.7390334606170654,
+      "learning_rate": 1.5693430656934307e-06,
+      "loss": 2.4195,
+      "step": 100
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "eval_loss": 2.55997371673584,
+      "eval_runtime": 37.0732,
+      "eval_samples_per_second": 26.974,
+      "eval_steps_per_second": 1.133,
+      "step": 100
+    },
+    {
+      "epoch": 0.11068493150684931,
+      "grad_norm": 2.8887312412261963,
+      "learning_rate": 1.5875912408759126e-06,
+      "loss": 2.4193,
+      "step": 101
+    },
+    {
+      "epoch": 0.11178082191780822,
+      "grad_norm": 3.0721421241760254,
+      "learning_rate": 1.6058394160583942e-06,
+      "loss": 2.4709,
+      "step": 102
+    },
+    {
+      "epoch": 0.11287671232876713,
+      "grad_norm": 2.728156566619873,
+      "learning_rate": 1.624087591240876e-06,
+      "loss": 2.5105,
+      "step": 103
+    },
+    {
+      "epoch": 0.11397260273972602,
+      "grad_norm": 3.1214346885681152,
+      "learning_rate": 1.6423357664233577e-06,
+      "loss": 2.4316,
+      "step": 104
+    },
+    {
+      "epoch": 0.11506849315068493,
+      "grad_norm": 2.9249517917633057,
+      "learning_rate": 1.6605839416058394e-06,
+      "loss": 2.3737,
+      "step": 105
+    },
+    {
+      "epoch": 0.11616438356164384,
+      "grad_norm": 2.6646173000335693,
+      "learning_rate": 1.6788321167883212e-06,
+      "loss": 2.4165,
+      "step": 106
+    },
+    {
+      "epoch": 0.11726027397260275,
+      "grad_norm": 3.108067750930786,
+      "learning_rate": 1.6970802919708029e-06,
+      "loss": 2.4217,
+      "step": 107
+    },
+    {
+      "epoch": 0.11835616438356164,
+      "grad_norm": 3.142338991165161,
+      "learning_rate": 1.715328467153285e-06,
+      "loss": 2.3527,
+      "step": 108
+    },
+    {
+      "epoch": 0.11945205479452055,
+      "grad_norm": 2.9078011512756348,
+      "learning_rate": 1.7335766423357666e-06,
+      "loss": 2.3927,
+      "step": 109
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "grad_norm": 3.40766978263855,
+      "learning_rate": 1.7518248175182485e-06,
+      "loss": 2.4021,
+      "step": 110
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "eval_loss": 2.5031485557556152,
+      "eval_runtime": 36.8147,
+      "eval_samples_per_second": 27.163,
+      "eval_steps_per_second": 1.141,
+      "step": 110
+    },
+    {
+      "epoch": 0.12164383561643835,
+      "grad_norm": 4.396062850952148,
+      "learning_rate": 1.77007299270073e-06,
+      "loss": 2.4373,
+      "step": 111
+    },
+    {
+      "epoch": 0.12273972602739726,
+      "grad_norm": 4.0441365242004395,
+      "learning_rate": 1.788321167883212e-06,
+      "loss": 2.4162,
+      "step": 112
+    },
+    {
+      "epoch": 0.12383561643835617,
+      "grad_norm": 2.8869969844818115,
+      "learning_rate": 1.8065693430656936e-06,
+      "loss": 2.4515,
+      "step": 113
+    },
+    {
+      "epoch": 0.12493150684931507,
+      "grad_norm": 3.4833669662475586,
+      "learning_rate": 1.8248175182481753e-06,
+      "loss": 2.3965,
+      "step": 114
+    },
+    {
+      "epoch": 0.12602739726027398,
+      "grad_norm": 2.9996068477630615,
+      "learning_rate": 1.8430656934306571e-06,
+      "loss": 2.3374,
+      "step": 115
+    },
+    {
+      "epoch": 0.12712328767123288,
+      "grad_norm": 2.701371192932129,
+      "learning_rate": 1.8613138686131388e-06,
+      "loss": 2.3608,
+      "step": 116
+    },
+    {
+      "epoch": 0.12821917808219177,
+      "grad_norm": 2.9491288661956787,
+      "learning_rate": 1.8795620437956206e-06,
+      "loss": 2.355,
+      "step": 117
+    },
+    {
+      "epoch": 0.1293150684931507,
+      "grad_norm": 2.9266560077667236,
+      "learning_rate": 1.8978102189781023e-06,
+      "loss": 2.3475,
+      "step": 118
+    },
+    {
+      "epoch": 0.1304109589041096,
+      "grad_norm": 3.0419797897338867,
+      "learning_rate": 1.916058394160584e-06,
+      "loss": 2.3475,
+      "step": 119
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "grad_norm": 2.4531214237213135,
+      "learning_rate": 1.934306569343066e-06,
+      "loss": 2.3724,
+      "step": 120
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "eval_loss": 2.455423593521118,
+      "eval_runtime": 37.0304,
+      "eval_samples_per_second": 27.005,
+      "eval_steps_per_second": 1.134,
+      "step": 120
+    },
+    {
+      "epoch": 0.1326027397260274,
+      "grad_norm": 3.572502851486206,
+      "learning_rate": 1.952554744525548e-06,
+      "loss": 2.2671,
+      "step": 121
+    },
+    {
+      "epoch": 0.1336986301369863,
+      "grad_norm": 3.5408504009246826,
+      "learning_rate": 1.9708029197080293e-06,
+      "loss": 2.3133,
+      "step": 122
+    },
+    {
+      "epoch": 0.13479452054794522,
+      "grad_norm": 2.488725423812866,
+      "learning_rate": 1.989051094890511e-06,
+      "loss": 2.3141,
+      "step": 123
+    },
+    {
+      "epoch": 0.1358904109589041,
+      "grad_norm": 3.241276264190674,
+      "learning_rate": 2.007299270072993e-06,
+      "loss": 2.318,
+      "step": 124
+    },
+    {
+      "epoch": 0.136986301369863,
+      "grad_norm": 2.853959798812866,
+      "learning_rate": 2.025547445255475e-06,
+      "loss": 2.3712,
+      "step": 125
+    },
+    {
+      "epoch": 0.13808219178082193,
+      "grad_norm": 2.9798169136047363,
+      "learning_rate": 2.0437956204379563e-06,
+      "loss": 2.3072,
+      "step": 126
+    },
+    {
+      "epoch": 0.13917808219178082,
+      "grad_norm": 2.8529601097106934,
+      "learning_rate": 2.062043795620438e-06,
+      "loss": 2.318,
+      "step": 127
+    },
+    {
+      "epoch": 0.14027397260273972,
+      "grad_norm": 3.475337028503418,
+      "learning_rate": 2.08029197080292e-06,
+      "loss": 2.31,
+      "step": 128
+    },
+    {
+      "epoch": 0.14136986301369864,
+      "grad_norm": 4.213955879211426,
+      "learning_rate": 2.0985401459854014e-06,
+      "loss": 2.3734,
+      "step": 129
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "grad_norm": 3.0833330154418945,
+      "learning_rate": 2.1167883211678833e-06,
+      "loss": 2.2602,
+      "step": 130
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "eval_loss": 2.4205682277679443,
+      "eval_runtime": 37.0798,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 130
+    },
+    {
+      "epoch": 0.14356164383561643,
+      "grad_norm": 3.411235809326172,
+      "learning_rate": 2.135036496350365e-06,
+      "loss": 2.3255,
+      "step": 131
+    },
+    {
+      "epoch": 0.14465753424657535,
+      "grad_norm": 2.882053852081299,
+      "learning_rate": 2.1532846715328466e-06,
+      "loss": 2.3483,
+      "step": 132
+    },
+    {
+      "epoch": 0.14575342465753424,
+      "grad_norm": 3.2436420917510986,
+      "learning_rate": 2.171532846715329e-06,
+      "loss": 2.2492,
+      "step": 133
+    },
+    {
+      "epoch": 0.14684931506849316,
+      "grad_norm": 2.929466962814331,
+      "learning_rate": 2.1897810218978103e-06,
+      "loss": 2.2647,
+      "step": 134
+    },
+    {
+      "epoch": 0.14794520547945206,
+      "grad_norm": 2.9344966411590576,
+      "learning_rate": 2.208029197080292e-06,
+      "loss": 2.3061,
+      "step": 135
+    },
+    {
+      "epoch": 0.14904109589041095,
+      "grad_norm": 3.2567203044891357,
+      "learning_rate": 2.226277372262774e-06,
+      "loss": 2.2746,
+      "step": 136
+    },
+    {
+      "epoch": 0.15013698630136987,
+      "grad_norm": 3.1494054794311523,
+      "learning_rate": 2.244525547445256e-06,
+      "loss": 2.2866,
+      "step": 137
+    },
+    {
+      "epoch": 0.15123287671232877,
+      "grad_norm": 2.723390817642212,
+      "learning_rate": 2.2627737226277373e-06,
+      "loss": 2.3109,
+      "step": 138
+    },
+    {
+      "epoch": 0.15232876712328766,
+      "grad_norm": 2.8904967308044434,
+      "learning_rate": 2.281021897810219e-06,
+      "loss": 2.301,
+      "step": 139
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "grad_norm": 2.8899483680725098,
+      "learning_rate": 2.299270072992701e-06,
+      "loss": 2.2637,
+      "step": 140
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "eval_loss": 2.3952724933624268,
+      "eval_runtime": 37.1272,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 140
+    },
+    {
+      "epoch": 0.15452054794520548,
+      "grad_norm": 3.5309066772460938,
+      "learning_rate": 2.3175182481751825e-06,
+      "loss": 2.2572,
+      "step": 141
+    },
+    {
+      "epoch": 0.15561643835616437,
+      "grad_norm": 2.8369667530059814,
+      "learning_rate": 2.3357664233576643e-06,
+      "loss": 2.2712,
+      "step": 142
+    },
+    {
+      "epoch": 0.1567123287671233,
+      "grad_norm": 3.629161834716797,
+      "learning_rate": 2.354014598540146e-06,
+      "loss": 2.2301,
+      "step": 143
+    },
+    {
+      "epoch": 0.1578082191780822,
+      "grad_norm": 3.5313000679016113,
+      "learning_rate": 2.3722627737226276e-06,
+      "loss": 2.2712,
+      "step": 144
+    },
+    {
+      "epoch": 0.1589041095890411,
+      "grad_norm": 3.1328771114349365,
+      "learning_rate": 2.39051094890511e-06,
+      "loss": 2.2706,
+      "step": 145
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 4.6992950439453125,
+      "learning_rate": 2.4087591240875918e-06,
+      "loss": 2.219,
+      "step": 146
+    },
+    {
+      "epoch": 0.1610958904109589,
+      "grad_norm": 5.230221748352051,
+      "learning_rate": 2.427007299270073e-06,
+      "loss": 2.2396,
+      "step": 147
+    },
+    {
+      "epoch": 0.16219178082191782,
+      "grad_norm": 3.5680761337280273,
+      "learning_rate": 2.445255474452555e-06,
+      "loss": 2.2434,
+      "step": 148
+    },
+    {
+      "epoch": 0.1632876712328767,
+      "grad_norm": 3.1642441749572754,
+      "learning_rate": 2.463503649635037e-06,
+      "loss": 2.2856,
+      "step": 149
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "grad_norm": 2.694085121154785,
+      "learning_rate": 2.4817518248175183e-06,
+      "loss": 2.2006,
+      "step": 150
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "eval_loss": 2.37199068069458,
+      "eval_runtime": 37.1342,
+      "eval_samples_per_second": 26.929,
+      "eval_steps_per_second": 1.131,
+      "step": 150
+    },
+    {
+      "epoch": 0.16547945205479453,
+      "grad_norm": 2.9868853092193604,
+      "learning_rate": 2.5e-06,
+      "loss": 2.1694,
+      "step": 151
+    },
+    {
+      "epoch": 0.16657534246575342,
+      "grad_norm": 2.793532371520996,
+      "learning_rate": 2.518248175182482e-06,
+      "loss": 2.2965,
+      "step": 152
+    },
+    {
+      "epoch": 0.16767123287671232,
+      "grad_norm": 2.60152530670166,
+      "learning_rate": 2.5364963503649635e-06,
+      "loss": 2.1954,
+      "step": 153
+    },
+    {
+      "epoch": 0.16876712328767124,
+      "grad_norm": 2.6192219257354736,
+      "learning_rate": 2.5547445255474458e-06,
+      "loss": 2.1909,
+      "step": 154
+    },
+    {
+      "epoch": 0.16986301369863013,
+      "grad_norm": 2.467344045639038,
+      "learning_rate": 2.5729927007299272e-06,
+      "loss": 2.2438,
+      "step": 155
+    },
+    {
+      "epoch": 0.17095890410958905,
+      "grad_norm": 2.772012948989868,
+      "learning_rate": 2.591240875912409e-06,
+      "loss": 2.2213,
+      "step": 156
+    },
+    {
+      "epoch": 0.17205479452054795,
+      "grad_norm": 2.613675355911255,
+      "learning_rate": 2.6094890510948905e-06,
+      "loss": 2.2213,
+      "step": 157
+    },
+    {
+      "epoch": 0.17315068493150684,
+      "grad_norm": 2.849245071411133,
+      "learning_rate": 2.627737226277373e-06,
+      "loss": 2.255,
+      "step": 158
+    },
+    {
+      "epoch": 0.17424657534246576,
+      "grad_norm": 2.8388969898223877,
+      "learning_rate": 2.6459854014598542e-06,
+      "loss": 2.2572,
+      "step": 159
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "grad_norm": 2.5219321250915527,
+      "learning_rate": 2.664233576642336e-06,
+      "loss": 2.2253,
+      "step": 160
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "eval_loss": 2.3509061336517334,
+      "eval_runtime": 37.0953,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 160
+    },
+    {
+      "epoch": 0.17643835616438355,
+      "grad_norm": 2.8460891246795654,
+      "learning_rate": 2.6824817518248175e-06,
+      "loss": 2.2009,
+      "step": 161
+    },
+    {
+      "epoch": 0.17753424657534247,
+      "grad_norm": 2.673377275466919,
+      "learning_rate": 2.7007299270072994e-06,
+      "loss": 2.2211,
+      "step": 162
+    },
+    {
+      "epoch": 0.17863013698630137,
+      "grad_norm": 2.8514182567596436,
+      "learning_rate": 2.718978102189781e-06,
+      "loss": 2.1955,
+      "step": 163
+    },
+    {
+      "epoch": 0.17972602739726026,
+      "grad_norm": 3.530385971069336,
+      "learning_rate": 2.737226277372263e-06,
+      "loss": 2.2536,
+      "step": 164
+    },
+    {
+      "epoch": 0.18082191780821918,
+      "grad_norm": 2.733394145965576,
+      "learning_rate": 2.755474452554745e-06,
+      "loss": 2.2367,
+      "step": 165
+    },
+    {
+      "epoch": 0.18191780821917808,
+      "grad_norm": 3.871150016784668,
+      "learning_rate": 2.7737226277372264e-06,
+      "loss": 2.1949,
+      "step": 166
+    },
+    {
+      "epoch": 0.18301369863013697,
+      "grad_norm": 3.7782530784606934,
+      "learning_rate": 2.7919708029197082e-06,
+      "loss": 2.2979,
+      "step": 167
+    },
+    {
+      "epoch": 0.1841095890410959,
+      "grad_norm": 3.094599485397339,
+      "learning_rate": 2.81021897810219e-06,
+      "loss": 2.2512,
+      "step": 168
+    },
+    {
+      "epoch": 0.1852054794520548,
+      "grad_norm": 3.2166848182678223,
+      "learning_rate": 2.828467153284672e-06,
+      "loss": 2.1784,
+      "step": 169
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "grad_norm": 3.4127073287963867,
+      "learning_rate": 2.8467153284671534e-06,
+      "loss": 2.2241,
+      "step": 170
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "eval_loss": 2.333245038986206,
+      "eval_runtime": 37.0402,
+      "eval_samples_per_second": 26.998,
+      "eval_steps_per_second": 1.134,
+      "step": 170
+    },
+    {
+      "epoch": 0.1873972602739726,
+      "grad_norm": 2.677579879760742,
+      "learning_rate": 2.8649635036496353e-06,
+      "loss": 2.135,
+      "step": 171
+    },
+    {
+      "epoch": 0.1884931506849315,
+      "grad_norm": 2.9602150917053223,
+      "learning_rate": 2.8832116788321167e-06,
+      "loss": 2.1916,
+      "step": 172
+    },
+    {
+      "epoch": 0.18958904109589042,
+      "grad_norm": 2.580122232437134,
+      "learning_rate": 2.901459854014599e-06,
+      "loss": 2.1477,
+      "step": 173
+    },
+    {
+      "epoch": 0.19068493150684931,
+      "grad_norm": 2.652217388153076,
+      "learning_rate": 2.9197080291970804e-06,
+      "loss": 2.2394,
+      "step": 174
+    },
+    {
+      "epoch": 0.1917808219178082,
+      "grad_norm": 2.6887567043304443,
+      "learning_rate": 2.9379562043795623e-06,
+      "loss": 2.2245,
+      "step": 175
+    },
+    {
+      "epoch": 0.19287671232876713,
+      "grad_norm": 2.768165111541748,
+      "learning_rate": 2.956204379562044e-06,
+      "loss": 2.2017,
+      "step": 176
+    },
+    {
+      "epoch": 0.19397260273972602,
+      "grad_norm": 2.9375691413879395,
+      "learning_rate": 2.9744525547445256e-06,
+      "loss": 2.2464,
+      "step": 177
+    },
+    {
+      "epoch": 0.19506849315068492,
+      "grad_norm": 2.8364248275756836,
+      "learning_rate": 2.992700729927008e-06,
+      "loss": 2.2155,
+      "step": 178
+    },
+    {
+      "epoch": 0.19616438356164384,
+      "grad_norm": 2.843487024307251,
+      "learning_rate": 3.0109489051094893e-06,
+      "loss": 2.1577,
+      "step": 179
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "grad_norm": 3.005230665206909,
+      "learning_rate": 3.029197080291971e-06,
+      "loss": 2.1549,
+      "step": 180
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "eval_loss": 2.3209409713745117,
+      "eval_runtime": 37.022,
+      "eval_samples_per_second": 27.011,
+      "eval_steps_per_second": 1.134,
+      "step": 180
+    },
+    {
+      "epoch": 0.19835616438356166,
+      "grad_norm": 2.490260601043701,
+      "learning_rate": 3.0474452554744526e-06,
+      "loss": 2.2313,
+      "step": 181
+    },
+    {
+      "epoch": 0.19945205479452055,
+      "grad_norm": 2.973118543624878,
+      "learning_rate": 3.065693430656935e-06,
+      "loss": 2.2316,
+      "step": 182
+    },
+    {
+      "epoch": 0.20054794520547944,
+      "grad_norm": 3.3308680057525635,
+      "learning_rate": 3.0839416058394163e-06,
+      "loss": 2.2087,
+      "step": 183
+    },
+    {
+      "epoch": 0.20164383561643837,
+      "grad_norm": 3.1440727710723877,
+      "learning_rate": 3.102189781021898e-06,
+      "loss": 2.2256,
+      "step": 184
+    },
+    {
+      "epoch": 0.20273972602739726,
+      "grad_norm": 2.5696325302124023,
+      "learning_rate": 3.1204379562043796e-06,
+      "loss": 2.2093,
+      "step": 185
+    },
+    {
+      "epoch": 0.20383561643835615,
+      "grad_norm": 3.1225037574768066,
+      "learning_rate": 3.1386861313868614e-06,
+      "loss": 2.1425,
+      "step": 186
+    },
+    {
+      "epoch": 0.20493150684931508,
+      "grad_norm": 3.650743246078491,
+      "learning_rate": 3.156934306569343e-06,
+      "loss": 2.1339,
+      "step": 187
+    },
+    {
+      "epoch": 0.20602739726027397,
+      "grad_norm": 2.924072027206421,
+      "learning_rate": 3.175182481751825e-06,
+      "loss": 2.2363,
+      "step": 188
+    },
+    {
+      "epoch": 0.20712328767123286,
+      "grad_norm": 2.9561989307403564,
+      "learning_rate": 3.193430656934307e-06,
+      "loss": 2.1934,
+      "step": 189
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "grad_norm": 2.731393814086914,
+      "learning_rate": 3.2116788321167884e-06,
+      "loss": 2.1718,
+      "step": 190
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "eval_loss": 2.3087782859802246,
+      "eval_runtime": 37.0161,
+      "eval_samples_per_second": 27.015,
+      "eval_steps_per_second": 1.135,
+      "step": 190
+    },
+    {
+      "epoch": 0.20931506849315068,
+      "grad_norm": 2.774442434310913,
+      "learning_rate": 3.2299270072992707e-06,
+      "loss": 2.1841,
+      "step": 191
+    },
+    {
+      "epoch": 0.2104109589041096,
+      "grad_norm": 3.0579135417938232,
+      "learning_rate": 3.248175182481752e-06,
+      "loss": 2.1677,
+      "step": 192
+    },
+    {
+      "epoch": 0.2115068493150685,
+      "grad_norm": 2.732984781265259,
+      "learning_rate": 3.266423357664234e-06,
+      "loss": 2.0975,
+      "step": 193
+    },
+    {
+      "epoch": 0.2126027397260274,
+      "grad_norm": 3.442659854888916,
+      "learning_rate": 3.2846715328467155e-06,
+      "loss": 2.1923,
+      "step": 194
+    },
+    {
+      "epoch": 0.2136986301369863,
+      "grad_norm": 3.064223527908325,
+      "learning_rate": 3.3029197080291973e-06,
+      "loss": 2.1707,
+      "step": 195
+    },
+    {
+      "epoch": 0.2147945205479452,
+      "grad_norm": 2.5138468742370605,
+      "learning_rate": 3.3211678832116788e-06,
+      "loss": 2.1933,
+      "step": 196
+    },
+    {
+      "epoch": 0.2158904109589041,
+      "grad_norm": 3.1076464653015137,
+      "learning_rate": 3.339416058394161e-06,
+      "loss": 2.1725,
+      "step": 197
+    },
+    {
+      "epoch": 0.21698630136986302,
+      "grad_norm": 2.7291388511657715,
+      "learning_rate": 3.3576642335766425e-06,
+      "loss": 2.1722,
+      "step": 198
+    },
+    {
+      "epoch": 0.21808219178082192,
+      "grad_norm": 2.805506467819214,
+      "learning_rate": 3.3759124087591243e-06,
+      "loss": 2.1688,
+      "step": 199
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "grad_norm": 2.6190085411071777,
+      "learning_rate": 3.3941605839416058e-06,
+      "loss": 2.1708,
+      "step": 200
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "eval_loss": 2.295546054840088,
+      "eval_runtime": 37.0133,
+      "eval_samples_per_second": 27.017,
+      "eval_steps_per_second": 1.135,
+      "step": 200
+    },
+    {
+      "epoch": 0.22027397260273973,
+      "grad_norm": 3.1270673274993896,
+      "learning_rate": 3.412408759124088e-06,
+      "loss": 2.1759,
+      "step": 201
+    },
+    {
+      "epoch": 0.22136986301369863,
+      "grad_norm": 3.494178533554077,
+      "learning_rate": 3.43065693430657e-06,
+      "loss": 2.1795,
+      "step": 202
+    },
+    {
+      "epoch": 0.22246575342465755,
+      "grad_norm": 2.7398736476898193,
+      "learning_rate": 3.4489051094890513e-06,
+      "loss": 2.1201,
+      "step": 203
+    },
+    {
+      "epoch": 0.22356164383561644,
+      "grad_norm": 2.981506824493408,
+      "learning_rate": 3.467153284671533e-06,
+      "loss": 2.1693,
+      "step": 204
+    },
+    {
+      "epoch": 0.22465753424657534,
+      "grad_norm": 2.4129178524017334,
+      "learning_rate": 3.4854014598540146e-06,
+      "loss": 2.1125,
+      "step": 205
+    },
+    {
+      "epoch": 0.22575342465753426,
+      "grad_norm": 2.701197385787964,
+      "learning_rate": 3.503649635036497e-06,
+      "loss": 2.1565,
+      "step": 206
+    },
+    {
+      "epoch": 0.22684931506849315,
+      "grad_norm": 2.533306837081909,
+      "learning_rate": 3.5218978102189783e-06,
+      "loss": 2.1487,
+      "step": 207
+    },
+    {
+      "epoch": 0.22794520547945205,
+      "grad_norm": 2.8030788898468018,
+      "learning_rate": 3.54014598540146e-06,
+      "loss": 2.1801,
+      "step": 208
+    },
+    {
+      "epoch": 0.22904109589041097,
+      "grad_norm": 2.895414352416992,
+      "learning_rate": 3.5583941605839416e-06,
+      "loss": 2.189,
+      "step": 209
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "grad_norm": 2.508390426635742,
+      "learning_rate": 3.576642335766424e-06,
+      "loss": 2.2376,
+      "step": 210
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "eval_loss": 2.2900516986846924,
+      "eval_runtime": 36.7645,
+      "eval_samples_per_second": 27.2,
+      "eval_steps_per_second": 1.142,
+      "step": 210
+    },
+    {
+      "epoch": 0.23123287671232876,
+      "grad_norm": 2.6393816471099854,
+      "learning_rate": 3.5948905109489054e-06,
+      "loss": 2.2267,
+      "step": 211
+    },
+    {
+      "epoch": 0.23232876712328768,
+      "grad_norm": 3.0687549114227295,
+      "learning_rate": 3.6131386861313872e-06,
+      "loss": 2.0744,
+      "step": 212
+    },
+    {
+      "epoch": 0.23342465753424657,
+      "grad_norm": 2.6342105865478516,
+      "learning_rate": 3.6313868613138687e-06,
+      "loss": 2.0737,
+      "step": 213
+    },
+    {
+      "epoch": 0.2345205479452055,
+      "grad_norm": 3.5662875175476074,
+      "learning_rate": 3.6496350364963505e-06,
+      "loss": 2.1147,
+      "step": 214
+    },
+    {
+      "epoch": 0.2356164383561644,
+      "grad_norm": 2.5357561111450195,
+      "learning_rate": 3.667883211678833e-06,
+      "loss": 2.1268,
+      "step": 215
+    },
+    {
+      "epoch": 0.23671232876712328,
+      "grad_norm": 3.439711093902588,
+      "learning_rate": 3.6861313868613142e-06,
+      "loss": 2.1118,
+      "step": 216
+    },
+    {
+      "epoch": 0.2378082191780822,
+      "grad_norm": 2.9410204887390137,
+      "learning_rate": 3.704379562043796e-06,
+      "loss": 2.1087,
+      "step": 217
+    },
+    {
+      "epoch": 0.2389041095890411,
+      "grad_norm": 3.3182294368743896,
+      "learning_rate": 3.7226277372262775e-06,
+      "loss": 2.196,
+      "step": 218
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 3.3834011554718018,
+      "learning_rate": 3.7408759124087594e-06,
+      "loss": 2.1633,
+      "step": 219
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "grad_norm": 3.114952325820923,
+      "learning_rate": 3.7591240875912412e-06,
+      "loss": 2.1107,
+      "step": 220
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "eval_loss": 2.277580976486206,
+      "eval_runtime": 36.9846,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 220
+    },
+    {
+      "epoch": 0.2421917808219178,
+      "grad_norm": 3.1151392459869385,
+      "learning_rate": 3.777372262773723e-06,
+      "loss": 2.1471,
+      "step": 221
+    },
+    {
+      "epoch": 0.2432876712328767,
+      "grad_norm": 3.798450469970703,
+      "learning_rate": 3.7956204379562045e-06,
+      "loss": 2.1324,
+      "step": 222
+    },
+    {
+      "epoch": 0.24438356164383562,
+      "grad_norm": 2.6187098026275635,
+      "learning_rate": 3.8138686131386864e-06,
+      "loss": 2.1061,
+      "step": 223
+    },
+    {
+      "epoch": 0.24547945205479452,
+      "grad_norm": 4.014018535614014,
+      "learning_rate": 3.832116788321168e-06,
+      "loss": 2.166,
+      "step": 224
+    },
+    {
+      "epoch": 0.2465753424657534,
+      "grad_norm": 3.0043938159942627,
+      "learning_rate": 3.85036496350365e-06,
+      "loss": 2.1197,
+      "step": 225
+    },
+    {
+      "epoch": 0.24767123287671233,
+      "grad_norm": 3.507066249847412,
+      "learning_rate": 3.868613138686132e-06,
+      "loss": 2.0848,
+      "step": 226
+    },
+    {
+      "epoch": 0.24876712328767123,
+      "grad_norm": 3.5713281631469727,
+      "learning_rate": 3.886861313868613e-06,
+      "loss": 2.1239,
+      "step": 227
+    },
+    {
+      "epoch": 0.24986301369863015,
+      "grad_norm": 3.961082696914673,
+      "learning_rate": 3.905109489051096e-06,
+      "loss": 2.1406,
+      "step": 228
+    },
+    {
+      "epoch": 0.25095890410958904,
+      "grad_norm": 3.1514859199523926,
+      "learning_rate": 3.923357664233577e-06,
+      "loss": 2.1689,
+      "step": 229
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "grad_norm": 3.8486833572387695,
+      "learning_rate": 3.9416058394160585e-06,
+      "loss": 2.1321,
+      "step": 230
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "eval_loss": 2.2620601654052734,
+      "eval_runtime": 37.058,
+      "eval_samples_per_second": 26.985,
+      "eval_steps_per_second": 1.133,
+      "step": 230
+    },
+    {
+      "epoch": 0.25315068493150683,
+      "grad_norm": 2.5626907348632812,
+      "learning_rate": 3.95985401459854e-06,
+      "loss": 2.1777,
+      "step": 231
+    },
+    {
+      "epoch": 0.25424657534246575,
+      "grad_norm": 4.105278015136719,
+      "learning_rate": 3.978102189781022e-06,
+      "loss": 2.2099,
+      "step": 232
+    },
+    {
+      "epoch": 0.2553424657534247,
+      "grad_norm": 2.9818739891052246,
+      "learning_rate": 3.996350364963504e-06,
+      "loss": 2.0705,
+      "step": 233
+    },
+    {
+      "epoch": 0.25643835616438354,
+      "grad_norm": 2.87030291557312,
+      "learning_rate": 4.014598540145986e-06,
+      "loss": 2.1453,
+      "step": 234
+    },
+    {
+      "epoch": 0.25753424657534246,
+      "grad_norm": 2.4847915172576904,
+      "learning_rate": 4.032846715328467e-06,
+      "loss": 2.0461,
+      "step": 235
+    },
+    {
+      "epoch": 0.2586301369863014,
+      "grad_norm": 3.451420783996582,
+      "learning_rate": 4.05109489051095e-06,
+      "loss": 2.1133,
+      "step": 236
+    },
+    {
+      "epoch": 0.25972602739726025,
+      "grad_norm": 2.954723358154297,
+      "learning_rate": 4.069343065693431e-06,
+      "loss": 2.1316,
+      "step": 237
+    },
+    {
+      "epoch": 0.2608219178082192,
+      "grad_norm": 2.5295896530151367,
+      "learning_rate": 4.0875912408759126e-06,
+      "loss": 2.0901,
+      "step": 238
+    },
+    {
+      "epoch": 0.2619178082191781,
+      "grad_norm": 2.6359453201293945,
+      "learning_rate": 4.105839416058395e-06,
+      "loss": 2.117,
+      "step": 239
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "grad_norm": 2.5236105918884277,
+      "learning_rate": 4.124087591240876e-06,
+      "loss": 2.1958,
+      "step": 240
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "eval_loss": 2.250295400619507,
+      "eval_runtime": 37.0811,
+      "eval_samples_per_second": 26.968,
+      "eval_steps_per_second": 1.133,
+      "step": 240
+    },
+    {
+      "epoch": 0.2641095890410959,
+      "grad_norm": 2.581918954849243,
+      "learning_rate": 4.1423357664233586e-06,
+      "loss": 2.1054,
+      "step": 241
+    },
+    {
+      "epoch": 0.2652054794520548,
+      "grad_norm": 2.501332998275757,
+      "learning_rate": 4.16058394160584e-06,
+      "loss": 2.0878,
+      "step": 242
+    },
+    {
+      "epoch": 0.2663013698630137,
+      "grad_norm": 2.6477770805358887,
+      "learning_rate": 4.1788321167883214e-06,
+      "loss": 2.1656,
+      "step": 243
+    },
+    {
+      "epoch": 0.2673972602739726,
+      "grad_norm": 2.7252438068389893,
+      "learning_rate": 4.197080291970803e-06,
+      "loss": 2.0497,
+      "step": 244
+    },
+    {
+      "epoch": 0.2684931506849315,
+      "grad_norm": 2.6056301593780518,
+      "learning_rate": 4.215328467153285e-06,
+      "loss": 2.1411,
+      "step": 245
+    },
+    {
+      "epoch": 0.26958904109589044,
+      "grad_norm": 2.6581931114196777,
+      "learning_rate": 4.233576642335767e-06,
+      "loss": 2.0829,
+      "step": 246
+    },
+    {
+      "epoch": 0.2706849315068493,
+      "grad_norm": 2.6316752433776855,
+      "learning_rate": 4.251824817518249e-06,
+      "loss": 2.0882,
+      "step": 247
+    },
+    {
+      "epoch": 0.2717808219178082,
+      "grad_norm": 2.803823947906494,
+      "learning_rate": 4.27007299270073e-06,
+      "loss": 2.1203,
+      "step": 248
+    },
+    {
+      "epoch": 0.27287671232876715,
+      "grad_norm": 2.9551174640655518,
+      "learning_rate": 4.288321167883212e-06,
+      "loss": 2.0897,
+      "step": 249
+    },
+    {
+      "epoch": 0.273972602739726,
+      "grad_norm": 2.4847161769866943,
+      "learning_rate": 4.306569343065693e-06,
+      "loss": 2.0845,
+      "step": 250
+    },
+    {
+      "epoch": 0.273972602739726,
+      "eval_loss": 2.2491259574890137,
+      "eval_runtime": 37.1024,
+      "eval_samples_per_second": 26.952,
+      "eval_steps_per_second": 1.132,
+      "step": 250
+    },
+    {
+      "epoch": 0.27506849315068493,
+      "grad_norm": 2.6398351192474365,
+      "learning_rate": 4.3248175182481755e-06,
+      "loss": 2.0479,
+      "step": 251
+    },
+    {
+      "epoch": 0.27616438356164386,
+      "grad_norm": 3.0948562622070312,
+      "learning_rate": 4.343065693430658e-06,
+      "loss": 2.1184,
+      "step": 252
+    },
+    {
+      "epoch": 0.2772602739726027,
+      "grad_norm": 2.5378012657165527,
+      "learning_rate": 4.361313868613139e-06,
+      "loss": 2.1099,
+      "step": 253
+    },
+    {
+      "epoch": 0.27835616438356164,
+      "grad_norm": 3.0282225608825684,
+      "learning_rate": 4.379562043795621e-06,
+      "loss": 2.0982,
+      "step": 254
+    },
+    {
+      "epoch": 0.27945205479452057,
+      "grad_norm": 2.6205334663391113,
+      "learning_rate": 4.397810218978102e-06,
+      "loss": 2.1285,
+      "step": 255
+    },
+    {
+      "epoch": 0.28054794520547943,
+      "grad_norm": 3.2056095600128174,
+      "learning_rate": 4.416058394160584e-06,
+      "loss": 2.1221,
+      "step": 256
+    },
+    {
+      "epoch": 0.28164383561643835,
+      "grad_norm": 3.1370127201080322,
+      "learning_rate": 4.434306569343066e-06,
+      "loss": 2.1016,
+      "step": 257
+    },
+    {
+      "epoch": 0.2827397260273973,
+      "grad_norm": 2.494168519973755,
+      "learning_rate": 4.452554744525548e-06,
+      "loss": 2.0754,
+      "step": 258
+    },
+    {
+      "epoch": 0.28383561643835614,
+      "grad_norm": 2.7979183197021484,
+      "learning_rate": 4.4708029197080295e-06,
+      "loss": 2.1218,
+      "step": 259
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "grad_norm": 2.755682945251465,
+      "learning_rate": 4.489051094890512e-06,
+      "loss": 2.1136,
+      "step": 260
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "eval_loss": 2.2433817386627197,
+      "eval_runtime": 37.1121,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 260
+    },
+    {
+      "epoch": 0.286027397260274,
+      "grad_norm": 2.3641891479492188,
+      "learning_rate": 4.507299270072993e-06,
+      "loss": 2.1488,
+      "step": 261
+    },
+    {
+      "epoch": 0.28712328767123285,
+      "grad_norm": 2.741292953491211,
+      "learning_rate": 4.525547445255475e-06,
+      "loss": 2.1048,
+      "step": 262
+    },
+    {
+      "epoch": 0.2882191780821918,
+      "grad_norm": 2.811483383178711,
+      "learning_rate": 4.543795620437956e-06,
+      "loss": 2.0996,
+      "step": 263
+    },
+    {
+      "epoch": 0.2893150684931507,
+      "grad_norm": 2.6705479621887207,
+      "learning_rate": 4.562043795620438e-06,
+      "loss": 2.0652,
+      "step": 264
+    },
+    {
+      "epoch": 0.29041095890410956,
+      "grad_norm": 2.9383747577667236,
+      "learning_rate": 4.580291970802921e-06,
+      "loss": 2.1574,
+      "step": 265
+    },
+    {
+      "epoch": 0.2915068493150685,
+      "grad_norm": 2.591373920440674,
+      "learning_rate": 4.598540145985402e-06,
+      "loss": 2.1655,
+      "step": 266
+    },
+    {
+      "epoch": 0.2926027397260274,
+      "grad_norm": 3.16955828666687,
+      "learning_rate": 4.6167883211678835e-06,
+      "loss": 2.1285,
+      "step": 267
+    },
+    {
+      "epoch": 0.2936986301369863,
+      "grad_norm": 2.66941499710083,
+      "learning_rate": 4.635036496350365e-06,
+      "loss": 2.0777,
+      "step": 268
+    },
+    {
+      "epoch": 0.2947945205479452,
+      "grad_norm": 2.768139600753784,
+      "learning_rate": 4.653284671532847e-06,
+      "loss": 2.0748,
+      "step": 269
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "grad_norm": 2.5951924324035645,
+      "learning_rate": 4.671532846715329e-06,
+      "loss": 2.1289,
+      "step": 270
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "eval_loss": 2.2393550872802734,
+      "eval_runtime": 37.0993,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 270
+    },
+    {
+      "epoch": 0.29698630136986304,
+      "grad_norm": 2.5422160625457764,
+      "learning_rate": 4.689781021897811e-06,
+      "loss": 2.1342,
+      "step": 271
+    },
+    {
+      "epoch": 0.2980821917808219,
+      "grad_norm": 2.540555477142334,
+      "learning_rate": 4.708029197080292e-06,
+      "loss": 2.097,
+      "step": 272
+    },
+    {
+      "epoch": 0.2991780821917808,
+      "grad_norm": 2.995366334915161,
+      "learning_rate": 4.726277372262774e-06,
+      "loss": 2.1452,
+      "step": 273
+    },
+    {
+      "epoch": 0.30027397260273975,
+      "grad_norm": 2.2561020851135254,
+      "learning_rate": 4.744525547445255e-06,
+      "loss": 2.1069,
+      "step": 274
+    },
+    {
+      "epoch": 0.3013698630136986,
+      "grad_norm": 2.53476619720459,
+      "learning_rate": 4.7627737226277375e-06,
+      "loss": 2.0381,
+      "step": 275
+    },
+    {
+      "epoch": 0.30246575342465754,
+      "grad_norm": 2.3415749073028564,
+      "learning_rate": 4.78102189781022e-06,
+      "loss": 2.1027,
+      "step": 276
+    },
+    {
+      "epoch": 0.30356164383561646,
+      "grad_norm": 2.9756431579589844,
+      "learning_rate": 4.799270072992701e-06,
+      "loss": 2.1136,
+      "step": 277
+    },
+    {
+      "epoch": 0.3046575342465753,
+      "grad_norm": 2.606208324432373,
+      "learning_rate": 4.8175182481751835e-06,
+      "loss": 2.0797,
+      "step": 278
+    },
+    {
+      "epoch": 0.30575342465753425,
+      "grad_norm": 2.5089313983917236,
+      "learning_rate": 4.835766423357665e-06,
+      "loss": 2.1026,
+      "step": 279
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "grad_norm": 2.5003838539123535,
+      "learning_rate": 4.854014598540146e-06,
+      "loss": 2.0923,
+      "step": 280
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "eval_loss": 2.2351036071777344,
+      "eval_runtime": 37.111,
+      "eval_samples_per_second": 26.946,
+      "eval_steps_per_second": 1.132,
+      "step": 280
+    },
+    {
+      "epoch": 0.30794520547945203,
+      "grad_norm": 2.403602361679077,
+      "learning_rate": 4.872262773722628e-06,
+      "loss": 2.0388,
+      "step": 281
+    },
+    {
+      "epoch": 0.30904109589041096,
+      "grad_norm": 2.7775959968566895,
+      "learning_rate": 4.89051094890511e-06,
+      "loss": 2.0561,
+      "step": 282
+    },
+    {
+      "epoch": 0.3101369863013699,
+      "grad_norm": 2.7128653526306152,
+      "learning_rate": 4.9087591240875915e-06,
+      "loss": 2.042,
+      "step": 283
+    },
+    {
+      "epoch": 0.31123287671232874,
+      "grad_norm": 2.7029991149902344,
+      "learning_rate": 4.927007299270074e-06,
+      "loss": 2.0814,
+      "step": 284
+    },
+    {
+      "epoch": 0.31232876712328766,
+      "grad_norm": 2.3942933082580566,
+      "learning_rate": 4.945255474452555e-06,
+      "loss": 2.0906,
+      "step": 285
+    },
+    {
+      "epoch": 0.3134246575342466,
+      "grad_norm": 2.5086541175842285,
+      "learning_rate": 4.963503649635037e-06,
+      "loss": 2.0625,
+      "step": 286
+    },
+    {
+      "epoch": 0.31452054794520545,
+      "grad_norm": 2.6668760776519775,
+      "learning_rate": 4.981751824817518e-06,
+      "loss": 2.1538,
+      "step": 287
+    },
+    {
+      "epoch": 0.3156164383561644,
+      "grad_norm": 2.5066750049591064,
+      "learning_rate": 5e-06,
+      "loss": 2.1423,
+      "step": 288
+    },
+    {
+      "epoch": 0.3167123287671233,
+      "grad_norm": 2.84967041015625,
+      "learning_rate": 4.997969130787978e-06,
+      "loss": 2.0943,
+      "step": 289
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "grad_norm": 2.151946544647217,
+      "learning_rate": 4.995938261575955e-06,
+      "loss": 2.1098,
+      "step": 290
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "eval_loss": 2.223247528076172,
+      "eval_runtime": 37.1755,
+      "eval_samples_per_second": 26.899,
+      "eval_steps_per_second": 1.13,
+      "step": 290
+    },
+    {
+      "epoch": 0.3189041095890411,
+      "grad_norm": 2.887166976928711,
+      "learning_rate": 4.993907392363932e-06,
+      "loss": 2.0281,
+      "step": 291
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 2.5930211544036865,
+      "learning_rate": 4.9918765231519095e-06,
+      "loss": 2.1063,
+      "step": 292
+    },
+    {
+      "epoch": 0.32109589041095893,
+      "grad_norm": 2.5568413734436035,
+      "learning_rate": 4.989845653939887e-06,
+      "loss": 2.0313,
+      "step": 293
+    },
+    {
+      "epoch": 0.3221917808219178,
+      "grad_norm": 2.536095380783081,
+      "learning_rate": 4.987814784727864e-06,
+      "loss": 2.0749,
+      "step": 294
+    },
+    {
+      "epoch": 0.3232876712328767,
+      "grad_norm": 2.3385651111602783,
+      "learning_rate": 4.985783915515841e-06,
+      "loss": 2.0666,
+      "step": 295
+    },
+    {
+      "epoch": 0.32438356164383564,
+      "grad_norm": 2.504826068878174,
+      "learning_rate": 4.983753046303819e-06,
+      "loss": 2.0677,
+      "step": 296
+    },
+    {
+      "epoch": 0.3254794520547945,
+      "grad_norm": 2.5289409160614014,
+      "learning_rate": 4.981722177091796e-06,
+      "loss": 2.0828,
+      "step": 297
+    },
+    {
+      "epoch": 0.3265753424657534,
+      "grad_norm": 2.803795099258423,
+      "learning_rate": 4.979691307879773e-06,
+      "loss": 2.0733,
+      "step": 298
+    },
+    {
+      "epoch": 0.32767123287671235,
+      "grad_norm": 2.651273012161255,
+      "learning_rate": 4.9776604386677504e-06,
+      "loss": 2.0987,
+      "step": 299
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "grad_norm": 2.4715864658355713,
+      "learning_rate": 4.975629569455728e-06,
+      "loss": 2.0733,
+      "step": 300
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "eval_loss": 2.214858293533325,
+      "eval_runtime": 37.0559,
+      "eval_samples_per_second": 26.986,
+      "eval_steps_per_second": 1.133,
+      "step": 300
+    },
+    {
+      "epoch": 0.32986301369863014,
+      "grad_norm": 2.2573442459106445,
+      "learning_rate": 4.973598700243705e-06,
+      "loss": 2.1291,
+      "step": 301
+    },
+    {
+      "epoch": 0.33095890410958906,
+      "grad_norm": 2.645966053009033,
+      "learning_rate": 4.971567831031681e-06,
+      "loss": 2.046,
+      "step": 302
+    },
+    {
+      "epoch": 0.3320547945205479,
+      "grad_norm": 2.361392021179199,
+      "learning_rate": 4.9695369618196595e-06,
+      "loss": 2.0447,
+      "step": 303
+    },
+    {
+      "epoch": 0.33315068493150685,
+      "grad_norm": 2.272714376449585,
+      "learning_rate": 4.967506092607636e-06,
+      "loss": 1.9272,
+      "step": 304
+    },
+    {
+      "epoch": 0.33424657534246577,
+      "grad_norm": 2.5846943855285645,
+      "learning_rate": 4.965475223395614e-06,
+      "loss": 2.0199,
+      "step": 305
+    },
+    {
+      "epoch": 0.33534246575342463,
+      "grad_norm": 2.38985013961792,
+      "learning_rate": 4.9634443541835905e-06,
+      "loss": 2.0394,
+      "step": 306
+    },
+    {
+      "epoch": 0.33643835616438356,
+      "grad_norm": 2.3526251316070557,
+      "learning_rate": 4.961413484971569e-06,
+      "loss": 2.0403,
+      "step": 307
+    },
+    {
+      "epoch": 0.3375342465753425,
+      "grad_norm": 2.399531602859497,
+      "learning_rate": 4.959382615759545e-06,
+      "loss": 2.0607,
+      "step": 308
+    },
+    {
+      "epoch": 0.33863013698630134,
+      "grad_norm": 2.50803804397583,
+      "learning_rate": 4.957351746547523e-06,
+      "loss": 2.0688,
+      "step": 309
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "grad_norm": 2.278088092803955,
+      "learning_rate": 4.9553208773355e-06,
+      "loss": 2.0453,
+      "step": 310
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "eval_loss": 2.2185235023498535,
+      "eval_runtime": 36.9712,
+      "eval_samples_per_second": 27.048,
+      "eval_steps_per_second": 1.136,
+      "step": 310
+    },
+    {
+      "epoch": 0.3408219178082192,
+      "grad_norm": 2.555300712585449,
+      "learning_rate": 4.953290008123478e-06,
+      "loss": 2.1726,
+      "step": 311
+    },
+    {
+      "epoch": 0.3419178082191781,
+      "grad_norm": 2.614870548248291,
+      "learning_rate": 4.951259138911454e-06,
+      "loss": 2.0659,
+      "step": 312
+    },
+    {
+      "epoch": 0.343013698630137,
+      "grad_norm": 2.1932661533355713,
+      "learning_rate": 4.9492282696994315e-06,
+      "loss": 1.9981,
+      "step": 313
+    },
+    {
+      "epoch": 0.3441095890410959,
+      "grad_norm": 2.707451581954956,
+      "learning_rate": 4.947197400487409e-06,
+      "loss": 2.1321,
+      "step": 314
+    },
+    {
+      "epoch": 0.3452054794520548,
+      "grad_norm": 2.4317946434020996,
+      "learning_rate": 4.945166531275386e-06,
+      "loss": 2.0548,
+      "step": 315
+    },
+    {
+      "epoch": 0.3463013698630137,
+      "grad_norm": 2.639883279800415,
+      "learning_rate": 4.943135662063363e-06,
+      "loss": 2.0512,
+      "step": 316
+    },
+    {
+      "epoch": 0.3473972602739726,
+      "grad_norm": 3.1968822479248047,
+      "learning_rate": 4.9411047928513406e-06,
+      "loss": 2.0611,
+      "step": 317
+    },
+    {
+      "epoch": 0.34849315068493153,
+      "grad_norm": 2.4361510276794434,
+      "learning_rate": 4.939073923639318e-06,
+      "loss": 2.0974,
+      "step": 318
+    },
+    {
+      "epoch": 0.3495890410958904,
+      "grad_norm": 2.8278262615203857,
+      "learning_rate": 4.937043054427295e-06,
+      "loss": 2.0992,
+      "step": 319
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "grad_norm": 3.01247239112854,
+      "learning_rate": 4.935012185215272e-06,
+      "loss": 2.0372,
+      "step": 320
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "eval_loss": 2.206442356109619,
+      "eval_runtime": 36.8797,
+      "eval_samples_per_second": 27.115,
+      "eval_steps_per_second": 1.139,
+      "step": 320
+    },
+    {
+      "epoch": 0.35178082191780824,
+      "grad_norm": 2.3372156620025635,
+      "learning_rate": 4.93298131600325e-06,
+      "loss": 2.0211,
+      "step": 321
+    },
+    {
+      "epoch": 0.3528767123287671,
+      "grad_norm": 2.91377329826355,
+      "learning_rate": 4.930950446791227e-06,
+      "loss": 2.0879,
+      "step": 322
+    },
+    {
+      "epoch": 0.353972602739726,
+      "grad_norm": 2.3207666873931885,
+      "learning_rate": 4.928919577579204e-06,
+      "loss": 2.056,
+      "step": 323
+    },
+    {
+      "epoch": 0.35506849315068495,
+      "grad_norm": 2.3260514736175537,
+      "learning_rate": 4.9268887083671815e-06,
+      "loss": 2.0684,
+      "step": 324
+    },
+    {
+      "epoch": 0.3561643835616438,
+      "grad_norm": 2.3070554733276367,
+      "learning_rate": 4.924857839155159e-06,
+      "loss": 2.0494,
+      "step": 325
+    },
+    {
+      "epoch": 0.35726027397260274,
+      "grad_norm": 2.189871072769165,
+      "learning_rate": 4.922826969943136e-06,
+      "loss": 2.0531,
+      "step": 326
+    },
+    {
+      "epoch": 0.35835616438356166,
+      "grad_norm": 2.553520917892456,
+      "learning_rate": 4.920796100731113e-06,
+      "loss": 2.0749,
+      "step": 327
+    },
+    {
+      "epoch": 0.3594520547945205,
+      "grad_norm": 2.62296199798584,
+      "learning_rate": 4.918765231519091e-06,
+      "loss": 1.9983,
+      "step": 328
+    },
+    {
+      "epoch": 0.36054794520547945,
+      "grad_norm": 2.4612369537353516,
+      "learning_rate": 4.916734362307068e-06,
+      "loss": 2.0641,
+      "step": 329
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "grad_norm": 2.656157970428467,
+      "learning_rate": 4.914703493095045e-06,
+      "loss": 2.0148,
+      "step": 330
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "eval_loss": 2.2054214477539062,
+      "eval_runtime": 36.886,
+      "eval_samples_per_second": 27.111,
+      "eval_steps_per_second": 1.139,
+      "step": 330
+    },
+    {
+      "epoch": 0.36273972602739724,
+      "grad_norm": 2.2823736667633057,
+      "learning_rate": 4.912672623883022e-06,
+      "loss": 2.0221,
+      "step": 331
+    },
+    {
+      "epoch": 0.36383561643835616,
+      "grad_norm": 2.8403289318084717,
+      "learning_rate": 4.910641754671e-06,
+      "loss": 2.0149,
+      "step": 332
+    },
+    {
+      "epoch": 0.3649315068493151,
+      "grad_norm": 2.453887462615967,
+      "learning_rate": 4.908610885458977e-06,
+      "loss": 2.0411,
+      "step": 333
+    },
+    {
+      "epoch": 0.36602739726027395,
+      "grad_norm": 2.6094300746917725,
+      "learning_rate": 4.906580016246954e-06,
+      "loss": 2.0293,
+      "step": 334
+    },
+    {
+      "epoch": 0.36712328767123287,
+      "grad_norm": 2.4412384033203125,
+      "learning_rate": 4.9045491470349315e-06,
+      "loss": 2.0778,
+      "step": 335
+    },
+    {
+      "epoch": 0.3682191780821918,
+      "grad_norm": 2.4435057640075684,
+      "learning_rate": 4.902518277822909e-06,
+      "loss": 2.0597,
+      "step": 336
+    },
+    {
+      "epoch": 0.3693150684931507,
+      "grad_norm": 2.2390244007110596,
+      "learning_rate": 4.900487408610886e-06,
+      "loss": 2.0654,
+      "step": 337
+    },
+    {
+      "epoch": 0.3704109589041096,
+      "grad_norm": 2.4243953227996826,
+      "learning_rate": 4.898456539398863e-06,
+      "loss": 2.0696,
+      "step": 338
+    },
+    {
+      "epoch": 0.3715068493150685,
+      "grad_norm": 2.375720262527466,
+      "learning_rate": 4.89642567018684e-06,
+      "loss": 2.0866,
+      "step": 339
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "grad_norm": 2.6379668712615967,
+      "learning_rate": 4.894394800974818e-06,
+      "loss": 2.1552,
+      "step": 340
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "eval_loss": 2.196237802505493,
+      "eval_runtime": 36.8953,
+      "eval_samples_per_second": 27.104,
+      "eval_steps_per_second": 1.138,
+      "step": 340
+    },
+    {
+      "epoch": 0.3736986301369863,
+      "grad_norm": 2.417145252227783,
+      "learning_rate": 4.892363931762794e-06,
+      "loss": 2.0024,
+      "step": 341
+    },
+    {
+      "epoch": 0.3747945205479452,
+      "grad_norm": 2.234590768814087,
+      "learning_rate": 4.8903330625507725e-06,
+      "loss": 2.0837,
+      "step": 342
+    },
+    {
+      "epoch": 0.37589041095890413,
+      "grad_norm": 2.404266595840454,
+      "learning_rate": 4.888302193338749e-06,
+      "loss": 2.018,
+      "step": 343
+    },
+    {
+      "epoch": 0.376986301369863,
+      "grad_norm": 2.364802837371826,
+      "learning_rate": 4.886271324126727e-06,
+      "loss": 1.994,
+      "step": 344
+    },
+    {
+      "epoch": 0.3780821917808219,
+      "grad_norm": 2.2262251377105713,
+      "learning_rate": 4.8842404549147034e-06,
+      "loss": 2.0929,
+      "step": 345
+    },
+    {
+      "epoch": 0.37917808219178084,
+      "grad_norm": 2.4854862689971924,
+      "learning_rate": 4.8822095857026816e-06,
+      "loss": 2.0801,
+      "step": 346
+    },
+    {
+      "epoch": 0.3802739726027397,
+      "grad_norm": 2.3861470222473145,
+      "learning_rate": 4.880178716490658e-06,
+      "loss": 2.0496,
+      "step": 347
+    },
+    {
+      "epoch": 0.38136986301369863,
+      "grad_norm": 2.372326612472534,
+      "learning_rate": 4.878147847278636e-06,
+      "loss": 1.9939,
+      "step": 348
+    },
+    {
+      "epoch": 0.38246575342465755,
+      "grad_norm": 2.3593149185180664,
+      "learning_rate": 4.8761169780666125e-06,
+      "loss": 2.0108,
+      "step": 349
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "grad_norm": 2.1802141666412354,
+      "learning_rate": 4.874086108854591e-06,
+      "loss": 1.983,
+      "step": 350
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "eval_loss": 2.1947567462921143,
+      "eval_runtime": 36.8798,
+      "eval_samples_per_second": 27.115,
+      "eval_steps_per_second": 1.139,
+      "step": 350
+    },
+    {
+      "epoch": 0.38465753424657534,
+      "grad_norm": 2.47067928314209,
+      "learning_rate": 4.872055239642567e-06,
+      "loss": 2.0002,
+      "step": 351
+    },
+    {
+      "epoch": 0.38575342465753426,
+      "grad_norm": 2.359609365463257,
+      "learning_rate": 4.870024370430544e-06,
+      "loss": 1.9799,
+      "step": 352
+    },
+    {
+      "epoch": 0.3868493150684931,
+      "grad_norm": 2.3711702823638916,
+      "learning_rate": 4.867993501218522e-06,
+      "loss": 2.0341,
+      "step": 353
+    },
+    {
+      "epoch": 0.38794520547945205,
+      "grad_norm": 2.246647834777832,
+      "learning_rate": 4.865962632006499e-06,
+      "loss": 2.0687,
+      "step": 354
+    },
+    {
+      "epoch": 0.38904109589041097,
+      "grad_norm": 2.3031938076019287,
+      "learning_rate": 4.863931762794476e-06,
+      "loss": 1.9994,
+      "step": 355
+    },
+    {
+      "epoch": 0.39013698630136984,
+      "grad_norm": 2.492159605026245,
+      "learning_rate": 4.8619008935824535e-06,
+      "loss": 2.0195,
+      "step": 356
+    },
+    {
+      "epoch": 0.39123287671232876,
+      "grad_norm": 2.2878570556640625,
+      "learning_rate": 4.859870024370431e-06,
+      "loss": 2.0589,
+      "step": 357
+    },
+    {
+      "epoch": 0.3923287671232877,
+      "grad_norm": 2.444096088409424,
+      "learning_rate": 4.857839155158408e-06,
+      "loss": 2.0796,
+      "step": 358
+    },
+    {
+      "epoch": 0.3934246575342466,
+      "grad_norm": 2.463857650756836,
+      "learning_rate": 4.855808285946385e-06,
+      "loss": 2.0824,
+      "step": 359
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "grad_norm": 2.0893168449401855,
+      "learning_rate": 4.8537774167343626e-06,
+      "loss": 2.0651,
+      "step": 360
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "eval_loss": 2.188523054122925,
+      "eval_runtime": 36.8679,
+      "eval_samples_per_second": 27.124,
+      "eval_steps_per_second": 1.139,
+      "step": 360
+    },
+    {
+      "epoch": 0.3956164383561644,
+      "grad_norm": 2.602850914001465,
+      "learning_rate": 4.85174654752234e-06,
+      "loss": 2.0507,
+      "step": 361
+    },
+    {
+      "epoch": 0.3967123287671233,
+      "grad_norm": 2.4185140132904053,
+      "learning_rate": 4.849715678310317e-06,
+      "loss": 2.0386,
+      "step": 362
+    },
+    {
+      "epoch": 0.3978082191780822,
+      "grad_norm": 2.1492245197296143,
+      "learning_rate": 4.847684809098294e-06,
+      "loss": 2.0376,
+      "step": 363
+    },
+    {
+      "epoch": 0.3989041095890411,
+      "grad_norm": 2.5710153579711914,
+      "learning_rate": 4.845653939886272e-06,
+      "loss": 2.0383,
+      "step": 364
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 2.4986608028411865,
+      "learning_rate": 4.843623070674249e-06,
+      "loss": 2.0464,
+      "step": 365
+    },
+    {
+      "epoch": 0.4010958904109589,
+      "grad_norm": 2.260388135910034,
+      "learning_rate": 4.841592201462226e-06,
+      "loss": 2.0217,
+      "step": 366
+    },
+    {
+      "epoch": 0.4021917808219178,
+      "grad_norm": 2.6500394344329834,
+      "learning_rate": 4.8395613322502035e-06,
+      "loss": 2.0777,
+      "step": 367
+    },
+    {
+      "epoch": 0.40328767123287673,
+      "grad_norm": 2.159274101257324,
+      "learning_rate": 4.837530463038181e-06,
+      "loss": 1.9747,
+      "step": 368
+    },
+    {
+      "epoch": 0.4043835616438356,
+      "grad_norm": 2.4623043537139893,
+      "learning_rate": 4.835499593826158e-06,
+      "loss": 2.039,
+      "step": 369
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "grad_norm": 2.2265007495880127,
+      "learning_rate": 4.833468724614135e-06,
+      "loss": 2.0618,
+      "step": 370
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "eval_loss": 2.184051275253296,
+      "eval_runtime": 36.9284,
+      "eval_samples_per_second": 27.079,
+      "eval_steps_per_second": 1.137,
+      "step": 370
+    },
+    {
+      "epoch": 0.40657534246575344,
+      "grad_norm": 2.300276517868042,
+      "learning_rate": 4.831437855402113e-06,
+      "loss": 2.0382,
+      "step": 371
+    },
+    {
+      "epoch": 0.4076712328767123,
+      "grad_norm": 2.341156482696533,
+      "learning_rate": 4.82940698619009e-06,
+      "loss": 1.9361,
+      "step": 372
+    },
+    {
+      "epoch": 0.40876712328767123,
+      "grad_norm": 2.4196999073028564,
+      "learning_rate": 4.827376116978067e-06,
+      "loss": 2.0087,
+      "step": 373
+    },
+    {
+      "epoch": 0.40986301369863015,
+      "grad_norm": 2.4650514125823975,
+      "learning_rate": 4.8253452477660444e-06,
+      "loss": 2.0687,
+      "step": 374
+    },
+    {
+      "epoch": 0.410958904109589,
+      "grad_norm": 2.393739700317383,
+      "learning_rate": 4.823314378554022e-06,
+      "loss": 2.0421,
+      "step": 375
+    },
+    {
+      "epoch": 0.41205479452054794,
+      "grad_norm": 2.2436327934265137,
+      "learning_rate": 4.821283509341998e-06,
+      "loss": 2.0358,
+      "step": 376
+    },
+    {
+      "epoch": 0.41315068493150686,
+      "grad_norm": 2.1801650524139404,
+      "learning_rate": 4.819252640129976e-06,
+      "loss": 2.0081,
+      "step": 377
+    },
+    {
+      "epoch": 0.41424657534246573,
+      "grad_norm": 2.300882339477539,
+      "learning_rate": 4.817221770917953e-06,
+      "loss": 2.0321,
+      "step": 378
+    },
+    {
+      "epoch": 0.41534246575342465,
+      "grad_norm": 2.2136058807373047,
+      "learning_rate": 4.815190901705931e-06,
+      "loss": 1.9999,
+      "step": 379
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "grad_norm": 2.168034076690674,
+      "learning_rate": 4.813160032493907e-06,
+      "loss": 2.0336,
+      "step": 380
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "eval_loss": 2.1770179271698,
+      "eval_runtime": 37.0015,
+      "eval_samples_per_second": 27.026,
+      "eval_steps_per_second": 1.135,
+      "step": 380
+    },
+    {
+      "epoch": 0.41753424657534244,
+      "grad_norm": 2.1604163646698,
+      "learning_rate": 4.811129163281885e-06,
+      "loss": 2.0212,
+      "step": 381
+    },
+    {
+      "epoch": 0.41863013698630136,
+      "grad_norm": 2.23923659324646,
+      "learning_rate": 4.809098294069862e-06,
+      "loss": 2.0495,
+      "step": 382
+    },
+    {
+      "epoch": 0.4197260273972603,
+      "grad_norm": 2.149949312210083,
+      "learning_rate": 4.80706742485784e-06,
+      "loss": 2.041,
+      "step": 383
+    },
+    {
+      "epoch": 0.4208219178082192,
+      "grad_norm": 2.49120831489563,
+      "learning_rate": 4.805036555645816e-06,
+      "loss": 2.0931,
+      "step": 384
+    },
+    {
+      "epoch": 0.42191780821917807,
+      "grad_norm": 2.2123563289642334,
+      "learning_rate": 4.8030056864337945e-06,
+      "loss": 2.0087,
+      "step": 385
+    },
+    {
+      "epoch": 0.423013698630137,
+      "grad_norm": 2.4235594272613525,
+      "learning_rate": 4.800974817221771e-06,
+      "loss": 2.0656,
+      "step": 386
+    },
+    {
+      "epoch": 0.4241095890410959,
+      "grad_norm": 2.158252239227295,
+      "learning_rate": 4.798943948009749e-06,
+      "loss": 2.0181,
+      "step": 387
+    },
+    {
+      "epoch": 0.4252054794520548,
+      "grad_norm": 2.2412970066070557,
+      "learning_rate": 4.7969130787977254e-06,
+      "loss": 2.0323,
+      "step": 388
+    },
+    {
+      "epoch": 0.4263013698630137,
+      "grad_norm": 2.2520956993103027,
+      "learning_rate": 4.794882209585703e-06,
+      "loss": 2.026,
+      "step": 389
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "grad_norm": 2.168954849243164,
+      "learning_rate": 4.79285134037368e-06,
+      "loss": 2.0459,
+      "step": 390
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "eval_loss": 2.1705472469329834,
+      "eval_runtime": 37.0913,
+      "eval_samples_per_second": 26.96,
+      "eval_steps_per_second": 1.132,
+      "step": 390
+    },
+    {
+      "epoch": 0.4284931506849315,
+      "grad_norm": 2.1397407054901123,
+      "learning_rate": 4.790820471161657e-06,
+      "loss": 1.9995,
+      "step": 391
+    },
+    {
+      "epoch": 0.4295890410958904,
+      "grad_norm": 2.1747751235961914,
+      "learning_rate": 4.7887896019496345e-06,
+      "loss": 2.0088,
+      "step": 392
+    },
+    {
+      "epoch": 0.43068493150684933,
+      "grad_norm": 2.2860991954803467,
+      "learning_rate": 4.786758732737612e-06,
+      "loss": 2.0134,
+      "step": 393
+    },
+    {
+      "epoch": 0.4317808219178082,
+      "grad_norm": 2.2133519649505615,
+      "learning_rate": 4.784727863525589e-06,
+      "loss": 2.0586,
+      "step": 394
+    },
+    {
+      "epoch": 0.4328767123287671,
+      "grad_norm": 2.2147040367126465,
+      "learning_rate": 4.782696994313566e-06,
+      "loss": 1.9748,
+      "step": 395
+    },
+    {
+      "epoch": 0.43397260273972604,
+      "grad_norm": 2.343418836593628,
+      "learning_rate": 4.780666125101544e-06,
+      "loss": 2.0223,
+      "step": 396
+    },
+    {
+      "epoch": 0.4350684931506849,
+      "grad_norm": 2.198935031890869,
+      "learning_rate": 4.778635255889521e-06,
+      "loss": 1.9627,
+      "step": 397
+    },
+    {
+      "epoch": 0.43616438356164383,
+      "grad_norm": 2.1314704418182373,
+      "learning_rate": 4.776604386677498e-06,
+      "loss": 1.9938,
+      "step": 398
+    },
+    {
+      "epoch": 0.43726027397260275,
+      "grad_norm": 2.3005123138427734,
+      "learning_rate": 4.7745735174654755e-06,
+      "loss": 2.0029,
+      "step": 399
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "grad_norm": 2.2826664447784424,
+      "learning_rate": 4.772542648253453e-06,
+      "loss": 2.0585,
+      "step": 400
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "eval_loss": 2.1691415309906006,
+      "eval_runtime": 37.0852,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 400
+    },
+    {
+      "epoch": 0.43945205479452054,
+      "grad_norm": 2.0547878742218018,
+      "learning_rate": 4.77051177904143e-06,
+      "loss": 2.0582,
+      "step": 401
+    },
+    {
+      "epoch": 0.44054794520547946,
+      "grad_norm": 2.200042486190796,
+      "learning_rate": 4.768480909829407e-06,
+      "loss": 1.9936,
+      "step": 402
+    },
+    {
+      "epoch": 0.44164383561643833,
+      "grad_norm": 2.354637861251831,
+      "learning_rate": 4.766450040617385e-06,
+      "loss": 2.0327,
+      "step": 403
+    },
+    {
+      "epoch": 0.44273972602739725,
+      "grad_norm": 2.2332348823547363,
+      "learning_rate": 4.764419171405362e-06,
+      "loss": 2.0052,
+      "step": 404
+    },
+    {
+      "epoch": 0.4438356164383562,
+      "grad_norm": 2.288060426712036,
+      "learning_rate": 4.762388302193339e-06,
+      "loss": 2.0472,
+      "step": 405
+    },
+    {
+      "epoch": 0.4449315068493151,
+      "grad_norm": 2.1368887424468994,
+      "learning_rate": 4.760357432981316e-06,
+      "loss": 2.0305,
+      "step": 406
+    },
+    {
+      "epoch": 0.44602739726027396,
+      "grad_norm": 2.362630844116211,
+      "learning_rate": 4.758326563769294e-06,
+      "loss": 2.002,
+      "step": 407
+    },
+    {
+      "epoch": 0.4471232876712329,
+      "grad_norm": 2.283419370651245,
+      "learning_rate": 4.756295694557271e-06,
+      "loss": 2.0271,
+      "step": 408
+    },
+    {
+      "epoch": 0.4482191780821918,
+      "grad_norm": 2.0400819778442383,
+      "learning_rate": 4.754264825345248e-06,
+      "loss": 2.0256,
+      "step": 409
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "grad_norm": 2.588416814804077,
+      "learning_rate": 4.7522339561332255e-06,
+      "loss": 1.9756,
+      "step": 410
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "eval_loss": 2.17041015625,
+      "eval_runtime": 36.958,
+      "eval_samples_per_second": 27.058,
+      "eval_steps_per_second": 1.136,
+      "step": 410
+    },
+    {
+      "epoch": 0.4504109589041096,
+      "grad_norm": 2.2397494316101074,
+      "learning_rate": 4.750203086921203e-06,
+      "loss": 1.9876,
+      "step": 411
+    },
+    {
+      "epoch": 0.4515068493150685,
+      "grad_norm": 2.2085509300231934,
+      "learning_rate": 4.74817221770918e-06,
+      "loss": 1.9464,
+      "step": 412
+    },
+    {
+      "epoch": 0.4526027397260274,
+      "grad_norm": 2.140974760055542,
+      "learning_rate": 4.7461413484971565e-06,
+      "loss": 2.0179,
+      "step": 413
+    },
+    {
+      "epoch": 0.4536986301369863,
+      "grad_norm": 2.124366044998169,
+      "learning_rate": 4.744110479285135e-06,
+      "loss": 2.0732,
+      "step": 414
+    },
+    {
+      "epoch": 0.4547945205479452,
+      "grad_norm": 2.2463996410369873,
+      "learning_rate": 4.742079610073111e-06,
+      "loss": 1.9616,
+      "step": 415
+    },
+    {
+      "epoch": 0.4558904109589041,
+      "grad_norm": 2.1249961853027344,
+      "learning_rate": 4.740048740861089e-06,
+      "loss": 1.9889,
+      "step": 416
+    },
+    {
+      "epoch": 0.456986301369863,
+      "grad_norm": 2.25842022895813,
+      "learning_rate": 4.738017871649066e-06,
+      "loss": 2.0033,
+      "step": 417
+    },
+    {
+      "epoch": 0.45808219178082193,
+      "grad_norm": 2.271935224533081,
+      "learning_rate": 4.735987002437044e-06,
+      "loss": 2.0148,
+      "step": 418
+    },
+    {
+      "epoch": 0.4591780821917808,
+      "grad_norm": 2.4797163009643555,
+      "learning_rate": 4.73395613322502e-06,
+      "loss": 2.0736,
+      "step": 419
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "grad_norm": 2.1901395320892334,
+      "learning_rate": 4.731925264012998e-06,
+      "loss": 2.0763,
+      "step": 420
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "eval_loss": 2.161660671234131,
+      "eval_runtime": 36.9099,
+      "eval_samples_per_second": 27.093,
+      "eval_steps_per_second": 1.138,
+      "step": 420
+    },
+    {
+      "epoch": 0.46136986301369864,
+      "grad_norm": 2.083536386489868,
+      "learning_rate": 4.729894394800975e-06,
+      "loss": 2.0299,
+      "step": 421
+    },
+    {
+      "epoch": 0.4624657534246575,
+      "grad_norm": 2.1118173599243164,
+      "learning_rate": 4.727863525588953e-06,
+      "loss": 1.9742,
+      "step": 422
+    },
+    {
+      "epoch": 0.46356164383561643,
+      "grad_norm": 2.195802927017212,
+      "learning_rate": 4.725832656376929e-06,
+      "loss": 1.9318,
+      "step": 423
+    },
+    {
+      "epoch": 0.46465753424657535,
+      "grad_norm": 2.318098545074463,
+      "learning_rate": 4.723801787164907e-06,
+      "loss": 1.9875,
+      "step": 424
+    },
+    {
+      "epoch": 0.4657534246575342,
+      "grad_norm": 2.1698496341705322,
+      "learning_rate": 4.721770917952885e-06,
+      "loss": 2.031,
+      "step": 425
+    },
+    {
+      "epoch": 0.46684931506849314,
+      "grad_norm": 2.283935785293579,
+      "learning_rate": 4.719740048740861e-06,
+      "loss": 1.9766,
+      "step": 426
+    },
+    {
+      "epoch": 0.46794520547945206,
+      "grad_norm": 2.4180757999420166,
+      "learning_rate": 4.717709179528839e-06,
+      "loss": 1.9976,
+      "step": 427
+    },
+    {
+      "epoch": 0.469041095890411,
+      "grad_norm": 2.2689030170440674,
+      "learning_rate": 4.715678310316816e-06,
+      "loss": 2.03,
+      "step": 428
+    },
+    {
+      "epoch": 0.47013698630136985,
+      "grad_norm": 2.423708200454712,
+      "learning_rate": 4.713647441104794e-06,
+      "loss": 2.0202,
+      "step": 429
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "grad_norm": 2.184705972671509,
+      "learning_rate": 4.71161657189277e-06,
+      "loss": 1.9655,
+      "step": 430
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "eval_loss": 2.1548638343811035,
+      "eval_runtime": 36.9426,
+      "eval_samples_per_second": 27.069,
+      "eval_steps_per_second": 1.137,
+      "step": 430
+    },
+    {
+      "epoch": 0.4723287671232877,
+      "grad_norm": 2.295851945877075,
+      "learning_rate": 4.709585702680748e-06,
+      "loss": 1.9561,
+      "step": 431
+    },
+    {
+      "epoch": 0.47342465753424656,
+      "grad_norm": 2.1762046813964844,
+      "learning_rate": 4.707554833468725e-06,
+      "loss": 2.0341,
+      "step": 432
+    },
+    {
+      "epoch": 0.4745205479452055,
+      "grad_norm": 2.4684343338012695,
+      "learning_rate": 4.705523964256703e-06,
+      "loss": 2.0898,
+      "step": 433
+    },
+    {
+      "epoch": 0.4756164383561644,
+      "grad_norm": 2.293318271636963,
+      "learning_rate": 4.703493095044679e-06,
+      "loss": 1.9794,
+      "step": 434
+    },
+    {
+      "epoch": 0.4767123287671233,
+      "grad_norm": 2.1884210109710693,
+      "learning_rate": 4.701462225832657e-06,
+      "loss": 1.9979,
+      "step": 435
+    },
+    {
+      "epoch": 0.4778082191780822,
+      "grad_norm": 2.412504196166992,
+      "learning_rate": 4.699431356620634e-06,
+      "loss": 1.9237,
+      "step": 436
+    },
+    {
+      "epoch": 0.4789041095890411,
+      "grad_norm": 2.192824602127075,
+      "learning_rate": 4.697400487408612e-06,
+      "loss": 2.0756,
+      "step": 437
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 2.368100643157959,
+      "learning_rate": 4.695369618196588e-06,
+      "loss": 2.028,
+      "step": 438
+    },
+    {
+      "epoch": 0.4810958904109589,
+      "grad_norm": 2.218454122543335,
+      "learning_rate": 4.693338748984566e-06,
+      "loss": 2.0049,
+      "step": 439
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "grad_norm": 2.356168031692505,
+      "learning_rate": 4.691307879772543e-06,
+      "loss": 1.9958,
+      "step": 440
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "eval_loss": 2.1540727615356445,
+      "eval_runtime": 36.9557,
+      "eval_samples_per_second": 27.059,
+      "eval_steps_per_second": 1.136,
+      "step": 440
+    },
+    {
+      "epoch": 0.4832876712328767,
+      "grad_norm": 2.09956693649292,
+      "learning_rate": 4.68927701056052e-06,
+      "loss": 1.9388,
+      "step": 441
+    },
+    {
+      "epoch": 0.4843835616438356,
+      "grad_norm": 2.5746614933013916,
+      "learning_rate": 4.6872461413484975e-06,
+      "loss": 2.0425,
+      "step": 442
+    },
+    {
+      "epoch": 0.48547945205479454,
+      "grad_norm": 2.2064783573150635,
+      "learning_rate": 4.685215272136475e-06,
+      "loss": 2.0082,
+      "step": 443
+    },
+    {
+      "epoch": 0.4865753424657534,
+      "grad_norm": 2.159029006958008,
+      "learning_rate": 4.683184402924452e-06,
+      "loss": 2.0055,
+      "step": 444
+    },
+    {
+      "epoch": 0.4876712328767123,
+      "grad_norm": 2.5222008228302,
+      "learning_rate": 4.681153533712429e-06,
+      "loss": 2.0007,
+      "step": 445
+    },
+    {
+      "epoch": 0.48876712328767125,
+      "grad_norm": 2.163487672805786,
+      "learning_rate": 4.679122664500407e-06,
+      "loss": 2.0719,
+      "step": 446
+    },
+    {
+      "epoch": 0.4898630136986301,
+      "grad_norm": 2.3383803367614746,
+      "learning_rate": 4.677091795288384e-06,
+      "loss": 1.9623,
+      "step": 447
+    },
+    {
+      "epoch": 0.49095890410958903,
+      "grad_norm": 2.1056675910949707,
+      "learning_rate": 4.675060926076361e-06,
+      "loss": 1.9443,
+      "step": 448
+    },
+    {
+      "epoch": 0.49205479452054796,
+      "grad_norm": 2.145270824432373,
+      "learning_rate": 4.673030056864338e-06,
+      "loss": 2.0469,
+      "step": 449
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "grad_norm": 2.578706741333008,
+      "learning_rate": 4.670999187652316e-06,
+      "loss": 1.9569,
+      "step": 450
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "eval_loss": 2.1527743339538574,
+      "eval_runtime": 36.8961,
+      "eval_samples_per_second": 27.103,
+      "eval_steps_per_second": 1.138,
+      "step": 450
+    },
+    {
+      "epoch": 0.49424657534246574,
+      "grad_norm": 2.1328542232513428,
+      "learning_rate": 4.668968318440293e-06,
+      "loss": 1.9559,
+      "step": 451
+    },
+    {
+      "epoch": 0.49534246575342467,
+      "grad_norm": 2.148820638656616,
+      "learning_rate": 4.66693744922827e-06,
+      "loss": 1.9545,
+      "step": 452
+    },
+    {
+      "epoch": 0.4964383561643836,
+      "grad_norm": 2.250340461730957,
+      "learning_rate": 4.6649065800162475e-06,
+      "loss": 2.0358,
+      "step": 453
+    },
+    {
+      "epoch": 0.49753424657534245,
+      "grad_norm": 2.452397584915161,
+      "learning_rate": 4.662875710804225e-06,
+      "loss": 1.9294,
+      "step": 454
+    },
+    {
+      "epoch": 0.4986301369863014,
+      "grad_norm": 2.140615701675415,
+      "learning_rate": 4.660844841592202e-06,
+      "loss": 2.0084,
+      "step": 455
+    },
+    {
+      "epoch": 0.4997260273972603,
+      "grad_norm": 2.176806926727295,
+      "learning_rate": 4.658813972380179e-06,
+      "loss": 1.9174,
+      "step": 456
+    },
+    {
+      "epoch": 0.5008219178082192,
+      "grad_norm": 2.349031925201416,
+      "learning_rate": 4.656783103168157e-06,
+      "loss": 2.0377,
+      "step": 457
+    },
+    {
+      "epoch": 0.5019178082191781,
+      "grad_norm": 2.436889171600342,
+      "learning_rate": 4.654752233956134e-06,
+      "loss": 1.9133,
+      "step": 458
+    },
+    {
+      "epoch": 0.503013698630137,
+      "grad_norm": 2.116610288619995,
+      "learning_rate": 4.652721364744111e-06,
+      "loss": 1.9993,
+      "step": 459
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "grad_norm": 2.186915636062622,
+      "learning_rate": 4.6506904955320885e-06,
+      "loss": 1.9744,
+      "step": 460
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "eval_loss": 2.1580278873443604,
+      "eval_runtime": 36.9097,
+      "eval_samples_per_second": 27.093,
+      "eval_steps_per_second": 1.138,
+      "step": 460
+    },
+    {
+      "epoch": 0.5052054794520547,
+      "grad_norm": 2.2009212970733643,
+      "learning_rate": 4.648659626320066e-06,
+      "loss": 2.0173,
+      "step": 461
+    },
+    {
+      "epoch": 0.5063013698630137,
+      "grad_norm": 2.117286443710327,
+      "learning_rate": 4.646628757108043e-06,
+      "loss": 2.0238,
+      "step": 462
+    },
+    {
+      "epoch": 0.5073972602739726,
+      "grad_norm": 2.238790273666382,
+      "learning_rate": 4.6445978878960194e-06,
+      "loss": 1.9885,
+      "step": 463
+    },
+    {
+      "epoch": 0.5084931506849315,
+      "grad_norm": 2.3493096828460693,
+      "learning_rate": 4.6425670186839976e-06,
+      "loss": 1.9593,
+      "step": 464
+    },
+    {
+      "epoch": 0.5095890410958904,
+      "grad_norm": 2.3276610374450684,
+      "learning_rate": 4.640536149471974e-06,
+      "loss": 1.9993,
+      "step": 465
+    },
+    {
+      "epoch": 0.5106849315068493,
+      "grad_norm": 2.3671715259552,
+      "learning_rate": 4.638505280259952e-06,
+      "loss": 2.0256,
+      "step": 466
+    },
+    {
+      "epoch": 0.5117808219178083,
+      "grad_norm": 2.251803159713745,
+      "learning_rate": 4.6364744110479285e-06,
+      "loss": 2.0436,
+      "step": 467
+    },
+    {
+      "epoch": 0.5128767123287671,
+      "grad_norm": 2.3023312091827393,
+      "learning_rate": 4.634443541835907e-06,
+      "loss": 1.9968,
+      "step": 468
+    },
+    {
+      "epoch": 0.513972602739726,
+      "grad_norm": 2.3660333156585693,
+      "learning_rate": 4.632412672623883e-06,
+      "loss": 2.0492,
+      "step": 469
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "grad_norm": 2.262179136276245,
+      "learning_rate": 4.630381803411861e-06,
+      "loss": 1.9698,
+      "step": 470
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "eval_loss": 2.1516528129577637,
+      "eval_runtime": 36.7956,
+      "eval_samples_per_second": 27.177,
+      "eval_steps_per_second": 1.141,
+      "step": 470
+    },
+    {
+      "epoch": 0.5161643835616438,
+      "grad_norm": 2.450770616531372,
+      "learning_rate": 4.628350934199838e-06,
+      "loss": 1.8799,
+      "step": 471
+    },
+    {
+      "epoch": 0.5172602739726028,
+      "grad_norm": 2.187458038330078,
+      "learning_rate": 4.626320064987816e-06,
+      "loss": 1.9829,
+      "step": 472
+    },
+    {
+      "epoch": 0.5183561643835617,
+      "grad_norm": 2.1726722717285156,
+      "learning_rate": 4.624289195775792e-06,
+      "loss": 2.012,
+      "step": 473
+    },
+    {
+      "epoch": 0.5194520547945205,
+      "grad_norm": 2.4177136421203613,
+      "learning_rate": 4.62225832656377e-06,
+      "loss": 1.9796,
+      "step": 474
+    },
+    {
+      "epoch": 0.5205479452054794,
+      "grad_norm": 2.2353951930999756,
+      "learning_rate": 4.620227457351747e-06,
+      "loss": 1.99,
+      "step": 475
+    },
+    {
+      "epoch": 0.5216438356164383,
+      "grad_norm": 2.369591236114502,
+      "learning_rate": 4.618196588139724e-06,
+      "loss": 1.9847,
+      "step": 476
+    },
+    {
+      "epoch": 0.5227397260273973,
+      "grad_norm": 2.546703577041626,
+      "learning_rate": 4.616165718927701e-06,
+      "loss": 2.0047,
+      "step": 477
+    },
+    {
+      "epoch": 0.5238356164383562,
+      "grad_norm": 2.3509185314178467,
+      "learning_rate": 4.6141348497156786e-06,
+      "loss": 1.9589,
+      "step": 478
+    },
+    {
+      "epoch": 0.5249315068493151,
+      "grad_norm": 2.3063488006591797,
+      "learning_rate": 4.612103980503656e-06,
+      "loss": 2.045,
+      "step": 479
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "grad_norm": 2.285572052001953,
+      "learning_rate": 4.610073111291633e-06,
+      "loss": 2.0218,
+      "step": 480
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "eval_loss": 2.1472301483154297,
+      "eval_runtime": 36.7689,
+      "eval_samples_per_second": 27.197,
+      "eval_steps_per_second": 1.142,
+      "step": 480
+    },
+    {
+      "epoch": 0.5271232876712328,
+      "grad_norm": 2.22987961769104,
+      "learning_rate": 4.60804224207961e-06,
+      "loss": 1.9955,
+      "step": 481
+    },
+    {
+      "epoch": 0.5282191780821918,
+      "grad_norm": 2.200918197631836,
+      "learning_rate": 4.606011372867588e-06,
+      "loss": 1.9768,
+      "step": 482
+    },
+    {
+      "epoch": 0.5293150684931507,
+      "grad_norm": 2.2441930770874023,
+      "learning_rate": 4.603980503655565e-06,
+      "loss": 2.0088,
+      "step": 483
+    },
+    {
+      "epoch": 0.5304109589041096,
+      "grad_norm": 2.380383014678955,
+      "learning_rate": 4.601949634443542e-06,
+      "loss": 1.9853,
+      "step": 484
+    },
+    {
+      "epoch": 0.5315068493150685,
+      "grad_norm": 2.317995071411133,
+      "learning_rate": 4.5999187652315195e-06,
+      "loss": 1.9183,
+      "step": 485
+    },
+    {
+      "epoch": 0.5326027397260275,
+      "grad_norm": 2.2462668418884277,
+      "learning_rate": 4.597887896019497e-06,
+      "loss": 1.9201,
+      "step": 486
+    },
+    {
+      "epoch": 0.5336986301369863,
+      "grad_norm": 2.1482582092285156,
+      "learning_rate": 4.595857026807474e-06,
+      "loss": 1.9722,
+      "step": 487
+    },
+    {
+      "epoch": 0.5347945205479452,
+      "grad_norm": 2.079831123352051,
+      "learning_rate": 4.593826157595451e-06,
+      "loss": 1.977,
+      "step": 488
+    },
+    {
+      "epoch": 0.5358904109589041,
+      "grad_norm": 2.2169578075408936,
+      "learning_rate": 4.591795288383429e-06,
+      "loss": 1.9615,
+      "step": 489
+    },
+    {
+      "epoch": 0.536986301369863,
+      "grad_norm": 2.087846279144287,
+      "learning_rate": 4.589764419171406e-06,
+      "loss": 2.0402,
+      "step": 490
+    },
+    {
+      "epoch": 0.536986301369863,
+      "eval_loss": 2.1416702270507812,
+      "eval_runtime": 37.1191,
+      "eval_samples_per_second": 26.94,
+      "eval_steps_per_second": 1.131,
+      "step": 490
+    },
+    {
+      "epoch": 0.538082191780822,
+      "grad_norm": 2.2161977291107178,
+      "learning_rate": 4.587733549959383e-06,
+      "loss": 1.9959,
+      "step": 491
+    },
+    {
+      "epoch": 0.5391780821917809,
+      "grad_norm": 2.19530987739563,
+      "learning_rate": 4.5857026807473604e-06,
+      "loss": 1.961,
+      "step": 492
+    },
+    {
+      "epoch": 0.5402739726027397,
+      "grad_norm": 2.1150615215301514,
+      "learning_rate": 4.583671811535338e-06,
+      "loss": 1.946,
+      "step": 493
+    },
+    {
+      "epoch": 0.5413698630136986,
+      "grad_norm": 2.186352491378784,
+      "learning_rate": 4.581640942323315e-06,
+      "loss": 1.9137,
+      "step": 494
+    },
+    {
+      "epoch": 0.5424657534246575,
+      "grad_norm": 2.233370780944824,
+      "learning_rate": 4.579610073111292e-06,
+      "loss": 2.021,
+      "step": 495
+    },
+    {
+      "epoch": 0.5435616438356164,
+      "grad_norm": 2.1326382160186768,
+      "learning_rate": 4.5775792038992695e-06,
+      "loss": 2.0242,
+      "step": 496
+    },
+    {
+      "epoch": 0.5446575342465754,
+      "grad_norm": 2.536837577819824,
+      "learning_rate": 4.575548334687247e-06,
+      "loss": 2.0185,
+      "step": 497
+    },
+    {
+      "epoch": 0.5457534246575343,
+      "grad_norm": 2.193598747253418,
+      "learning_rate": 4.573517465475224e-06,
+      "loss": 2.0533,
+      "step": 498
+    },
+    {
+      "epoch": 0.5468493150684931,
+      "grad_norm": 2.1346311569213867,
+      "learning_rate": 4.571486596263201e-06,
+      "loss": 1.9564,
+      "step": 499
+    },
+    {
+      "epoch": 0.547945205479452,
+      "grad_norm": 2.4294722080230713,
+      "learning_rate": 4.569455727051178e-06,
+      "loss": 2.0003,
+      "step": 500
+    },
+    {
+      "epoch": 0.547945205479452,
+      "eval_loss": 2.1430764198303223,
+      "eval_runtime": 37.0591,
+      "eval_samples_per_second": 26.984,
+      "eval_steps_per_second": 1.133,
+      "step": 500
+    },
+    {
+      "epoch": 0.549041095890411,
+      "grad_norm": 2.2394182682037354,
+      "learning_rate": 4.567424857839156e-06,
+      "loss": 2.0055,
+      "step": 501
+    },
+    {
+      "epoch": 0.5501369863013699,
+      "grad_norm": 2.2543067932128906,
+      "learning_rate": 4.565393988627132e-06,
+      "loss": 1.9837,
+      "step": 502
+    },
+    {
+      "epoch": 0.5512328767123288,
+      "grad_norm": 2.2805464267730713,
+      "learning_rate": 4.5633631194151105e-06,
+      "loss": 2.0495,
+      "step": 503
+    },
+    {
+      "epoch": 0.5523287671232877,
+      "grad_norm": 2.2838146686553955,
+      "learning_rate": 4.561332250203087e-06,
+      "loss": 1.9866,
+      "step": 504
+    },
+    {
+      "epoch": 0.5534246575342465,
+      "grad_norm": 2.286147356033325,
+      "learning_rate": 4.559301380991065e-06,
+      "loss": 1.9805,
+      "step": 505
+    },
+    {
+      "epoch": 0.5545205479452054,
+      "grad_norm": 2.064378023147583,
+      "learning_rate": 4.5572705117790414e-06,
+      "loss": 1.9259,
+      "step": 506
+    },
+    {
+      "epoch": 0.5556164383561644,
+      "grad_norm": 2.4020493030548096,
+      "learning_rate": 4.5552396425670196e-06,
+      "loss": 1.9731,
+      "step": 507
+    },
+    {
+      "epoch": 0.5567123287671233,
+      "grad_norm": 2.11594557762146,
+      "learning_rate": 4.553208773354996e-06,
+      "loss": 2.0126,
+      "step": 508
+    },
+    {
+      "epoch": 0.5578082191780822,
+      "grad_norm": 2.4520673751831055,
+      "learning_rate": 4.551177904142974e-06,
+      "loss": 1.9396,
+      "step": 509
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "grad_norm": 2.302771806716919,
+      "learning_rate": 4.5491470349309505e-06,
+      "loss": 2.0361,
+      "step": 510
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "eval_loss": 2.1408755779266357,
+      "eval_runtime": 36.9881,
+      "eval_samples_per_second": 27.036,
+      "eval_steps_per_second": 1.136,
+      "step": 510
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 2.250927686691284,
+      "learning_rate": 4.547116165718929e-06,
+      "loss": 1.9616,
+      "step": 511
+    },
+    {
+      "epoch": 0.5610958904109589,
+      "grad_norm": 2.4257938861846924,
+      "learning_rate": 4.545085296506905e-06,
+      "loss": 1.9836,
+      "step": 512
+    },
+    {
+      "epoch": 0.5621917808219178,
+      "grad_norm": 2.08693265914917,
+      "learning_rate": 4.543054427294882e-06,
+      "loss": 1.9541,
+      "step": 513
+    },
+    {
+      "epoch": 0.5632876712328767,
+      "grad_norm": 2.133875846862793,
+      "learning_rate": 4.54102355808286e-06,
+      "loss": 1.9377,
+      "step": 514
+    },
+    {
+      "epoch": 0.5643835616438356,
+      "grad_norm": 2.1217594146728516,
+      "learning_rate": 4.538992688870837e-06,
+      "loss": 1.9369,
+      "step": 515
+    },
+    {
+      "epoch": 0.5654794520547946,
+      "grad_norm": 2.117994546890259,
+      "learning_rate": 4.536961819658814e-06,
+      "loss": 1.9666,
+      "step": 516
+    },
+    {
+      "epoch": 0.5665753424657535,
+      "grad_norm": 2.1514523029327393,
+      "learning_rate": 4.5349309504467915e-06,
+      "loss": 1.9704,
+      "step": 517
+    },
+    {
+      "epoch": 0.5676712328767123,
+      "grad_norm": 2.2309377193450928,
+      "learning_rate": 4.532900081234769e-06,
+      "loss": 1.9858,
+      "step": 518
+    },
+    {
+      "epoch": 0.5687671232876712,
+      "grad_norm": 2.100398540496826,
+      "learning_rate": 4.530869212022746e-06,
+      "loss": 1.9573,
+      "step": 519
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "grad_norm": 2.122840166091919,
+      "learning_rate": 4.528838342810723e-06,
+      "loss": 1.9813,
+      "step": 520
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "eval_loss": 2.135481357574463,
+      "eval_runtime": 36.893,
+      "eval_samples_per_second": 27.105,
+      "eval_steps_per_second": 1.138,
+      "step": 520
+    },
+    {
+      "epoch": 0.570958904109589,
+      "grad_norm": 2.1103999614715576,
+      "learning_rate": 4.526807473598701e-06,
+      "loss": 1.9691,
+      "step": 521
+    },
+    {
+      "epoch": 0.572054794520548,
+      "grad_norm": 2.059232473373413,
+      "learning_rate": 4.524776604386678e-06,
+      "loss": 1.9687,
+      "step": 522
+    },
+    {
+      "epoch": 0.5731506849315069,
+      "grad_norm": 2.207000732421875,
+      "learning_rate": 4.522745735174655e-06,
+      "loss": 1.949,
+      "step": 523
+    },
+    {
+      "epoch": 0.5742465753424657,
+      "grad_norm": 2.264000177383423,
+      "learning_rate": 4.520714865962632e-06,
+      "loss": 2.0066,
+      "step": 524
+    },
+    {
+      "epoch": 0.5753424657534246,
+      "grad_norm": 2.121701717376709,
+      "learning_rate": 4.51868399675061e-06,
+      "loss": 1.9879,
+      "step": 525
+    },
+    {
+      "epoch": 0.5764383561643835,
+      "grad_norm": 2.0745644569396973,
+      "learning_rate": 4.516653127538587e-06,
+      "loss": 1.9924,
+      "step": 526
+    },
+    {
+      "epoch": 0.5775342465753425,
+      "grad_norm": 2.2257349491119385,
+      "learning_rate": 4.514622258326564e-06,
+      "loss": 1.9525,
+      "step": 527
+    },
+    {
+      "epoch": 0.5786301369863014,
+      "grad_norm": 2.1678507328033447,
+      "learning_rate": 4.5125913891145415e-06,
+      "loss": 1.9378,
+      "step": 528
+    },
+    {
+      "epoch": 0.5797260273972603,
+      "grad_norm": 2.2800841331481934,
+      "learning_rate": 4.510560519902519e-06,
+      "loss": 1.9685,
+      "step": 529
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "grad_norm": 2.1471869945526123,
+      "learning_rate": 4.508529650690496e-06,
+      "loss": 2.0027,
+      "step": 530
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "eval_loss": 2.1354892253875732,
+      "eval_runtime": 36.8434,
+      "eval_samples_per_second": 27.142,
+      "eval_steps_per_second": 1.14,
+      "step": 530
+    },
+    {
+      "epoch": 0.581917808219178,
+      "grad_norm": 2.2238621711730957,
+      "learning_rate": 4.506498781478473e-06,
+      "loss": 1.9258,
+      "step": 531
+    },
+    {
+      "epoch": 0.583013698630137,
+      "grad_norm": 2.2224624156951904,
+      "learning_rate": 4.504467912266451e-06,
+      "loss": 1.9168,
+      "step": 532
+    },
+    {
+      "epoch": 0.5841095890410959,
+      "grad_norm": 1.9850294589996338,
+      "learning_rate": 4.502437043054428e-06,
+      "loss": 1.9656,
+      "step": 533
+    },
+    {
+      "epoch": 0.5852054794520548,
+      "grad_norm": 2.1303517818450928,
+      "learning_rate": 4.500406173842405e-06,
+      "loss": 2.0349,
+      "step": 534
+    },
+    {
+      "epoch": 0.5863013698630137,
+      "grad_norm": 2.2114832401275635,
+      "learning_rate": 4.4983753046303824e-06,
+      "loss": 1.9878,
+      "step": 535
+    },
+    {
+      "epoch": 0.5873972602739727,
+      "grad_norm": 1.9375886917114258,
+      "learning_rate": 4.49634443541836e-06,
+      "loss": 1.9724,
+      "step": 536
+    },
+    {
+      "epoch": 0.5884931506849315,
+      "grad_norm": 2.185518741607666,
+      "learning_rate": 4.494313566206336e-06,
+      "loss": 1.9584,
+      "step": 537
+    },
+    {
+      "epoch": 0.5895890410958904,
+      "grad_norm": 2.125638484954834,
+      "learning_rate": 4.492282696994314e-06,
+      "loss": 2.0203,
+      "step": 538
+    },
+    {
+      "epoch": 0.5906849315068493,
+      "grad_norm": 1.9456262588500977,
+      "learning_rate": 4.490251827782291e-06,
+      "loss": 1.9523,
+      "step": 539
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "grad_norm": 2.088125705718994,
+      "learning_rate": 4.488220958570269e-06,
+      "loss": 1.9271,
+      "step": 540
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "eval_loss": 2.134958505630493,
+      "eval_runtime": 36.8342,
+      "eval_samples_per_second": 27.149,
+      "eval_steps_per_second": 1.14,
+      "step": 540
+    },
+    {
+      "epoch": 0.5928767123287672,
+      "grad_norm": 2.1423559188842773,
+      "learning_rate": 4.486190089358245e-06,
+      "loss": 1.9467,
+      "step": 541
+    },
+    {
+      "epoch": 0.5939726027397261,
+      "grad_norm": 2.2169737815856934,
+      "learning_rate": 4.484159220146223e-06,
+      "loss": 1.8901,
+      "step": 542
+    },
+    {
+      "epoch": 0.5950684931506849,
+      "grad_norm": 2.419947385787964,
+      "learning_rate": 4.4821283509342e-06,
+      "loss": 2.0064,
+      "step": 543
+    },
+    {
+      "epoch": 0.5961643835616438,
+      "grad_norm": 2.0396435260772705,
+      "learning_rate": 4.480097481722178e-06,
+      "loss": 1.9819,
+      "step": 544
+    },
+    {
+      "epoch": 0.5972602739726027,
+      "grad_norm": 2.210700511932373,
+      "learning_rate": 4.478066612510154e-06,
+      "loss": 1.9578,
+      "step": 545
+    },
+    {
+      "epoch": 0.5983561643835617,
+      "grad_norm": 2.159698009490967,
+      "learning_rate": 4.4760357432981325e-06,
+      "loss": 1.9313,
+      "step": 546
+    },
+    {
+      "epoch": 0.5994520547945206,
+      "grad_norm": 2.125058650970459,
+      "learning_rate": 4.474004874086109e-06,
+      "loss": 1.9549,
+      "step": 547
+    },
+    {
+      "epoch": 0.6005479452054795,
+      "grad_norm": 2.477320432662964,
+      "learning_rate": 4.471974004874087e-06,
+      "loss": 2.0883,
+      "step": 548
+    },
+    {
+      "epoch": 0.6016438356164383,
+      "grad_norm": 2.214901924133301,
+      "learning_rate": 4.4699431356620635e-06,
+      "loss": 1.9184,
+      "step": 549
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "grad_norm": 2.1001312732696533,
+      "learning_rate": 4.467912266450041e-06,
+      "loss": 2.0231,
+      "step": 550
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "eval_loss": 2.1315157413482666,
+      "eval_runtime": 36.8131,
+      "eval_samples_per_second": 27.164,
+      "eval_steps_per_second": 1.141,
+      "step": 550
+    },
+    {
+      "epoch": 0.6038356164383561,
+      "grad_norm": 2.1030538082122803,
+      "learning_rate": 4.465881397238018e-06,
+      "loss": 1.967,
+      "step": 551
+    },
+    {
+      "epoch": 0.6049315068493151,
+      "grad_norm": 2.0390465259552,
+      "learning_rate": 4.463850528025995e-06,
+      "loss": 1.9615,
+      "step": 552
+    },
+    {
+      "epoch": 0.606027397260274,
+      "grad_norm": 2.1331627368927,
+      "learning_rate": 4.4618196588139726e-06,
+      "loss": 1.9667,
+      "step": 553
+    },
+    {
+      "epoch": 0.6071232876712329,
+      "grad_norm": 2.1987671852111816,
+      "learning_rate": 4.45978878960195e-06,
+      "loss": 1.9922,
+      "step": 554
+    },
+    {
+      "epoch": 0.6082191780821918,
+      "grad_norm": 2.127434253692627,
+      "learning_rate": 4.457757920389927e-06,
+      "loss": 1.9913,
+      "step": 555
+    },
+    {
+      "epoch": 0.6093150684931506,
+      "grad_norm": 2.2486324310302734,
+      "learning_rate": 4.455727051177904e-06,
+      "loss": 1.9239,
+      "step": 556
+    },
+    {
+      "epoch": 0.6104109589041096,
+      "grad_norm": 2.127575397491455,
+      "learning_rate": 4.453696181965882e-06,
+      "loss": 1.9521,
+      "step": 557
+    },
+    {
+      "epoch": 0.6115068493150685,
+      "grad_norm": 2.0798254013061523,
+      "learning_rate": 4.451665312753859e-06,
+      "loss": 2.0085,
+      "step": 558
+    },
+    {
+      "epoch": 0.6126027397260274,
+      "grad_norm": 2.287574291229248,
+      "learning_rate": 4.449634443541836e-06,
+      "loss": 1.98,
+      "step": 559
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "grad_norm": 2.2056541442871094,
+      "learning_rate": 4.4476035743298135e-06,
+      "loss": 2.0282,
+      "step": 560
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "eval_loss": 2.1287026405334473,
+      "eval_runtime": 36.7788,
+      "eval_samples_per_second": 27.19,
+      "eval_steps_per_second": 1.142,
+      "step": 560
+    },
+    {
+      "epoch": 0.6147945205479453,
+      "grad_norm": 2.169157028198242,
+      "learning_rate": 4.445572705117791e-06,
+      "loss": 1.9308,
+      "step": 561
+    },
+    {
+      "epoch": 0.6158904109589041,
+      "grad_norm": 2.1035149097442627,
+      "learning_rate": 4.443541835905768e-06,
+      "loss": 1.9507,
+      "step": 562
+    },
+    {
+      "epoch": 0.616986301369863,
+      "grad_norm": 2.1174328327178955,
+      "learning_rate": 4.441510966693745e-06,
+      "loss": 2.0162,
+      "step": 563
+    },
+    {
+      "epoch": 0.6180821917808219,
+      "grad_norm": 2.201641321182251,
+      "learning_rate": 4.439480097481723e-06,
+      "loss": 1.9573,
+      "step": 564
+    },
+    {
+      "epoch": 0.6191780821917808,
+      "grad_norm": 2.1763603687286377,
+      "learning_rate": 4.4374492282697e-06,
+      "loss": 2.0149,
+      "step": 565
+    },
+    {
+      "epoch": 0.6202739726027398,
+      "grad_norm": 2.2866439819335938,
+      "learning_rate": 4.435418359057677e-06,
+      "loss": 1.8836,
+      "step": 566
+    },
+    {
+      "epoch": 0.6213698630136987,
+      "grad_norm": 2.2060160636901855,
+      "learning_rate": 4.433387489845654e-06,
+      "loss": 1.9806,
+      "step": 567
+    },
+    {
+      "epoch": 0.6224657534246575,
+      "grad_norm": 2.0614662170410156,
+      "learning_rate": 4.431356620633632e-06,
+      "loss": 1.9821,
+      "step": 568
+    },
+    {
+      "epoch": 0.6235616438356164,
+      "grad_norm": 2.1055402755737305,
+      "learning_rate": 4.429325751421609e-06,
+      "loss": 1.9365,
+      "step": 569
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "grad_norm": 1.9907679557800293,
+      "learning_rate": 4.427294882209586e-06,
+      "loss": 1.9429,
+      "step": 570
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "eval_loss": 2.1286206245422363,
+      "eval_runtime": 36.7748,
+      "eval_samples_per_second": 27.193,
+      "eval_steps_per_second": 1.142,
+      "step": 570
+    },
+    {
+      "epoch": 0.6257534246575343,
+      "grad_norm": 2.0169243812561035,
+      "learning_rate": 4.4252640129975635e-06,
+      "loss": 1.9643,
+      "step": 571
+    },
+    {
+      "epoch": 0.6268493150684932,
+      "grad_norm": 2.1082401275634766,
+      "learning_rate": 4.423233143785541e-06,
+      "loss": 1.9009,
+      "step": 572
+    },
+    {
+      "epoch": 0.6279452054794521,
+      "grad_norm": 2.202542543411255,
+      "learning_rate": 4.421202274573518e-06,
+      "loss": 1.9694,
+      "step": 573
+    },
+    {
+      "epoch": 0.6290410958904109,
+      "grad_norm": 2.028489828109741,
+      "learning_rate": 4.4191714053614945e-06,
+      "loss": 1.9697,
+      "step": 574
+    },
+    {
+      "epoch": 0.6301369863013698,
+      "grad_norm": 2.134274482727051,
+      "learning_rate": 4.417140536149473e-06,
+      "loss": 1.8672,
+      "step": 575
+    },
+    {
+      "epoch": 0.6312328767123287,
+      "grad_norm": 2.063542366027832,
+      "learning_rate": 4.415109666937449e-06,
+      "loss": 1.909,
+      "step": 576
+    },
+    {
+      "epoch": 0.6323287671232877,
+      "grad_norm": 1.9747039079666138,
+      "learning_rate": 4.413078797725427e-06,
+      "loss": 1.9795,
+      "step": 577
+    },
+    {
+      "epoch": 0.6334246575342466,
+      "grad_norm": 2.063673496246338,
+      "learning_rate": 4.411047928513404e-06,
+      "loss": 1.9283,
+      "step": 578
+    },
+    {
+      "epoch": 0.6345205479452055,
+      "grad_norm": 2.024611711502075,
+      "learning_rate": 4.409017059301382e-06,
+      "loss": 1.9682,
+      "step": 579
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "grad_norm": 2.0471599102020264,
+      "learning_rate": 4.406986190089358e-06,
+      "loss": 1.969,
+      "step": 580
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "eval_loss": 2.1233692169189453,
+      "eval_runtime": 36.8185,
+      "eval_samples_per_second": 27.16,
+      "eval_steps_per_second": 1.141,
+      "step": 580
+    },
+    {
+      "epoch": 0.6367123287671232,
+      "grad_norm": 2.1038553714752197,
+      "learning_rate": 4.404955320877336e-06,
+      "loss": 1.9915,
+      "step": 581
+    },
+    {
+      "epoch": 0.6378082191780822,
+      "grad_norm": 2.1222152709960938,
+      "learning_rate": 4.402924451665313e-06,
+      "loss": 1.9892,
+      "step": 582
+    },
+    {
+      "epoch": 0.6389041095890411,
+      "grad_norm": 2.09137225151062,
+      "learning_rate": 4.400893582453291e-06,
+      "loss": 1.934,
+      "step": 583
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 2.2158281803131104,
+      "learning_rate": 4.398862713241267e-06,
+      "loss": 1.9188,
+      "step": 584
+    },
+    {
+      "epoch": 0.6410958904109589,
+      "grad_norm": 2.34619140625,
+      "learning_rate": 4.396831844029245e-06,
+      "loss": 1.9495,
+      "step": 585
+    },
+    {
+      "epoch": 0.6421917808219179,
+      "grad_norm": 2.2732315063476562,
+      "learning_rate": 4.394800974817222e-06,
+      "loss": 1.9632,
+      "step": 586
+    },
+    {
+      "epoch": 0.6432876712328767,
+      "grad_norm": 2.153338670730591,
+      "learning_rate": 4.392770105605199e-06,
+      "loss": 1.9551,
+      "step": 587
+    },
+    {
+      "epoch": 0.6443835616438356,
+      "grad_norm": 2.2201521396636963,
+      "learning_rate": 4.390739236393176e-06,
+      "loss": 1.9677,
+      "step": 588
+    },
+    {
+      "epoch": 0.6454794520547945,
+      "grad_norm": 1.9894570112228394,
+      "learning_rate": 4.388708367181154e-06,
+      "loss": 1.9826,
+      "step": 589
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "grad_norm": 2.145061492919922,
+      "learning_rate": 4.386677497969131e-06,
+      "loss": 1.9826,
+      "step": 590
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "eval_loss": 2.122312545776367,
+      "eval_runtime": 37.1153,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 590
+    },
+    {
+      "epoch": 0.6476712328767124,
+      "grad_norm": 2.257237672805786,
+      "learning_rate": 4.384646628757108e-06,
+      "loss": 1.9722,
+      "step": 591
+    },
+    {
+      "epoch": 0.6487671232876713,
+      "grad_norm": 2.2406375408172607,
+      "learning_rate": 4.3826157595450855e-06,
+      "loss": 2.0141,
+      "step": 592
+    },
+    {
+      "epoch": 0.6498630136986301,
+      "grad_norm": 1.9656853675842285,
+      "learning_rate": 4.380584890333063e-06,
+      "loss": 1.9435,
+      "step": 593
+    },
+    {
+      "epoch": 0.650958904109589,
+      "grad_norm": 2.1063120365142822,
+      "learning_rate": 4.37855402112104e-06,
+      "loss": 1.9001,
+      "step": 594
+    },
+    {
+      "epoch": 0.6520547945205479,
+      "grad_norm": 2.1594269275665283,
+      "learning_rate": 4.376523151909017e-06,
+      "loss": 1.8596,
+      "step": 595
+    },
+    {
+      "epoch": 0.6531506849315069,
+      "grad_norm": 2.2261908054351807,
+      "learning_rate": 4.3744922826969946e-06,
+      "loss": 2.016,
+      "step": 596
+    },
+    {
+      "epoch": 0.6542465753424658,
+      "grad_norm": 1.9940093755722046,
+      "learning_rate": 4.372461413484972e-06,
+      "loss": 1.9293,
+      "step": 597
+    },
+    {
+      "epoch": 0.6553424657534247,
+      "grad_norm": 2.62186861038208,
+      "learning_rate": 4.370430544272949e-06,
+      "loss": 1.9656,
+      "step": 598
+    },
+    {
+      "epoch": 0.6564383561643835,
+      "grad_norm": 2.3531494140625,
+      "learning_rate": 4.368399675060926e-06,
+      "loss": 1.9984,
+      "step": 599
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "grad_norm": 2.362769365310669,
+      "learning_rate": 4.366368805848904e-06,
+      "loss": 1.9971,
+      "step": 600
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "eval_loss": 2.1172940731048584,
+      "eval_runtime": 37.0498,
+      "eval_samples_per_second": 26.991,
+      "eval_steps_per_second": 1.134,
+      "step": 600
+    },
+    {
+      "epoch": 0.6586301369863014,
+      "grad_norm": 2.3806607723236084,
+      "learning_rate": 4.364337936636881e-06,
+      "loss": 1.9566,
+      "step": 601
+    },
+    {
+      "epoch": 0.6597260273972603,
+      "grad_norm": 2.5531814098358154,
+      "learning_rate": 4.362307067424858e-06,
+      "loss": 1.9002,
+      "step": 602
+    },
+    {
+      "epoch": 0.6608219178082192,
+      "grad_norm": 2.283364772796631,
+      "learning_rate": 4.3602761982128355e-06,
+      "loss": 1.9633,
+      "step": 603
+    },
+    {
+      "epoch": 0.6619178082191781,
+      "grad_norm": 2.180171489715576,
+      "learning_rate": 4.358245329000813e-06,
+      "loss": 1.9277,
+      "step": 604
+    },
+    {
+      "epoch": 0.663013698630137,
+      "grad_norm": 2.0771350860595703,
+      "learning_rate": 4.35621445978879e-06,
+      "loss": 1.9484,
+      "step": 605
+    },
+    {
+      "epoch": 0.6641095890410958,
+      "grad_norm": 2.3772811889648438,
+      "learning_rate": 4.354183590576767e-06,
+      "loss": 1.9564,
+      "step": 606
+    },
+    {
+      "epoch": 0.6652054794520548,
+      "grad_norm": 2.205486297607422,
+      "learning_rate": 4.352152721364745e-06,
+      "loss": 1.9389,
+      "step": 607
+    },
+    {
+      "epoch": 0.6663013698630137,
+      "grad_norm": 2.085190773010254,
+      "learning_rate": 4.350121852152722e-06,
+      "loss": 1.8948,
+      "step": 608
+    },
+    {
+      "epoch": 0.6673972602739726,
+      "grad_norm": 2.4398982524871826,
+      "learning_rate": 4.348090982940699e-06,
+      "loss": 1.9736,
+      "step": 609
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "grad_norm": 2.194087028503418,
+      "learning_rate": 4.3460601137286764e-06,
+      "loss": 1.9607,
+      "step": 610
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "eval_loss": 2.115041971206665,
+      "eval_runtime": 36.9738,
+      "eval_samples_per_second": 27.046,
+      "eval_steps_per_second": 1.136,
+      "step": 610
+    },
+    {
+      "epoch": 0.6695890410958905,
+      "grad_norm": 2.143054485321045,
+      "learning_rate": 4.344029244516654e-06,
+      "loss": 1.9733,
+      "step": 611
+    },
+    {
+      "epoch": 0.6706849315068493,
+      "grad_norm": 2.2484617233276367,
+      "learning_rate": 4.341998375304631e-06,
+      "loss": 1.9417,
+      "step": 612
+    },
+    {
+      "epoch": 0.6717808219178082,
+      "grad_norm": 2.0430805683135986,
+      "learning_rate": 4.339967506092607e-06,
+      "loss": 2.0192,
+      "step": 613
+    },
+    {
+      "epoch": 0.6728767123287671,
+      "grad_norm": 2.061089277267456,
+      "learning_rate": 4.3379366368805855e-06,
+      "loss": 1.9543,
+      "step": 614
+    },
+    {
+      "epoch": 0.673972602739726,
+      "grad_norm": 2.298022747039795,
+      "learning_rate": 4.335905767668562e-06,
+      "loss": 1.9579,
+      "step": 615
+    },
+    {
+      "epoch": 0.675068493150685,
+      "grad_norm": 2.1418962478637695,
+      "learning_rate": 4.33387489845654e-06,
+      "loss": 1.938,
+      "step": 616
+    },
+    {
+      "epoch": 0.6761643835616439,
+      "grad_norm": 2.1005501747131348,
+      "learning_rate": 4.3318440292445165e-06,
+      "loss": 1.972,
+      "step": 617
+    },
+    {
+      "epoch": 0.6772602739726027,
+      "grad_norm": 2.1665523052215576,
+      "learning_rate": 4.329813160032495e-06,
+      "loss": 1.9535,
+      "step": 618
+    },
+    {
+      "epoch": 0.6783561643835616,
+      "grad_norm": 2.144580841064453,
+      "learning_rate": 4.327782290820471e-06,
+      "loss": 1.9188,
+      "step": 619
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "grad_norm": 1.9855908155441284,
+      "learning_rate": 4.325751421608449e-06,
+      "loss": 1.912,
+      "step": 620
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "eval_loss": 2.118319272994995,
+      "eval_runtime": 36.9677,
+      "eval_samples_per_second": 27.051,
+      "eval_steps_per_second": 1.136,
+      "step": 620
+    },
+    {
+      "epoch": 0.6805479452054795,
+      "grad_norm": 2.395538091659546,
+      "learning_rate": 4.323720552396426e-06,
+      "loss": 1.9886,
+      "step": 621
+    },
+    {
+      "epoch": 0.6816438356164384,
+      "grad_norm": 2.190574884414673,
+      "learning_rate": 4.321689683184404e-06,
+      "loss": 1.9326,
+      "step": 622
+    },
+    {
+      "epoch": 0.6827397260273973,
+      "grad_norm": 2.1826560497283936,
+      "learning_rate": 4.31965881397238e-06,
+      "loss": 1.9278,
+      "step": 623
+    },
+    {
+      "epoch": 0.6838356164383562,
+      "grad_norm": 2.151097536087036,
+      "learning_rate": 4.3176279447603574e-06,
+      "loss": 1.9359,
+      "step": 624
+    },
+    {
+      "epoch": 0.684931506849315,
+      "grad_norm": 2.1830122470855713,
+      "learning_rate": 4.315597075548335e-06,
+      "loss": 1.9728,
+      "step": 625
+    },
+    {
+      "epoch": 0.686027397260274,
+      "grad_norm": 2.203325033187866,
+      "learning_rate": 4.313566206336312e-06,
+      "loss": 1.9638,
+      "step": 626
+    },
+    {
+      "epoch": 0.6871232876712329,
+      "grad_norm": 2.7240309715270996,
+      "learning_rate": 4.311535337124289e-06,
+      "loss": 1.9907,
+      "step": 627
+    },
+    {
+      "epoch": 0.6882191780821918,
+      "grad_norm": 2.133599281311035,
+      "learning_rate": 4.3095044679122665e-06,
+      "loss": 1.9352,
+      "step": 628
+    },
+    {
+      "epoch": 0.6893150684931507,
+      "grad_norm": 2.0318551063537598,
+      "learning_rate": 4.307473598700244e-06,
+      "loss": 1.9432,
+      "step": 629
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "grad_norm": 2.502101182937622,
+      "learning_rate": 4.305442729488221e-06,
+      "loss": 1.9641,
+      "step": 630
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "eval_loss": 2.1158926486968994,
+      "eval_runtime": 36.8818,
+      "eval_samples_per_second": 27.114,
+      "eval_steps_per_second": 1.139,
+      "step": 630
+    },
+    {
+      "epoch": 0.6915068493150685,
+      "grad_norm": 2.3620944023132324,
+      "learning_rate": 4.303411860276198e-06,
+      "loss": 1.8841,
+      "step": 631
+    },
+    {
+      "epoch": 0.6926027397260274,
+      "grad_norm": 2.3139240741729736,
+      "learning_rate": 4.301380991064176e-06,
+      "loss": 1.9896,
+      "step": 632
+    },
+    {
+      "epoch": 0.6936986301369863,
+      "grad_norm": 2.314877510070801,
+      "learning_rate": 4.299350121852153e-06,
+      "loss": 1.9655,
+      "step": 633
+    },
+    {
+      "epoch": 0.6947945205479452,
+      "grad_norm": 2.036371946334839,
+      "learning_rate": 4.29731925264013e-06,
+      "loss": 1.9917,
+      "step": 634
+    },
+    {
+      "epoch": 0.6958904109589041,
+      "grad_norm": 2.1859662532806396,
+      "learning_rate": 4.2952883834281075e-06,
+      "loss": 1.9805,
+      "step": 635
+    },
+    {
+      "epoch": 0.6969863013698631,
+      "grad_norm": 2.2317097187042236,
+      "learning_rate": 4.293257514216085e-06,
+      "loss": 1.9402,
+      "step": 636
+    },
+    {
+      "epoch": 0.6980821917808219,
+      "grad_norm": 2.3249905109405518,
+      "learning_rate": 4.291226645004062e-06,
+      "loss": 1.9013,
+      "step": 637
+    },
+    {
+      "epoch": 0.6991780821917808,
+      "grad_norm": 2.199758768081665,
+      "learning_rate": 4.289195775792039e-06,
+      "loss": 1.9783,
+      "step": 638
+    },
+    {
+      "epoch": 0.7002739726027397,
+      "grad_norm": 2.3598549365997314,
+      "learning_rate": 4.287164906580017e-06,
+      "loss": 1.9936,
+      "step": 639
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "grad_norm": 2.1705522537231445,
+      "learning_rate": 4.285134037367994e-06,
+      "loss": 1.9304,
+      "step": 640
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "eval_loss": 2.1055190563201904,
+      "eval_runtime": 36.8673,
+      "eval_samples_per_second": 27.124,
+      "eval_steps_per_second": 1.139,
+      "step": 640
+    },
+    {
+      "epoch": 0.7024657534246576,
+      "grad_norm": 2.3253612518310547,
+      "learning_rate": 4.283103168155971e-06,
+      "loss": 1.9196,
+      "step": 641
+    },
+    {
+      "epoch": 0.7035616438356165,
+      "grad_norm": 2.3054087162017822,
+      "learning_rate": 4.281072298943948e-06,
+      "loss": 1.9053,
+      "step": 642
+    },
+    {
+      "epoch": 0.7046575342465753,
+      "grad_norm": 2.262057304382324,
+      "learning_rate": 4.279041429731926e-06,
+      "loss": 1.8999,
+      "step": 643
+    },
+    {
+      "epoch": 0.7057534246575342,
+      "grad_norm": 2.421391248703003,
+      "learning_rate": 4.277010560519903e-06,
+      "loss": 1.9438,
+      "step": 644
+    },
+    {
+      "epoch": 0.7068493150684931,
+      "grad_norm": 2.2841498851776123,
+      "learning_rate": 4.27497969130788e-06,
+      "loss": 1.9187,
+      "step": 645
+    },
+    {
+      "epoch": 0.707945205479452,
+      "grad_norm": 2.3226938247680664,
+      "learning_rate": 4.2729488220958575e-06,
+      "loss": 1.9275,
+      "step": 646
+    },
+    {
+      "epoch": 0.709041095890411,
+      "grad_norm": 2.2306301593780518,
+      "learning_rate": 4.270917952883835e-06,
+      "loss": 1.9684,
+      "step": 647
+    },
+    {
+      "epoch": 0.7101369863013699,
+      "grad_norm": 2.0551652908325195,
+      "learning_rate": 4.268887083671812e-06,
+      "loss": 1.9849,
+      "step": 648
+    },
+    {
+      "epoch": 0.7112328767123288,
+      "grad_norm": 2.2952210903167725,
+      "learning_rate": 4.266856214459789e-06,
+      "loss": 1.9026,
+      "step": 649
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "grad_norm": 2.1630797386169434,
+      "learning_rate": 4.264825345247766e-06,
+      "loss": 2.0184,
+      "step": 650
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "eval_loss": 2.1037871837615967,
+      "eval_runtime": 36.8221,
+      "eval_samples_per_second": 27.158,
+      "eval_steps_per_second": 1.141,
+      "step": 650
+    },
+    {
+      "epoch": 0.7134246575342466,
+      "grad_norm": 1.964150309562683,
+      "learning_rate": 4.262794476035744e-06,
+      "loss": 1.8932,
+      "step": 651
+    },
+    {
+      "epoch": 0.7145205479452055,
+      "grad_norm": 2.1474685668945312,
+      "learning_rate": 4.26076360682372e-06,
+      "loss": 1.9383,
+      "step": 652
+    },
+    {
+      "epoch": 0.7156164383561644,
+      "grad_norm": 2.199824094772339,
+      "learning_rate": 4.2587327376116984e-06,
+      "loss": 1.8599,
+      "step": 653
+    },
+    {
+      "epoch": 0.7167123287671233,
+      "grad_norm": 2.12156343460083,
+      "learning_rate": 4.256701868399675e-06,
+      "loss": 1.9162,
+      "step": 654
+    },
+    {
+      "epoch": 0.7178082191780822,
+      "grad_norm": 2.3832154273986816,
+      "learning_rate": 4.254670999187653e-06,
+      "loss": 1.9613,
+      "step": 655
+    },
+    {
+      "epoch": 0.718904109589041,
+      "grad_norm": 2.063906192779541,
+      "learning_rate": 4.252640129975629e-06,
+      "loss": 1.9421,
+      "step": 656
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 2.184842586517334,
+      "learning_rate": 4.2506092607636075e-06,
+      "loss": 1.9471,
+      "step": 657
+    },
+    {
+      "epoch": 0.7210958904109589,
+      "grad_norm": 2.1340036392211914,
+      "learning_rate": 4.248578391551584e-06,
+      "loss": 1.931,
+      "step": 658
+    },
+    {
+      "epoch": 0.7221917808219178,
+      "grad_norm": 2.117544651031494,
+      "learning_rate": 4.246547522339562e-06,
+      "loss": 1.8698,
+      "step": 659
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "grad_norm": 2.308344841003418,
+      "learning_rate": 4.2445166531275385e-06,
+      "loss": 1.9217,
+      "step": 660
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "eval_loss": 2.1086323261260986,
+      "eval_runtime": 36.8082,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 660
+    },
+    {
+      "epoch": 0.7243835616438357,
+      "grad_norm": 2.2239017486572266,
+      "learning_rate": 4.242485783915517e-06,
+      "loss": 1.9286,
+      "step": 661
+    },
+    {
+      "epoch": 0.7254794520547945,
+      "grad_norm": 2.263697624206543,
+      "learning_rate": 4.240454914703493e-06,
+      "loss": 1.8821,
+      "step": 662
+    },
+    {
+      "epoch": 0.7265753424657534,
+      "grad_norm": 2.4305059909820557,
+      "learning_rate": 4.23842404549147e-06,
+      "loss": 1.8996,
+      "step": 663
+    },
+    {
+      "epoch": 0.7276712328767123,
+      "grad_norm": 2.2333695888519287,
+      "learning_rate": 4.236393176279448e-06,
+      "loss": 1.8906,
+      "step": 664
+    },
+    {
+      "epoch": 0.7287671232876712,
+      "grad_norm": 2.160982608795166,
+      "learning_rate": 4.234362307067425e-06,
+      "loss": 1.879,
+      "step": 665
+    },
+    {
+      "epoch": 0.7298630136986302,
+      "grad_norm": 2.3707094192504883,
+      "learning_rate": 4.232331437855402e-06,
+      "loss": 1.9577,
+      "step": 666
+    },
+    {
+      "epoch": 0.7309589041095891,
+      "grad_norm": 2.0502827167510986,
+      "learning_rate": 4.2303005686433795e-06,
+      "loss": 1.9801,
+      "step": 667
+    },
+    {
+      "epoch": 0.7320547945205479,
+      "grad_norm": 2.0648176670074463,
+      "learning_rate": 4.228269699431357e-06,
+      "loss": 1.978,
+      "step": 668
+    },
+    {
+      "epoch": 0.7331506849315068,
+      "grad_norm": 2.346316337585449,
+      "learning_rate": 4.226238830219334e-06,
+      "loss": 1.9387,
+      "step": 669
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "grad_norm": 1.9734236001968384,
+      "learning_rate": 4.224207961007311e-06,
+      "loss": 1.9396,
+      "step": 670
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "eval_loss": 2.1104705333709717,
+      "eval_runtime": 36.8144,
+      "eval_samples_per_second": 27.163,
+      "eval_steps_per_second": 1.141,
+      "step": 670
+    },
+    {
+      "epoch": 0.7353424657534247,
+      "grad_norm": 2.3170549869537354,
+      "learning_rate": 4.2221770917952886e-06,
+      "loss": 1.9398,
+      "step": 671
+    },
+    {
+      "epoch": 0.7364383561643836,
+      "grad_norm": 2.0437264442443848,
+      "learning_rate": 4.220146222583266e-06,
+      "loss": 1.9284,
+      "step": 672
+    },
+    {
+      "epoch": 0.7375342465753425,
+      "grad_norm": 2.0808510780334473,
+      "learning_rate": 4.218115353371243e-06,
+      "loss": 1.9425,
+      "step": 673
+    },
+    {
+      "epoch": 0.7386301369863014,
+      "grad_norm": 2.0854742527008057,
+      "learning_rate": 4.21608448415922e-06,
+      "loss": 1.9358,
+      "step": 674
+    },
+    {
+      "epoch": 0.7397260273972602,
+      "grad_norm": 2.131622552871704,
+      "learning_rate": 4.214053614947198e-06,
+      "loss": 1.9048,
+      "step": 675
+    },
+    {
+      "epoch": 0.7408219178082192,
+      "grad_norm": 2.276359796524048,
+      "learning_rate": 4.212022745735175e-06,
+      "loss": 1.9871,
+      "step": 676
+    },
+    {
+      "epoch": 0.7419178082191781,
+      "grad_norm": 2.2194812297821045,
+      "learning_rate": 4.209991876523152e-06,
+      "loss": 1.882,
+      "step": 677
+    },
+    {
+      "epoch": 0.743013698630137,
+      "grad_norm": 2.181891679763794,
+      "learning_rate": 4.2079610073111295e-06,
+      "loss": 1.9546,
+      "step": 678
+    },
+    {
+      "epoch": 0.7441095890410959,
+      "grad_norm": 2.0956673622131348,
+      "learning_rate": 4.205930138099107e-06,
+      "loss": 1.8836,
+      "step": 679
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "grad_norm": 2.057894229888916,
+      "learning_rate": 4.203899268887084e-06,
+      "loss": 1.9842,
+      "step": 680
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "eval_loss": 2.104182720184326,
+      "eval_runtime": 36.8189,
+      "eval_samples_per_second": 27.16,
+      "eval_steps_per_second": 1.141,
+      "step": 680
+    },
+    {
+      "epoch": 0.7463013698630137,
+      "grad_norm": 2.0965662002563477,
+      "learning_rate": 4.201868399675061e-06,
+      "loss": 1.9378,
+      "step": 681
+    },
+    {
+      "epoch": 0.7473972602739726,
+      "grad_norm": 2.1843457221984863,
+      "learning_rate": 4.199837530463039e-06,
+      "loss": 1.9784,
+      "step": 682
+    },
+    {
+      "epoch": 0.7484931506849315,
+      "grad_norm": 2.2044246196746826,
+      "learning_rate": 4.197806661251016e-06,
+      "loss": 1.8642,
+      "step": 683
+    },
+    {
+      "epoch": 0.7495890410958904,
+      "grad_norm": 2.202115058898926,
+      "learning_rate": 4.195775792038993e-06,
+      "loss": 1.929,
+      "step": 684
+    },
+    {
+      "epoch": 0.7506849315068493,
+      "grad_norm": 2.3671462535858154,
+      "learning_rate": 4.19374492282697e-06,
+      "loss": 1.9172,
+      "step": 685
+    },
+    {
+      "epoch": 0.7517808219178083,
+      "grad_norm": 2.0602803230285645,
+      "learning_rate": 4.191714053614948e-06,
+      "loss": 1.8503,
+      "step": 686
+    },
+    {
+      "epoch": 0.7528767123287671,
+      "grad_norm": 2.12145733833313,
+      "learning_rate": 4.189683184402924e-06,
+      "loss": 1.9511,
+      "step": 687
+    },
+    {
+      "epoch": 0.753972602739726,
+      "grad_norm": 2.1049914360046387,
+      "learning_rate": 4.187652315190902e-06,
+      "loss": 1.9043,
+      "step": 688
+    },
+    {
+      "epoch": 0.7550684931506849,
+      "grad_norm": 2.235497236251831,
+      "learning_rate": 4.185621445978879e-06,
+      "loss": 1.9784,
+      "step": 689
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "grad_norm": 2.173222303390503,
+      "learning_rate": 4.183590576766857e-06,
+      "loss": 1.9052,
+      "step": 690
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "eval_loss": 2.0982251167297363,
+      "eval_runtime": 37.1446,
+      "eval_samples_per_second": 26.922,
+      "eval_steps_per_second": 1.131,
+      "step": 690
+    },
+    {
+      "epoch": 0.7572602739726028,
+      "grad_norm": 2.0693562030792236,
+      "learning_rate": 4.181559707554833e-06,
+      "loss": 1.9223,
+      "step": 691
+    },
+    {
+      "epoch": 0.7583561643835617,
+      "grad_norm": 2.209419012069702,
+      "learning_rate": 4.179528838342811e-06,
+      "loss": 1.9709,
+      "step": 692
+    },
+    {
+      "epoch": 0.7594520547945205,
+      "grad_norm": 2.066041946411133,
+      "learning_rate": 4.177497969130788e-06,
+      "loss": 1.889,
+      "step": 693
+    },
+    {
+      "epoch": 0.7605479452054794,
+      "grad_norm": 2.4114325046539307,
+      "learning_rate": 4.175467099918766e-06,
+      "loss": 1.9059,
+      "step": 694
+    },
+    {
+      "epoch": 0.7616438356164383,
+      "grad_norm": 2.0185043811798096,
+      "learning_rate": 4.173436230706742e-06,
+      "loss": 1.9877,
+      "step": 695
+    },
+    {
+      "epoch": 0.7627397260273973,
+      "grad_norm": 2.23392915725708,
+      "learning_rate": 4.1714053614947205e-06,
+      "loss": 1.9199,
+      "step": 696
+    },
+    {
+      "epoch": 0.7638356164383562,
+      "grad_norm": 2.2158737182617188,
+      "learning_rate": 4.169374492282697e-06,
+      "loss": 1.9001,
+      "step": 697
+    },
+    {
+      "epoch": 0.7649315068493151,
+      "grad_norm": 2.1511690616607666,
+      "learning_rate": 4.167343623070675e-06,
+      "loss": 1.9166,
+      "step": 698
+    },
+    {
+      "epoch": 0.766027397260274,
+      "grad_norm": 2.374741554260254,
+      "learning_rate": 4.165312753858652e-06,
+      "loss": 1.9684,
+      "step": 699
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "grad_norm": 2.0409998893737793,
+      "learning_rate": 4.163281884646629e-06,
+      "loss": 1.9177,
+      "step": 700
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "eval_loss": 2.096895933151245,
+      "eval_runtime": 37.066,
+      "eval_samples_per_second": 26.979,
+      "eval_steps_per_second": 1.133,
+      "step": 700
+    },
+    {
+      "epoch": 0.7682191780821918,
+      "grad_norm": 2.3173422813415527,
+      "learning_rate": 4.161251015434607e-06,
+      "loss": 1.9071,
+      "step": 701
+    },
+    {
+      "epoch": 0.7693150684931507,
+      "grad_norm": 2.077842950820923,
+      "learning_rate": 4.159220146222583e-06,
+      "loss": 1.935,
+      "step": 702
+    },
+    {
+      "epoch": 0.7704109589041096,
+      "grad_norm": 2.0518617630004883,
+      "learning_rate": 4.157189277010561e-06,
+      "loss": 1.8942,
+      "step": 703
+    },
+    {
+      "epoch": 0.7715068493150685,
+      "grad_norm": 1.9950882196426392,
+      "learning_rate": 4.155158407798538e-06,
+      "loss": 1.917,
+      "step": 704
+    },
+    {
+      "epoch": 0.7726027397260274,
+      "grad_norm": 2.3222219944000244,
+      "learning_rate": 4.153127538586516e-06,
+      "loss": 1.9551,
+      "step": 705
+    },
+    {
+      "epoch": 0.7736986301369863,
+      "grad_norm": 2.129333972930908,
+      "learning_rate": 4.151096669374492e-06,
+      "loss": 1.9204,
+      "step": 706
+    },
+    {
+      "epoch": 0.7747945205479452,
+      "grad_norm": 2.0357553958892822,
+      "learning_rate": 4.1490658001624705e-06,
+      "loss": 1.9399,
+      "step": 707
+    },
+    {
+      "epoch": 0.7758904109589041,
+      "grad_norm": 2.233694553375244,
+      "learning_rate": 4.147034930950447e-06,
+      "loss": 1.9182,
+      "step": 708
+    },
+    {
+      "epoch": 0.776986301369863,
+      "grad_norm": 2.1391239166259766,
+      "learning_rate": 4.145004061738425e-06,
+      "loss": 1.9157,
+      "step": 709
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "grad_norm": 2.162320613861084,
+      "learning_rate": 4.1429731925264015e-06,
+      "loss": 1.9115,
+      "step": 710
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "eval_loss": 2.0954039096832275,
+      "eval_runtime": 36.9862,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 710
+    },
+    {
+      "epoch": 0.7791780821917809,
+      "grad_norm": 2.2382194995880127,
+      "learning_rate": 4.140942323314379e-06,
+      "loss": 1.8428,
+      "step": 711
+    },
+    {
+      "epoch": 0.7802739726027397,
+      "grad_norm": 2.3286185264587402,
+      "learning_rate": 4.138911454102356e-06,
+      "loss": 1.8906,
+      "step": 712
+    },
+    {
+      "epoch": 0.7813698630136986,
+      "grad_norm": 2.182708501815796,
+      "learning_rate": 4.136880584890333e-06,
+      "loss": 1.9132,
+      "step": 713
+    },
+    {
+      "epoch": 0.7824657534246575,
+      "grad_norm": 1.9978337287902832,
+      "learning_rate": 4.1348497156783106e-06,
+      "loss": 1.9433,
+      "step": 714
+    },
+    {
+      "epoch": 0.7835616438356164,
+      "grad_norm": 2.3092565536499023,
+      "learning_rate": 4.132818846466288e-06,
+      "loss": 1.9374,
+      "step": 715
+    },
+    {
+      "epoch": 0.7846575342465754,
+      "grad_norm": 2.0351054668426514,
+      "learning_rate": 4.130787977254265e-06,
+      "loss": 1.9361,
+      "step": 716
+    },
+    {
+      "epoch": 0.7857534246575343,
+      "grad_norm": 2.14770770072937,
+      "learning_rate": 4.128757108042242e-06,
+      "loss": 1.9553,
+      "step": 717
+    },
+    {
+      "epoch": 0.7868493150684932,
+      "grad_norm": 2.0613555908203125,
+      "learning_rate": 4.12672623883022e-06,
+      "loss": 1.9506,
+      "step": 718
+    },
+    {
+      "epoch": 0.787945205479452,
+      "grad_norm": 2.1312081813812256,
+      "learning_rate": 4.124695369618197e-06,
+      "loss": 2.0079,
+      "step": 719
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "grad_norm": 2.088552474975586,
+      "learning_rate": 4.122664500406174e-06,
+      "loss": 1.9166,
+      "step": 720
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "eval_loss": 2.093032121658325,
+      "eval_runtime": 36.8821,
+      "eval_samples_per_second": 27.113,
+      "eval_steps_per_second": 1.139,
+      "step": 720
+    },
+    {
+      "epoch": 0.7901369863013699,
+      "grad_norm": 2.1748929023742676,
+      "learning_rate": 4.1206336311941515e-06,
+      "loss": 1.9212,
+      "step": 721
+    },
+    {
+      "epoch": 0.7912328767123288,
+      "grad_norm": 2.1492016315460205,
+      "learning_rate": 4.118602761982129e-06,
+      "loss": 1.9452,
+      "step": 722
+    },
+    {
+      "epoch": 0.7923287671232877,
+      "grad_norm": 2.3364970684051514,
+      "learning_rate": 4.116571892770106e-06,
+      "loss": 1.8821,
+      "step": 723
+    },
+    {
+      "epoch": 0.7934246575342466,
+      "grad_norm": 2.318488597869873,
+      "learning_rate": 4.114541023558083e-06,
+      "loss": 1.8597,
+      "step": 724
+    },
+    {
+      "epoch": 0.7945205479452054,
+      "grad_norm": 2.153841257095337,
+      "learning_rate": 4.112510154346061e-06,
+      "loss": 1.9843,
+      "step": 725
+    },
+    {
+      "epoch": 0.7956164383561644,
+      "grad_norm": 2.3464651107788086,
+      "learning_rate": 4.110479285134038e-06,
+      "loss": 1.9105,
+      "step": 726
+    },
+    {
+      "epoch": 0.7967123287671233,
+      "grad_norm": 2.34413743019104,
+      "learning_rate": 4.108448415922015e-06,
+      "loss": 1.9285,
+      "step": 727
+    },
+    {
+      "epoch": 0.7978082191780822,
+      "grad_norm": 2.0822196006774902,
+      "learning_rate": 4.1064175467099924e-06,
+      "loss": 1.9536,
+      "step": 728
+    },
+    {
+      "epoch": 0.7989041095890411,
+      "grad_norm": 2.1978065967559814,
+      "learning_rate": 4.10438667749797e-06,
+      "loss": 1.9487,
+      "step": 729
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 2.4218783378601074,
+      "learning_rate": 4.102355808285947e-06,
+      "loss": 1.8113,
+      "step": 730
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 2.0963101387023926,
+      "eval_runtime": 36.8853,
+      "eval_samples_per_second": 27.111,
+      "eval_steps_per_second": 1.139,
+      "step": 730
+    },
+    {
+      "epoch": 0.8010958904109589,
+      "grad_norm": 2.3887739181518555,
+      "learning_rate": 4.100324939073924e-06,
+      "loss": 2.0024,
+      "step": 731
+    },
+    {
+      "epoch": 0.8021917808219178,
+      "grad_norm": 2.296579360961914,
+      "learning_rate": 4.0982940698619015e-06,
+      "loss": 1.9092,
+      "step": 732
+    },
+    {
+      "epoch": 0.8032876712328767,
+      "grad_norm": 2.3881773948669434,
+      "learning_rate": 4.096263200649879e-06,
+      "loss": 1.9112,
+      "step": 733
+    },
+    {
+      "epoch": 0.8043835616438356,
+      "grad_norm": 2.40120530128479,
+      "learning_rate": 4.094232331437856e-06,
+      "loss": 1.9191,
+      "step": 734
+    },
+    {
+      "epoch": 0.8054794520547945,
+      "grad_norm": 2.184706211090088,
+      "learning_rate": 4.092201462225833e-06,
+      "loss": 1.8857,
+      "step": 735
+    },
+    {
+      "epoch": 0.8065753424657535,
+      "grad_norm": 2.3338732719421387,
+      "learning_rate": 4.090170593013811e-06,
+      "loss": 1.8889,
+      "step": 736
+    },
+    {
+      "epoch": 0.8076712328767123,
+      "grad_norm": 2.4065468311309814,
+      "learning_rate": 4.088139723801787e-06,
+      "loss": 1.9309,
+      "step": 737
+    },
+    {
+      "epoch": 0.8087671232876712,
+      "grad_norm": 2.044992446899414,
+      "learning_rate": 4.086108854589765e-06,
+      "loss": 1.8727,
+      "step": 738
+    },
+    {
+      "epoch": 0.8098630136986301,
+      "grad_norm": 2.23860239982605,
+      "learning_rate": 4.084077985377742e-06,
+      "loss": 1.8936,
+      "step": 739
+    },
+    {
+      "epoch": 0.810958904109589,
+      "grad_norm": 2.1324410438537598,
+      "learning_rate": 4.08204711616572e-06,
+      "loss": 1.9035,
+      "step": 740
+    },
+    {
+      "epoch": 0.810958904109589,
+      "eval_loss": 2.092885732650757,
+      "eval_runtime": 36.7991,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 740
+    },
+    {
+      "epoch": 0.812054794520548,
+      "grad_norm": 2.337340831756592,
+      "learning_rate": 4.080016246953696e-06,
+      "loss": 1.9747,
+      "step": 741
+    },
+    {
+      "epoch": 0.8131506849315069,
+      "grad_norm": 2.262801170349121,
+      "learning_rate": 4.077985377741674e-06,
+      "loss": 1.8817,
+      "step": 742
+    },
+    {
+      "epoch": 0.8142465753424658,
+      "grad_norm": 2.0494258403778076,
+      "learning_rate": 4.075954508529651e-06,
+      "loss": 1.9493,
+      "step": 743
+    },
+    {
+      "epoch": 0.8153424657534246,
+      "grad_norm": 2.085829019546509,
+      "learning_rate": 4.073923639317629e-06,
+      "loss": 1.9611,
+      "step": 744
+    },
+    {
+      "epoch": 0.8164383561643835,
+      "grad_norm": 2.105806827545166,
+      "learning_rate": 4.071892770105605e-06,
+      "loss": 1.9779,
+      "step": 745
+    },
+    {
+      "epoch": 0.8175342465753425,
+      "grad_norm": 2.2664732933044434,
+      "learning_rate": 4.069861900893583e-06,
+      "loss": 1.9009,
+      "step": 746
+    },
+    {
+      "epoch": 0.8186301369863014,
+      "grad_norm": 1.9899768829345703,
+      "learning_rate": 4.06783103168156e-06,
+      "loss": 1.8585,
+      "step": 747
+    },
+    {
+      "epoch": 0.8197260273972603,
+      "grad_norm": 2.242236614227295,
+      "learning_rate": 4.065800162469538e-06,
+      "loss": 1.8904,
+      "step": 748
+    },
+    {
+      "epoch": 0.8208219178082192,
+      "grad_norm": 2.4481139183044434,
+      "learning_rate": 4.063769293257514e-06,
+      "loss": 1.8922,
+      "step": 749
+    },
+    {
+      "epoch": 0.821917808219178,
+      "grad_norm": 2.231722354888916,
+      "learning_rate": 4.061738424045492e-06,
+      "loss": 1.9292,
+      "step": 750
+    },
+    {
+      "epoch": 0.821917808219178,
+      "eval_loss": 2.0909852981567383,
+      "eval_runtime": 36.8255,
+      "eval_samples_per_second": 27.155,
+      "eval_steps_per_second": 1.141,
+      "step": 750
+    },
+    {
+      "epoch": 0.823013698630137,
+      "grad_norm": 2.323070764541626,
+      "learning_rate": 4.059707554833469e-06,
+      "loss": 1.9053,
+      "step": 751
+    },
+    {
+      "epoch": 0.8241095890410959,
+      "grad_norm": 2.137535572052002,
+      "learning_rate": 4.057676685621446e-06,
+      "loss": 1.9272,
+      "step": 752
+    },
+    {
+      "epoch": 0.8252054794520548,
+      "grad_norm": 2.2087643146514893,
+      "learning_rate": 4.0556458164094235e-06,
+      "loss": 1.9541,
+      "step": 753
+    },
+    {
+      "epoch": 0.8263013698630137,
+      "grad_norm": 2.04811692237854,
+      "learning_rate": 4.053614947197401e-06,
+      "loss": 1.922,
+      "step": 754
+    },
+    {
+      "epoch": 0.8273972602739726,
+      "grad_norm": 2.062622308731079,
+      "learning_rate": 4.051584077985378e-06,
+      "loss": 1.8388,
+      "step": 755
+    },
+    {
+      "epoch": 0.8284931506849315,
+      "grad_norm": 2.14631986618042,
+      "learning_rate": 4.049553208773355e-06,
+      "loss": 1.9153,
+      "step": 756
+    },
+    {
+      "epoch": 0.8295890410958904,
+      "grad_norm": 2.353529214859009,
+      "learning_rate": 4.047522339561333e-06,
+      "loss": 1.9623,
+      "step": 757
+    },
+    {
+      "epoch": 0.8306849315068493,
+      "grad_norm": 2.122342348098755,
+      "learning_rate": 4.04549147034931e-06,
+      "loss": 1.9095,
+      "step": 758
+    },
+    {
+      "epoch": 0.8317808219178082,
+      "grad_norm": 2.2540230751037598,
+      "learning_rate": 4.043460601137287e-06,
+      "loss": 1.9068,
+      "step": 759
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "grad_norm": 2.164384365081787,
+      "learning_rate": 4.041429731925264e-06,
+      "loss": 2.0239,
+      "step": 760
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "eval_loss": 2.0904648303985596,
+      "eval_runtime": 36.7617,
+      "eval_samples_per_second": 27.202,
+      "eval_steps_per_second": 1.142,
+      "step": 760
+    },
+    {
+      "epoch": 0.8339726027397261,
+      "grad_norm": 2.2487905025482178,
+      "learning_rate": 4.039398862713242e-06,
+      "loss": 1.9109,
+      "step": 761
+    },
+    {
+      "epoch": 0.8350684931506849,
+      "grad_norm": 2.0947930812835693,
+      "learning_rate": 4.037367993501219e-06,
+      "loss": 1.9133,
+      "step": 762
+    },
+    {
+      "epoch": 0.8361643835616438,
+      "grad_norm": 2.2259862422943115,
+      "learning_rate": 4.035337124289196e-06,
+      "loss": 1.9821,
+      "step": 763
+    },
+    {
+      "epoch": 0.8372602739726027,
+      "grad_norm": 2.0616455078125,
+      "learning_rate": 4.0333062550771735e-06,
+      "loss": 1.9037,
+      "step": 764
+    },
+    {
+      "epoch": 0.8383561643835616,
+      "grad_norm": 2.2704474925994873,
+      "learning_rate": 4.031275385865151e-06,
+      "loss": 1.8664,
+      "step": 765
+    },
+    {
+      "epoch": 0.8394520547945206,
+      "grad_norm": 2.0059332847595215,
+      "learning_rate": 4.029244516653128e-06,
+      "loss": 1.908,
+      "step": 766
+    },
+    {
+      "epoch": 0.8405479452054795,
+      "grad_norm": 2.0462095737457275,
+      "learning_rate": 4.027213647441105e-06,
+      "loss": 1.9397,
+      "step": 767
+    },
+    {
+      "epoch": 0.8416438356164384,
+      "grad_norm": 2.150207996368408,
+      "learning_rate": 4.025182778229083e-06,
+      "loss": 1.8505,
+      "step": 768
+    },
+    {
+      "epoch": 0.8427397260273972,
+      "grad_norm": 2.174863815307617,
+      "learning_rate": 4.02315190901706e-06,
+      "loss": 1.8895,
+      "step": 769
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "grad_norm": 2.1907355785369873,
+      "learning_rate": 4.021121039805037e-06,
+      "loss": 1.9156,
+      "step": 770
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "eval_loss": 2.0900955200195312,
+      "eval_runtime": 36.8004,
+      "eval_samples_per_second": 27.174,
+      "eval_steps_per_second": 1.141,
+      "step": 770
+    },
+    {
+      "epoch": 0.8449315068493151,
+      "grad_norm": 2.2831835746765137,
+      "learning_rate": 4.0190901705930144e-06,
+      "loss": 1.9115,
+      "step": 771
+    },
+    {
+      "epoch": 0.846027397260274,
+      "grad_norm": 2.039520263671875,
+      "learning_rate": 4.017059301380992e-06,
+      "loss": 2.0047,
+      "step": 772
+    },
+    {
+      "epoch": 0.8471232876712329,
+      "grad_norm": 2.147209405899048,
+      "learning_rate": 4.015028432168969e-06,
+      "loss": 1.8845,
+      "step": 773
+    },
+    {
+      "epoch": 0.8482191780821918,
+      "grad_norm": 2.168997287750244,
+      "learning_rate": 4.012997562956945e-06,
+      "loss": 1.9606,
+      "step": 774
+    },
+    {
+      "epoch": 0.8493150684931506,
+      "grad_norm": 2.038966417312622,
+      "learning_rate": 4.0109666937449235e-06,
+      "loss": 1.8734,
+      "step": 775
+    },
+    {
+      "epoch": 0.8504109589041096,
+      "grad_norm": 2.0531768798828125,
+      "learning_rate": 4.0089358245329e-06,
+      "loss": 1.889,
+      "step": 776
+    },
+    {
+      "epoch": 0.8515068493150685,
+      "grad_norm": 2.348008155822754,
+      "learning_rate": 4.006904955320878e-06,
+      "loss": 1.8972,
+      "step": 777
+    },
+    {
+      "epoch": 0.8526027397260274,
+      "grad_norm": 2.1647253036499023,
+      "learning_rate": 4.0048740861088545e-06,
+      "loss": 1.8652,
+      "step": 778
+    },
+    {
+      "epoch": 0.8536986301369863,
+      "grad_norm": 2.1971640586853027,
+      "learning_rate": 4.002843216896833e-06,
+      "loss": 1.9495,
+      "step": 779
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "grad_norm": 2.169632911682129,
+      "learning_rate": 4.000812347684809e-06,
+      "loss": 1.9307,
+      "step": 780
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "eval_loss": 2.087860107421875,
+      "eval_runtime": 36.7982,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 780
+    },
+    {
+      "epoch": 0.8558904109589041,
+      "grad_norm": 2.4424400329589844,
+      "learning_rate": 3.998781478472787e-06,
+      "loss": 1.8849,
+      "step": 781
+    },
+    {
+      "epoch": 0.856986301369863,
+      "grad_norm": 2.2226672172546387,
+      "learning_rate": 3.996750609260764e-06,
+      "loss": 1.8975,
+      "step": 782
+    },
+    {
+      "epoch": 0.8580821917808219,
+      "grad_norm": 2.1898491382598877,
+      "learning_rate": 3.994719740048742e-06,
+      "loss": 1.8751,
+      "step": 783
+    },
+    {
+      "epoch": 0.8591780821917808,
+      "grad_norm": 2.466963052749634,
+      "learning_rate": 3.992688870836718e-06,
+      "loss": 2.0155,
+      "step": 784
+    },
+    {
+      "epoch": 0.8602739726027397,
+      "grad_norm": 2.184039831161499,
+      "learning_rate": 3.990658001624696e-06,
+      "loss": 1.9133,
+      "step": 785
+    },
+    {
+      "epoch": 0.8613698630136987,
+      "grad_norm": 2.089228868484497,
+      "learning_rate": 3.988627132412673e-06,
+      "loss": 1.8707,
+      "step": 786
+    },
+    {
+      "epoch": 0.8624657534246576,
+      "grad_norm": 2.0700666904449463,
+      "learning_rate": 3.98659626320065e-06,
+      "loss": 1.9487,
+      "step": 787
+    },
+    {
+      "epoch": 0.8635616438356164,
+      "grad_norm": 2.1495554447174072,
+      "learning_rate": 3.984565393988627e-06,
+      "loss": 1.9037,
+      "step": 788
+    },
+    {
+      "epoch": 0.8646575342465753,
+      "grad_norm": 2.0916988849639893,
+      "learning_rate": 3.9825345247766046e-06,
+      "loss": 1.9096,
+      "step": 789
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "grad_norm": 2.072387933731079,
+      "learning_rate": 3.980503655564582e-06,
+      "loss": 1.9139,
+      "step": 790
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "eval_loss": 2.0876142978668213,
+      "eval_runtime": 37.1068,
+      "eval_samples_per_second": 26.949,
+      "eval_steps_per_second": 1.132,
+      "step": 790
+    },
+    {
+      "epoch": 0.8668493150684932,
+      "grad_norm": 2.1206343173980713,
+      "learning_rate": 3.978472786352559e-06,
+      "loss": 1.9331,
+      "step": 791
+    },
+    {
+      "epoch": 0.8679452054794521,
+      "grad_norm": 2.20263934135437,
+      "learning_rate": 3.976441917140536e-06,
+      "loss": 1.9297,
+      "step": 792
+    },
+    {
+      "epoch": 0.869041095890411,
+      "grad_norm": 2.027554750442505,
+      "learning_rate": 3.974411047928514e-06,
+      "loss": 1.919,
+      "step": 793
+    },
+    {
+      "epoch": 0.8701369863013698,
+      "grad_norm": 2.0291616916656494,
+      "learning_rate": 3.972380178716491e-06,
+      "loss": 1.9682,
+      "step": 794
+    },
+    {
+      "epoch": 0.8712328767123287,
+      "grad_norm": 2.2719762325286865,
+      "learning_rate": 3.970349309504468e-06,
+      "loss": 1.9407,
+      "step": 795
+    },
+    {
+      "epoch": 0.8723287671232877,
+      "grad_norm": 2.0452375411987305,
+      "learning_rate": 3.9683184402924455e-06,
+      "loss": 1.9487,
+      "step": 796
+    },
+    {
+      "epoch": 0.8734246575342466,
+      "grad_norm": 2.170441150665283,
+      "learning_rate": 3.966287571080423e-06,
+      "loss": 1.8651,
+      "step": 797
+    },
+    {
+      "epoch": 0.8745205479452055,
+      "grad_norm": 2.3370516300201416,
+      "learning_rate": 3.9642567018684e-06,
+      "loss": 1.8599,
+      "step": 798
+    },
+    {
+      "epoch": 0.8756164383561644,
+      "grad_norm": 2.336737632751465,
+      "learning_rate": 3.962225832656377e-06,
+      "loss": 1.9304,
+      "step": 799
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "grad_norm": 1.9708369970321655,
+      "learning_rate": 3.960194963444355e-06,
+      "loss": 1.9125,
+      "step": 800
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "eval_loss": 2.0812273025512695,
+      "eval_runtime": 37.0752,
+      "eval_samples_per_second": 26.972,
+      "eval_steps_per_second": 1.133,
+      "step": 800
+    },
+    {
+      "epoch": 0.8778082191780822,
+      "grad_norm": 2.154219627380371,
+      "learning_rate": 3.958164094232332e-06,
+      "loss": 1.8304,
+      "step": 801
+    },
+    {
+      "epoch": 0.8789041095890411,
+      "grad_norm": 2.237786054611206,
+      "learning_rate": 3.956133225020309e-06,
+      "loss": 1.9147,
+      "step": 802
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 2.07143235206604,
+      "learning_rate": 3.954102355808286e-06,
+      "loss": 1.9128,
+      "step": 803
+    },
+    {
+      "epoch": 0.8810958904109589,
+      "grad_norm": 2.2780392169952393,
+      "learning_rate": 3.952071486596264e-06,
+      "loss": 1.9145,
+      "step": 804
+    },
+    {
+      "epoch": 0.8821917808219178,
+      "grad_norm": 2.2396292686462402,
+      "learning_rate": 3.950040617384241e-06,
+      "loss": 1.9049,
+      "step": 805
+    },
+    {
+      "epoch": 0.8832876712328767,
+      "grad_norm": 2.426511287689209,
+      "learning_rate": 3.948009748172218e-06,
+      "loss": 1.8984,
+      "step": 806
+    },
+    {
+      "epoch": 0.8843835616438356,
+      "grad_norm": 2.1874325275421143,
+      "learning_rate": 3.9459788789601955e-06,
+      "loss": 1.8888,
+      "step": 807
+    },
+    {
+      "epoch": 0.8854794520547945,
+      "grad_norm": 2.038098096847534,
+      "learning_rate": 3.943948009748173e-06,
+      "loss": 1.9582,
+      "step": 808
+    },
+    {
+      "epoch": 0.8865753424657534,
+      "grad_norm": 2.3748130798339844,
+      "learning_rate": 3.94191714053615e-06,
+      "loss": 1.9091,
+      "step": 809
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "grad_norm": 2.2457494735717773,
+      "learning_rate": 3.939886271324127e-06,
+      "loss": 1.9067,
+      "step": 810
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "eval_loss": 2.0766079425811768,
+      "eval_runtime": 36.9896,
+      "eval_samples_per_second": 27.035,
+      "eval_steps_per_second": 1.135,
+      "step": 810
+    },
+    {
+      "epoch": 0.8887671232876713,
+      "grad_norm": 2.156644821166992,
+      "learning_rate": 3.937855402112104e-06,
+      "loss": 1.9347,
+      "step": 811
+    },
+    {
+      "epoch": 0.8898630136986302,
+      "grad_norm": 2.3991472721099854,
+      "learning_rate": 3.935824532900082e-06,
+      "loss": 1.8645,
+      "step": 812
+    },
+    {
+      "epoch": 0.890958904109589,
+      "grad_norm": 2.252729654312134,
+      "learning_rate": 3.933793663688058e-06,
+      "loss": 1.8537,
+      "step": 813
+    },
+    {
+      "epoch": 0.8920547945205479,
+      "grad_norm": 2.2917964458465576,
+      "learning_rate": 3.9317627944760365e-06,
+      "loss": 1.9271,
+      "step": 814
+    },
+    {
+      "epoch": 0.8931506849315068,
+      "grad_norm": 2.217310667037964,
+      "learning_rate": 3.929731925264013e-06,
+      "loss": 1.9406,
+      "step": 815
+    },
+    {
+      "epoch": 0.8942465753424658,
+      "grad_norm": 2.167931079864502,
+      "learning_rate": 3.927701056051991e-06,
+      "loss": 1.8554,
+      "step": 816
+    },
+    {
+      "epoch": 0.8953424657534247,
+      "grad_norm": 2.18337082862854,
+      "learning_rate": 3.9256701868399674e-06,
+      "loss": 1.8795,
+      "step": 817
+    },
+    {
+      "epoch": 0.8964383561643836,
+      "grad_norm": 2.2348973751068115,
+      "learning_rate": 3.9236393176279456e-06,
+      "loss": 1.8831,
+      "step": 818
+    },
+    {
+      "epoch": 0.8975342465753424,
+      "grad_norm": 2.1822657585144043,
+      "learning_rate": 3.921608448415922e-06,
+      "loss": 1.8728,
+      "step": 819
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "grad_norm": 2.5253243446350098,
+      "learning_rate": 3.9195775792039e-06,
+      "loss": 1.8633,
+      "step": 820
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "eval_loss": 2.0769834518432617,
+      "eval_runtime": 36.838,
+      "eval_samples_per_second": 27.146,
+      "eval_steps_per_second": 1.14,
+      "step": 820
+    },
+    {
+      "epoch": 0.8997260273972603,
+      "grad_norm": 2.2109973430633545,
+      "learning_rate": 3.9175467099918765e-06,
+      "loss": 1.912,
+      "step": 821
+    },
+    {
+      "epoch": 0.9008219178082192,
+      "grad_norm": 2.053692102432251,
+      "learning_rate": 3.915515840779855e-06,
+      "loss": 1.8928,
+      "step": 822
+    },
+    {
+      "epoch": 0.9019178082191781,
+      "grad_norm": 2.1418213844299316,
+      "learning_rate": 3.913484971567831e-06,
+      "loss": 1.9399,
+      "step": 823
+    },
+    {
+      "epoch": 0.903013698630137,
+      "grad_norm": 2.3615458011627197,
+      "learning_rate": 3.911454102355808e-06,
+      "loss": 1.9293,
+      "step": 824
+    },
+    {
+      "epoch": 0.9041095890410958,
+      "grad_norm": 2.098254680633545,
+      "learning_rate": 3.909423233143786e-06,
+      "loss": 1.8301,
+      "step": 825
+    },
+    {
+      "epoch": 0.9052054794520548,
+      "grad_norm": 2.4473679065704346,
+      "learning_rate": 3.907392363931763e-06,
+      "loss": 1.9254,
+      "step": 826
+    },
+    {
+      "epoch": 0.9063013698630137,
+      "grad_norm": 2.252843141555786,
+      "learning_rate": 3.90536149471974e-06,
+      "loss": 1.9155,
+      "step": 827
+    },
+    {
+      "epoch": 0.9073972602739726,
+      "grad_norm": 2.1574337482452393,
+      "learning_rate": 3.9033306255077175e-06,
+      "loss": 1.9809,
+      "step": 828
+    },
+    {
+      "epoch": 0.9084931506849315,
+      "grad_norm": 2.2899720668792725,
+      "learning_rate": 3.901299756295695e-06,
+      "loss": 2.0045,
+      "step": 829
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "grad_norm": 2.227285146713257,
+      "learning_rate": 3.899268887083672e-06,
+      "loss": 2.026,
+      "step": 830
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "eval_loss": 2.0764760971069336,
+      "eval_runtime": 36.9195,
+      "eval_samples_per_second": 27.086,
+      "eval_steps_per_second": 1.138,
+      "step": 830
+    },
+    {
+      "epoch": 0.9106849315068493,
+      "grad_norm": 2.1935312747955322,
+      "learning_rate": 3.897238017871649e-06,
+      "loss": 1.8922,
+      "step": 831
+    },
+    {
+      "epoch": 0.9117808219178082,
+      "grad_norm": 2.129692316055298,
+      "learning_rate": 3.8952071486596266e-06,
+      "loss": 1.8981,
+      "step": 832
+    },
+    {
+      "epoch": 0.9128767123287671,
+      "grad_norm": 2.4435646533966064,
+      "learning_rate": 3.893176279447604e-06,
+      "loss": 1.91,
+      "step": 833
+    },
+    {
+      "epoch": 0.913972602739726,
+      "grad_norm": 2.1556732654571533,
+      "learning_rate": 3.891145410235581e-06,
+      "loss": 1.8688,
+      "step": 834
+    },
+    {
+      "epoch": 0.915068493150685,
+      "grad_norm": 2.175891160964966,
+      "learning_rate": 3.889114541023558e-06,
+      "loss": 1.9777,
+      "step": 835
+    },
+    {
+      "epoch": 0.9161643835616439,
+      "grad_norm": 2.220466375350952,
+      "learning_rate": 3.887083671811536e-06,
+      "loss": 1.9152,
+      "step": 836
+    },
+    {
+      "epoch": 0.9172602739726028,
+      "grad_norm": 2.331397294998169,
+      "learning_rate": 3.885052802599513e-06,
+      "loss": 1.892,
+      "step": 837
+    },
+    {
+      "epoch": 0.9183561643835616,
+      "grad_norm": 2.2434773445129395,
+      "learning_rate": 3.88302193338749e-06,
+      "loss": 1.841,
+      "step": 838
+    },
+    {
+      "epoch": 0.9194520547945205,
+      "grad_norm": 2.3248002529144287,
+      "learning_rate": 3.8809910641754675e-06,
+      "loss": 1.9657,
+      "step": 839
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "grad_norm": 2.090252637863159,
+      "learning_rate": 3.878960194963445e-06,
+      "loss": 1.9121,
+      "step": 840
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "eval_loss": 2.0798065662384033,
+      "eval_runtime": 36.8557,
+      "eval_samples_per_second": 27.133,
+      "eval_steps_per_second": 1.14,
+      "step": 840
+    },
+    {
+      "epoch": 0.9216438356164384,
+      "grad_norm": 2.6968672275543213,
+      "learning_rate": 3.876929325751422e-06,
+      "loss": 1.9003,
+      "step": 841
+    },
+    {
+      "epoch": 0.9227397260273973,
+      "grad_norm": 2.356321096420288,
+      "learning_rate": 3.874898456539399e-06,
+      "loss": 1.9471,
+      "step": 842
+    },
+    {
+      "epoch": 0.9238356164383562,
+      "grad_norm": 2.2678606510162354,
+      "learning_rate": 3.872867587327377e-06,
+      "loss": 1.8488,
+      "step": 843
+    },
+    {
+      "epoch": 0.924931506849315,
+      "grad_norm": 2.1122310161590576,
+      "learning_rate": 3.870836718115354e-06,
+      "loss": 1.9247,
+      "step": 844
+    },
+    {
+      "epoch": 0.9260273972602739,
+      "grad_norm": 2.1219136714935303,
+      "learning_rate": 3.868805848903331e-06,
+      "loss": 1.8829,
+      "step": 845
+    },
+    {
+      "epoch": 0.9271232876712329,
+      "grad_norm": 2.175607204437256,
+      "learning_rate": 3.8667749796913084e-06,
+      "loss": 1.887,
+      "step": 846
+    },
+    {
+      "epoch": 0.9282191780821918,
+      "grad_norm": 2.156214714050293,
+      "learning_rate": 3.864744110479286e-06,
+      "loss": 1.9119,
+      "step": 847
+    },
+    {
+      "epoch": 0.9293150684931507,
+      "grad_norm": 2.158461570739746,
+      "learning_rate": 3.862713241267262e-06,
+      "loss": 1.9092,
+      "step": 848
+    },
+    {
+      "epoch": 0.9304109589041096,
+      "grad_norm": 2.180551767349243,
+      "learning_rate": 3.86068237205524e-06,
+      "loss": 1.946,
+      "step": 849
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "grad_norm": 2.1030542850494385,
+      "learning_rate": 3.858651502843217e-06,
+      "loss": 1.9484,
+      "step": 850
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "eval_loss": 2.0767791271209717,
+      "eval_runtime": 36.8047,
+      "eval_samples_per_second": 27.17,
+      "eval_steps_per_second": 1.141,
+      "step": 850
+    },
+    {
+      "epoch": 0.9326027397260274,
+      "grad_norm": 2.240769386291504,
+      "learning_rate": 3.856620633631195e-06,
+      "loss": 1.9114,
+      "step": 851
+    },
+    {
+      "epoch": 0.9336986301369863,
+      "grad_norm": 2.1964619159698486,
+      "learning_rate": 3.854589764419171e-06,
+      "loss": 1.8339,
+      "step": 852
+    },
+    {
+      "epoch": 0.9347945205479452,
+      "grad_norm": 2.2875845432281494,
+      "learning_rate": 3.852558895207149e-06,
+      "loss": 1.8942,
+      "step": 853
+    },
+    {
+      "epoch": 0.9358904109589041,
+      "grad_norm": 2.1462507247924805,
+      "learning_rate": 3.850528025995126e-06,
+      "loss": 1.9838,
+      "step": 854
+    },
+    {
+      "epoch": 0.936986301369863,
+      "grad_norm": 2.0982913970947266,
+      "learning_rate": 3.848497156783104e-06,
+      "loss": 1.9351,
+      "step": 855
+    },
+    {
+      "epoch": 0.938082191780822,
+      "grad_norm": 2.260263442993164,
+      "learning_rate": 3.84646628757108e-06,
+      "loss": 1.8745,
+      "step": 856
+    },
+    {
+      "epoch": 0.9391780821917808,
+      "grad_norm": 2.091778039932251,
+      "learning_rate": 3.8444354183590585e-06,
+      "loss": 1.8888,
+      "step": 857
+    },
+    {
+      "epoch": 0.9402739726027397,
+      "grad_norm": 2.054974317550659,
+      "learning_rate": 3.842404549147035e-06,
+      "loss": 1.8801,
+      "step": 858
+    },
+    {
+      "epoch": 0.9413698630136986,
+      "grad_norm": 2.079641580581665,
+      "learning_rate": 3.840373679935013e-06,
+      "loss": 1.8879,
+      "step": 859
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "grad_norm": 2.3441810607910156,
+      "learning_rate": 3.8383428107229894e-06,
+      "loss": 1.8571,
+      "step": 860
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "eval_loss": 2.0766854286193848,
+      "eval_runtime": 36.8414,
+      "eval_samples_per_second": 27.143,
+      "eval_steps_per_second": 1.14,
+      "step": 860
+    },
+    {
+      "epoch": 0.9435616438356165,
+      "grad_norm": 2.123061180114746,
+      "learning_rate": 3.836311941510967e-06,
+      "loss": 1.8922,
+      "step": 861
+    },
+    {
+      "epoch": 0.9446575342465754,
+      "grad_norm": 2.1573386192321777,
+      "learning_rate": 3.834281072298944e-06,
+      "loss": 1.8713,
+      "step": 862
+    },
+    {
+      "epoch": 0.9457534246575342,
+      "grad_norm": 2.225186824798584,
+      "learning_rate": 3.832250203086921e-06,
+      "loss": 1.925,
+      "step": 863
+    },
+    {
+      "epoch": 0.9468493150684931,
+      "grad_norm": 2.1539998054504395,
+      "learning_rate": 3.8302193338748985e-06,
+      "loss": 1.8721,
+      "step": 864
+    },
+    {
+      "epoch": 0.947945205479452,
+      "grad_norm": 2.075580358505249,
+      "learning_rate": 3.828188464662876e-06,
+      "loss": 1.9489,
+      "step": 865
+    },
+    {
+      "epoch": 0.949041095890411,
+      "grad_norm": 2.1789908409118652,
+      "learning_rate": 3.826157595450853e-06,
+      "loss": 1.9883,
+      "step": 866
+    },
+    {
+      "epoch": 0.9501369863013699,
+      "grad_norm": 2.0838615894317627,
+      "learning_rate": 3.82412672623883e-06,
+      "loss": 1.9149,
+      "step": 867
+    },
+    {
+      "epoch": 0.9512328767123288,
+      "grad_norm": 2.1321213245391846,
+      "learning_rate": 3.822095857026808e-06,
+      "loss": 1.8505,
+      "step": 868
+    },
+    {
+      "epoch": 0.9523287671232876,
+      "grad_norm": 2.1419479846954346,
+      "learning_rate": 3.820064987814785e-06,
+      "loss": 1.9226,
+      "step": 869
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "grad_norm": 2.290163040161133,
+      "learning_rate": 3.818034118602762e-06,
+      "loss": 1.8737,
+      "step": 870
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "eval_loss": 2.073598623275757,
+      "eval_runtime": 36.7896,
+      "eval_samples_per_second": 27.182,
+      "eval_steps_per_second": 1.142,
+      "step": 870
+    },
+    {
+      "epoch": 0.9545205479452055,
+      "grad_norm": 2.1951074600219727,
+      "learning_rate": 3.8160032493907395e-06,
+      "loss": 1.8699,
+      "step": 871
+    },
+    {
+      "epoch": 0.9556164383561644,
+      "grad_norm": 2.25960111618042,
+      "learning_rate": 3.813972380178717e-06,
+      "loss": 1.8813,
+      "step": 872
+    },
+    {
+      "epoch": 0.9567123287671233,
+      "grad_norm": 2.4376907348632812,
+      "learning_rate": 3.811941510966694e-06,
+      "loss": 1.8782,
+      "step": 873
+    },
+    {
+      "epoch": 0.9578082191780822,
+      "grad_norm": 1.980772614479065,
+      "learning_rate": 3.809910641754671e-06,
+      "loss": 1.9205,
+      "step": 874
+    },
+    {
+      "epoch": 0.958904109589041,
+      "grad_norm": 2.0699028968811035,
+      "learning_rate": 3.8078797725426486e-06,
+      "loss": 1.8674,
+      "step": 875
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 2.2765867710113525,
+      "learning_rate": 3.8058489033306254e-06,
+      "loss": 1.81,
+      "step": 876
+    },
+    {
+      "epoch": 0.9610958904109589,
+      "grad_norm": 2.0998799800872803,
+      "learning_rate": 3.803818034118603e-06,
+      "loss": 1.9061,
+      "step": 877
+    },
+    {
+      "epoch": 0.9621917808219178,
+      "grad_norm": 2.2362914085388184,
+      "learning_rate": 3.80178716490658e-06,
+      "loss": 1.8798,
+      "step": 878
+    },
+    {
+      "epoch": 0.9632876712328767,
+      "grad_norm": 2.0789835453033447,
+      "learning_rate": 3.7997562956945577e-06,
+      "loss": 1.9491,
+      "step": 879
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "grad_norm": 2.0458688735961914,
+      "learning_rate": 3.7977254264825345e-06,
+      "loss": 1.931,
+      "step": 880
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "eval_loss": 2.0672221183776855,
+      "eval_runtime": 36.8218,
+      "eval_samples_per_second": 27.158,
+      "eval_steps_per_second": 1.141,
+      "step": 880
+    },
+    {
+      "epoch": 0.9654794520547946,
+      "grad_norm": 2.361626625061035,
+      "learning_rate": 3.7956945572705122e-06,
+      "loss": 1.8815,
+      "step": 881
+    },
+    {
+      "epoch": 0.9665753424657534,
+      "grad_norm": 2.294160842895508,
+      "learning_rate": 3.793663688058489e-06,
+      "loss": 1.8649,
+      "step": 882
+    },
+    {
+      "epoch": 0.9676712328767123,
+      "grad_norm": 2.2242612838745117,
+      "learning_rate": 3.791632818846467e-06,
+      "loss": 1.9535,
+      "step": 883
+    },
+    {
+      "epoch": 0.9687671232876712,
+      "grad_norm": 2.220269203186035,
+      "learning_rate": 3.7896019496344436e-06,
+      "loss": 1.9506,
+      "step": 884
+    },
+    {
+      "epoch": 0.9698630136986301,
+      "grad_norm": 2.112518072128296,
+      "learning_rate": 3.787571080422421e-06,
+      "loss": 1.9487,
+      "step": 885
+    },
+    {
+      "epoch": 0.9709589041095891,
+      "grad_norm": 2.17695951461792,
+      "learning_rate": 3.785540211210398e-06,
+      "loss": 1.9467,
+      "step": 886
+    },
+    {
+      "epoch": 0.972054794520548,
+      "grad_norm": 2.0990567207336426,
+      "learning_rate": 3.7835093419983755e-06,
+      "loss": 1.9216,
+      "step": 887
+    },
+    {
+      "epoch": 0.9731506849315068,
+      "grad_norm": 2.222566604614258,
+      "learning_rate": 3.7814784727863527e-06,
+      "loss": 1.9043,
+      "step": 888
+    },
+    {
+      "epoch": 0.9742465753424657,
+      "grad_norm": 2.09890079498291,
+      "learning_rate": 3.77944760357433e-06,
+      "loss": 1.884,
+      "step": 889
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "grad_norm": 2.136141061782837,
+      "learning_rate": 3.7774167343623073e-06,
+      "loss": 2.0032,
+      "step": 890
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "eval_loss": 2.0665438175201416,
+      "eval_runtime": 37.1207,
+      "eval_samples_per_second": 26.939,
+      "eval_steps_per_second": 1.131,
+      "step": 890
+    },
+    {
+      "epoch": 0.9764383561643836,
+      "grad_norm": 2.267444372177124,
+      "learning_rate": 3.7753858651502846e-06,
+      "loss": 1.947,
+      "step": 891
+    },
+    {
+      "epoch": 0.9775342465753425,
+      "grad_norm": 2.1141815185546875,
+      "learning_rate": 3.773354995938262e-06,
+      "loss": 1.8705,
+      "step": 892
+    },
+    {
+      "epoch": 0.9786301369863014,
+      "grad_norm": 2.1036782264709473,
+      "learning_rate": 3.771324126726239e-06,
+      "loss": 1.9381,
+      "step": 893
+    },
+    {
+      "epoch": 0.9797260273972602,
+      "grad_norm": 2.081982135772705,
+      "learning_rate": 3.7692932575142164e-06,
+      "loss": 1.9079,
+      "step": 894
+    },
+    {
+      "epoch": 0.9808219178082191,
+      "grad_norm": 2.0815446376800537,
+      "learning_rate": 3.7672623883021937e-06,
+      "loss": 1.9417,
+      "step": 895
+    },
+    {
+      "epoch": 0.9819178082191781,
+      "grad_norm": 2.0531511306762695,
+      "learning_rate": 3.765231519090171e-06,
+      "loss": 1.876,
+      "step": 896
+    },
+    {
+      "epoch": 0.983013698630137,
+      "grad_norm": 2.212940216064453,
+      "learning_rate": 3.7632006498781482e-06,
+      "loss": 1.8008,
+      "step": 897
+    },
+    {
+      "epoch": 0.9841095890410959,
+      "grad_norm": 2.1541531085968018,
+      "learning_rate": 3.761169780666125e-06,
+      "loss": 1.9039,
+      "step": 898
+    },
+    {
+      "epoch": 0.9852054794520548,
+      "grad_norm": 2.113884925842285,
+      "learning_rate": 3.7591389114541028e-06,
+      "loss": 1.8562,
+      "step": 899
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "grad_norm": 2.1503701210021973,
+      "learning_rate": 3.7571080422420796e-06,
+      "loss": 1.9052,
+      "step": 900
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "eval_loss": 2.0664408206939697,
+      "eval_runtime": 37.0876,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 900
+    },
+    {
+      "epoch": 0.9873972602739726,
+      "grad_norm": 2.1584348678588867,
+      "learning_rate": 3.7550771730300573e-06,
+      "loss": 1.9296,
+      "step": 901
+    },
+    {
+      "epoch": 0.9884931506849315,
+      "grad_norm": 2.164318084716797,
+      "learning_rate": 3.753046303818034e-06,
+      "loss": 1.8834,
+      "step": 902
+    },
+    {
+      "epoch": 0.9895890410958904,
+      "grad_norm": 2.0222678184509277,
+      "learning_rate": 3.751015434606012e-06,
+      "loss": 1.9426,
+      "step": 903
+    },
+    {
+      "epoch": 0.9906849315068493,
+      "grad_norm": 2.116835832595825,
+      "learning_rate": 3.7489845653939887e-06,
+      "loss": 1.9195,
+      "step": 904
+    },
+    {
+      "epoch": 0.9917808219178083,
+      "grad_norm": 2.2708873748779297,
+      "learning_rate": 3.7469536961819664e-06,
+      "loss": 1.9631,
+      "step": 905
+    },
+    {
+      "epoch": 0.9928767123287672,
+      "grad_norm": 2.1174001693725586,
+      "learning_rate": 3.7449228269699433e-06,
+      "loss": 1.8602,
+      "step": 906
+    },
+    {
+      "epoch": 0.993972602739726,
+      "grad_norm": 2.091830015182495,
+      "learning_rate": 3.742891957757921e-06,
+      "loss": 1.8906,
+      "step": 907
+    },
+    {
+      "epoch": 0.9950684931506849,
+      "grad_norm": 2.007119655609131,
+      "learning_rate": 3.740861088545898e-06,
+      "loss": 1.9122,
+      "step": 908
+    },
+    {
+      "epoch": 0.9961643835616438,
+      "grad_norm": 2.070396661758423,
+      "learning_rate": 3.7388302193338755e-06,
+      "loss": 1.9106,
+      "step": 909
+    },
+    {
+      "epoch": 0.9972602739726028,
+      "grad_norm": 2.080331325531006,
+      "learning_rate": 3.7367993501218524e-06,
+      "loss": 1.934,
+      "step": 910
+    },
+    {
+      "epoch": 0.9972602739726028,
+      "eval_loss": 2.0694408416748047,
+      "eval_runtime": 36.9953,
+      "eval_samples_per_second": 27.03,
+      "eval_steps_per_second": 1.135,
+      "step": 910
+    },
+    {
+      "epoch": 0.9983561643835617,
+      "grad_norm": 2.0920615196228027,
+      "learning_rate": 3.7347684809098297e-06,
+      "loss": 1.8679,
+      "step": 911
+    },
+    {
+      "epoch": 0.9994520547945206,
+      "grad_norm": 2.112732172012329,
+      "learning_rate": 3.732737611697807e-06,
+      "loss": 1.8723,
+      "step": 912
+    },
+    {
+      "epoch": 1.0005479452054795,
+      "grad_norm": 2.1449220180511475,
+      "learning_rate": 3.7307067424857842e-06,
+      "loss": 1.8509,
+      "step": 913
+    },
+    {
+      "epoch": 1.0016438356164383,
+      "grad_norm": 2.2179951667785645,
+      "learning_rate": 3.7286758732737615e-06,
+      "loss": 1.7911,
+      "step": 914
+    },
+    {
+      "epoch": 1.0027397260273974,
+      "grad_norm": 2.1392273902893066,
+      "learning_rate": 3.7266450040617388e-06,
+      "loss": 1.8475,
+      "step": 915
+    },
+    {
+      "epoch": 1.0038356164383562,
+      "grad_norm": 2.190265417098999,
+      "learning_rate": 3.724614134849716e-06,
+      "loss": 1.8394,
+      "step": 916
+    },
+    {
+      "epoch": 1.004931506849315,
+      "grad_norm": 2.1833760738372803,
+      "learning_rate": 3.7225832656376933e-06,
+      "loss": 1.9696,
+      "step": 917
+    },
+    {
+      "epoch": 1.006027397260274,
+      "grad_norm": 2.2504308223724365,
+      "learning_rate": 3.7205523964256706e-06,
+      "loss": 1.8135,
+      "step": 918
+    },
+    {
+      "epoch": 1.0071232876712328,
+      "grad_norm": 2.3909389972686768,
+      "learning_rate": 3.718521527213648e-06,
+      "loss": 1.9331,
+      "step": 919
+    },
+    {
+      "epoch": 1.0082191780821919,
+      "grad_norm": 2.3367602825164795,
+      "learning_rate": 3.716490658001625e-06,
+      "loss": 1.9026,
+      "step": 920
+    },
+    {
+      "epoch": 1.0082191780821919,
+      "eval_loss": 2.066087484359741,
+      "eval_runtime": 36.8701,
+      "eval_samples_per_second": 27.122,
+      "eval_steps_per_second": 1.139,
+      "step": 920
+    },
+    {
+      "epoch": 1.0093150684931507,
+      "grad_norm": 2.1732144355773926,
+      "learning_rate": 3.7144597887896024e-06,
+      "loss": 1.8888,
+      "step": 921
+    },
+    {
+      "epoch": 1.0104109589041095,
+      "grad_norm": 2.4632201194763184,
+      "learning_rate": 3.7124289195775797e-06,
+      "loss": 1.7725,
+      "step": 922
+    },
+    {
+      "epoch": 1.0115068493150685,
+      "grad_norm": 2.4887173175811768,
+      "learning_rate": 3.710398050365557e-06,
+      "loss": 1.9125,
+      "step": 923
+    },
+    {
+      "epoch": 1.0126027397260273,
+      "grad_norm": 2.4684619903564453,
+      "learning_rate": 3.708367181153534e-06,
+      "loss": 1.8829,
+      "step": 924
+    },
+    {
+      "epoch": 1.0136986301369864,
+      "grad_norm": 2.348788022994995,
+      "learning_rate": 3.7063363119415115e-06,
+      "loss": 1.8542,
+      "step": 925
+    },
+    {
+      "epoch": 1.0147945205479452,
+      "grad_norm": 2.4432766437530518,
+      "learning_rate": 3.7043054427294884e-06,
+      "loss": 1.903,
+      "step": 926
+    },
+    {
+      "epoch": 1.0158904109589042,
+      "grad_norm": 2.425910234451294,
+      "learning_rate": 3.702274573517466e-06,
+      "loss": 1.8532,
+      "step": 927
+    },
+    {
+      "epoch": 1.016986301369863,
+      "grad_norm": 2.447777032852173,
+      "learning_rate": 3.700243704305443e-06,
+      "loss": 1.8277,
+      "step": 928
+    },
+    {
+      "epoch": 1.0180821917808218,
+      "grad_norm": 2.1830408573150635,
+      "learning_rate": 3.6982128350934206e-06,
+      "loss": 1.8546,
+      "step": 929
+    },
+    {
+      "epoch": 1.0191780821917809,
+      "grad_norm": 2.2692818641662598,
+      "learning_rate": 3.6961819658813975e-06,
+      "loss": 1.8356,
+      "step": 930
+    },
+    {
+      "epoch": 1.0191780821917809,
+      "eval_loss": 2.0714614391326904,
+      "eval_runtime": 36.8661,
+      "eval_samples_per_second": 27.125,
+      "eval_steps_per_second": 1.139,
+      "step": 930
+    },
+    {
+      "epoch": 1.0202739726027397,
+      "grad_norm": 2.5044169425964355,
+      "learning_rate": 3.694151096669375e-06,
+      "loss": 1.8173,
+      "step": 931
+    },
+    {
+      "epoch": 1.0213698630136987,
+      "grad_norm": 2.419637680053711,
+      "learning_rate": 3.692120227457352e-06,
+      "loss": 1.9418,
+      "step": 932
+    },
+    {
+      "epoch": 1.0224657534246575,
+      "grad_norm": 2.282288074493408,
+      "learning_rate": 3.6900893582453297e-06,
+      "loss": 1.8698,
+      "step": 933
+    },
+    {
+      "epoch": 1.0235616438356165,
+      "grad_norm": 2.3215551376342773,
+      "learning_rate": 3.6880584890333066e-06,
+      "loss": 1.8935,
+      "step": 934
+    },
+    {
+      "epoch": 1.0246575342465754,
+      "grad_norm": 2.5938258171081543,
+      "learning_rate": 3.6860276198212834e-06,
+      "loss": 1.79,
+      "step": 935
+    },
+    {
+      "epoch": 1.0257534246575342,
+      "grad_norm": 2.215024471282959,
+      "learning_rate": 3.683996750609261e-06,
+      "loss": 1.8768,
+      "step": 936
+    },
+    {
+      "epoch": 1.0268493150684932,
+      "grad_norm": 2.239609956741333,
+      "learning_rate": 3.681965881397238e-06,
+      "loss": 1.8887,
+      "step": 937
+    },
+    {
+      "epoch": 1.027945205479452,
+      "grad_norm": 2.3134827613830566,
+      "learning_rate": 3.6799350121852157e-06,
+      "loss": 1.8897,
+      "step": 938
+    },
+    {
+      "epoch": 1.029041095890411,
+      "grad_norm": 2.4100260734558105,
+      "learning_rate": 3.6779041429731925e-06,
+      "loss": 1.909,
+      "step": 939
+    },
+    {
+      "epoch": 1.0301369863013699,
+      "grad_norm": 2.26204776763916,
+      "learning_rate": 3.6758732737611702e-06,
+      "loss": 1.8514,
+      "step": 940
+    },
+    {
+      "epoch": 1.0301369863013699,
+      "eval_loss": 2.0666027069091797,
+      "eval_runtime": 36.8336,
+      "eval_samples_per_second": 27.149,
+      "eval_steps_per_second": 1.14,
+      "step": 940
+    },
+    {
+      "epoch": 1.0312328767123287,
+      "grad_norm": 2.22210955619812,
+      "learning_rate": 3.673842404549147e-06,
+      "loss": 1.8743,
+      "step": 941
+    },
+    {
+      "epoch": 1.0323287671232877,
+      "grad_norm": 2.2371468544006348,
+      "learning_rate": 3.671811535337125e-06,
+      "loss": 1.8573,
+      "step": 942
+    },
+    {
+      "epoch": 1.0334246575342465,
+      "grad_norm": 2.346745014190674,
+      "learning_rate": 3.6697806661251016e-06,
+      "loss": 1.8614,
+      "step": 943
+    },
+    {
+      "epoch": 1.0345205479452055,
+      "grad_norm": 2.4765074253082275,
+      "learning_rate": 3.6677497969130793e-06,
+      "loss": 1.8654,
+      "step": 944
+    },
+    {
+      "epoch": 1.0356164383561643,
+      "grad_norm": 2.3023204803466797,
+      "learning_rate": 3.665718927701056e-06,
+      "loss": 1.9901,
+      "step": 945
+    },
+    {
+      "epoch": 1.0367123287671234,
+      "grad_norm": 2.322421073913574,
+      "learning_rate": 3.663688058489034e-06,
+      "loss": 1.8775,
+      "step": 946
+    },
+    {
+      "epoch": 1.0378082191780822,
+      "grad_norm": 2.2324275970458984,
+      "learning_rate": 3.6616571892770107e-06,
+      "loss": 1.9303,
+      "step": 947
+    },
+    {
+      "epoch": 1.038904109589041,
+      "grad_norm": 2.372067928314209,
+      "learning_rate": 3.659626320064988e-06,
+      "loss": 1.8673,
+      "step": 948
+    },
+    {
+      "epoch": 1.04,
+      "grad_norm": 2.2404983043670654,
+      "learning_rate": 3.6575954508529653e-06,
+      "loss": 1.8025,
+      "step": 949
+    },
+    {
+      "epoch": 1.0410958904109588,
+      "grad_norm": 2.180673599243164,
+      "learning_rate": 3.6555645816409426e-06,
+      "loss": 1.8563,
+      "step": 950
+    },
+    {
+      "epoch": 1.0410958904109588,
+      "eval_loss": 2.061262845993042,
+      "eval_runtime": 36.7727,
+      "eval_samples_per_second": 27.194,
+      "eval_steps_per_second": 1.142,
+      "step": 950
+    },
+    {
+      "epoch": 1.0421917808219179,
+      "grad_norm": 2.299043893814087,
+      "learning_rate": 3.65353371242892e-06,
+      "loss": 1.8453,
+      "step": 951
+    },
+    {
+      "epoch": 1.0432876712328767,
+      "grad_norm": 2.4878089427948,
+      "learning_rate": 3.651502843216897e-06,
+      "loss": 1.8901,
+      "step": 952
+    },
+    {
+      "epoch": 1.0443835616438357,
+      "grad_norm": 2.150404930114746,
+      "learning_rate": 3.6494719740048744e-06,
+      "loss": 1.7371,
+      "step": 953
+    },
+    {
+      "epoch": 1.0454794520547945,
+      "grad_norm": 2.177750587463379,
+      "learning_rate": 3.6474411047928517e-06,
+      "loss": 1.8814,
+      "step": 954
+    },
+    {
+      "epoch": 1.0465753424657533,
+      "grad_norm": 2.230797052383423,
+      "learning_rate": 3.645410235580829e-06,
+      "loss": 1.8323,
+      "step": 955
+    },
+    {
+      "epoch": 1.0476712328767124,
+      "grad_norm": 2.2484378814697266,
+      "learning_rate": 3.6433793663688062e-06,
+      "loss": 1.8749,
+      "step": 956
+    },
+    {
+      "epoch": 1.0487671232876712,
+      "grad_norm": 2.201279640197754,
+      "learning_rate": 3.6413484971567835e-06,
+      "loss": 1.8518,
+      "step": 957
+    },
+    {
+      "epoch": 1.0498630136986302,
+      "grad_norm": 2.1125736236572266,
+      "learning_rate": 3.6393176279447608e-06,
+      "loss": 1.846,
+      "step": 958
+    },
+    {
+      "epoch": 1.050958904109589,
+      "grad_norm": 2.3078789710998535,
+      "learning_rate": 3.637286758732738e-06,
+      "loss": 1.8791,
+      "step": 959
+    },
+    {
+      "epoch": 1.0520547945205478,
+      "grad_norm": 2.2911148071289062,
+      "learning_rate": 3.6352558895207153e-06,
+      "loss": 1.8901,
+      "step": 960
+    },
+    {
+      "epoch": 1.0520547945205478,
+      "eval_loss": 2.0583066940307617,
+      "eval_runtime": 36.7535,
+      "eval_samples_per_second": 27.208,
+      "eval_steps_per_second": 1.143,
+      "step": 960
+    },
+    {
+      "epoch": 1.0531506849315069,
+      "grad_norm": 2.3990273475646973,
+      "learning_rate": 3.633225020308692e-06,
+      "loss": 1.9071,
+      "step": 961
+    },
+    {
+      "epoch": 1.0542465753424657,
+      "grad_norm": 2.358685255050659,
+      "learning_rate": 3.63119415109667e-06,
+      "loss": 1.8982,
+      "step": 962
+    },
+    {
+      "epoch": 1.0553424657534247,
+      "grad_norm": 2.1145992279052734,
+      "learning_rate": 3.6291632818846467e-06,
+      "loss": 1.8649,
+      "step": 963
+    },
+    {
+      "epoch": 1.0564383561643835,
+      "grad_norm": 2.2302935123443604,
+      "learning_rate": 3.6271324126726244e-06,
+      "loss": 1.8911,
+      "step": 964
+    },
+    {
+      "epoch": 1.0575342465753426,
+      "grad_norm": 2.344797372817993,
+      "learning_rate": 3.6251015434606013e-06,
+      "loss": 1.8,
+      "step": 965
+    },
+    {
+      "epoch": 1.0586301369863014,
+      "grad_norm": 2.2514073848724365,
+      "learning_rate": 3.623070674248579e-06,
+      "loss": 1.8128,
+      "step": 966
+    },
+    {
+      "epoch": 1.0597260273972602,
+      "grad_norm": 2.541221857070923,
+      "learning_rate": 3.621039805036556e-06,
+      "loss": 1.8504,
+      "step": 967
+    },
+    {
+      "epoch": 1.0608219178082192,
+      "grad_norm": 2.3211770057678223,
+      "learning_rate": 3.6190089358245335e-06,
+      "loss": 1.8329,
+      "step": 968
+    },
+    {
+      "epoch": 1.061917808219178,
+      "grad_norm": 2.1008756160736084,
+      "learning_rate": 3.6169780666125104e-06,
+      "loss": 1.8972,
+      "step": 969
+    },
+    {
+      "epoch": 1.063013698630137,
+      "grad_norm": 2.4148898124694824,
+      "learning_rate": 3.614947197400488e-06,
+      "loss": 1.8748,
+      "step": 970
+    },
+    {
+      "epoch": 1.063013698630137,
+      "eval_loss": 2.0592916011810303,
+      "eval_runtime": 36.7663,
+      "eval_samples_per_second": 27.199,
+      "eval_steps_per_second": 1.142,
+      "step": 970
+    },
+    {
+      "epoch": 1.0641095890410959,
+      "grad_norm": 2.4116220474243164,
+      "learning_rate": 3.612916328188465e-06,
+      "loss": 1.8915,
+      "step": 971
+    },
+    {
+      "epoch": 1.0652054794520547,
+      "grad_norm": 2.257411241531372,
+      "learning_rate": 3.610885458976442e-06,
+      "loss": 1.7901,
+      "step": 972
+    },
+    {
+      "epoch": 1.0663013698630137,
+      "grad_norm": 2.166875123977661,
+      "learning_rate": 3.6088545897644195e-06,
+      "loss": 1.8969,
+      "step": 973
+    },
+    {
+      "epoch": 1.0673972602739725,
+      "grad_norm": 2.464716911315918,
+      "learning_rate": 3.6068237205523963e-06,
+      "loss": 1.8844,
+      "step": 974
+    },
+    {
+      "epoch": 1.0684931506849316,
+      "grad_norm": 2.337334156036377,
+      "learning_rate": 3.604792851340374e-06,
+      "loss": 1.8192,
+      "step": 975
+    },
+    {
+      "epoch": 1.0695890410958904,
+      "grad_norm": 2.117663621902466,
+      "learning_rate": 3.602761982128351e-06,
+      "loss": 1.8233,
+      "step": 976
+    },
+    {
+      "epoch": 1.0706849315068494,
+      "grad_norm": 2.3353586196899414,
+      "learning_rate": 3.6007311129163286e-06,
+      "loss": 1.9133,
+      "step": 977
+    },
+    {
+      "epoch": 1.0717808219178082,
+      "grad_norm": 2.444279909133911,
+      "learning_rate": 3.5987002437043054e-06,
+      "loss": 1.8803,
+      "step": 978
+    },
+    {
+      "epoch": 1.072876712328767,
+      "grad_norm": 2.3458383083343506,
+      "learning_rate": 3.596669374492283e-06,
+      "loss": 1.8108,
+      "step": 979
+    },
+    {
+      "epoch": 1.073972602739726,
+      "grad_norm": 2.207977771759033,
+      "learning_rate": 3.59463850528026e-06,
+      "loss": 1.8423,
+      "step": 980
+    },
+    {
+      "epoch": 1.073972602739726,
+      "eval_loss": 2.0578773021698,
+      "eval_runtime": 36.8123,
+      "eval_samples_per_second": 27.165,
+      "eval_steps_per_second": 1.141,
+      "step": 980
+    },
+    {
+      "epoch": 1.0750684931506849,
+      "grad_norm": 2.378127336502075,
+      "learning_rate": 3.5926076360682377e-06,
+      "loss": 1.9268,
+      "step": 981
+    },
+    {
+      "epoch": 1.076164383561644,
+      "grad_norm": 2.248105525970459,
+      "learning_rate": 3.5905767668562145e-06,
+      "loss": 1.8111,
+      "step": 982
+    },
+    {
+      "epoch": 1.0772602739726027,
+      "grad_norm": 2.275592088699341,
+      "learning_rate": 3.5885458976441922e-06,
+      "loss": 1.9099,
+      "step": 983
+    },
+    {
+      "epoch": 1.0783561643835617,
+      "grad_norm": 3.108377456665039,
+      "learning_rate": 3.586515028432169e-06,
+      "loss": 1.8728,
+      "step": 984
+    },
+    {
+      "epoch": 1.0794520547945206,
+      "grad_norm": 2.2702999114990234,
+      "learning_rate": 3.5844841592201464e-06,
+      "loss": 1.9007,
+      "step": 985
+    },
+    {
+      "epoch": 1.0805479452054794,
+      "grad_norm": 3.218647003173828,
+      "learning_rate": 3.5824532900081237e-06,
+      "loss": 1.9157,
+      "step": 986
+    },
+    {
+      "epoch": 1.0816438356164384,
+      "grad_norm": 2.3393027782440186,
+      "learning_rate": 3.580422420796101e-06,
+      "loss": 1.7936,
+      "step": 987
+    },
+    {
+      "epoch": 1.0827397260273972,
+      "grad_norm": 2.420541286468506,
+      "learning_rate": 3.578391551584078e-06,
+      "loss": 1.8608,
+      "step": 988
+    },
+    {
+      "epoch": 1.0838356164383562,
+      "grad_norm": 2.3354556560516357,
+      "learning_rate": 3.5763606823720555e-06,
+      "loss": 1.8684,
+      "step": 989
+    },
+    {
+      "epoch": 1.084931506849315,
+      "grad_norm": 2.200282573699951,
+      "learning_rate": 3.5743298131600328e-06,
+      "loss": 1.7814,
+      "step": 990
+    },
+    {
+      "epoch": 1.084931506849315,
+      "eval_loss": 2.0565249919891357,
+      "eval_runtime": 36.952,
+      "eval_samples_per_second": 27.062,
+      "eval_steps_per_second": 1.137,
+      "step": 990
+    },
+    {
+      "epoch": 1.086027397260274,
+      "grad_norm": 2.149033308029175,
+      "learning_rate": 3.57229894394801e-06,
+      "loss": 1.8708,
+      "step": 991
+    },
+    {
+      "epoch": 1.087123287671233,
+      "grad_norm": 2.4663679599761963,
+      "learning_rate": 3.5702680747359873e-06,
+      "loss": 1.8947,
+      "step": 992
+    },
+    {
+      "epoch": 1.0882191780821917,
+      "grad_norm": 2.3142261505126953,
+      "learning_rate": 3.5682372055239646e-06,
+      "loss": 1.8727,
+      "step": 993
+    },
+    {
+      "epoch": 1.0893150684931507,
+      "grad_norm": 2.2223033905029297,
+      "learning_rate": 3.566206336311942e-06,
+      "loss": 1.8203,
+      "step": 994
+    },
+    {
+      "epoch": 1.0904109589041096,
+      "grad_norm": 2.423855781555176,
+      "learning_rate": 3.564175467099919e-06,
+      "loss": 1.8687,
+      "step": 995
+    },
+    {
+      "epoch": 1.0915068493150686,
+      "grad_norm": 2.4022302627563477,
+      "learning_rate": 3.5621445978878964e-06,
+      "loss": 1.8215,
+      "step": 996
+    },
+    {
+      "epoch": 1.0926027397260274,
+      "grad_norm": 2.378236770629883,
+      "learning_rate": 3.5601137286758737e-06,
+      "loss": 1.7994,
+      "step": 997
+    },
+    {
+      "epoch": 1.0936986301369862,
+      "grad_norm": 2.4499096870422363,
+      "learning_rate": 3.5580828594638505e-06,
+      "loss": 1.8779,
+      "step": 998
+    },
+    {
+      "epoch": 1.0947945205479452,
+      "grad_norm": 2.3181755542755127,
+      "learning_rate": 3.5560519902518282e-06,
+      "loss": 1.8341,
+      "step": 999
+    },
+    {
+      "epoch": 1.095890410958904,
+      "grad_norm": 2.2406623363494873,
+      "learning_rate": 3.554021121039805e-06,
+      "loss": 1.874,
+      "step": 1000
+    },
+    {
+      "epoch": 1.095890410958904,
+      "eval_loss": 2.0523672103881836,
+      "eval_runtime": 37.0454,
+      "eval_samples_per_second": 26.994,
+      "eval_steps_per_second": 1.134,
+      "step": 1000
+    },
+    {
+      "epoch": 1.096986301369863,
+      "grad_norm": 2.254667043685913,
+      "learning_rate": 3.551990251827783e-06,
+      "loss": 1.9017,
+      "step": 1001
+    },
+    {
+      "epoch": 1.098082191780822,
+      "grad_norm": 2.2846896648406982,
+      "learning_rate": 3.5499593826157596e-06,
+      "loss": 1.8505,
+      "step": 1002
+    },
+    {
+      "epoch": 1.099178082191781,
+      "grad_norm": 2.183469772338867,
+      "learning_rate": 3.5479285134037373e-06,
+      "loss": 1.8266,
+      "step": 1003
+    },
+    {
+      "epoch": 1.1002739726027397,
+      "grad_norm": 2.151798963546753,
+      "learning_rate": 3.545897644191714e-06,
+      "loss": 1.8487,
+      "step": 1004
+    },
+    {
+      "epoch": 1.1013698630136985,
+      "grad_norm": 2.2039146423339844,
+      "learning_rate": 3.543866774979692e-06,
+      "loss": 1.8327,
+      "step": 1005
+    },
+    {
+      "epoch": 1.1024657534246576,
+      "grad_norm": 2.2091665267944336,
+      "learning_rate": 3.5418359057676687e-06,
+      "loss": 1.8214,
+      "step": 1006
+    },
+    {
+      "epoch": 1.1035616438356164,
+      "grad_norm": 2.2821404933929443,
+      "learning_rate": 3.5398050365556464e-06,
+      "loss": 1.8591,
+      "step": 1007
+    },
+    {
+      "epoch": 1.1046575342465754,
+      "grad_norm": 2.1726973056793213,
+      "learning_rate": 3.5377741673436233e-06,
+      "loss": 1.8265,
+      "step": 1008
+    },
+    {
+      "epoch": 1.1057534246575342,
+      "grad_norm": 2.2870233058929443,
+      "learning_rate": 3.535743298131601e-06,
+      "loss": 1.8783,
+      "step": 1009
+    },
+    {
+      "epoch": 1.106849315068493,
+      "grad_norm": 2.148080587387085,
+      "learning_rate": 3.533712428919578e-06,
+      "loss": 1.8272,
+      "step": 1010
+    },
+    {
+      "epoch": 1.106849315068493,
+      "eval_loss": 2.049504041671753,
+      "eval_runtime": 36.8717,
+      "eval_samples_per_second": 27.121,
+      "eval_steps_per_second": 1.139,
+      "step": 1010
+    },
+    {
+      "epoch": 1.107945205479452,
+      "grad_norm": 2.182035446166992,
+      "learning_rate": 3.5316815597075547e-06,
+      "loss": 1.8771,
+      "step": 1011
+    },
+    {
+      "epoch": 1.1090410958904109,
+      "grad_norm": 2.135085344314575,
+      "learning_rate": 3.5296506904955324e-06,
+      "loss": 1.9031,
+      "step": 1012
+    },
+    {
+      "epoch": 1.11013698630137,
+      "grad_norm": 2.1017255783081055,
+      "learning_rate": 3.5276198212835093e-06,
+      "loss": 1.8341,
+      "step": 1013
+    },
+    {
+      "epoch": 1.1112328767123287,
+      "grad_norm": 2.2773115634918213,
+      "learning_rate": 3.525588952071487e-06,
+      "loss": 1.8564,
+      "step": 1014
+    },
+    {
+      "epoch": 1.1123287671232878,
+      "grad_norm": 2.2773115634918213,
+      "learning_rate": 3.525588952071487e-06,
+      "loss": 1.8971,
+      "step": 1015
+    },
+    {
+      "epoch": 1.1134246575342466,
+      "grad_norm": 2.273375988006592,
+      "learning_rate": 3.523558082859464e-06,
+      "loss": 1.9074,
+      "step": 1016
+    },
+    {
+      "epoch": 1.1145205479452054,
+      "grad_norm": 2.2055182456970215,
+      "learning_rate": 3.5215272136474415e-06,
+      "loss": 1.8724,
+      "step": 1017
+    },
+    {
+      "epoch": 1.1156164383561644,
+      "grad_norm": 2.246126651763916,
+      "learning_rate": 3.5194963444354184e-06,
+      "loss": 1.8387,
+      "step": 1018
+    },
+    {
+      "epoch": 1.1167123287671232,
+      "grad_norm": 2.3076183795928955,
+      "learning_rate": 3.517465475223396e-06,
+      "loss": 1.8387,
+      "step": 1019
+    },
+    {
+      "epoch": 1.1178082191780823,
+      "grad_norm": 2.3365488052368164,
+      "learning_rate": 3.515434606011373e-06,
+      "loss": 1.8063,
+      "step": 1020
+    },
+    {
+      "epoch": 1.1178082191780823,
+      "eval_loss": 2.0528781414031982,
+      "eval_runtime": 36.8885,
+      "eval_samples_per_second": 27.109,
+      "eval_steps_per_second": 1.139,
+      "step": 1020
+    },
+    {
+      "epoch": 1.118904109589041,
+      "grad_norm": 2.319324493408203,
+      "learning_rate": 3.5134037367993506e-06,
+      "loss": 1.8843,
+      "step": 1021
+    },
+    {
+      "epoch": 1.12,
+      "grad_norm": 2.238720655441284,
+      "learning_rate": 3.5113728675873275e-06,
+      "loss": 1.9375,
+      "step": 1022
+    },
+    {
+      "epoch": 1.121095890410959,
+      "grad_norm": 2.2824833393096924,
+      "learning_rate": 3.5093419983753047e-06,
+      "loss": 1.8749,
+      "step": 1023
+    },
+    {
+      "epoch": 1.1221917808219177,
+      "grad_norm": 2.5120770931243896,
+      "learning_rate": 3.507311129163282e-06,
+      "loss": 1.8774,
+      "step": 1024
+    },
+    {
+      "epoch": 1.1232876712328768,
+      "grad_norm": 2.1762094497680664,
+      "learning_rate": 3.5052802599512593e-06,
+      "loss": 1.8018,
+      "step": 1025
+    },
+    {
+      "epoch": 1.1243835616438356,
+      "grad_norm": 2.2530343532562256,
+      "learning_rate": 3.5032493907392366e-06,
+      "loss": 1.8451,
+      "step": 1026
+    },
+    {
+      "epoch": 1.1254794520547946,
+      "grad_norm": 2.21649169921875,
+      "learning_rate": 3.501218521527214e-06,
+      "loss": 1.8613,
+      "step": 1027
+    },
+    {
+      "epoch": 1.1265753424657534,
+      "grad_norm": 2.306431293487549,
+      "learning_rate": 3.499187652315191e-06,
+      "loss": 1.7947,
+      "step": 1028
+    },
+    {
+      "epoch": 1.1276712328767124,
+      "grad_norm": 2.203005313873291,
+      "learning_rate": 3.4971567831031684e-06,
+      "loss": 1.8483,
+      "step": 1029
+    },
+    {
+      "epoch": 1.1287671232876713,
+      "grad_norm": 2.1645009517669678,
+      "learning_rate": 3.4951259138911457e-06,
+      "loss": 1.8915,
+      "step": 1030
+    },
+    {
+      "epoch": 1.1287671232876713,
+      "eval_loss": 2.053393602371216,
+      "eval_runtime": 36.855,
+      "eval_samples_per_second": 27.133,
+      "eval_steps_per_second": 1.14,
+      "step": 1030
+    },
+    {
+      "epoch": 1.12986301369863,
+      "grad_norm": 2.2863385677337646,
+      "learning_rate": 3.493095044679123e-06,
+      "loss": 1.8073,
+      "step": 1031
+    },
+    {
+      "epoch": 1.130958904109589,
+      "grad_norm": 2.1439812183380127,
+      "learning_rate": 3.4910641754671002e-06,
+      "loss": 1.9191,
+      "step": 1032
+    },
+    {
+      "epoch": 1.132054794520548,
+      "grad_norm": 2.224228858947754,
+      "learning_rate": 3.4890333062550775e-06,
+      "loss": 1.7478,
+      "step": 1033
+    },
+    {
+      "epoch": 1.133150684931507,
+      "grad_norm": 2.319059133529663,
+      "learning_rate": 3.4870024370430548e-06,
+      "loss": 1.8779,
+      "step": 1034
+    },
+    {
+      "epoch": 1.1342465753424658,
+      "grad_norm": 2.5003139972686768,
+      "learning_rate": 3.484971567831032e-06,
+      "loss": 1.776,
+      "step": 1035
+    },
+    {
+      "epoch": 1.1353424657534246,
+      "grad_norm": 2.2032580375671387,
+      "learning_rate": 3.482940698619009e-06,
+      "loss": 1.8221,
+      "step": 1036
+    },
+    {
+      "epoch": 1.1364383561643836,
+      "grad_norm": 2.5253560543060303,
+      "learning_rate": 3.4809098294069866e-06,
+      "loss": 1.8952,
+      "step": 1037
+    },
+    {
+      "epoch": 1.1375342465753424,
+      "grad_norm": 2.3215854167938232,
+      "learning_rate": 3.4788789601949634e-06,
+      "loss": 1.847,
+      "step": 1038
+    },
+    {
+      "epoch": 1.1386301369863014,
+      "grad_norm": 2.2543280124664307,
+      "learning_rate": 3.476848090982941e-06,
+      "loss": 1.8689,
+      "step": 1039
+    },
+    {
+      "epoch": 1.1397260273972603,
+      "grad_norm": 2.1592042446136475,
+      "learning_rate": 3.474817221770918e-06,
+      "loss": 1.8375,
+      "step": 1040
+    },
+    {
+      "epoch": 1.1397260273972603,
+      "eval_loss": 2.0524044036865234,
+      "eval_runtime": 36.8163,
+      "eval_samples_per_second": 27.162,
+      "eval_steps_per_second": 1.141,
+      "step": 1040
+    },
+    {
+      "epoch": 1.140821917808219,
+      "grad_norm": 2.2874810695648193,
+      "learning_rate": 3.4727863525588957e-06,
+      "loss": 1.9442,
+      "step": 1041
+    },
+    {
+      "epoch": 1.141917808219178,
+      "grad_norm": 2.288395404815674,
+      "learning_rate": 3.4707554833468725e-06,
+      "loss": 1.8881,
+      "step": 1042
+    },
+    {
+      "epoch": 1.143013698630137,
+      "grad_norm": 2.2264153957366943,
+      "learning_rate": 3.4687246141348502e-06,
+      "loss": 1.8518,
+      "step": 1043
+    },
+    {
+      "epoch": 1.144109589041096,
+      "grad_norm": 2.4413652420043945,
+      "learning_rate": 3.466693744922827e-06,
+      "loss": 1.8146,
+      "step": 1044
+    },
+    {
+      "epoch": 1.1452054794520548,
+      "grad_norm": 2.3597898483276367,
+      "learning_rate": 3.464662875710805e-06,
+      "loss": 1.8421,
+      "step": 1045
+    },
+    {
+      "epoch": 1.1463013698630138,
+      "grad_norm": 2.2489402294158936,
+      "learning_rate": 3.4626320064987817e-06,
+      "loss": 1.8643,
+      "step": 1046
+    },
+    {
+      "epoch": 1.1473972602739726,
+      "grad_norm": 2.3958940505981445,
+      "learning_rate": 3.4606011372867594e-06,
+      "loss": 1.8522,
+      "step": 1047
+    },
+    {
+      "epoch": 1.1484931506849314,
+      "grad_norm": 2.295349597930908,
+      "learning_rate": 3.458570268074736e-06,
+      "loss": 1.8677,
+      "step": 1048
+    },
+    {
+      "epoch": 1.1495890410958904,
+      "grad_norm": 2.2571280002593994,
+      "learning_rate": 3.4565393988627135e-06,
+      "loss": 1.8393,
+      "step": 1049
+    },
+    {
+      "epoch": 1.1506849315068493,
+      "grad_norm": 2.173983335494995,
+      "learning_rate": 3.4545085296506908e-06,
+      "loss": 1.8267,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1506849315068493,
+      "eval_loss": 2.0489604473114014,
+      "eval_runtime": 36.7678,
+      "eval_samples_per_second": 27.198,
+      "eval_steps_per_second": 1.142,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1517808219178083,
+      "grad_norm": 2.3395209312438965,
+      "learning_rate": 3.452477660438668e-06,
+      "loss": 1.8032,
+      "step": 1051
+    },
+    {
+      "epoch": 1.152876712328767,
+      "grad_norm": 2.1894757747650146,
+      "learning_rate": 3.4504467912266453e-06,
+      "loss": 1.8834,
+      "step": 1052
+    },
+    {
+      "epoch": 1.1539726027397261,
+      "grad_norm": 2.1978437900543213,
+      "learning_rate": 3.4484159220146226e-06,
+      "loss": 1.7897,
+      "step": 1053
+    },
+    {
+      "epoch": 1.155068493150685,
+      "grad_norm": 2.275792360305786,
+      "learning_rate": 3.4463850528026e-06,
+      "loss": 1.8566,
+      "step": 1054
+    },
+    {
+      "epoch": 1.1561643835616437,
+      "grad_norm": 2.2659800052642822,
+      "learning_rate": 3.444354183590577e-06,
+      "loss": 1.8307,
+      "step": 1055
+    },
+    {
+      "epoch": 1.1572602739726028,
+      "grad_norm": 2.232149600982666,
+      "learning_rate": 3.4423233143785544e-06,
+      "loss": 1.8701,
+      "step": 1056
+    },
+    {
+      "epoch": 1.1583561643835616,
+      "grad_norm": 2.3636281490325928,
+      "learning_rate": 3.4402924451665317e-06,
+      "loss": 1.8008,
+      "step": 1057
+    },
+    {
+      "epoch": 1.1594520547945206,
+      "grad_norm": 2.271533489227295,
+      "learning_rate": 3.438261575954509e-06,
+      "loss": 1.7751,
+      "step": 1058
+    },
+    {
+      "epoch": 1.1605479452054794,
+      "grad_norm": 2.3286566734313965,
+      "learning_rate": 3.4362307067424862e-06,
+      "loss": 1.865,
+      "step": 1059
+    },
+    {
+      "epoch": 1.1616438356164385,
+      "grad_norm": 2.46665096282959,
+      "learning_rate": 3.434199837530463e-06,
+      "loss": 1.9001,
+      "step": 1060
+    },
+    {
+      "epoch": 1.1616438356164385,
+      "eval_loss": 2.0477516651153564,
+      "eval_runtime": 36.7818,
+      "eval_samples_per_second": 27.187,
+      "eval_steps_per_second": 1.142,
+      "step": 1060
+    },
+    {
+      "epoch": 1.1627397260273973,
+      "grad_norm": 2.307744264602661,
+      "learning_rate": 3.432168968318441e-06,
+      "loss": 1.8637,
+      "step": 1061
+    },
+    {
+      "epoch": 1.163835616438356,
+      "grad_norm": 2.3935985565185547,
+      "learning_rate": 3.4301380991064176e-06,
+      "loss": 1.7867,
+      "step": 1062
+    },
+    {
+      "epoch": 1.1649315068493151,
+      "grad_norm": 2.290971040725708,
+      "learning_rate": 3.4281072298943953e-06,
+      "loss": 1.8708,
+      "step": 1063
+    },
+    {
+      "epoch": 1.166027397260274,
+      "grad_norm": 2.191629409790039,
+      "learning_rate": 3.426076360682372e-06,
+      "loss": 1.9031,
+      "step": 1064
+    },
+    {
+      "epoch": 1.167123287671233,
+      "grad_norm": 2.284426689147949,
+      "learning_rate": 3.42404549147035e-06,
+      "loss": 1.9026,
+      "step": 1065
+    },
+    {
+      "epoch": 1.1682191780821918,
+      "grad_norm": 2.2567849159240723,
+      "learning_rate": 3.4220146222583267e-06,
+      "loss": 1.8507,
+      "step": 1066
+    },
+    {
+      "epoch": 1.1693150684931506,
+      "grad_norm": 2.2093160152435303,
+      "learning_rate": 3.4199837530463044e-06,
+      "loss": 1.8001,
+      "step": 1067
+    },
+    {
+      "epoch": 1.1704109589041096,
+      "grad_norm": 2.092560291290283,
+      "learning_rate": 3.4179528838342813e-06,
+      "loss": 1.8393,
+      "step": 1068
+    },
+    {
+      "epoch": 1.1715068493150684,
+      "grad_norm": 2.313749074935913,
+      "learning_rate": 3.415922014622259e-06,
+      "loss": 1.9385,
+      "step": 1069
+    },
+    {
+      "epoch": 1.1726027397260275,
+      "grad_norm": 2.488673210144043,
+      "learning_rate": 3.413891145410236e-06,
+      "loss": 1.8924,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1726027397260275,
+      "eval_loss": 2.0482401847839355,
+      "eval_runtime": 36.7919,
+      "eval_samples_per_second": 27.18,
+      "eval_steps_per_second": 1.142,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1736986301369863,
+      "grad_norm": 2.0804412364959717,
+      "learning_rate": 3.4118602761982135e-06,
+      "loss": 1.881,
+      "step": 1071
+    },
+    {
+      "epoch": 1.174794520547945,
+      "grad_norm": 2.165426254272461,
+      "learning_rate": 3.4098294069861904e-06,
+      "loss": 1.9035,
+      "step": 1072
+    },
+    {
+      "epoch": 1.1758904109589041,
+      "grad_norm": 2.3034675121307373,
+      "learning_rate": 3.4077985377741673e-06,
+      "loss": 1.8554,
+      "step": 1073
+    },
+    {
+      "epoch": 1.176986301369863,
+      "grad_norm": 2.303159713745117,
+      "learning_rate": 3.405767668562145e-06,
+      "loss": 1.8255,
+      "step": 1074
+    },
+    {
+      "epoch": 1.178082191780822,
+      "grad_norm": 2.2402753829956055,
+      "learning_rate": 3.403736799350122e-06,
+      "loss": 1.9087,
+      "step": 1075
+    },
+    {
+      "epoch": 1.1791780821917808,
+      "grad_norm": 2.251638650894165,
+      "learning_rate": 3.4017059301380995e-06,
+      "loss": 1.8878,
+      "step": 1076
+    },
+    {
+      "epoch": 1.1802739726027398,
+      "grad_norm": 2.2661657333374023,
+      "learning_rate": 3.3996750609260764e-06,
+      "loss": 1.8244,
+      "step": 1077
+    },
+    {
+      "epoch": 1.1813698630136986,
+      "grad_norm": 2.3280904293060303,
+      "learning_rate": 3.397644191714054e-06,
+      "loss": 1.9328,
+      "step": 1078
+    },
+    {
+      "epoch": 1.1824657534246574,
+      "grad_norm": 2.315779447555542,
+      "learning_rate": 3.395613322502031e-06,
+      "loss": 1.8289,
+      "step": 1079
+    },
+    {
+      "epoch": 1.1835616438356165,
+      "grad_norm": 2.136967182159424,
+      "learning_rate": 3.3935824532900086e-06,
+      "loss": 1.8307,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1835616438356165,
+      "eval_loss": 2.0473902225494385,
+      "eval_runtime": 36.8094,
+      "eval_samples_per_second": 27.167,
+      "eval_steps_per_second": 1.141,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1846575342465753,
+      "grad_norm": 2.3701720237731934,
+      "learning_rate": 3.3915515840779855e-06,
+      "loss": 1.8123,
+      "step": 1081
+    },
+    {
+      "epoch": 1.1857534246575343,
+      "grad_norm": 2.2662320137023926,
+      "learning_rate": 3.389520714865963e-06,
+      "loss": 1.85,
+      "step": 1082
+    },
+    {
+      "epoch": 1.1868493150684931,
+      "grad_norm": 2.2151782512664795,
+      "learning_rate": 3.38748984565394e-06,
+      "loss": 1.845,
+      "step": 1083
+    },
+    {
+      "epoch": 1.1879452054794521,
+      "grad_norm": 2.2209115028381348,
+      "learning_rate": 3.3854589764419177e-06,
+      "loss": 1.8787,
+      "step": 1084
+    },
+    {
+      "epoch": 1.189041095890411,
+      "grad_norm": 2.4760477542877197,
+      "learning_rate": 3.3834281072298946e-06,
+      "loss": 1.818,
+      "step": 1085
+    },
+    {
+      "epoch": 1.1901369863013698,
+      "grad_norm": 2.309034585952759,
+      "learning_rate": 3.381397238017872e-06,
+      "loss": 1.8387,
+      "step": 1086
+    },
+    {
+      "epoch": 1.1912328767123288,
+      "grad_norm": 2.2174596786499023,
+      "learning_rate": 3.379366368805849e-06,
+      "loss": 1.8135,
+      "step": 1087
+    },
+    {
+      "epoch": 1.1923287671232876,
+      "grad_norm": 2.3043556213378906,
+      "learning_rate": 3.3773354995938264e-06,
+      "loss": 1.9156,
+      "step": 1088
+    },
+    {
+      "epoch": 1.1934246575342466,
+      "grad_norm": 2.2240524291992188,
+      "learning_rate": 3.3753046303818037e-06,
+      "loss": 1.8636,
+      "step": 1089
+    },
+    {
+      "epoch": 1.1945205479452055,
+      "grad_norm": 2.1108691692352295,
+      "learning_rate": 3.373273761169781e-06,
+      "loss": 1.8458,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1945205479452055,
+      "eval_loss": 2.04378604888916,
+      "eval_runtime": 36.9058,
+      "eval_samples_per_second": 27.096,
+      "eval_steps_per_second": 1.138,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1956164383561645,
+      "grad_norm": 2.1893398761749268,
+      "learning_rate": 3.3712428919577582e-06,
+      "loss": 1.8446,
+      "step": 1091
+    },
+    {
+      "epoch": 1.1967123287671233,
+      "grad_norm": 2.305298089981079,
+      "learning_rate": 3.3692120227457355e-06,
+      "loss": 1.8283,
+      "step": 1092
+    },
+    {
+      "epoch": 1.1978082191780821,
+      "grad_norm": 2.364433765411377,
+      "learning_rate": 3.3671811535337128e-06,
+      "loss": 1.8708,
+      "step": 1093
+    },
+    {
+      "epoch": 1.1989041095890411,
+      "grad_norm": 2.259892225265503,
+      "learning_rate": 3.36515028432169e-06,
+      "loss": 1.7966,
+      "step": 1094
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 2.2487263679504395,
+      "learning_rate": 3.3631194151096673e-06,
+      "loss": 1.8238,
+      "step": 1095
+    },
+    {
+      "epoch": 1.201095890410959,
+      "grad_norm": 2.1592557430267334,
+      "learning_rate": 3.3610885458976446e-06,
+      "loss": 1.8651,
+      "step": 1096
+    },
+    {
+      "epoch": 1.2021917808219178,
+      "grad_norm": 2.257737159729004,
+      "learning_rate": 3.359057676685622e-06,
+      "loss": 1.8926,
+      "step": 1097
+    },
+    {
+      "epoch": 1.2032876712328768,
+      "grad_norm": 2.0702707767486572,
+      "learning_rate": 3.357026807473599e-06,
+      "loss": 1.8833,
+      "step": 1098
+    },
+    {
+      "epoch": 1.2043835616438356,
+      "grad_norm": 2.1568455696105957,
+      "learning_rate": 3.354995938261576e-06,
+      "loss": 1.7732,
+      "step": 1099
+    },
+    {
+      "epoch": 1.2054794520547945,
+      "grad_norm": 2.294715404510498,
+      "learning_rate": 3.3529650690495537e-06,
+      "loss": 1.8205,
+      "step": 1100
+    },
+    {
+      "epoch": 1.2054794520547945,
+      "eval_loss": 2.044201612472534,
+      "eval_runtime": 36.9861,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 1100
+    },
+    {
+      "epoch": 1.2065753424657535,
+      "grad_norm": 2.3234570026397705,
+      "learning_rate": 3.3509341998375305e-06,
+      "loss": 1.7935,
+      "step": 1101
+    },
+    {
+      "epoch": 1.2076712328767123,
+      "grad_norm": 2.307224750518799,
+      "learning_rate": 3.3489033306255082e-06,
+      "loss": 1.8254,
+      "step": 1102
+    },
+    {
+      "epoch": 1.208767123287671,
+      "grad_norm": 2.3725457191467285,
+      "learning_rate": 3.346872461413485e-06,
+      "loss": 1.8493,
+      "step": 1103
+    },
+    {
+      "epoch": 1.2098630136986301,
+      "grad_norm": 2.2057435512542725,
+      "learning_rate": 3.344841592201463e-06,
+      "loss": 1.8241,
+      "step": 1104
+    },
+    {
+      "epoch": 1.210958904109589,
+      "grad_norm": 2.232947826385498,
+      "learning_rate": 3.3428107229894397e-06,
+      "loss": 1.8057,
+      "step": 1105
+    },
+    {
+      "epoch": 1.212054794520548,
+      "grad_norm": 2.4268925189971924,
+      "learning_rate": 3.3407798537774174e-06,
+      "loss": 1.8334,
+      "step": 1106
+    },
+    {
+      "epoch": 1.2131506849315068,
+      "grad_norm": 2.3134701251983643,
+      "learning_rate": 3.338748984565394e-06,
+      "loss": 1.8516,
+      "step": 1107
+    },
+    {
+      "epoch": 1.2142465753424658,
+      "grad_norm": 2.375873327255249,
+      "learning_rate": 3.336718115353372e-06,
+      "loss": 1.8942,
+      "step": 1108
+    },
+    {
+      "epoch": 1.2153424657534246,
+      "grad_norm": 2.622528076171875,
+      "learning_rate": 3.3346872461413488e-06,
+      "loss": 1.9027,
+      "step": 1109
+    },
+    {
+      "epoch": 1.2164383561643834,
+      "grad_norm": 2.290865898132324,
+      "learning_rate": 3.3326563769293256e-06,
+      "loss": 1.8258,
+      "step": 1110
+    },
+    {
+      "epoch": 1.2164383561643834,
+      "eval_loss": 2.0408973693847656,
+      "eval_runtime": 36.8596,
+      "eval_samples_per_second": 27.13,
+      "eval_steps_per_second": 1.139,
+      "step": 1110
+    },
+    {
+      "epoch": 1.2175342465753425,
+      "grad_norm": 2.1596908569335938,
+      "learning_rate": 3.3306255077173033e-06,
+      "loss": 1.8066,
+      "step": 1111
+    },
+    {
+      "epoch": 1.2186301369863013,
+      "grad_norm": 2.4652187824249268,
+      "learning_rate": 3.32859463850528e-06,
+      "loss": 1.8372,
+      "step": 1112
+    },
+    {
+      "epoch": 1.2197260273972603,
+      "grad_norm": 2.282440185546875,
+      "learning_rate": 3.326563769293258e-06,
+      "loss": 1.8535,
+      "step": 1113
+    },
+    {
+      "epoch": 1.2208219178082191,
+      "grad_norm": 2.2390308380126953,
+      "learning_rate": 3.3245329000812347e-06,
+      "loss": 1.7719,
+      "step": 1114
+    },
+    {
+      "epoch": 1.2219178082191782,
+      "grad_norm": 2.368565320968628,
+      "learning_rate": 3.3225020308692124e-06,
+      "loss": 1.8207,
+      "step": 1115
+    },
+    {
+      "epoch": 1.223013698630137,
+      "grad_norm": 2.463942766189575,
+      "learning_rate": 3.3204711616571893e-06,
+      "loss": 1.8402,
+      "step": 1116
+    },
+    {
+      "epoch": 1.2241095890410958,
+      "grad_norm": 2.266096830368042,
+      "learning_rate": 3.318440292445167e-06,
+      "loss": 1.7854,
+      "step": 1117
+    },
+    {
+      "epoch": 1.2252054794520548,
+      "grad_norm": 2.3463494777679443,
+      "learning_rate": 3.316409423233144e-06,
+      "loss": 1.8449,
+      "step": 1118
+    },
+    {
+      "epoch": 1.2263013698630136,
+      "grad_norm": 2.381727933883667,
+      "learning_rate": 3.3143785540211215e-06,
+      "loss": 1.8085,
+      "step": 1119
+    },
+    {
+      "epoch": 1.2273972602739727,
+      "grad_norm": 2.3251113891601562,
+      "learning_rate": 3.3123476848090984e-06,
+      "loss": 1.8922,
+      "step": 1120
+    },
+    {
+      "epoch": 1.2273972602739727,
+      "eval_loss": 2.0425429344177246,
+      "eval_runtime": 36.851,
+      "eval_samples_per_second": 27.136,
+      "eval_steps_per_second": 1.14,
+      "step": 1120
+    },
+    {
+      "epoch": 1.2284931506849315,
+      "grad_norm": 2.2314298152923584,
+      "learning_rate": 3.310316815597076e-06,
+      "loss": 1.8474,
+      "step": 1121
+    },
+    {
+      "epoch": 1.2295890410958905,
+      "grad_norm": 2.1847472190856934,
+      "learning_rate": 3.308285946385053e-06,
+      "loss": 1.8454,
+      "step": 1122
+    },
+    {
+      "epoch": 1.2306849315068493,
+      "grad_norm": 2.249422788619995,
+      "learning_rate": 3.30625507717303e-06,
+      "loss": 1.9257,
+      "step": 1123
+    },
+    {
+      "epoch": 1.2317808219178081,
+      "grad_norm": 2.223930835723877,
+      "learning_rate": 3.3042242079610075e-06,
+      "loss": 1.7816,
+      "step": 1124
+    },
+    {
+      "epoch": 1.2328767123287672,
+      "grad_norm": 2.3209826946258545,
+      "learning_rate": 3.3021933387489847e-06,
+      "loss": 1.8773,
+      "step": 1125
+    },
+    {
+      "epoch": 1.233972602739726,
+      "grad_norm": 2.2288055419921875,
+      "learning_rate": 3.300162469536962e-06,
+      "loss": 1.8931,
+      "step": 1126
+    },
+    {
+      "epoch": 1.235068493150685,
+      "grad_norm": 2.201449155807495,
+      "learning_rate": 3.2981316003249393e-06,
+      "loss": 1.8149,
+      "step": 1127
+    },
+    {
+      "epoch": 1.2361643835616438,
+      "grad_norm": 2.215824842453003,
+      "learning_rate": 3.2961007311129166e-06,
+      "loss": 1.8949,
+      "step": 1128
+    },
+    {
+      "epoch": 1.2372602739726029,
+      "grad_norm": 2.133671522140503,
+      "learning_rate": 3.294069861900894e-06,
+      "loss": 1.8669,
+      "step": 1129
+    },
+    {
+      "epoch": 1.2383561643835617,
+      "grad_norm": 2.2195792198181152,
+      "learning_rate": 3.292038992688871e-06,
+      "loss": 1.9245,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2383561643835617,
+      "eval_loss": 2.0442771911621094,
+      "eval_runtime": 36.8077,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2394520547945205,
+      "grad_norm": 2.310866117477417,
+      "learning_rate": 3.2900081234768484e-06,
+      "loss": 1.8523,
+      "step": 1131
+    },
+    {
+      "epoch": 1.2405479452054795,
+      "grad_norm": 2.318300247192383,
+      "learning_rate": 3.2879772542648257e-06,
+      "loss": 1.8665,
+      "step": 1132
+    },
+    {
+      "epoch": 1.2416438356164383,
+      "grad_norm": 2.082564115524292,
+      "learning_rate": 3.285946385052803e-06,
+      "loss": 1.8962,
+      "step": 1133
+    },
+    {
+      "epoch": 1.2427397260273974,
+      "grad_norm": 2.1982104778289795,
+      "learning_rate": 3.2839155158407802e-06,
+      "loss": 1.7722,
+      "step": 1134
+    },
+    {
+      "epoch": 1.2438356164383562,
+      "grad_norm": 2.330782890319824,
+      "learning_rate": 3.2818846466287575e-06,
+      "loss": 1.8787,
+      "step": 1135
+    },
+    {
+      "epoch": 1.244931506849315,
+      "grad_norm": 2.136747121810913,
+      "learning_rate": 3.2798537774167344e-06,
+      "loss": 1.7739,
+      "step": 1136
+    },
+    {
+      "epoch": 1.246027397260274,
+      "grad_norm": 2.205862283706665,
+      "learning_rate": 3.277822908204712e-06,
+      "loss": 1.9127,
+      "step": 1137
+    },
+    {
+      "epoch": 1.2471232876712328,
+      "grad_norm": 2.2060093879699707,
+      "learning_rate": 3.275792038992689e-06,
+      "loss": 1.8379,
+      "step": 1138
+    },
+    {
+      "epoch": 1.2482191780821918,
+      "grad_norm": 2.255772829055786,
+      "learning_rate": 3.2737611697806666e-06,
+      "loss": 1.7931,
+      "step": 1139
+    },
+    {
+      "epoch": 1.2493150684931507,
+      "grad_norm": 2.240095615386963,
+      "learning_rate": 3.2717303005686435e-06,
+      "loss": 1.8466,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2493150684931507,
+      "eval_loss": 2.0429375171661377,
+      "eval_runtime": 36.7983,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2504109589041095,
+      "grad_norm": 2.1491949558258057,
+      "learning_rate": 3.269699431356621e-06,
+      "loss": 1.8137,
+      "step": 1141
+    },
+    {
+      "epoch": 1.2515068493150685,
+      "grad_norm": 2.3854610919952393,
+      "learning_rate": 3.267668562144598e-06,
+      "loss": 1.8486,
+      "step": 1142
+    },
+    {
+      "epoch": 1.2526027397260273,
+      "grad_norm": 2.3498709201812744,
+      "learning_rate": 3.2656376929325757e-06,
+      "loss": 1.837,
+      "step": 1143
+    },
+    {
+      "epoch": 1.2536986301369863,
+      "grad_norm": 2.1651649475097656,
+      "learning_rate": 3.2636068237205526e-06,
+      "loss": 1.8,
+      "step": 1144
+    },
+    {
+      "epoch": 1.2547945205479452,
+      "grad_norm": 2.2500200271606445,
+      "learning_rate": 3.2615759545085303e-06,
+      "loss": 1.8691,
+      "step": 1145
+    },
+    {
+      "epoch": 1.2558904109589042,
+      "grad_norm": 2.2178192138671875,
+      "learning_rate": 3.259545085296507e-06,
+      "loss": 1.8345,
+      "step": 1146
+    },
+    {
+      "epoch": 1.256986301369863,
+      "grad_norm": 2.238114595413208,
+      "learning_rate": 3.257514216084484e-06,
+      "loss": 1.8935,
+      "step": 1147
+    },
+    {
+      "epoch": 1.2580821917808218,
+      "grad_norm": 2.37042498588562,
+      "learning_rate": 3.2554833468724617e-06,
+      "loss": 1.7872,
+      "step": 1148
+    },
+    {
+      "epoch": 1.2591780821917808,
+      "grad_norm": 2.1831908226013184,
+      "learning_rate": 3.2534524776604385e-06,
+      "loss": 1.8913,
+      "step": 1149
+    },
+    {
+      "epoch": 1.2602739726027397,
+      "grad_norm": 2.251382827758789,
+      "learning_rate": 3.2514216084484162e-06,
+      "loss": 1.8611,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2602739726027397,
+      "eval_loss": 2.041306257247925,
+      "eval_runtime": 36.8199,
+      "eval_samples_per_second": 27.159,
+      "eval_steps_per_second": 1.141,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2613698630136987,
+      "grad_norm": 2.2690420150756836,
+      "learning_rate": 3.249390739236393e-06,
+      "loss": 1.8283,
+      "step": 1151
+    },
+    {
+      "epoch": 1.2624657534246575,
+      "grad_norm": 2.255833387374878,
+      "learning_rate": 3.2473598700243708e-06,
+      "loss": 1.8563,
+      "step": 1152
+    },
+    {
+      "epoch": 1.2635616438356165,
+      "grad_norm": 2.1362855434417725,
+      "learning_rate": 3.2453290008123476e-06,
+      "loss": 1.809,
+      "step": 1153
+    },
+    {
+      "epoch": 1.2646575342465753,
+      "grad_norm": 2.1227219104766846,
+      "learning_rate": 3.2432981316003253e-06,
+      "loss": 1.848,
+      "step": 1154
+    },
+    {
+      "epoch": 1.2657534246575342,
+      "grad_norm": 2.3639719486236572,
+      "learning_rate": 3.241267262388302e-06,
+      "loss": 1.8472,
+      "step": 1155
+    },
+    {
+      "epoch": 1.2668493150684932,
+      "grad_norm": 2.334404468536377,
+      "learning_rate": 3.23923639317628e-06,
+      "loss": 1.8073,
+      "step": 1156
+    },
+    {
+      "epoch": 1.267945205479452,
+      "grad_norm": 2.231656312942505,
+      "learning_rate": 3.2372055239642567e-06,
+      "loss": 1.8446,
+      "step": 1157
+    },
+    {
+      "epoch": 1.269041095890411,
+      "grad_norm": 2.371375560760498,
+      "learning_rate": 3.2351746547522344e-06,
+      "loss": 1.8002,
+      "step": 1158
+    },
+    {
+      "epoch": 1.2701369863013698,
+      "grad_norm": 2.5212879180908203,
+      "learning_rate": 3.2331437855402113e-06,
+      "loss": 1.863,
+      "step": 1159
+    },
+    {
+      "epoch": 1.2712328767123289,
+      "grad_norm": 2.3285911083221436,
+      "learning_rate": 3.2311129163281885e-06,
+      "loss": 1.9101,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2712328767123289,
+      "eval_loss": 2.038487195968628,
+      "eval_runtime": 36.8002,
+      "eval_samples_per_second": 27.174,
+      "eval_steps_per_second": 1.141,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2723287671232877,
+      "grad_norm": 2.1316046714782715,
+      "learning_rate": 3.229082047116166e-06,
+      "loss": 1.7952,
+      "step": 1161
+    },
+    {
+      "epoch": 1.2734246575342465,
+      "grad_norm": 2.4078125953674316,
+      "learning_rate": 3.227051177904143e-06,
+      "loss": 1.869,
+      "step": 1162
+    },
+    {
+      "epoch": 1.2745205479452055,
+      "grad_norm": 2.363672971725464,
+      "learning_rate": 3.2250203086921204e-06,
+      "loss": 1.8478,
+      "step": 1163
+    },
+    {
+      "epoch": 1.2756164383561643,
+      "grad_norm": 2.240558385848999,
+      "learning_rate": 3.2229894394800977e-06,
+      "loss": 1.8483,
+      "step": 1164
+    },
+    {
+      "epoch": 1.2767123287671232,
+      "grad_norm": 2.9142000675201416,
+      "learning_rate": 3.220958570268075e-06,
+      "loss": 1.8789,
+      "step": 1165
+    },
+    {
+      "epoch": 1.2778082191780822,
+      "grad_norm": 2.1930558681488037,
+      "learning_rate": 3.218927701056052e-06,
+      "loss": 1.7659,
+      "step": 1166
+    },
+    {
+      "epoch": 1.2789041095890412,
+      "grad_norm": 2.312471866607666,
+      "learning_rate": 3.2168968318440295e-06,
+      "loss": 1.8374,
+      "step": 1167
+    },
+    {
+      "epoch": 1.28,
+      "grad_norm": 2.6222236156463623,
+      "learning_rate": 3.2148659626320068e-06,
+      "loss": 1.7539,
+      "step": 1168
+    },
+    {
+      "epoch": 1.2810958904109588,
+      "grad_norm": 2.3013322353363037,
+      "learning_rate": 3.212835093419984e-06,
+      "loss": 1.9011,
+      "step": 1169
+    },
+    {
+      "epoch": 1.2821917808219179,
+      "grad_norm": 2.215240716934204,
+      "learning_rate": 3.2108042242079613e-06,
+      "loss": 1.8032,
+      "step": 1170
+    },
+    {
+      "epoch": 1.2821917808219179,
+      "eval_loss": 2.0365872383117676,
+      "eval_runtime": 36.8265,
+      "eval_samples_per_second": 27.154,
+      "eval_steps_per_second": 1.14,
+      "step": 1170
+    },
+    {
+      "epoch": 1.2832876712328767,
+      "grad_norm": 2.3620481491088867,
+      "learning_rate": 3.2087733549959386e-06,
+      "loss": 1.8258,
+      "step": 1171
+    },
+    {
+      "epoch": 1.2843835616438355,
+      "grad_norm": 2.2650208473205566,
+      "learning_rate": 3.206742485783916e-06,
+      "loss": 1.8888,
+      "step": 1172
+    },
+    {
+      "epoch": 1.2854794520547945,
+      "grad_norm": 2.2564077377319336,
+      "learning_rate": 3.2047116165718927e-06,
+      "loss": 1.7874,
+      "step": 1173
+    },
+    {
+      "epoch": 1.2865753424657536,
+      "grad_norm": 2.2855281829833984,
+      "learning_rate": 3.2026807473598704e-06,
+      "loss": 1.8003,
+      "step": 1174
+    },
+    {
+      "epoch": 1.2876712328767124,
+      "grad_norm": 2.353893756866455,
+      "learning_rate": 3.2006498781478473e-06,
+      "loss": 1.8466,
+      "step": 1175
+    },
+    {
+      "epoch": 1.2887671232876712,
+      "grad_norm": 2.1893045902252197,
+      "learning_rate": 3.198619008935825e-06,
+      "loss": 1.8652,
+      "step": 1176
+    },
+    {
+      "epoch": 1.2898630136986302,
+      "grad_norm": 2.6629793643951416,
+      "learning_rate": 3.196588139723802e-06,
+      "loss": 1.797,
+      "step": 1177
+    },
+    {
+      "epoch": 1.290958904109589,
+      "grad_norm": 2.2427361011505127,
+      "learning_rate": 3.1945572705117795e-06,
+      "loss": 1.8294,
+      "step": 1178
+    },
+    {
+      "epoch": 1.2920547945205478,
+      "grad_norm": 2.321878671646118,
+      "learning_rate": 3.1925264012997564e-06,
+      "loss": 1.8494,
+      "step": 1179
+    },
+    {
+      "epoch": 1.2931506849315069,
+      "grad_norm": 2.6928281784057617,
+      "learning_rate": 3.190495532087734e-06,
+      "loss": 1.8574,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2931506849315069,
+      "eval_loss": 2.0426652431488037,
+      "eval_runtime": 36.8482,
+      "eval_samples_per_second": 27.138,
+      "eval_steps_per_second": 1.14,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2942465753424657,
+      "grad_norm": 2.382164239883423,
+      "learning_rate": 3.188464662875711e-06,
+      "loss": 1.8518,
+      "step": 1181
+    },
+    {
+      "epoch": 1.2953424657534247,
+      "grad_norm": 2.4109256267547607,
+      "learning_rate": 3.1864337936636886e-06,
+      "loss": 1.9083,
+      "step": 1182
+    },
+    {
+      "epoch": 1.2964383561643835,
+      "grad_norm": 2.468062400817871,
+      "learning_rate": 3.1844029244516655e-06,
+      "loss": 1.8247,
+      "step": 1183
+    },
+    {
+      "epoch": 1.2975342465753426,
+      "grad_norm": 2.459378242492676,
+      "learning_rate": 3.182372055239643e-06,
+      "loss": 1.791,
+      "step": 1184
+    },
+    {
+      "epoch": 1.2986301369863014,
+      "grad_norm": 2.3331072330474854,
+      "learning_rate": 3.18034118602762e-06,
+      "loss": 1.7869,
+      "step": 1185
+    },
+    {
+      "epoch": 1.2997260273972602,
+      "grad_norm": 2.339428424835205,
+      "learning_rate": 3.1783103168155973e-06,
+      "loss": 1.8729,
+      "step": 1186
+    },
+    {
+      "epoch": 1.3008219178082192,
+      "grad_norm": 2.3033406734466553,
+      "learning_rate": 3.1762794476035746e-06,
+      "loss": 1.8455,
+      "step": 1187
+    },
+    {
+      "epoch": 1.301917808219178,
+      "grad_norm": 2.2480714321136475,
+      "learning_rate": 3.174248578391552e-06,
+      "loss": 1.8233,
+      "step": 1188
+    },
+    {
+      "epoch": 1.303013698630137,
+      "grad_norm": 2.2953624725341797,
+      "learning_rate": 3.172217709179529e-06,
+      "loss": 1.8261,
+      "step": 1189
+    },
+    {
+      "epoch": 1.3041095890410959,
+      "grad_norm": 2.494307518005371,
+      "learning_rate": 3.1701868399675064e-06,
+      "loss": 1.8153,
+      "step": 1190
+    },
+    {
+      "epoch": 1.3041095890410959,
+      "eval_loss": 2.037719964981079,
+      "eval_runtime": 36.9788,
+      "eval_samples_per_second": 27.043,
+      "eval_steps_per_second": 1.136,
+      "step": 1190
+    },
+    {
+      "epoch": 1.305205479452055,
+      "grad_norm": 2.314159631729126,
+      "learning_rate": 3.1681559707554837e-06,
+      "loss": 1.8406,
+      "step": 1191
+    },
+    {
+      "epoch": 1.3063013698630137,
+      "grad_norm": 2.4262588024139404,
+      "learning_rate": 3.166125101543461e-06,
+      "loss": 1.7803,
+      "step": 1192
+    },
+    {
+      "epoch": 1.3073972602739725,
+      "grad_norm": 2.3800835609436035,
+      "learning_rate": 3.1640942323314382e-06,
+      "loss": 1.8521,
+      "step": 1193
+    },
+    {
+      "epoch": 1.3084931506849315,
+      "grad_norm": 2.234551429748535,
+      "learning_rate": 3.1620633631194155e-06,
+      "loss": 1.845,
+      "step": 1194
+    },
+    {
+      "epoch": 1.3095890410958904,
+      "grad_norm": 2.167635440826416,
+      "learning_rate": 3.1600324939073928e-06,
+      "loss": 1.789,
+      "step": 1195
+    },
+    {
+      "epoch": 1.3106849315068494,
+      "grad_norm": 2.320247173309326,
+      "learning_rate": 3.15800162469537e-06,
+      "loss": 1.8311,
+      "step": 1196
+    },
+    {
+      "epoch": 1.3117808219178082,
+      "grad_norm": 2.2715561389923096,
+      "learning_rate": 3.155970755483347e-06,
+      "loss": 1.8308,
+      "step": 1197
+    },
+    {
+      "epoch": 1.3128767123287672,
+      "grad_norm": 2.2535455226898193,
+      "learning_rate": 3.1539398862713246e-06,
+      "loss": 1.8358,
+      "step": 1198
+    },
+    {
+      "epoch": 1.313972602739726,
+      "grad_norm": 2.2399163246154785,
+      "learning_rate": 3.1519090170593015e-06,
+      "loss": 1.8513,
+      "step": 1199
+    },
+    {
+      "epoch": 1.3150684931506849,
+      "grad_norm": 2.2984132766723633,
+      "learning_rate": 3.149878147847279e-06,
+      "loss": 1.8377,
+      "step": 1200
+    },
+    {
+      "epoch": 1.3150684931506849,
+      "eval_loss": 2.0334064960479736,
+      "eval_runtime": 37.0346,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1200
+    },
+    {
+      "epoch": 1.316164383561644,
+      "grad_norm": 2.1296892166137695,
+      "learning_rate": 3.147847278635256e-06,
+      "loss": 1.9023,
+      "step": 1201
+    },
+    {
+      "epoch": 1.3172602739726027,
+      "grad_norm": 2.1034350395202637,
+      "learning_rate": 3.1458164094232337e-06,
+      "loss": 1.917,
+      "step": 1202
+    },
+    {
+      "epoch": 1.3183561643835615,
+      "grad_norm": 2.258863687515259,
+      "learning_rate": 3.1437855402112106e-06,
+      "loss": 1.8093,
+      "step": 1203
+    },
+    {
+      "epoch": 1.3194520547945205,
+      "grad_norm": 2.20119309425354,
+      "learning_rate": 3.1417546709991883e-06,
+      "loss": 1.7796,
+      "step": 1204
+    },
+    {
+      "epoch": 1.3205479452054796,
+      "grad_norm": 2.3005869388580322,
+      "learning_rate": 3.139723801787165e-06,
+      "loss": 1.8496,
+      "step": 1205
+    },
+    {
+      "epoch": 1.3216438356164384,
+      "grad_norm": 2.2009031772613525,
+      "learning_rate": 3.137692932575143e-06,
+      "loss": 1.8244,
+      "step": 1206
+    },
+    {
+      "epoch": 1.3227397260273972,
+      "grad_norm": 2.3467633724212646,
+      "learning_rate": 3.1356620633631197e-06,
+      "loss": 1.8803,
+      "step": 1207
+    },
+    {
+      "epoch": 1.3238356164383562,
+      "grad_norm": 2.318711519241333,
+      "learning_rate": 3.1336311941510974e-06,
+      "loss": 1.7999,
+      "step": 1208
+    },
+    {
+      "epoch": 1.324931506849315,
+      "grad_norm": 2.182602882385254,
+      "learning_rate": 3.1316003249390742e-06,
+      "loss": 1.836,
+      "step": 1209
+    },
+    {
+      "epoch": 1.3260273972602739,
+      "grad_norm": 2.217298984527588,
+      "learning_rate": 3.129569455727051e-06,
+      "loss": 1.8776,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3260273972602739,
+      "eval_loss": 2.033573627471924,
+      "eval_runtime": 36.8029,
+      "eval_samples_per_second": 27.172,
+      "eval_steps_per_second": 1.141,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3271232876712329,
+      "grad_norm": 2.2814879417419434,
+      "learning_rate": 3.1275385865150288e-06,
+      "loss": 1.8739,
+      "step": 1211
+    },
+    {
+      "epoch": 1.3282191780821917,
+      "grad_norm": 2.1692471504211426,
+      "learning_rate": 3.1255077173030056e-06,
+      "loss": 1.8539,
+      "step": 1212
+    },
+    {
+      "epoch": 1.3293150684931507,
+      "grad_norm": 2.2472000122070312,
+      "learning_rate": 3.1234768480909833e-06,
+      "loss": 1.8527,
+      "step": 1213
+    },
+    {
+      "epoch": 1.3304109589041095,
+      "grad_norm": 2.1969852447509766,
+      "learning_rate": 3.12144597887896e-06,
+      "loss": 1.8429,
+      "step": 1214
+    },
+    {
+      "epoch": 1.3315068493150686,
+      "grad_norm": 2.141629219055176,
+      "learning_rate": 3.119415109666938e-06,
+      "loss": 1.868,
+      "step": 1215
+    },
+    {
+      "epoch": 1.3326027397260274,
+      "grad_norm": 2.282562494277954,
+      "learning_rate": 3.1173842404549147e-06,
+      "loss": 1.8913,
+      "step": 1216
+    },
+    {
+      "epoch": 1.3336986301369862,
+      "grad_norm": 2.2375850677490234,
+      "learning_rate": 3.1153533712428924e-06,
+      "loss": 1.8462,
+      "step": 1217
+    },
+    {
+      "epoch": 1.3347945205479452,
+      "grad_norm": 2.3529837131500244,
+      "learning_rate": 3.1133225020308693e-06,
+      "loss": 1.8445,
+      "step": 1218
+    },
+    {
+      "epoch": 1.335890410958904,
+      "grad_norm": 2.385857343673706,
+      "learning_rate": 3.111291632818847e-06,
+      "loss": 1.89,
+      "step": 1219
+    },
+    {
+      "epoch": 1.336986301369863,
+      "grad_norm": 2.301253080368042,
+      "learning_rate": 3.109260763606824e-06,
+      "loss": 1.8882,
+      "step": 1220
+    },
+    {
+      "epoch": 1.336986301369863,
+      "eval_loss": 2.040025234222412,
+      "eval_runtime": 36.839,
+      "eval_samples_per_second": 27.145,
+      "eval_steps_per_second": 1.14,
+      "step": 1220
+    },
+    {
+      "epoch": 1.3380821917808219,
+      "grad_norm": 2.3772335052490234,
+      "learning_rate": 3.1072298943948015e-06,
+      "loss": 1.8545,
+      "step": 1221
+    },
+    {
+      "epoch": 1.339178082191781,
+      "grad_norm": 2.387830972671509,
+      "learning_rate": 3.1051990251827784e-06,
+      "loss": 1.8653,
+      "step": 1222
+    },
+    {
+      "epoch": 1.3402739726027397,
+      "grad_norm": 2.4285190105438232,
+      "learning_rate": 3.1031681559707557e-06,
+      "loss": 1.8268,
+      "step": 1223
+    },
+    {
+      "epoch": 1.3413698630136985,
+      "grad_norm": 2.1892075538635254,
+      "learning_rate": 3.101137286758733e-06,
+      "loss": 1.8818,
+      "step": 1224
+    },
+    {
+      "epoch": 1.3424657534246576,
+      "grad_norm": 2.2510695457458496,
+      "learning_rate": 3.09910641754671e-06,
+      "loss": 1.879,
+      "step": 1225
+    },
+    {
+      "epoch": 1.3435616438356164,
+      "grad_norm": 2.505671977996826,
+      "learning_rate": 3.0970755483346875e-06,
+      "loss": 1.872,
+      "step": 1226
+    },
+    {
+      "epoch": 1.3446575342465754,
+      "grad_norm": 2.5315067768096924,
+      "learning_rate": 3.0950446791226648e-06,
+      "loss": 1.8651,
+      "step": 1227
+    },
+    {
+      "epoch": 1.3457534246575342,
+      "grad_norm": 2.223600387573242,
+      "learning_rate": 3.093013809910642e-06,
+      "loss": 1.9167,
+      "step": 1228
+    },
+    {
+      "epoch": 1.3468493150684933,
+      "grad_norm": 2.357802391052246,
+      "learning_rate": 3.0909829406986193e-06,
+      "loss": 1.8593,
+      "step": 1229
+    },
+    {
+      "epoch": 1.347945205479452,
+      "grad_norm": 2.3102364540100098,
+      "learning_rate": 3.0889520714865966e-06,
+      "loss": 1.8994,
+      "step": 1230
+    },
+    {
+      "epoch": 1.347945205479452,
+      "eval_loss": 2.0332930088043213,
+      "eval_runtime": 36.989,
+      "eval_samples_per_second": 27.035,
+      "eval_steps_per_second": 1.135,
+      "step": 1230
+    },
+    {
+      "epoch": 1.3490410958904109,
+      "grad_norm": 2.085080623626709,
+      "learning_rate": 3.086921202274574e-06,
+      "loss": 1.8741,
+      "step": 1231
+    },
+    {
+      "epoch": 1.35013698630137,
+      "grad_norm": 2.267305612564087,
+      "learning_rate": 3.084890333062551e-06,
+      "loss": 1.8426,
+      "step": 1232
+    },
+    {
+      "epoch": 1.3512328767123287,
+      "grad_norm": 2.181033134460449,
+      "learning_rate": 3.0828594638505284e-06,
+      "loss": 1.8199,
+      "step": 1233
+    },
+    {
+      "epoch": 1.3523287671232875,
+      "grad_norm": 2.288386821746826,
+      "learning_rate": 3.0808285946385057e-06,
+      "loss": 1.8469,
+      "step": 1234
+    },
+    {
+      "epoch": 1.3534246575342466,
+      "grad_norm": 2.1516127586364746,
+      "learning_rate": 3.078797725426483e-06,
+      "loss": 1.8056,
+      "step": 1235
+    },
+    {
+      "epoch": 1.3545205479452056,
+      "grad_norm": 2.2699198722839355,
+      "learning_rate": 3.07676685621446e-06,
+      "loss": 1.7535,
+      "step": 1236
+    },
+    {
+      "epoch": 1.3556164383561644,
+      "grad_norm": 2.193251132965088,
+      "learning_rate": 3.0747359870024375e-06,
+      "loss": 1.8407,
+      "step": 1237
+    },
+    {
+      "epoch": 1.3567123287671232,
+      "grad_norm": 2.2217774391174316,
+      "learning_rate": 3.0727051177904144e-06,
+      "loss": 1.8062,
+      "step": 1238
+    },
+    {
+      "epoch": 1.3578082191780823,
+      "grad_norm": 2.2376203536987305,
+      "learning_rate": 3.070674248578392e-06,
+      "loss": 1.829,
+      "step": 1239
+    },
+    {
+      "epoch": 1.358904109589041,
+      "grad_norm": 2.387559413909912,
+      "learning_rate": 3.068643379366369e-06,
+      "loss": 1.8459,
+      "step": 1240
+    },
+    {
+      "epoch": 1.358904109589041,
+      "eval_loss": 2.029298782348633,
+      "eval_runtime": 37.0491,
+      "eval_samples_per_second": 26.991,
+      "eval_steps_per_second": 1.134,
+      "step": 1240
+    },
+    {
+      "epoch": 1.3599999999999999,
+      "grad_norm": 2.1940178871154785,
+      "learning_rate": 3.0666125101543466e-06,
+      "loss": 1.831,
+      "step": 1241
+    },
+    {
+      "epoch": 1.361095890410959,
+      "grad_norm": 2.145514488220215,
+      "learning_rate": 3.0645816409423235e-06,
+      "loss": 1.7919,
+      "step": 1242
+    },
+    {
+      "epoch": 1.362191780821918,
+      "grad_norm": 2.319164276123047,
+      "learning_rate": 3.062550771730301e-06,
+      "loss": 1.8472,
+      "step": 1243
+    },
+    {
+      "epoch": 1.3632876712328768,
+      "grad_norm": 2.218986749649048,
+      "learning_rate": 3.060519902518278e-06,
+      "loss": 1.8184,
+      "step": 1244
+    },
+    {
+      "epoch": 1.3643835616438356,
+      "grad_norm": 2.3755338191986084,
+      "learning_rate": 3.0584890333062557e-06,
+      "loss": 1.8166,
+      "step": 1245
+    },
+    {
+      "epoch": 1.3654794520547946,
+      "grad_norm": 2.3186213970184326,
+      "learning_rate": 3.0564581640942326e-06,
+      "loss": 1.7599,
+      "step": 1246
+    },
+    {
+      "epoch": 1.3665753424657534,
+      "grad_norm": 2.193958282470703,
+      "learning_rate": 3.0544272948822094e-06,
+      "loss": 1.8404,
+      "step": 1247
+    },
+    {
+      "epoch": 1.3676712328767122,
+      "grad_norm": 2.40419602394104,
+      "learning_rate": 3.052396425670187e-06,
+      "loss": 1.821,
+      "step": 1248
+    },
+    {
+      "epoch": 1.3687671232876713,
+      "grad_norm": 2.445542573928833,
+      "learning_rate": 3.050365556458164e-06,
+      "loss": 1.8518,
+      "step": 1249
+    },
+    {
+      "epoch": 1.36986301369863,
+      "grad_norm": 2.2776567935943604,
+      "learning_rate": 3.0483346872461417e-06,
+      "loss": 1.8121,
+      "step": 1250
+    },
+    {
+      "epoch": 1.36986301369863,
+      "eval_loss": 2.030043125152588,
+      "eval_runtime": 37.0721,
+      "eval_samples_per_second": 26.974,
+      "eval_steps_per_second": 1.133,
+      "step": 1250
+    },
+    {
+      "epoch": 1.370958904109589,
+      "grad_norm": 2.420448064804077,
+      "learning_rate": 3.0463038180341185e-06,
+      "loss": 1.8168,
+      "step": 1251
+    },
+    {
+      "epoch": 1.372054794520548,
+      "grad_norm": 2.252065658569336,
+      "learning_rate": 3.0442729488220962e-06,
+      "loss": 1.8757,
+      "step": 1252
+    },
+    {
+      "epoch": 1.373150684931507,
+      "grad_norm": 2.4772560596466064,
+      "learning_rate": 3.042242079610073e-06,
+      "loss": 1.8711,
+      "step": 1253
+    },
+    {
+      "epoch": 1.3742465753424657,
+      "grad_norm": 2.3986129760742188,
+      "learning_rate": 3.0402112103980508e-06,
+      "loss": 1.8636,
+      "step": 1254
+    },
+    {
+      "epoch": 1.3753424657534246,
+      "grad_norm": 2.257274866104126,
+      "learning_rate": 3.0381803411860276e-06,
+      "loss": 1.7858,
+      "step": 1255
+    },
+    {
+      "epoch": 1.3764383561643836,
+      "grad_norm": 2.2373499870300293,
+      "learning_rate": 3.0361494719740053e-06,
+      "loss": 1.817,
+      "step": 1256
+    },
+    {
+      "epoch": 1.3775342465753424,
+      "grad_norm": 2.292388916015625,
+      "learning_rate": 3.034118602761982e-06,
+      "loss": 1.827,
+      "step": 1257
+    },
+    {
+      "epoch": 1.3786301369863014,
+      "grad_norm": 2.4618642330169678,
+      "learning_rate": 3.03208773354996e-06,
+      "loss": 1.8218,
+      "step": 1258
+    },
+    {
+      "epoch": 1.3797260273972602,
+      "grad_norm": 2.271958112716675,
+      "learning_rate": 3.0300568643379367e-06,
+      "loss": 1.8092,
+      "step": 1259
+    },
+    {
+      "epoch": 1.3808219178082193,
+      "grad_norm": 2.3990538120269775,
+      "learning_rate": 3.028025995125914e-06,
+      "loss": 1.8633,
+      "step": 1260
+    },
+    {
+      "epoch": 1.3808219178082193,
+      "eval_loss": 2.028458595275879,
+      "eval_runtime": 37.0659,
+      "eval_samples_per_second": 26.979,
+      "eval_steps_per_second": 1.133,
+      "step": 1260
+    },
+    {
+      "epoch": 1.381917808219178,
+      "grad_norm": 2.208803653717041,
+      "learning_rate": 3.0259951259138913e-06,
+      "loss": 1.8463,
+      "step": 1261
+    },
+    {
+      "epoch": 1.383013698630137,
+      "grad_norm": 2.3269028663635254,
+      "learning_rate": 3.0239642567018686e-06,
+      "loss": 1.7784,
+      "step": 1262
+    },
+    {
+      "epoch": 1.384109589041096,
+      "grad_norm": 2.367910623550415,
+      "learning_rate": 3.021933387489846e-06,
+      "loss": 1.8281,
+      "step": 1263
+    },
+    {
+      "epoch": 1.3852054794520547,
+      "grad_norm": 2.273686647415161,
+      "learning_rate": 3.019902518277823e-06,
+      "loss": 1.7832,
+      "step": 1264
+    },
+    {
+      "epoch": 1.3863013698630138,
+      "grad_norm": 2.421964168548584,
+      "learning_rate": 3.0178716490658004e-06,
+      "loss": 1.8407,
+      "step": 1265
+    },
+    {
+      "epoch": 1.3873972602739726,
+      "grad_norm": 2.2627696990966797,
+      "learning_rate": 3.0158407798537777e-06,
+      "loss": 1.8718,
+      "step": 1266
+    },
+    {
+      "epoch": 1.3884931506849316,
+      "grad_norm": 2.1981618404388428,
+      "learning_rate": 3.013809910641755e-06,
+      "loss": 1.7763,
+      "step": 1267
+    },
+    {
+      "epoch": 1.3895890410958904,
+      "grad_norm": 2.245253562927246,
+      "learning_rate": 3.0117790414297322e-06,
+      "loss": 1.7915,
+      "step": 1268
+    },
+    {
+      "epoch": 1.3906849315068492,
+      "grad_norm": 2.2292444705963135,
+      "learning_rate": 3.0097481722177095e-06,
+      "loss": 1.7909,
+      "step": 1269
+    },
+    {
+      "epoch": 1.3917808219178083,
+      "grad_norm": 2.2290430068969727,
+      "learning_rate": 3.0077173030056868e-06,
+      "loss": 1.8364,
+      "step": 1270
+    },
+    {
+      "epoch": 1.3917808219178083,
+      "eval_loss": 2.0290327072143555,
+      "eval_runtime": 37.0666,
+      "eval_samples_per_second": 26.978,
+      "eval_steps_per_second": 1.133,
+      "step": 1270
+    },
+    {
+      "epoch": 1.392876712328767,
+      "grad_norm": 2.220881700515747,
+      "learning_rate": 3.005686433793664e-06,
+      "loss": 1.7968,
+      "step": 1271
+    },
+    {
+      "epoch": 1.393972602739726,
+      "grad_norm": 2.375332832336426,
+      "learning_rate": 3.0036555645816413e-06,
+      "loss": 1.8502,
+      "step": 1272
+    },
+    {
+      "epoch": 1.395068493150685,
+      "grad_norm": 2.3065083026885986,
+      "learning_rate": 3.001624695369618e-06,
+      "loss": 1.8116,
+      "step": 1273
+    },
+    {
+      "epoch": 1.396164383561644,
+      "grad_norm": 2.7262511253356934,
+      "learning_rate": 2.999593826157596e-06,
+      "loss": 1.8641,
+      "step": 1274
+    },
+    {
+      "epoch": 1.3972602739726028,
+      "grad_norm": 2.3381216526031494,
+      "learning_rate": 2.9975629569455727e-06,
+      "loss": 1.8128,
+      "step": 1275
+    },
+    {
+      "epoch": 1.3983561643835616,
+      "grad_norm": 2.1994192600250244,
+      "learning_rate": 2.9955320877335504e-06,
+      "loss": 1.8563,
+      "step": 1276
+    },
+    {
+      "epoch": 1.3994520547945206,
+      "grad_norm": 2.3810958862304688,
+      "learning_rate": 2.9935012185215273e-06,
+      "loss": 1.8196,
+      "step": 1277
+    },
+    {
+      "epoch": 1.4005479452054794,
+      "grad_norm": 2.4343955516815186,
+      "learning_rate": 2.991470349309505e-06,
+      "loss": 1.8564,
+      "step": 1278
+    },
+    {
+      "epoch": 1.4016438356164382,
+      "grad_norm": 2.1929988861083984,
+      "learning_rate": 2.989439480097482e-06,
+      "loss": 1.8119,
+      "step": 1279
+    },
+    {
+      "epoch": 1.4027397260273973,
+      "grad_norm": 2.2305309772491455,
+      "learning_rate": 2.9874086108854595e-06,
+      "loss": 1.8289,
+      "step": 1280
+    },
+    {
+      "epoch": 1.4027397260273973,
+      "eval_loss": 2.032022476196289,
+      "eval_runtime": 37.084,
+      "eval_samples_per_second": 26.966,
+      "eval_steps_per_second": 1.133,
+      "step": 1280
+    },
+    {
+      "epoch": 1.403835616438356,
+      "grad_norm": 2.6255314350128174,
+      "learning_rate": 2.9853777416734364e-06,
+      "loss": 1.8176,
+      "step": 1281
+    },
+    {
+      "epoch": 1.4049315068493151,
+      "grad_norm": 2.238891363143921,
+      "learning_rate": 2.983346872461414e-06,
+      "loss": 1.8679,
+      "step": 1282
+    },
+    {
+      "epoch": 1.406027397260274,
+      "grad_norm": 2.9972007274627686,
+      "learning_rate": 2.981316003249391e-06,
+      "loss": 1.8839,
+      "step": 1283
+    },
+    {
+      "epoch": 1.407123287671233,
+      "grad_norm": 2.3255932331085205,
+      "learning_rate": 2.9792851340373678e-06,
+      "loss": 1.8451,
+      "step": 1284
+    },
+    {
+      "epoch": 1.4082191780821918,
+      "grad_norm": 2.290395498275757,
+      "learning_rate": 2.9772542648253455e-06,
+      "loss": 1.8244,
+      "step": 1285
+    },
+    {
+      "epoch": 1.4093150684931506,
+      "grad_norm": 2.288728713989258,
+      "learning_rate": 2.9752233956133223e-06,
+      "loss": 1.8628,
+      "step": 1286
+    },
+    {
+      "epoch": 1.4104109589041096,
+      "grad_norm": 2.177381992340088,
+      "learning_rate": 2.9731925264013e-06,
+      "loss": 1.8111,
+      "step": 1287
+    },
+    {
+      "epoch": 1.4115068493150684,
+      "grad_norm": 2.214956760406494,
+      "learning_rate": 2.971161657189277e-06,
+      "loss": 1.8377,
+      "step": 1288
+    },
+    {
+      "epoch": 1.4126027397260275,
+      "grad_norm": 2.3859663009643555,
+      "learning_rate": 2.9691307879772546e-06,
+      "loss": 1.8154,
+      "step": 1289
+    },
+    {
+      "epoch": 1.4136986301369863,
+      "grad_norm": 2.0873706340789795,
+      "learning_rate": 2.9670999187652314e-06,
+      "loss": 1.8098,
+      "step": 1290
+    },
+    {
+      "epoch": 1.4136986301369863,
+      "eval_loss": 2.029996633529663,
+      "eval_runtime": 37.0979,
+      "eval_samples_per_second": 26.956,
+      "eval_steps_per_second": 1.132,
+      "step": 1290
+    },
+    {
+      "epoch": 1.4147945205479453,
+      "grad_norm": 2.2632675170898438,
+      "learning_rate": 2.965069049553209e-06,
+      "loss": 1.7974,
+      "step": 1291
+    },
+    {
+      "epoch": 1.415890410958904,
+      "grad_norm": 2.208515167236328,
+      "learning_rate": 2.963038180341186e-06,
+      "loss": 1.8856,
+      "step": 1292
+    },
+    {
+      "epoch": 1.416986301369863,
+      "grad_norm": 2.176525115966797,
+      "learning_rate": 2.9610073111291637e-06,
+      "loss": 1.7783,
+      "step": 1293
+    },
+    {
+      "epoch": 1.418082191780822,
+      "grad_norm": 2.226682186126709,
+      "learning_rate": 2.9589764419171405e-06,
+      "loss": 1.865,
+      "step": 1294
+    },
+    {
+      "epoch": 1.4191780821917808,
+      "grad_norm": 2.3544883728027344,
+      "learning_rate": 2.9569455727051182e-06,
+      "loss": 1.8808,
+      "step": 1295
+    },
+    {
+      "epoch": 1.4202739726027398,
+      "grad_norm": 2.1729962825775146,
+      "learning_rate": 2.954914703493095e-06,
+      "loss": 1.7927,
+      "step": 1296
+    },
+    {
+      "epoch": 1.4213698630136986,
+      "grad_norm": 2.168168067932129,
+      "learning_rate": 2.9528838342810724e-06,
+      "loss": 1.8253,
+      "step": 1297
+    },
+    {
+      "epoch": 1.4224657534246576,
+      "grad_norm": 2.2479910850524902,
+      "learning_rate": 2.9508529650690496e-06,
+      "loss": 1.7976,
+      "step": 1298
+    },
+    {
+      "epoch": 1.4235616438356165,
+      "grad_norm": 2.1733837127685547,
+      "learning_rate": 2.948822095857027e-06,
+      "loss": 1.8015,
+      "step": 1299
+    },
+    {
+      "epoch": 1.4246575342465753,
+      "grad_norm": 2.2553391456604004,
+      "learning_rate": 2.946791226645004e-06,
+      "loss": 1.8545,
+      "step": 1300
+    },
+    {
+      "epoch": 1.4246575342465753,
+      "eval_loss": 2.0245048999786377,
+      "eval_runtime": 37.0943,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 1300
+    },
+    {
+      "epoch": 1.4257534246575343,
+      "grad_norm": 2.2817769050598145,
+      "learning_rate": 2.9447603574329815e-06,
+      "loss": 1.8017,
+      "step": 1301
+    },
+    {
+      "epoch": 1.426849315068493,
+      "grad_norm": 2.170865297317505,
+      "learning_rate": 2.9427294882209587e-06,
+      "loss": 1.7738,
+      "step": 1302
+    },
+    {
+      "epoch": 1.427945205479452,
+      "grad_norm": 2.296919107437134,
+      "learning_rate": 2.940698619008936e-06,
+      "loss": 1.8098,
+      "step": 1303
+    },
+    {
+      "epoch": 1.429041095890411,
+      "grad_norm": 2.476266622543335,
+      "learning_rate": 2.9386677497969133e-06,
+      "loss": 1.8229,
+      "step": 1304
+    },
+    {
+      "epoch": 1.43013698630137,
+      "grad_norm": 2.400918483734131,
+      "learning_rate": 2.9366368805848906e-06,
+      "loss": 1.8057,
+      "step": 1305
+    },
+    {
+      "epoch": 1.4312328767123288,
+      "grad_norm": 2.393892765045166,
+      "learning_rate": 2.934606011372868e-06,
+      "loss": 1.8568,
+      "step": 1306
+    },
+    {
+      "epoch": 1.4323287671232876,
+      "grad_norm": 2.2551307678222656,
+      "learning_rate": 2.932575142160845e-06,
+      "loss": 1.8493,
+      "step": 1307
+    },
+    {
+      "epoch": 1.4334246575342466,
+      "grad_norm": 2.3091611862182617,
+      "learning_rate": 2.9305442729488224e-06,
+      "loss": 1.8297,
+      "step": 1308
+    },
+    {
+      "epoch": 1.4345205479452054,
+      "grad_norm": 2.231614828109741,
+      "learning_rate": 2.9285134037367997e-06,
+      "loss": 1.8345,
+      "step": 1309
+    },
+    {
+      "epoch": 1.4356164383561643,
+      "grad_norm": 2.24942946434021,
+      "learning_rate": 2.9264825345247765e-06,
+      "loss": 1.8284,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4356164383561643,
+      "eval_loss": 2.024843215942383,
+      "eval_runtime": 36.7795,
+      "eval_samples_per_second": 27.189,
+      "eval_steps_per_second": 1.142,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4367123287671233,
+      "grad_norm": 2.2798643112182617,
+      "learning_rate": 2.9244516653127542e-06,
+      "loss": 1.8554,
+      "step": 1311
+    },
+    {
+      "epoch": 1.4378082191780823,
+      "grad_norm": 2.342576265335083,
+      "learning_rate": 2.922420796100731e-06,
+      "loss": 1.8296,
+      "step": 1312
+    },
+    {
+      "epoch": 1.4389041095890411,
+      "grad_norm": 2.268016815185547,
+      "learning_rate": 2.9203899268887088e-06,
+      "loss": 1.7467,
+      "step": 1313
+    },
+    {
+      "epoch": 1.44,
+      "grad_norm": 2.2520217895507812,
+      "learning_rate": 2.9183590576766856e-06,
+      "loss": 1.8748,
+      "step": 1314
+    },
+    {
+      "epoch": 1.441095890410959,
+      "grad_norm": 2.2091598510742188,
+      "learning_rate": 2.9163281884646633e-06,
+      "loss": 1.7715,
+      "step": 1315
+    },
+    {
+      "epoch": 1.4421917808219178,
+      "grad_norm": 2.197575092315674,
+      "learning_rate": 2.91429731925264e-06,
+      "loss": 1.8265,
+      "step": 1316
+    },
+    {
+      "epoch": 1.4432876712328766,
+      "grad_norm": 2.261122465133667,
+      "learning_rate": 2.912266450040618e-06,
+      "loss": 1.8758,
+      "step": 1317
+    },
+    {
+      "epoch": 1.4443835616438356,
+      "grad_norm": 2.3291778564453125,
+      "learning_rate": 2.9102355808285947e-06,
+      "loss": 1.8213,
+      "step": 1318
+    },
+    {
+      "epoch": 1.4454794520547944,
+      "grad_norm": 2.2182250022888184,
+      "learning_rate": 2.9082047116165724e-06,
+      "loss": 1.822,
+      "step": 1319
+    },
+    {
+      "epoch": 1.4465753424657535,
+      "grad_norm": 2.295978546142578,
+      "learning_rate": 2.9061738424045493e-06,
+      "loss": 1.8008,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4465753424657535,
+      "eval_loss": 2.0230674743652344,
+      "eval_runtime": 37.0013,
+      "eval_samples_per_second": 27.026,
+      "eval_steps_per_second": 1.135,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4476712328767123,
+      "grad_norm": 2.3187007904052734,
+      "learning_rate": 2.904142973192527e-06,
+      "loss": 1.7923,
+      "step": 1321
+    },
+    {
+      "epoch": 1.4487671232876713,
+      "grad_norm": 2.242851972579956,
+      "learning_rate": 2.902112103980504e-06,
+      "loss": 1.8874,
+      "step": 1322
+    },
+    {
+      "epoch": 1.4498630136986301,
+      "grad_norm": 2.2560081481933594,
+      "learning_rate": 2.900081234768481e-06,
+      "loss": 1.8284,
+      "step": 1323
+    },
+    {
+      "epoch": 1.450958904109589,
+      "grad_norm": 2.3017799854278564,
+      "learning_rate": 2.8980503655564584e-06,
+      "loss": 1.8263,
+      "step": 1324
+    },
+    {
+      "epoch": 1.452054794520548,
+      "grad_norm": 2.3016278743743896,
+      "learning_rate": 2.8960194963444357e-06,
+      "loss": 1.8174,
+      "step": 1325
+    },
+    {
+      "epoch": 1.4531506849315068,
+      "grad_norm": 2.1965372562408447,
+      "learning_rate": 2.893988627132413e-06,
+      "loss": 1.7812,
+      "step": 1326
+    },
+    {
+      "epoch": 1.4542465753424658,
+      "grad_norm": 2.302969455718994,
+      "learning_rate": 2.8919577579203902e-06,
+      "loss": 1.8115,
+      "step": 1327
+    },
+    {
+      "epoch": 1.4553424657534246,
+      "grad_norm": 2.338066339492798,
+      "learning_rate": 2.8899268887083675e-06,
+      "loss": 1.7535,
+      "step": 1328
+    },
+    {
+      "epoch": 1.4564383561643837,
+      "grad_norm": 2.446503162384033,
+      "learning_rate": 2.8878960194963448e-06,
+      "loss": 1.8052,
+      "step": 1329
+    },
+    {
+      "epoch": 1.4575342465753425,
+      "grad_norm": 2.2319273948669434,
+      "learning_rate": 2.885865150284322e-06,
+      "loss": 1.8226,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4575342465753425,
+      "eval_loss": 2.022364854812622,
+      "eval_runtime": 37.0901,
+      "eval_samples_per_second": 26.961,
+      "eval_steps_per_second": 1.132,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4586301369863013,
+      "grad_norm": 2.391080856323242,
+      "learning_rate": 2.8838342810722993e-06,
+      "loss": 1.8529,
+      "step": 1331
+    },
+    {
+      "epoch": 1.4597260273972603,
+      "grad_norm": 2.274463415145874,
+      "learning_rate": 2.8818034118602766e-06,
+      "loss": 1.8043,
+      "step": 1332
+    },
+    {
+      "epoch": 1.4608219178082191,
+      "grad_norm": 2.4040963649749756,
+      "learning_rate": 2.879772542648254e-06,
+      "loss": 1.7842,
+      "step": 1333
+    },
+    {
+      "epoch": 1.4619178082191782,
+      "grad_norm": 2.268629312515259,
+      "learning_rate": 2.8777416734362307e-06,
+      "loss": 1.859,
+      "step": 1334
+    },
+    {
+      "epoch": 1.463013698630137,
+      "grad_norm": 2.4727842807769775,
+      "learning_rate": 2.8757108042242084e-06,
+      "loss": 1.7615,
+      "step": 1335
+    },
+    {
+      "epoch": 1.464109589041096,
+      "grad_norm": 2.179358959197998,
+      "learning_rate": 2.8736799350121853e-06,
+      "loss": 1.9085,
+      "step": 1336
+    },
+    {
+      "epoch": 1.4652054794520548,
+      "grad_norm": 2.220917224884033,
+      "learning_rate": 2.871649065800163e-06,
+      "loss": 1.8618,
+      "step": 1337
+    },
+    {
+      "epoch": 1.4663013698630136,
+      "grad_norm": 2.3807291984558105,
+      "learning_rate": 2.86961819658814e-06,
+      "loss": 1.88,
+      "step": 1338
+    },
+    {
+      "epoch": 1.4673972602739727,
+      "grad_norm": 2.272507667541504,
+      "learning_rate": 2.8675873273761175e-06,
+      "loss": 1.8045,
+      "step": 1339
+    },
+    {
+      "epoch": 1.4684931506849315,
+      "grad_norm": 2.1832377910614014,
+      "learning_rate": 2.8655564581640944e-06,
+      "loss": 1.8273,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4684931506849315,
+      "eval_loss": 2.0218358039855957,
+      "eval_runtime": 37.1223,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4695890410958903,
+      "grad_norm": 2.2655251026153564,
+      "learning_rate": 2.863525588952072e-06,
+      "loss": 1.8162,
+      "step": 1341
+    },
+    {
+      "epoch": 1.4706849315068493,
+      "grad_norm": 2.191852331161499,
+      "learning_rate": 2.861494719740049e-06,
+      "loss": 1.8778,
+      "step": 1342
+    },
+    {
+      "epoch": 1.4717808219178083,
+      "grad_norm": 2.1869187355041504,
+      "learning_rate": 2.8594638505280266e-06,
+      "loss": 1.7695,
+      "step": 1343
+    },
+    {
+      "epoch": 1.4728767123287672,
+      "grad_norm": 2.248108148574829,
+      "learning_rate": 2.8574329813160035e-06,
+      "loss": 1.7759,
+      "step": 1344
+    },
+    {
+      "epoch": 1.473972602739726,
+      "grad_norm": 2.2599704265594482,
+      "learning_rate": 2.855402112103981e-06,
+      "loss": 1.8214,
+      "step": 1345
+    },
+    {
+      "epoch": 1.475068493150685,
+      "grad_norm": 2.2920355796813965,
+      "learning_rate": 2.853371242891958e-06,
+      "loss": 1.8061,
+      "step": 1346
+    },
+    {
+      "epoch": 1.4761643835616438,
+      "grad_norm": 2.26715087890625,
+      "learning_rate": 2.851340373679935e-06,
+      "loss": 1.8244,
+      "step": 1347
+    },
+    {
+      "epoch": 1.4772602739726026,
+      "grad_norm": 2.349169969558716,
+      "learning_rate": 2.8493095044679126e-06,
+      "loss": 1.8401,
+      "step": 1348
+    },
+    {
+      "epoch": 1.4783561643835617,
+      "grad_norm": 2.3571698665618896,
+      "learning_rate": 2.8472786352558894e-06,
+      "loss": 1.8913,
+      "step": 1349
+    },
+    {
+      "epoch": 1.4794520547945205,
+      "grad_norm": 2.334057331085205,
+      "learning_rate": 2.845247766043867e-06,
+      "loss": 1.8434,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4794520547945205,
+      "eval_loss": 2.0251200199127197,
+      "eval_runtime": 37.1298,
+      "eval_samples_per_second": 26.933,
+      "eval_steps_per_second": 1.131,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4805479452054795,
+      "grad_norm": 2.363516330718994,
+      "learning_rate": 2.843216896831844e-06,
+      "loss": 1.7846,
+      "step": 1351
+    },
+    {
+      "epoch": 1.4816438356164383,
+      "grad_norm": 2.372403860092163,
+      "learning_rate": 2.8411860276198217e-06,
+      "loss": 1.7791,
+      "step": 1352
+    },
+    {
+      "epoch": 1.4827397260273973,
+      "grad_norm": 2.2666590213775635,
+      "learning_rate": 2.8391551584077985e-06,
+      "loss": 1.7969,
+      "step": 1353
+    },
+    {
+      "epoch": 1.4838356164383562,
+      "grad_norm": 2.3651468753814697,
+      "learning_rate": 2.8371242891957762e-06,
+      "loss": 1.8565,
+      "step": 1354
+    },
+    {
+      "epoch": 1.484931506849315,
+      "grad_norm": 2.3559513092041016,
+      "learning_rate": 2.835093419983753e-06,
+      "loss": 1.8765,
+      "step": 1355
+    },
+    {
+      "epoch": 1.486027397260274,
+      "grad_norm": 2.232041358947754,
+      "learning_rate": 2.833062550771731e-06,
+      "loss": 1.805,
+      "step": 1356
+    },
+    {
+      "epoch": 1.4871232876712328,
+      "grad_norm": 2.2423202991485596,
+      "learning_rate": 2.8310316815597076e-06,
+      "loss": 1.8137,
+      "step": 1357
+    },
+    {
+      "epoch": 1.4882191780821918,
+      "grad_norm": 2.3454222679138184,
+      "learning_rate": 2.8290008123476853e-06,
+      "loss": 1.7839,
+      "step": 1358
+    },
+    {
+      "epoch": 1.4893150684931507,
+      "grad_norm": 2.351266384124756,
+      "learning_rate": 2.826969943135662e-06,
+      "loss": 1.8208,
+      "step": 1359
+    },
+    {
+      "epoch": 1.4904109589041097,
+      "grad_norm": 2.3025388717651367,
+      "learning_rate": 2.8249390739236395e-06,
+      "loss": 1.8059,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4904109589041097,
+      "eval_loss": 2.0235962867736816,
+      "eval_runtime": 37.1359,
+      "eval_samples_per_second": 26.928,
+      "eval_steps_per_second": 1.131,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4915068493150685,
+      "grad_norm": 2.2277750968933105,
+      "learning_rate": 2.8229082047116167e-06,
+      "loss": 1.8238,
+      "step": 1361
+    },
+    {
+      "epoch": 1.4926027397260273,
+      "grad_norm": 2.318286180496216,
+      "learning_rate": 2.820877335499594e-06,
+      "loss": 1.7867,
+      "step": 1362
+    },
+    {
+      "epoch": 1.4936986301369863,
+      "grad_norm": 2.296259880065918,
+      "learning_rate": 2.8188464662875713e-06,
+      "loss": 1.8068,
+      "step": 1363
+    },
+    {
+      "epoch": 1.4947945205479451,
+      "grad_norm": 2.2617714405059814,
+      "learning_rate": 2.8168155970755486e-06,
+      "loss": 1.7607,
+      "step": 1364
+    },
+    {
+      "epoch": 1.4958904109589042,
+      "grad_norm": 2.318294048309326,
+      "learning_rate": 2.814784727863526e-06,
+      "loss": 1.7872,
+      "step": 1365
+    },
+    {
+      "epoch": 1.496986301369863,
+      "grad_norm": 2.260286808013916,
+      "learning_rate": 2.812753858651503e-06,
+      "loss": 1.8013,
+      "step": 1366
+    },
+    {
+      "epoch": 1.498082191780822,
+      "grad_norm": 2.222074508666992,
+      "learning_rate": 2.8107229894394804e-06,
+      "loss": 1.8338,
+      "step": 1367
+    },
+    {
+      "epoch": 1.4991780821917808,
+      "grad_norm": 2.6403002738952637,
+      "learning_rate": 2.8086921202274577e-06,
+      "loss": 1.8126,
+      "step": 1368
+    },
+    {
+      "epoch": 1.5002739726027396,
+      "grad_norm": 2.2439680099487305,
+      "learning_rate": 2.806661251015435e-06,
+      "loss": 1.7669,
+      "step": 1369
+    },
+    {
+      "epoch": 1.5013698630136987,
+      "grad_norm": 2.3914687633514404,
+      "learning_rate": 2.8046303818034122e-06,
+      "loss": 1.8421,
+      "step": 1370
+    },
+    {
+      "epoch": 1.5013698630136987,
+      "eval_loss": 2.0222702026367188,
+      "eval_runtime": 37.1598,
+      "eval_samples_per_second": 26.911,
+      "eval_steps_per_second": 1.13,
+      "step": 1370
+    },
+    {
+      "epoch": 1.5024657534246575,
+      "grad_norm": 2.3510093688964844,
+      "learning_rate": 2.802599512591389e-06,
+      "loss": 1.8902,
+      "step": 1371
+    },
+    {
+      "epoch": 1.5035616438356163,
+      "grad_norm": 2.4427363872528076,
+      "learning_rate": 2.8005686433793668e-06,
+      "loss": 1.9376,
+      "step": 1372
+    },
+    {
+      "epoch": 1.5046575342465753,
+      "grad_norm": 2.299356460571289,
+      "learning_rate": 2.7985377741673436e-06,
+      "loss": 1.812,
+      "step": 1373
+    },
+    {
+      "epoch": 1.5057534246575344,
+      "grad_norm": 2.3738906383514404,
+      "learning_rate": 2.7965069049553213e-06,
+      "loss": 1.7666,
+      "step": 1374
+    },
+    {
+      "epoch": 1.5068493150684932,
+      "grad_norm": 2.3620493412017822,
+      "learning_rate": 2.794476035743298e-06,
+      "loss": 1.824,
+      "step": 1375
+    },
+    {
+      "epoch": 1.507945205479452,
+      "grad_norm": 2.539048671722412,
+      "learning_rate": 2.792445166531276e-06,
+      "loss": 1.8712,
+      "step": 1376
+    },
+    {
+      "epoch": 1.509041095890411,
+      "grad_norm": 2.2931456565856934,
+      "learning_rate": 2.7904142973192527e-06,
+      "loss": 1.8414,
+      "step": 1377
+    },
+    {
+      "epoch": 1.5101369863013698,
+      "grad_norm": 2.4368224143981934,
+      "learning_rate": 2.7883834281072304e-06,
+      "loss": 1.8419,
+      "step": 1378
+    },
+    {
+      "epoch": 1.5112328767123286,
+      "grad_norm": 2.298302173614502,
+      "learning_rate": 2.7863525588952073e-06,
+      "loss": 1.7846,
+      "step": 1379
+    },
+    {
+      "epoch": 1.5123287671232877,
+      "grad_norm": 2.3726806640625,
+      "learning_rate": 2.784321689683185e-06,
+      "loss": 1.7733,
+      "step": 1380
+    },
+    {
+      "epoch": 1.5123287671232877,
+      "eval_loss": 2.0201807022094727,
+      "eval_runtime": 37.1012,
+      "eval_samples_per_second": 26.953,
+      "eval_steps_per_second": 1.132,
+      "step": 1380
+    },
+    {
+      "epoch": 1.5134246575342467,
+      "grad_norm": 2.6515002250671387,
+      "learning_rate": 2.782290820471162e-06,
+      "loss": 1.7764,
+      "step": 1381
+    },
+    {
+      "epoch": 1.5145205479452055,
+      "grad_norm": 2.273272752761841,
+      "learning_rate": 2.7802599512591395e-06,
+      "loss": 1.7747,
+      "step": 1382
+    },
+    {
+      "epoch": 1.5156164383561643,
+      "grad_norm": 2.570009708404541,
+      "learning_rate": 2.7782290820471164e-06,
+      "loss": 1.8612,
+      "step": 1383
+    },
+    {
+      "epoch": 1.5167123287671234,
+      "grad_norm": 2.567878007888794,
+      "learning_rate": 2.7761982128350932e-06,
+      "loss": 1.8402,
+      "step": 1384
+    },
+    {
+      "epoch": 1.5178082191780822,
+      "grad_norm": 2.3768470287323,
+      "learning_rate": 2.774167343623071e-06,
+      "loss": 1.8473,
+      "step": 1385
+    },
+    {
+      "epoch": 1.518904109589041,
+      "grad_norm": 2.1383397579193115,
+      "learning_rate": 2.772136474411048e-06,
+      "loss": 1.8593,
+      "step": 1386
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 2.4437034130096436,
+      "learning_rate": 2.7701056051990255e-06,
+      "loss": 1.792,
+      "step": 1387
+    },
+    {
+      "epoch": 1.521095890410959,
+      "grad_norm": 2.433933734893799,
+      "learning_rate": 2.7680747359870023e-06,
+      "loss": 1.7312,
+      "step": 1388
+    },
+    {
+      "epoch": 1.5221917808219179,
+      "grad_norm": 2.340994119644165,
+      "learning_rate": 2.76604386677498e-06,
+      "loss": 1.8326,
+      "step": 1389
+    },
+    {
+      "epoch": 1.5232876712328767,
+      "grad_norm": 2.4730165004730225,
+      "learning_rate": 2.764012997562957e-06,
+      "loss": 1.8129,
+      "step": 1390
+    },
+    {
+      "epoch": 1.5232876712328767,
+      "eval_loss": 2.0158324241638184,
+      "eval_runtime": 37.0997,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 1390
+    },
+    {
+      "epoch": 1.5243835616438357,
+      "grad_norm": 2.2898151874542236,
+      "learning_rate": 2.7619821283509346e-06,
+      "loss": 1.8726,
+      "step": 1391
+    },
+    {
+      "epoch": 1.5254794520547945,
+      "grad_norm": 2.255439519882202,
+      "learning_rate": 2.7599512591389114e-06,
+      "loss": 1.7587,
+      "step": 1392
+    },
+    {
+      "epoch": 1.5265753424657533,
+      "grad_norm": 2.4269862174987793,
+      "learning_rate": 2.757920389926889e-06,
+      "loss": 1.8498,
+      "step": 1393
+    },
+    {
+      "epoch": 1.5276712328767124,
+      "grad_norm": 2.3411412239074707,
+      "learning_rate": 2.755889520714866e-06,
+      "loss": 1.8103,
+      "step": 1394
+    },
+    {
+      "epoch": 1.5287671232876714,
+      "grad_norm": 2.2859277725219727,
+      "learning_rate": 2.7538586515028437e-06,
+      "loss": 1.7909,
+      "step": 1395
+    },
+    {
+      "epoch": 1.52986301369863,
+      "grad_norm": 2.2675201892852783,
+      "learning_rate": 2.7518277822908205e-06,
+      "loss": 1.7672,
+      "step": 1396
+    },
+    {
+      "epoch": 1.530958904109589,
+      "grad_norm": 2.240610361099243,
+      "learning_rate": 2.749796913078798e-06,
+      "loss": 1.8162,
+      "step": 1397
+    },
+    {
+      "epoch": 1.532054794520548,
+      "grad_norm": 2.272045373916626,
+      "learning_rate": 2.747766043866775e-06,
+      "loss": 1.8593,
+      "step": 1398
+    },
+    {
+      "epoch": 1.5331506849315069,
+      "grad_norm": 2.1736183166503906,
+      "learning_rate": 2.7457351746547524e-06,
+      "loss": 1.8154,
+      "step": 1399
+    },
+    {
+      "epoch": 1.5342465753424657,
+      "grad_norm": 2.197085380554199,
+      "learning_rate": 2.7437043054427297e-06,
+      "loss": 1.8758,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5342465753424657,
+      "eval_loss": 2.017298698425293,
+      "eval_runtime": 37.1666,
+      "eval_samples_per_second": 26.906,
+      "eval_steps_per_second": 1.13,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5353424657534247,
+      "grad_norm": 2.2274084091186523,
+      "learning_rate": 2.741673436230707e-06,
+      "loss": 1.9241,
+      "step": 1401
+    },
+    {
+      "epoch": 1.5364383561643835,
+      "grad_norm": 2.3320305347442627,
+      "learning_rate": 2.739642567018684e-06,
+      "loss": 1.8209,
+      "step": 1402
+    },
+    {
+      "epoch": 1.5375342465753423,
+      "grad_norm": 2.370651960372925,
+      "learning_rate": 2.7376116978066615e-06,
+      "loss": 1.7897,
+      "step": 1403
+    },
+    {
+      "epoch": 1.5386301369863014,
+      "grad_norm": 2.387906312942505,
+      "learning_rate": 2.7355808285946388e-06,
+      "loss": 1.8385,
+      "step": 1404
+    },
+    {
+      "epoch": 1.5397260273972604,
+      "grad_norm": 2.340228796005249,
+      "learning_rate": 2.733549959382616e-06,
+      "loss": 1.8279,
+      "step": 1405
+    },
+    {
+      "epoch": 1.5408219178082192,
+      "grad_norm": 2.338785409927368,
+      "learning_rate": 2.7315190901705933e-06,
+      "loss": 1.7887,
+      "step": 1406
+    },
+    {
+      "epoch": 1.541917808219178,
+      "grad_norm": 2.350057601928711,
+      "learning_rate": 2.7294882209585706e-06,
+      "loss": 1.8772,
+      "step": 1407
+    },
+    {
+      "epoch": 1.543013698630137,
+      "grad_norm": 2.1129322052001953,
+      "learning_rate": 2.727457351746548e-06,
+      "loss": 1.8393,
+      "step": 1408
+    },
+    {
+      "epoch": 1.5441095890410959,
+      "grad_norm": 2.282553195953369,
+      "learning_rate": 2.725426482534525e-06,
+      "loss": 1.7893,
+      "step": 1409
+    },
+    {
+      "epoch": 1.5452054794520547,
+      "grad_norm": 2.3291871547698975,
+      "learning_rate": 2.723395613322502e-06,
+      "loss": 1.7921,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5452054794520547,
+      "eval_loss": 2.0166046619415283,
+      "eval_runtime": 36.7906,
+      "eval_samples_per_second": 27.181,
+      "eval_steps_per_second": 1.142,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5463013698630137,
+      "grad_norm": 2.3050696849823,
+      "learning_rate": 2.7213647441104797e-06,
+      "loss": 1.8292,
+      "step": 1411
+    },
+    {
+      "epoch": 1.5473972602739727,
+      "grad_norm": 2.2762458324432373,
+      "learning_rate": 2.7193338748984565e-06,
+      "loss": 1.811,
+      "step": 1412
+    },
+    {
+      "epoch": 1.5484931506849315,
+      "grad_norm": 2.5544228553771973,
+      "learning_rate": 2.7173030056864342e-06,
+      "loss": 1.8203,
+      "step": 1413
+    },
+    {
+      "epoch": 1.5495890410958904,
+      "grad_norm": 2.6392276287078857,
+      "learning_rate": 2.715272136474411e-06,
+      "loss": 1.9724,
+      "step": 1414
+    },
+    {
+      "epoch": 1.5506849315068494,
+      "grad_norm": 2.445366382598877,
+      "learning_rate": 2.713241267262389e-06,
+      "loss": 1.7589,
+      "step": 1415
+    },
+    {
+      "epoch": 1.5517808219178082,
+      "grad_norm": 2.5305662155151367,
+      "learning_rate": 2.7112103980503656e-06,
+      "loss": 1.8311,
+      "step": 1416
+    },
+    {
+      "epoch": 1.552876712328767,
+      "grad_norm": 2.405482769012451,
+      "learning_rate": 2.7091795288383433e-06,
+      "loss": 1.7893,
+      "step": 1417
+    },
+    {
+      "epoch": 1.553972602739726,
+      "grad_norm": 2.258361577987671,
+      "learning_rate": 2.70714865962632e-06,
+      "loss": 1.8489,
+      "step": 1418
+    },
+    {
+      "epoch": 1.555068493150685,
+      "grad_norm": 2.1798577308654785,
+      "learning_rate": 2.705117790414298e-06,
+      "loss": 1.7983,
+      "step": 1419
+    },
+    {
+      "epoch": 1.5561643835616439,
+      "grad_norm": 2.1985363960266113,
+      "learning_rate": 2.7030869212022747e-06,
+      "loss": 1.7934,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5561643835616439,
+      "eval_loss": 2.0219907760620117,
+      "eval_runtime": 36.9416,
+      "eval_samples_per_second": 27.07,
+      "eval_steps_per_second": 1.137,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5572602739726027,
+      "grad_norm": 2.273252248764038,
+      "learning_rate": 2.7010560519902516e-06,
+      "loss": 1.8571,
+      "step": 1421
+    },
+    {
+      "epoch": 1.5583561643835617,
+      "grad_norm": 2.22920298576355,
+      "learning_rate": 2.6990251827782293e-06,
+      "loss": 1.7785,
+      "step": 1422
+    },
+    {
+      "epoch": 1.5594520547945205,
+      "grad_norm": 2.1784098148345947,
+      "learning_rate": 2.696994313566206e-06,
+      "loss": 1.8053,
+      "step": 1423
+    },
+    {
+      "epoch": 1.5605479452054793,
+      "grad_norm": 2.223404884338379,
+      "learning_rate": 2.694963444354184e-06,
+      "loss": 1.7965,
+      "step": 1424
+    },
+    {
+      "epoch": 1.5616438356164384,
+      "grad_norm": 2.2004787921905518,
+      "learning_rate": 2.6929325751421607e-06,
+      "loss": 1.8851,
+      "step": 1425
+    },
+    {
+      "epoch": 1.5627397260273974,
+      "grad_norm": 2.202467441558838,
+      "learning_rate": 2.6909017059301384e-06,
+      "loss": 1.7884,
+      "step": 1426
+    },
+    {
+      "epoch": 1.563835616438356,
+      "grad_norm": 2.337430238723755,
+      "learning_rate": 2.6888708367181153e-06,
+      "loss": 1.8543,
+      "step": 1427
+    },
+    {
+      "epoch": 1.564931506849315,
+      "grad_norm": 2.322848081588745,
+      "learning_rate": 2.686839967506093e-06,
+      "loss": 1.7607,
+      "step": 1428
+    },
+    {
+      "epoch": 1.566027397260274,
+      "grad_norm": 2.277765989303589,
+      "learning_rate": 2.68480909829407e-06,
+      "loss": 1.8354,
+      "step": 1429
+    },
+    {
+      "epoch": 1.5671232876712329,
+      "grad_norm": 2.3229448795318604,
+      "learning_rate": 2.6827782290820475e-06,
+      "loss": 1.7277,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5671232876712329,
+      "eval_loss": 2.0220985412597656,
+      "eval_runtime": 37.0054,
+      "eval_samples_per_second": 27.023,
+      "eval_steps_per_second": 1.135,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5682191780821917,
+      "grad_norm": 2.49545955657959,
+      "learning_rate": 2.6807473598700244e-06,
+      "loss": 1.7351,
+      "step": 1431
+    },
+    {
+      "epoch": 1.5693150684931507,
+      "grad_norm": 2.530553102493286,
+      "learning_rate": 2.678716490658002e-06,
+      "loss": 1.8614,
+      "step": 1432
+    },
+    {
+      "epoch": 1.5704109589041098,
+      "grad_norm": 2.3352842330932617,
+      "learning_rate": 2.676685621445979e-06,
+      "loss": 1.8218,
+      "step": 1433
+    },
+    {
+      "epoch": 1.5715068493150683,
+      "grad_norm": 2.3533058166503906,
+      "learning_rate": 2.674654752233956e-06,
+      "loss": 1.8282,
+      "step": 1434
+    },
+    {
+      "epoch": 1.5726027397260274,
+      "grad_norm": 2.5509746074676514,
+      "learning_rate": 2.6726238830219335e-06,
+      "loss": 1.8333,
+      "step": 1435
+    },
+    {
+      "epoch": 1.5736986301369864,
+      "grad_norm": 2.185175895690918,
+      "learning_rate": 2.6705930138099107e-06,
+      "loss": 1.8299,
+      "step": 1436
+    },
+    {
+      "epoch": 1.5747945205479452,
+      "grad_norm": 2.2338390350341797,
+      "learning_rate": 2.668562144597888e-06,
+      "loss": 1.857,
+      "step": 1437
+    },
+    {
+      "epoch": 1.575890410958904,
+      "grad_norm": 2.4094536304473877,
+      "learning_rate": 2.6665312753858653e-06,
+      "loss": 1.7333,
+      "step": 1438
+    },
+    {
+      "epoch": 1.576986301369863,
+      "grad_norm": 2.4650957584381104,
+      "learning_rate": 2.6645004061738426e-06,
+      "loss": 1.8237,
+      "step": 1439
+    },
+    {
+      "epoch": 1.5780821917808219,
+      "grad_norm": 2.427415132522583,
+      "learning_rate": 2.66246953696182e-06,
+      "loss": 1.8206,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5780821917808219,
+      "eval_loss": 2.0222699642181396,
+      "eval_runtime": 37.0252,
+      "eval_samples_per_second": 27.009,
+      "eval_steps_per_second": 1.134,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5791780821917807,
+      "grad_norm": 2.3878679275512695,
+      "learning_rate": 2.660438667749797e-06,
+      "loss": 1.775,
+      "step": 1441
+    },
+    {
+      "epoch": 1.5802739726027397,
+      "grad_norm": 2.317819595336914,
+      "learning_rate": 2.6584077985377744e-06,
+      "loss": 1.7479,
+      "step": 1442
+    },
+    {
+      "epoch": 1.5813698630136988,
+      "grad_norm": 2.34951114654541,
+      "learning_rate": 2.6563769293257517e-06,
+      "loss": 1.8025,
+      "step": 1443
+    },
+    {
+      "epoch": 1.5824657534246576,
+      "grad_norm": 2.2849795818328857,
+      "learning_rate": 2.654346060113729e-06,
+      "loss": 1.8152,
+      "step": 1444
+    },
+    {
+      "epoch": 1.5835616438356164,
+      "grad_norm": 2.211489677429199,
+      "learning_rate": 2.6523151909017062e-06,
+      "loss": 1.7666,
+      "step": 1445
+    },
+    {
+      "epoch": 1.5846575342465754,
+      "grad_norm": 2.1876230239868164,
+      "learning_rate": 2.6502843216896835e-06,
+      "loss": 1.7933,
+      "step": 1446
+    },
+    {
+      "epoch": 1.5857534246575342,
+      "grad_norm": 2.367628335952759,
+      "learning_rate": 2.6482534524776603e-06,
+      "loss": 1.7756,
+      "step": 1447
+    },
+    {
+      "epoch": 1.586849315068493,
+      "grad_norm": 2.1198348999023438,
+      "learning_rate": 2.646222583265638e-06,
+      "loss": 1.8908,
+      "step": 1448
+    },
+    {
+      "epoch": 1.587945205479452,
+      "grad_norm": 2.192788600921631,
+      "learning_rate": 2.644191714053615e-06,
+      "loss": 1.8432,
+      "step": 1449
+    },
+    {
+      "epoch": 1.589041095890411,
+      "grad_norm": 2.4480655193328857,
+      "learning_rate": 2.6421608448415926e-06,
+      "loss": 1.8223,
+      "step": 1450
+    },
+    {
+      "epoch": 1.589041095890411,
+      "eval_loss": 2.019976854324341,
+      "eval_runtime": 37.0117,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1450
+    },
+    {
+      "epoch": 1.59013698630137,
+      "grad_norm": 2.573078155517578,
+      "learning_rate": 2.6401299756295694e-06,
+      "loss": 1.7881,
+      "step": 1451
+    },
+    {
+      "epoch": 1.5912328767123287,
+      "grad_norm": 2.2820355892181396,
+      "learning_rate": 2.638099106417547e-06,
+      "loss": 1.7703,
+      "step": 1452
+    },
+    {
+      "epoch": 1.5923287671232877,
+      "grad_norm": 2.526477575302124,
+      "learning_rate": 2.636068237205524e-06,
+      "loss": 1.8281,
+      "step": 1453
+    },
+    {
+      "epoch": 1.5934246575342466,
+      "grad_norm": 2.873647689819336,
+      "learning_rate": 2.6340373679935017e-06,
+      "loss": 1.7771,
+      "step": 1454
+    },
+    {
+      "epoch": 1.5945205479452054,
+      "grad_norm": 2.3782413005828857,
+      "learning_rate": 2.6320064987814785e-06,
+      "loss": 1.8864,
+      "step": 1455
+    },
+    {
+      "epoch": 1.5956164383561644,
+      "grad_norm": 2.3153343200683594,
+      "learning_rate": 2.6299756295694562e-06,
+      "loss": 1.734,
+      "step": 1456
+    },
+    {
+      "epoch": 1.5967123287671234,
+      "grad_norm": 2.631880283355713,
+      "learning_rate": 2.627944760357433e-06,
+      "loss": 1.8295,
+      "step": 1457
+    },
+    {
+      "epoch": 1.5978082191780822,
+      "grad_norm": 2.5827910900115967,
+      "learning_rate": 2.6259138911454104e-06,
+      "loss": 1.8738,
+      "step": 1458
+    },
+    {
+      "epoch": 1.598904109589041,
+      "grad_norm": 2.3519935607910156,
+      "learning_rate": 2.6238830219333877e-06,
+      "loss": 1.8832,
+      "step": 1459
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 2.5236215591430664,
+      "learning_rate": 2.621852152721365e-06,
+      "loss": 1.8595,
+      "step": 1460
+    },
+    {
+      "epoch": 1.6,
+      "eval_loss": 2.0170512199401855,
+      "eval_runtime": 37.0302,
+      "eval_samples_per_second": 27.005,
+      "eval_steps_per_second": 1.134,
+      "step": 1460
+    },
+    {
+      "epoch": 1.601095890410959,
+      "grad_norm": 2.725083112716675,
+      "learning_rate": 2.619821283509342e-06,
+      "loss": 1.7476,
+      "step": 1461
+    },
+    {
+      "epoch": 1.6021917808219177,
+      "grad_norm": 2.212644100189209,
+      "learning_rate": 2.6177904142973195e-06,
+      "loss": 1.8475,
+      "step": 1462
+    },
+    {
+      "epoch": 1.6032876712328767,
+      "grad_norm": 2.346808910369873,
+      "learning_rate": 2.6157595450852968e-06,
+      "loss": 1.7762,
+      "step": 1463
+    },
+    {
+      "epoch": 1.6043835616438358,
+      "grad_norm": 2.702331304550171,
+      "learning_rate": 2.613728675873274e-06,
+      "loss": 1.8296,
+      "step": 1464
+    },
+    {
+      "epoch": 1.6054794520547944,
+      "grad_norm": 2.4175899028778076,
+      "learning_rate": 2.6116978066612513e-06,
+      "loss": 1.8476,
+      "step": 1465
+    },
+    {
+      "epoch": 1.6065753424657534,
+      "grad_norm": 2.325457811355591,
+      "learning_rate": 2.6096669374492286e-06,
+      "loss": 1.8706,
+      "step": 1466
+    },
+    {
+      "epoch": 1.6076712328767124,
+      "grad_norm": 2.383842945098877,
+      "learning_rate": 2.607636068237206e-06,
+      "loss": 1.8246,
+      "step": 1467
+    },
+    {
+      "epoch": 1.6087671232876712,
+      "grad_norm": 2.289515256881714,
+      "learning_rate": 2.605605199025183e-06,
+      "loss": 1.7728,
+      "step": 1468
+    },
+    {
+      "epoch": 1.60986301369863,
+      "grad_norm": 2.5116689205169678,
+      "learning_rate": 2.6035743298131604e-06,
+      "loss": 1.8359,
+      "step": 1469
+    },
+    {
+      "epoch": 1.610958904109589,
+      "grad_norm": 2.279695749282837,
+      "learning_rate": 2.6015434606011377e-06,
+      "loss": 1.8155,
+      "step": 1470
+    },
+    {
+      "epoch": 1.610958904109589,
+      "eval_loss": 2.0142691135406494,
+      "eval_runtime": 37.0119,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1470
+    },
+    {
+      "epoch": 1.612054794520548,
+      "grad_norm": 2.4397690296173096,
+      "learning_rate": 2.5995125913891145e-06,
+      "loss": 1.8424,
+      "step": 1471
+    },
+    {
+      "epoch": 1.6131506849315067,
+      "grad_norm": 2.433577299118042,
+      "learning_rate": 2.5974817221770922e-06,
+      "loss": 1.8196,
+      "step": 1472
+    },
+    {
+      "epoch": 1.6142465753424657,
+      "grad_norm": 2.365459442138672,
+      "learning_rate": 2.595450852965069e-06,
+      "loss": 1.8124,
+      "step": 1473
+    },
+    {
+      "epoch": 1.6153424657534248,
+      "grad_norm": 2.253133535385132,
+      "learning_rate": 2.593419983753047e-06,
+      "loss": 1.8443,
+      "step": 1474
+    },
+    {
+      "epoch": 1.6164383561643836,
+      "grad_norm": 2.227625608444214,
+      "learning_rate": 2.5913891145410236e-06,
+      "loss": 1.8524,
+      "step": 1475
+    },
+    {
+      "epoch": 1.6175342465753424,
+      "grad_norm": 2.403702974319458,
+      "learning_rate": 2.5893582453290013e-06,
+      "loss": 1.7983,
+      "step": 1476
+    },
+    {
+      "epoch": 1.6186301369863014,
+      "grad_norm": 2.24507999420166,
+      "learning_rate": 2.587327376116978e-06,
+      "loss": 1.8563,
+      "step": 1477
+    },
+    {
+      "epoch": 1.6197260273972602,
+      "grad_norm": 2.284700393676758,
+      "learning_rate": 2.585296506904956e-06,
+      "loss": 1.7958,
+      "step": 1478
+    },
+    {
+      "epoch": 1.620821917808219,
+      "grad_norm": 2.288215398788452,
+      "learning_rate": 2.5832656376929327e-06,
+      "loss": 1.8021,
+      "step": 1479
+    },
+    {
+      "epoch": 1.621917808219178,
+      "grad_norm": 2.2312939167022705,
+      "learning_rate": 2.5812347684809104e-06,
+      "loss": 1.7856,
+      "step": 1480
+    },
+    {
+      "epoch": 1.621917808219178,
+      "eval_loss": 2.01674222946167,
+      "eval_runtime": 36.9943,
+      "eval_samples_per_second": 27.031,
+      "eval_steps_per_second": 1.135,
+      "step": 1480
+    },
+    {
+      "epoch": 1.6230136986301371,
+      "grad_norm": 2.319363832473755,
+      "learning_rate": 2.5792038992688873e-06,
+      "loss": 1.7263,
+      "step": 1481
+    },
+    {
+      "epoch": 1.624109589041096,
+      "grad_norm": 2.4342033863067627,
+      "learning_rate": 2.577173030056865e-06,
+      "loss": 1.897,
+      "step": 1482
+    },
+    {
+      "epoch": 1.6252054794520547,
+      "grad_norm": 2.3052282333374023,
+      "learning_rate": 2.575142160844842e-06,
+      "loss": 1.7165,
+      "step": 1483
+    },
+    {
+      "epoch": 1.6263013698630138,
+      "grad_norm": 2.2476744651794434,
+      "learning_rate": 2.5731112916328187e-06,
+      "loss": 1.8347,
+      "step": 1484
+    },
+    {
+      "epoch": 1.6273972602739726,
+      "grad_norm": 2.3128387928009033,
+      "learning_rate": 2.5710804224207964e-06,
+      "loss": 1.8403,
+      "step": 1485
+    },
+    {
+      "epoch": 1.6284931506849314,
+      "grad_norm": 2.3483974933624268,
+      "learning_rate": 2.5690495532087733e-06,
+      "loss": 1.781,
+      "step": 1486
+    },
+    {
+      "epoch": 1.6295890410958904,
+      "grad_norm": 2.328298568725586,
+      "learning_rate": 2.567018683996751e-06,
+      "loss": 1.7356,
+      "step": 1487
+    },
+    {
+      "epoch": 1.6306849315068495,
+      "grad_norm": 2.2756922245025635,
+      "learning_rate": 2.564987814784728e-06,
+      "loss": 1.8449,
+      "step": 1488
+    },
+    {
+      "epoch": 1.6317808219178083,
+      "grad_norm": 2.363874912261963,
+      "learning_rate": 2.5629569455727055e-06,
+      "loss": 1.7767,
+      "step": 1489
+    },
+    {
+      "epoch": 1.632876712328767,
+      "grad_norm": 2.2872767448425293,
+      "learning_rate": 2.5609260763606824e-06,
+      "loss": 1.8052,
+      "step": 1490
+    },
+    {
+      "epoch": 1.632876712328767,
+      "eval_loss": 2.0134243965148926,
+      "eval_runtime": 37.029,
+      "eval_samples_per_second": 27.006,
+      "eval_steps_per_second": 1.134,
+      "step": 1490
+    },
+    {
+      "epoch": 1.633972602739726,
+      "grad_norm": 2.354851007461548,
+      "learning_rate": 2.55889520714866e-06,
+      "loss": 1.7953,
+      "step": 1491
+    },
+    {
+      "epoch": 1.635068493150685,
+      "grad_norm": 2.4440531730651855,
+      "learning_rate": 2.556864337936637e-06,
+      "loss": 1.9063,
+      "step": 1492
+    },
+    {
+      "epoch": 1.6361643835616437,
+      "grad_norm": 2.296835422515869,
+      "learning_rate": 2.5548334687246146e-06,
+      "loss": 1.7961,
+      "step": 1493
+    },
+    {
+      "epoch": 1.6372602739726028,
+      "grad_norm": 2.7421159744262695,
+      "learning_rate": 2.5528025995125915e-06,
+      "loss": 1.809,
+      "step": 1494
+    },
+    {
+      "epoch": 1.6383561643835618,
+      "grad_norm": 2.819125175476074,
+      "learning_rate": 2.550771730300569e-06,
+      "loss": 1.853,
+      "step": 1495
+    },
+    {
+      "epoch": 1.6394520547945204,
+      "grad_norm": 2.2536680698394775,
+      "learning_rate": 2.548740861088546e-06,
+      "loss": 1.8297,
+      "step": 1496
+    },
+    {
+      "epoch": 1.6405479452054794,
+      "grad_norm": 2.7130825519561768,
+      "learning_rate": 2.5467099918765233e-06,
+      "loss": 1.843,
+      "step": 1497
+    },
+    {
+      "epoch": 1.6416438356164385,
+      "grad_norm": 3.0625038146972656,
+      "learning_rate": 2.5446791226645006e-06,
+      "loss": 1.7462,
+      "step": 1498
+    },
+    {
+      "epoch": 1.6427397260273973,
+      "grad_norm": 2.3095271587371826,
+      "learning_rate": 2.542648253452478e-06,
+      "loss": 1.7363,
+      "step": 1499
+    },
+    {
+      "epoch": 1.643835616438356,
+      "grad_norm": 2.536547899246216,
+      "learning_rate": 2.540617384240455e-06,
+      "loss": 1.8422,
+      "step": 1500
+    },
+    {
+      "epoch": 1.643835616438356,
+      "eval_loss": 2.0126097202301025,
+      "eval_runtime": 37.0469,
+      "eval_samples_per_second": 26.993,
+      "eval_steps_per_second": 1.134,
+      "step": 1500
+    },
+    {
+      "epoch": 1.644931506849315,
+      "grad_norm": 2.7454309463500977,
+      "learning_rate": 2.5385865150284324e-06,
+      "loss": 1.8183,
+      "step": 1501
+    },
+    {
+      "epoch": 1.6460273972602741,
+      "grad_norm": 2.3443901538848877,
+      "learning_rate": 2.5365556458164097e-06,
+      "loss": 1.7827,
+      "step": 1502
+    },
+    {
+      "epoch": 1.6471232876712327,
+      "grad_norm": 2.2412657737731934,
+      "learning_rate": 2.534524776604387e-06,
+      "loss": 1.797,
+      "step": 1503
+    },
+    {
+      "epoch": 1.6482191780821918,
+      "grad_norm": 2.6193699836730957,
+      "learning_rate": 2.5324939073923642e-06,
+      "loss": 1.8103,
+      "step": 1504
+    },
+    {
+      "epoch": 1.6493150684931508,
+      "grad_norm": 2.363919258117676,
+      "learning_rate": 2.5304630381803415e-06,
+      "loss": 1.6987,
+      "step": 1505
+    },
+    {
+      "epoch": 1.6504109589041096,
+      "grad_norm": 2.303164482116699,
+      "learning_rate": 2.5284321689683188e-06,
+      "loss": 1.8005,
+      "step": 1506
+    },
+    {
+      "epoch": 1.6515068493150684,
+      "grad_norm": 2.444174289703369,
+      "learning_rate": 2.526401299756296e-06,
+      "loss": 1.8009,
+      "step": 1507
+    },
+    {
+      "epoch": 1.6526027397260274,
+      "grad_norm": 2.418454647064209,
+      "learning_rate": 2.524370430544273e-06,
+      "loss": 1.782,
+      "step": 1508
+    },
+    {
+      "epoch": 1.6536986301369863,
+      "grad_norm": 2.48840069770813,
+      "learning_rate": 2.5223395613322506e-06,
+      "loss": 1.8143,
+      "step": 1509
+    },
+    {
+      "epoch": 1.654794520547945,
+      "grad_norm": 2.2814347743988037,
+      "learning_rate": 2.5203086921202274e-06,
+      "loss": 1.7687,
+      "step": 1510
+    },
+    {
+      "epoch": 1.654794520547945,
+      "eval_loss": 2.0148253440856934,
+      "eval_runtime": 36.7833,
+      "eval_samples_per_second": 27.186,
+      "eval_steps_per_second": 1.142,
+      "step": 1510
+    },
+    {
+      "epoch": 1.655890410958904,
+      "grad_norm": 2.228933572769165,
+      "learning_rate": 2.518277822908205e-06,
+      "loss": 1.8416,
+      "step": 1511
+    },
+    {
+      "epoch": 1.6569863013698631,
+      "grad_norm": 2.3696742057800293,
+      "learning_rate": 2.516246953696182e-06,
+      "loss": 1.7868,
+      "step": 1512
+    },
+    {
+      "epoch": 1.658082191780822,
+      "grad_norm": 2.216360092163086,
+      "learning_rate": 2.5142160844841597e-06,
+      "loss": 1.8312,
+      "step": 1513
+    },
+    {
+      "epoch": 1.6591780821917808,
+      "grad_norm": 2.454345226287842,
+      "learning_rate": 2.5121852152721365e-06,
+      "loss": 1.8309,
+      "step": 1514
+    },
+    {
+      "epoch": 1.6602739726027398,
+      "grad_norm": 2.2106776237487793,
+      "learning_rate": 2.5101543460601142e-06,
+      "loss": 1.8155,
+      "step": 1515
+    },
+    {
+      "epoch": 1.6613698630136986,
+      "grad_norm": 2.289006233215332,
+      "learning_rate": 2.508123476848091e-06,
+      "loss": 1.8162,
+      "step": 1516
+    },
+    {
+      "epoch": 1.6624657534246574,
+      "grad_norm": 2.297246217727661,
+      "learning_rate": 2.506092607636069e-06,
+      "loss": 1.7559,
+      "step": 1517
+    },
+    {
+      "epoch": 1.6635616438356164,
+      "grad_norm": 2.3909482955932617,
+      "learning_rate": 2.5040617384240457e-06,
+      "loss": 1.8389,
+      "step": 1518
+    },
+    {
+      "epoch": 1.6646575342465755,
+      "grad_norm": 2.398171901702881,
+      "learning_rate": 2.5020308692120234e-06,
+      "loss": 1.7785,
+      "step": 1519
+    },
+    {
+      "epoch": 1.6657534246575343,
+      "grad_norm": 2.377707004547119,
+      "learning_rate": 2.5e-06,
+      "loss": 1.768,
+      "step": 1520
+    },
+    {
+      "epoch": 1.6657534246575343,
+      "eval_loss": 2.012815475463867,
+      "eval_runtime": 36.9603,
+      "eval_samples_per_second": 27.056,
+      "eval_steps_per_second": 1.136,
+      "step": 1520
+    },
+    {
+      "epoch": 1.666849315068493,
+      "grad_norm": 2.336160898208618,
+      "learning_rate": 2.4979691307879775e-06,
+      "loss": 1.8653,
+      "step": 1521
+    },
+    {
+      "epoch": 1.6679452054794521,
+      "grad_norm": 2.288424253463745,
+      "learning_rate": 2.4959382615759548e-06,
+      "loss": 1.7798,
+      "step": 1522
+    },
+    {
+      "epoch": 1.669041095890411,
+      "grad_norm": 2.2599472999572754,
+      "learning_rate": 2.493907392363932e-06,
+      "loss": 1.7736,
+      "step": 1523
+    },
+    {
+      "epoch": 1.6701369863013698,
+      "grad_norm": 2.242276430130005,
+      "learning_rate": 2.4918765231519093e-06,
+      "loss": 1.8753,
+      "step": 1524
+    },
+    {
+      "epoch": 1.6712328767123288,
+      "grad_norm": 2.1897733211517334,
+      "learning_rate": 2.4898456539398866e-06,
+      "loss": 1.8353,
+      "step": 1525
+    },
+    {
+      "epoch": 1.6723287671232878,
+      "grad_norm": 2.254009246826172,
+      "learning_rate": 2.487814784727864e-06,
+      "loss": 1.7825,
+      "step": 1526
+    },
+    {
+      "epoch": 1.6734246575342466,
+      "grad_norm": 2.3704216480255127,
+      "learning_rate": 2.4857839155158407e-06,
+      "loss": 1.8742,
+      "step": 1527
+    },
+    {
+      "epoch": 1.6745205479452054,
+      "grad_norm": 2.1590118408203125,
+      "learning_rate": 2.483753046303818e-06,
+      "loss": 1.7974,
+      "step": 1528
+    },
+    {
+      "epoch": 1.6756164383561645,
+      "grad_norm": 2.2389416694641113,
+      "learning_rate": 2.4817221770917953e-06,
+      "loss": 1.8007,
+      "step": 1529
+    },
+    {
+      "epoch": 1.6767123287671233,
+      "grad_norm": 2.324131488800049,
+      "learning_rate": 2.4796913078797725e-06,
+      "loss": 1.8079,
+      "step": 1530
+    },
+    {
+      "epoch": 1.6767123287671233,
+      "eval_loss": 2.013923168182373,
+      "eval_runtime": 37.0828,
+      "eval_samples_per_second": 26.967,
+      "eval_steps_per_second": 1.133,
+      "step": 1530
+    },
+    {
+      "epoch": 1.677808219178082,
+      "grad_norm": 2.406456708908081,
+      "learning_rate": 2.47766043866775e-06,
+      "loss": 1.8356,
+      "step": 1531
+    },
+    {
+      "epoch": 1.6789041095890411,
+      "grad_norm": 2.289919137954712,
+      "learning_rate": 2.475629569455727e-06,
+      "loss": 1.8239,
+      "step": 1532
+    },
+    {
+      "epoch": 1.6800000000000002,
+      "grad_norm": 2.2617733478546143,
+      "learning_rate": 2.4735987002437044e-06,
+      "loss": 1.8025,
+      "step": 1533
+    },
+    {
+      "epoch": 1.6810958904109587,
+      "grad_norm": 2.133652925491333,
+      "learning_rate": 2.4715678310316816e-06,
+      "loss": 1.7743,
+      "step": 1534
+    },
+    {
+      "epoch": 1.6821917808219178,
+      "grad_norm": 2.215322256088257,
+      "learning_rate": 2.469536961819659e-06,
+      "loss": 1.7979,
+      "step": 1535
+    },
+    {
+      "epoch": 1.6832876712328768,
+      "grad_norm": 2.336021900177002,
+      "learning_rate": 2.467506092607636e-06,
+      "loss": 1.8027,
+      "step": 1536
+    },
+    {
+      "epoch": 1.6843835616438356,
+      "grad_norm": 2.3567142486572266,
+      "learning_rate": 2.4654752233956135e-06,
+      "loss": 1.7877,
+      "step": 1537
+    },
+    {
+      "epoch": 1.6854794520547944,
+      "grad_norm": 2.2642927169799805,
+      "learning_rate": 2.4634443541835907e-06,
+      "loss": 1.7695,
+      "step": 1538
+    },
+    {
+      "epoch": 1.6865753424657535,
+      "grad_norm": 2.353303909301758,
+      "learning_rate": 2.461413484971568e-06,
+      "loss": 1.867,
+      "step": 1539
+    },
+    {
+      "epoch": 1.6876712328767123,
+      "grad_norm": 2.252074956893921,
+      "learning_rate": 2.4593826157595453e-06,
+      "loss": 1.8009,
+      "step": 1540
+    },
+    {
+      "epoch": 1.6876712328767123,
+      "eval_loss": 2.0113608837127686,
+      "eval_runtime": 37.0875,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 1540
+    },
+    {
+      "epoch": 1.688767123287671,
+      "grad_norm": 2.2303407192230225,
+      "learning_rate": 2.4573517465475226e-06,
+      "loss": 1.8881,
+      "step": 1541
+    },
+    {
+      "epoch": 1.6898630136986301,
+      "grad_norm": 2.3359081745147705,
+      "learning_rate": 2.4553208773355e-06,
+      "loss": 1.7388,
+      "step": 1542
+    },
+    {
+      "epoch": 1.6909589041095892,
+      "grad_norm": 2.3783583641052246,
+      "learning_rate": 2.453290008123477e-06,
+      "loss": 1.7993,
+      "step": 1543
+    },
+    {
+      "epoch": 1.692054794520548,
+      "grad_norm": 2.159264326095581,
+      "learning_rate": 2.4512591389114544e-06,
+      "loss": 1.81,
+      "step": 1544
+    },
+    {
+      "epoch": 1.6931506849315068,
+      "grad_norm": 2.2232203483581543,
+      "learning_rate": 2.4492282696994317e-06,
+      "loss": 1.8095,
+      "step": 1545
+    },
+    {
+      "epoch": 1.6942465753424658,
+      "grad_norm": 2.1755056381225586,
+      "learning_rate": 2.447197400487409e-06,
+      "loss": 1.8149,
+      "step": 1546
+    },
+    {
+      "epoch": 1.6953424657534246,
+      "grad_norm": 2.325871467590332,
+      "learning_rate": 2.4451665312753862e-06,
+      "loss": 1.8154,
+      "step": 1547
+    },
+    {
+      "epoch": 1.6964383561643834,
+      "grad_norm": 2.286848545074463,
+      "learning_rate": 2.4431356620633635e-06,
+      "loss": 1.8409,
+      "step": 1548
+    },
+    {
+      "epoch": 1.6975342465753425,
+      "grad_norm": 2.359510898590088,
+      "learning_rate": 2.4411047928513408e-06,
+      "loss": 1.843,
+      "step": 1549
+    },
+    {
+      "epoch": 1.6986301369863015,
+      "grad_norm": 2.474428176879883,
+      "learning_rate": 2.439073923639318e-06,
+      "loss": 1.8089,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6986301369863015,
+      "eval_loss": 2.0077731609344482,
+      "eval_runtime": 37.1092,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6997260273972603,
+      "grad_norm": 2.3662912845611572,
+      "learning_rate": 2.4370430544272953e-06,
+      "loss": 1.8176,
+      "step": 1551
+    },
+    {
+      "epoch": 1.7008219178082191,
+      "grad_norm": 2.448948860168457,
+      "learning_rate": 2.435012185215272e-06,
+      "loss": 1.7776,
+      "step": 1552
+    },
+    {
+      "epoch": 1.7019178082191782,
+      "grad_norm": 2.333739757537842,
+      "learning_rate": 2.4329813160032495e-06,
+      "loss": 1.7552,
+      "step": 1553
+    },
+    {
+      "epoch": 1.703013698630137,
+      "grad_norm": 2.170365571975708,
+      "learning_rate": 2.4309504467912267e-06,
+      "loss": 1.7759,
+      "step": 1554
+    },
+    {
+      "epoch": 1.7041095890410958,
+      "grad_norm": 2.2343010902404785,
+      "learning_rate": 2.428919577579204e-06,
+      "loss": 1.8073,
+      "step": 1555
+    },
+    {
+      "epoch": 1.7052054794520548,
+      "grad_norm": 2.2245287895202637,
+      "learning_rate": 2.4268887083671813e-06,
+      "loss": 1.7826,
+      "step": 1556
+    },
+    {
+      "epoch": 1.7063013698630138,
+      "grad_norm": 2.339942216873169,
+      "learning_rate": 2.4248578391551586e-06,
+      "loss": 1.8766,
+      "step": 1557
+    },
+    {
+      "epoch": 1.7073972602739727,
+      "grad_norm": 2.1585116386413574,
+      "learning_rate": 2.422826969943136e-06,
+      "loss": 1.7605,
+      "step": 1558
+    },
+    {
+      "epoch": 1.7084931506849315,
+      "grad_norm": 2.2108142375946045,
+      "learning_rate": 2.420796100731113e-06,
+      "loss": 1.7855,
+      "step": 1559
+    },
+    {
+      "epoch": 1.7095890410958905,
+      "grad_norm": 2.192457675933838,
+      "learning_rate": 2.4187652315190904e-06,
+      "loss": 1.7489,
+      "step": 1560
+    },
+    {
+      "epoch": 1.7095890410958905,
+      "eval_loss": 2.0085604190826416,
+      "eval_runtime": 37.1105,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1560
+    },
+    {
+      "epoch": 1.7106849315068493,
+      "grad_norm": 2.291106939315796,
+      "learning_rate": 2.4167343623070677e-06,
+      "loss": 1.7981,
+      "step": 1561
+    },
+    {
+      "epoch": 1.7117808219178081,
+      "grad_norm": 2.509237766265869,
+      "learning_rate": 2.414703493095045e-06,
+      "loss": 1.8146,
+      "step": 1562
+    },
+    {
+      "epoch": 1.7128767123287671,
+      "grad_norm": 2.497162342071533,
+      "learning_rate": 2.4126726238830222e-06,
+      "loss": 1.807,
+      "step": 1563
+    },
+    {
+      "epoch": 1.7139726027397262,
+      "grad_norm": 2.256833076477051,
+      "learning_rate": 2.410641754670999e-06,
+      "loss": 1.8304,
+      "step": 1564
+    },
+    {
+      "epoch": 1.7150684931506848,
+      "grad_norm": 2.2523772716522217,
+      "learning_rate": 2.4086108854589763e-06,
+      "loss": 1.7794,
+      "step": 1565
+    },
+    {
+      "epoch": 1.7161643835616438,
+      "grad_norm": 2.289182424545288,
+      "learning_rate": 2.4065800162469536e-06,
+      "loss": 1.8425,
+      "step": 1566
+    },
+    {
+      "epoch": 1.7172602739726028,
+      "grad_norm": 2.734487771987915,
+      "learning_rate": 2.404549147034931e-06,
+      "loss": 1.7791,
+      "step": 1567
+    },
+    {
+      "epoch": 1.7183561643835616,
+      "grad_norm": 2.3307888507843018,
+      "learning_rate": 2.402518277822908e-06,
+      "loss": 1.8064,
+      "step": 1568
+    },
+    {
+      "epoch": 1.7194520547945205,
+      "grad_norm": 2.5016298294067383,
+      "learning_rate": 2.4004874086108854e-06,
+      "loss": 1.8544,
+      "step": 1569
+    },
+    {
+      "epoch": 1.7205479452054795,
+      "grad_norm": 2.142246961593628,
+      "learning_rate": 2.3984565393988627e-06,
+      "loss": 1.7653,
+      "step": 1570
+    },
+    {
+      "epoch": 1.7205479452054795,
+      "eval_loss": 2.0068864822387695,
+      "eval_runtime": 37.1078,
+      "eval_samples_per_second": 26.949,
+      "eval_steps_per_second": 1.132,
+      "step": 1570
+    },
+    {
+      "epoch": 1.7216438356164385,
+      "grad_norm": 2.5692505836486816,
+      "learning_rate": 2.39642567018684e-06,
+      "loss": 1.8444,
+      "step": 1571
+    },
+    {
+      "epoch": 1.7227397260273971,
+      "grad_norm": 2.2596213817596436,
+      "learning_rate": 2.3943948009748173e-06,
+      "loss": 1.7694,
+      "step": 1572
+    },
+    {
+      "epoch": 1.7238356164383561,
+      "grad_norm": 2.230984926223755,
+      "learning_rate": 2.3923639317627945e-06,
+      "loss": 1.786,
+      "step": 1573
+    },
+    {
+      "epoch": 1.7249315068493152,
+      "grad_norm": 2.366178274154663,
+      "learning_rate": 2.390333062550772e-06,
+      "loss": 1.8138,
+      "step": 1574
+    },
+    {
+      "epoch": 1.726027397260274,
+      "grad_norm": 2.23783016204834,
+      "learning_rate": 2.388302193338749e-06,
+      "loss": 1.7615,
+      "step": 1575
+    },
+    {
+      "epoch": 1.7271232876712328,
+      "grad_norm": 2.3766677379608154,
+      "learning_rate": 2.3862713241267264e-06,
+      "loss": 1.6811,
+      "step": 1576
+    },
+    {
+      "epoch": 1.7282191780821918,
+      "grad_norm": 2.3434040546417236,
+      "learning_rate": 2.3842404549147037e-06,
+      "loss": 1.8806,
+      "step": 1577
+    },
+    {
+      "epoch": 1.7293150684931506,
+      "grad_norm": 2.2444276809692383,
+      "learning_rate": 2.382209585702681e-06,
+      "loss": 1.7968,
+      "step": 1578
+    },
+    {
+      "epoch": 1.7304109589041095,
+      "grad_norm": 2.2524919509887695,
+      "learning_rate": 2.380178716490658e-06,
+      "loss": 1.7602,
+      "step": 1579
+    },
+    {
+      "epoch": 1.7315068493150685,
+      "grad_norm": 2.3310587406158447,
+      "learning_rate": 2.3781478472786355e-06,
+      "loss": 1.8414,
+      "step": 1580
+    },
+    {
+      "epoch": 1.7315068493150685,
+      "eval_loss": 1.9981482028961182,
+      "eval_runtime": 37.0827,
+      "eval_samples_per_second": 26.967,
+      "eval_steps_per_second": 1.133,
+      "step": 1580
+    },
+    {
+      "epoch": 1.7326027397260275,
+      "grad_norm": 2.4081802368164062,
+      "learning_rate": 2.3761169780666128e-06,
+      "loss": 1.8002,
+      "step": 1581
+    },
+    {
+      "epoch": 1.7336986301369863,
+      "grad_norm": 2.293506622314453,
+      "learning_rate": 2.37408610885459e-06,
+      "loss": 1.8365,
+      "step": 1582
+    },
+    {
+      "epoch": 1.7347945205479451,
+      "grad_norm": 2.2015790939331055,
+      "learning_rate": 2.3720552396425673e-06,
+      "loss": 1.8571,
+      "step": 1583
+    },
+    {
+      "epoch": 1.7358904109589042,
+      "grad_norm": 2.308809518814087,
+      "learning_rate": 2.3700243704305446e-06,
+      "loss": 1.7821,
+      "step": 1584
+    },
+    {
+      "epoch": 1.736986301369863,
+      "grad_norm": 2.376842975616455,
+      "learning_rate": 2.367993501218522e-06,
+      "loss": 1.8701,
+      "step": 1585
+    },
+    {
+      "epoch": 1.7380821917808218,
+      "grad_norm": 2.4102022647857666,
+      "learning_rate": 2.365962632006499e-06,
+      "loss": 1.8055,
+      "step": 1586
+    },
+    {
+      "epoch": 1.7391780821917808,
+      "grad_norm": 2.2734079360961914,
+      "learning_rate": 2.3639317627944764e-06,
+      "loss": 1.8117,
+      "step": 1587
+    },
+    {
+      "epoch": 1.7402739726027399,
+      "grad_norm": 2.3426599502563477,
+      "learning_rate": 2.3619008935824537e-06,
+      "loss": 1.8337,
+      "step": 1588
+    },
+    {
+      "epoch": 1.7413698630136987,
+      "grad_norm": 2.340383768081665,
+      "learning_rate": 2.3598700243704305e-06,
+      "loss": 1.8247,
+      "step": 1589
+    },
+    {
+      "epoch": 1.7424657534246575,
+      "grad_norm": 2.2561380863189697,
+      "learning_rate": 2.357839155158408e-06,
+      "loss": 1.8249,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7424657534246575,
+      "eval_loss": 1.9979513883590698,
+      "eval_runtime": 37.0948,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7435616438356165,
+      "grad_norm": 2.2459654808044434,
+      "learning_rate": 2.355808285946385e-06,
+      "loss": 1.8273,
+      "step": 1591
+    },
+    {
+      "epoch": 1.7446575342465753,
+      "grad_norm": 2.2017974853515625,
+      "learning_rate": 2.3537774167343624e-06,
+      "loss": 1.828,
+      "step": 1592
+    },
+    {
+      "epoch": 1.7457534246575341,
+      "grad_norm": 2.286304235458374,
+      "learning_rate": 2.3517465475223396e-06,
+      "loss": 1.8497,
+      "step": 1593
+    },
+    {
+      "epoch": 1.7468493150684932,
+      "grad_norm": 2.296494722366333,
+      "learning_rate": 2.349715678310317e-06,
+      "loss": 1.8155,
+      "step": 1594
+    },
+    {
+      "epoch": 1.7479452054794522,
+      "grad_norm": 2.2989203929901123,
+      "learning_rate": 2.347684809098294e-06,
+      "loss": 1.8214,
+      "step": 1595
+    },
+    {
+      "epoch": 1.749041095890411,
+      "grad_norm": 2.3034965991973877,
+      "learning_rate": 2.3456539398862715e-06,
+      "loss": 1.7792,
+      "step": 1596
+    },
+    {
+      "epoch": 1.7501369863013698,
+      "grad_norm": 2.3823719024658203,
+      "learning_rate": 2.3436230706742487e-06,
+      "loss": 1.8326,
+      "step": 1597
+    },
+    {
+      "epoch": 1.7512328767123289,
+      "grad_norm": 2.231095552444458,
+      "learning_rate": 2.341592201462226e-06,
+      "loss": 1.7713,
+      "step": 1598
+    },
+    {
+      "epoch": 1.7523287671232877,
+      "grad_norm": 2.287524938583374,
+      "learning_rate": 2.3395613322502033e-06,
+      "loss": 1.8143,
+      "step": 1599
+    },
+    {
+      "epoch": 1.7534246575342465,
+      "grad_norm": 2.267552614212036,
+      "learning_rate": 2.3375304630381806e-06,
+      "loss": 1.7883,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7534246575342465,
+      "eval_loss": 1.998821496963501,
+      "eval_runtime": 37.1222,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7545205479452055,
+      "grad_norm": 2.2986714839935303,
+      "learning_rate": 2.335499593826158e-06,
+      "loss": 1.7615,
+      "step": 1601
+    },
+    {
+      "epoch": 1.7556164383561645,
+      "grad_norm": 2.2913246154785156,
+      "learning_rate": 2.333468724614135e-06,
+      "loss": 1.7584,
+      "step": 1602
+    },
+    {
+      "epoch": 1.7567123287671231,
+      "grad_norm": 2.2486486434936523,
+      "learning_rate": 2.3314378554021124e-06,
+      "loss": 1.8089,
+      "step": 1603
+    },
+    {
+      "epoch": 1.7578082191780822,
+      "grad_norm": 2.319581985473633,
+      "learning_rate": 2.3294069861900897e-06,
+      "loss": 1.8364,
+      "step": 1604
+    },
+    {
+      "epoch": 1.7589041095890412,
+      "grad_norm": 2.2775609493255615,
+      "learning_rate": 2.327376116978067e-06,
+      "loss": 1.7661,
+      "step": 1605
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 2.3445472717285156,
+      "learning_rate": 2.3253452477660442e-06,
+      "loss": 1.7583,
+      "step": 1606
+    },
+    {
+      "epoch": 1.7610958904109588,
+      "grad_norm": 2.420865058898926,
+      "learning_rate": 2.3233143785540215e-06,
+      "loss": 1.8411,
+      "step": 1607
+    },
+    {
+      "epoch": 1.7621917808219179,
+      "grad_norm": 2.294821262359619,
+      "learning_rate": 2.3212835093419988e-06,
+      "loss": 1.7812,
+      "step": 1608
+    },
+    {
+      "epoch": 1.7632876712328767,
+      "grad_norm": 2.332784414291382,
+      "learning_rate": 2.319252640129976e-06,
+      "loss": 1.8252,
+      "step": 1609
+    },
+    {
+      "epoch": 1.7643835616438355,
+      "grad_norm": 2.433619499206543,
+      "learning_rate": 2.3172217709179533e-06,
+      "loss": 1.7709,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7643835616438355,
+      "eval_loss": 2.0010287761688232,
+      "eval_runtime": 36.7858,
+      "eval_samples_per_second": 27.184,
+      "eval_steps_per_second": 1.142,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7654794520547945,
+      "grad_norm": 2.532618522644043,
+      "learning_rate": 2.3151909017059306e-06,
+      "loss": 1.8677,
+      "step": 1611
+    },
+    {
+      "epoch": 1.7665753424657535,
+      "grad_norm": 2.354578733444214,
+      "learning_rate": 2.313160032493908e-06,
+      "loss": 1.8412,
+      "step": 1612
+    },
+    {
+      "epoch": 1.7676712328767124,
+      "grad_norm": 2.382774591445923,
+      "learning_rate": 2.311129163281885e-06,
+      "loss": 1.7543,
+      "step": 1613
+    },
+    {
+      "epoch": 1.7687671232876712,
+      "grad_norm": 2.3443398475646973,
+      "learning_rate": 2.309098294069862e-06,
+      "loss": 1.8267,
+      "step": 1614
+    },
+    {
+      "epoch": 1.7698630136986302,
+      "grad_norm": 2.3804502487182617,
+      "learning_rate": 2.3070674248578393e-06,
+      "loss": 1.7582,
+      "step": 1615
+    },
+    {
+      "epoch": 1.770958904109589,
+      "grad_norm": 2.3163580894470215,
+      "learning_rate": 2.3050365556458166e-06,
+      "loss": 1.7539,
+      "step": 1616
+    },
+    {
+      "epoch": 1.7720547945205478,
+      "grad_norm": 2.605738639831543,
+      "learning_rate": 2.303005686433794e-06,
+      "loss": 1.8124,
+      "step": 1617
+    },
+    {
+      "epoch": 1.7731506849315068,
+      "grad_norm": 2.3859915733337402,
+      "learning_rate": 2.300974817221771e-06,
+      "loss": 1.8271,
+      "step": 1618
+    },
+    {
+      "epoch": 1.7742465753424659,
+      "grad_norm": 2.3715479373931885,
+      "learning_rate": 2.2989439480097484e-06,
+      "loss": 1.7823,
+      "step": 1619
+    },
+    {
+      "epoch": 1.7753424657534247,
+      "grad_norm": 2.441100597381592,
+      "learning_rate": 2.2969130787977257e-06,
+      "loss": 1.7461,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7753424657534247,
+      "eval_loss": 2.002152442932129,
+      "eval_runtime": 37.0246,
+      "eval_samples_per_second": 27.009,
+      "eval_steps_per_second": 1.134,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7764383561643835,
+      "grad_norm": 2.4154021739959717,
+      "learning_rate": 2.294882209585703e-06,
+      "loss": 1.8012,
+      "step": 1621
+    },
+    {
+      "epoch": 1.7775342465753425,
+      "grad_norm": 2.333400249481201,
+      "learning_rate": 2.2928513403736802e-06,
+      "loss": 1.8222,
+      "step": 1622
+    },
+    {
+      "epoch": 1.7786301369863013,
+      "grad_norm": 2.3183937072753906,
+      "learning_rate": 2.2908204711616575e-06,
+      "loss": 1.7561,
+      "step": 1623
+    },
+    {
+      "epoch": 1.7797260273972602,
+      "grad_norm": 2.302823066711426,
+      "learning_rate": 2.2887896019496348e-06,
+      "loss": 1.7975,
+      "step": 1624
+    },
+    {
+      "epoch": 1.7808219178082192,
+      "grad_norm": 2.3702917098999023,
+      "learning_rate": 2.286758732737612e-06,
+      "loss": 1.8478,
+      "step": 1625
+    },
+    {
+      "epoch": 1.7819178082191782,
+      "grad_norm": 2.4997549057006836,
+      "learning_rate": 2.284727863525589e-06,
+      "loss": 1.7855,
+      "step": 1626
+    },
+    {
+      "epoch": 1.783013698630137,
+      "grad_norm": 2.296191930770874,
+      "learning_rate": 2.282696994313566e-06,
+      "loss": 1.7922,
+      "step": 1627
+    },
+    {
+      "epoch": 1.7841095890410958,
+      "grad_norm": 2.2567873001098633,
+      "learning_rate": 2.2806661251015434e-06,
+      "loss": 1.8455,
+      "step": 1628
+    },
+    {
+      "epoch": 1.7852054794520549,
+      "grad_norm": 2.393336534500122,
+      "learning_rate": 2.2786352558895207e-06,
+      "loss": 1.8141,
+      "step": 1629
+    },
+    {
+      "epoch": 1.7863013698630137,
+      "grad_norm": 2.610466480255127,
+      "learning_rate": 2.276604386677498e-06,
+      "loss": 1.7875,
+      "step": 1630
+    },
+    {
+      "epoch": 1.7863013698630137,
+      "eval_loss": 1.9985637664794922,
+      "eval_runtime": 37.1001,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 1630
+    },
+    {
+      "epoch": 1.7873972602739725,
+      "grad_norm": 2.2163283824920654,
+      "learning_rate": 2.2745735174654753e-06,
+      "loss": 1.7479,
+      "step": 1631
+    },
+    {
+      "epoch": 1.7884931506849315,
+      "grad_norm": 2.2641165256500244,
+      "learning_rate": 2.2725426482534526e-06,
+      "loss": 1.7749,
+      "step": 1632
+    },
+    {
+      "epoch": 1.7895890410958906,
+      "grad_norm": 2.2786946296691895,
+      "learning_rate": 2.27051177904143e-06,
+      "loss": 1.8403,
+      "step": 1633
+    },
+    {
+      "epoch": 1.7906849315068492,
+      "grad_norm": 2.660356044769287,
+      "learning_rate": 2.268480909829407e-06,
+      "loss": 1.7994,
+      "step": 1634
+    },
+    {
+      "epoch": 1.7917808219178082,
+      "grad_norm": 2.322181224822998,
+      "learning_rate": 2.2664500406173844e-06,
+      "loss": 1.7917,
+      "step": 1635
+    },
+    {
+      "epoch": 1.7928767123287672,
+      "grad_norm": 2.3020682334899902,
+      "learning_rate": 2.2644191714053617e-06,
+      "loss": 1.7983,
+      "step": 1636
+    },
+    {
+      "epoch": 1.793972602739726,
+      "grad_norm": 2.5280368328094482,
+      "learning_rate": 2.262388302193339e-06,
+      "loss": 1.8378,
+      "step": 1637
+    },
+    {
+      "epoch": 1.7950684931506848,
+      "grad_norm": 2.3975653648376465,
+      "learning_rate": 2.260357432981316e-06,
+      "loss": 1.8081,
+      "step": 1638
+    },
+    {
+      "epoch": 1.7961643835616439,
+      "grad_norm": 2.419806957244873,
+      "learning_rate": 2.2583265637692935e-06,
+      "loss": 1.793,
+      "step": 1639
+    },
+    {
+      "epoch": 1.7972602739726027,
+      "grad_norm": 2.4601471424102783,
+      "learning_rate": 2.2562956945572708e-06,
+      "loss": 1.8116,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7972602739726027,
+      "eval_loss": 1.9951634407043457,
+      "eval_runtime": 37.1095,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7983561643835615,
+      "grad_norm": 2.17830753326416,
+      "learning_rate": 2.254264825345248e-06,
+      "loss": 1.7845,
+      "step": 1641
+    },
+    {
+      "epoch": 1.7994520547945205,
+      "grad_norm": 2.3332247734069824,
+      "learning_rate": 2.2522339561332253e-06,
+      "loss": 1.7945,
+      "step": 1642
+    },
+    {
+      "epoch": 1.8005479452054796,
+      "grad_norm": 2.3841960430145264,
+      "learning_rate": 2.2502030869212026e-06,
+      "loss": 1.7994,
+      "step": 1643
+    },
+    {
+      "epoch": 1.8016438356164384,
+      "grad_norm": 2.3669073581695557,
+      "learning_rate": 2.24817221770918e-06,
+      "loss": 1.8053,
+      "step": 1644
+    },
+    {
+      "epoch": 1.8027397260273972,
+      "grad_norm": 2.406858205795288,
+      "learning_rate": 2.246141348497157e-06,
+      "loss": 1.8647,
+      "step": 1645
+    },
+    {
+      "epoch": 1.8038356164383562,
+      "grad_norm": 2.2854161262512207,
+      "learning_rate": 2.2441104792851344e-06,
+      "loss": 1.8312,
+      "step": 1646
+    },
+    {
+      "epoch": 1.804931506849315,
+      "grad_norm": 2.252976655960083,
+      "learning_rate": 2.2420796100731117e-06,
+      "loss": 1.7834,
+      "step": 1647
+    },
+    {
+      "epoch": 1.8060273972602738,
+      "grad_norm": 2.4131624698638916,
+      "learning_rate": 2.240048740861089e-06,
+      "loss": 1.8204,
+      "step": 1648
+    },
+    {
+      "epoch": 1.8071232876712329,
+      "grad_norm": 2.4727940559387207,
+      "learning_rate": 2.2380178716490662e-06,
+      "loss": 1.8326,
+      "step": 1649
+    },
+    {
+      "epoch": 1.808219178082192,
+      "grad_norm": 2.4432709217071533,
+      "learning_rate": 2.2359870024370435e-06,
+      "loss": 1.6911,
+      "step": 1650
+    },
+    {
+      "epoch": 1.808219178082192,
+      "eval_loss": 1.996634840965271,
+      "eval_runtime": 37.072,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 1650
+    },
+    {
+      "epoch": 1.8093150684931507,
+      "grad_norm": 2.3810362815856934,
+      "learning_rate": 2.2339561332250204e-06,
+      "loss": 1.8256,
+      "step": 1651
+    },
+    {
+      "epoch": 1.8104109589041095,
+      "grad_norm": 2.4489142894744873,
+      "learning_rate": 2.2319252640129976e-06,
+      "loss": 1.8368,
+      "step": 1652
+    },
+    {
+      "epoch": 1.8115068493150686,
+      "grad_norm": 2.458686113357544,
+      "learning_rate": 2.229894394800975e-06,
+      "loss": 1.8176,
+      "step": 1653
+    },
+    {
+      "epoch": 1.8126027397260274,
+      "grad_norm": 2.3977842330932617,
+      "learning_rate": 2.227863525588952e-06,
+      "loss": 1.7516,
+      "step": 1654
+    },
+    {
+      "epoch": 1.8136986301369862,
+      "grad_norm": 2.3777663707733154,
+      "learning_rate": 2.2258326563769295e-06,
+      "loss": 1.732,
+      "step": 1655
+    },
+    {
+      "epoch": 1.8147945205479452,
+      "grad_norm": 2.2874817848205566,
+      "learning_rate": 2.2238017871649067e-06,
+      "loss": 1.749,
+      "step": 1656
+    },
+    {
+      "epoch": 1.8158904109589042,
+      "grad_norm": 2.288114547729492,
+      "learning_rate": 2.221770917952884e-06,
+      "loss": 1.7357,
+      "step": 1657
+    },
+    {
+      "epoch": 1.816986301369863,
+      "grad_norm": 2.3532187938690186,
+      "learning_rate": 2.2197400487408613e-06,
+      "loss": 1.7953,
+      "step": 1658
+    },
+    {
+      "epoch": 1.8180821917808219,
+      "grad_norm": 2.2829396724700928,
+      "learning_rate": 2.2177091795288386e-06,
+      "loss": 1.8358,
+      "step": 1659
+    },
+    {
+      "epoch": 1.819178082191781,
+      "grad_norm": 2.3282570838928223,
+      "learning_rate": 2.215678310316816e-06,
+      "loss": 1.7448,
+      "step": 1660
+    },
+    {
+      "epoch": 1.819178082191781,
+      "eval_loss": 1.999374270439148,
+      "eval_runtime": 37.0739,
+      "eval_samples_per_second": 26.973,
+      "eval_steps_per_second": 1.133,
+      "step": 1660
+    },
+    {
+      "epoch": 1.8202739726027397,
+      "grad_norm": 2.474106788635254,
+      "learning_rate": 2.213647441104793e-06,
+      "loss": 1.7979,
+      "step": 1661
+    },
+    {
+      "epoch": 1.8213698630136985,
+      "grad_norm": 2.159498453140259,
+      "learning_rate": 2.2116165718927704e-06,
+      "loss": 1.8075,
+      "step": 1662
+    },
+    {
+      "epoch": 1.8224657534246576,
+      "grad_norm": 2.2503573894500732,
+      "learning_rate": 2.2095857026807473e-06,
+      "loss": 1.7742,
+      "step": 1663
+    },
+    {
+      "epoch": 1.8235616438356166,
+      "grad_norm": 2.4465060234069824,
+      "learning_rate": 2.2075548334687245e-06,
+      "loss": 1.8067,
+      "step": 1664
+    },
+    {
+      "epoch": 1.8246575342465754,
+      "grad_norm": 2.462158679962158,
+      "learning_rate": 2.205523964256702e-06,
+      "loss": 1.7456,
+      "step": 1665
+    },
+    {
+      "epoch": 1.8257534246575342,
+      "grad_norm": 2.354051113128662,
+      "learning_rate": 2.203493095044679e-06,
+      "loss": 1.7688,
+      "step": 1666
+    },
+    {
+      "epoch": 1.8268493150684932,
+      "grad_norm": 2.3165128231048584,
+      "learning_rate": 2.2014622258326564e-06,
+      "loss": 1.7986,
+      "step": 1667
+    },
+    {
+      "epoch": 1.827945205479452,
+      "grad_norm": 2.3131113052368164,
+      "learning_rate": 2.1994313566206336e-06,
+      "loss": 1.8348,
+      "step": 1668
+    },
+    {
+      "epoch": 1.8290410958904109,
+      "grad_norm": 2.303086996078491,
+      "learning_rate": 2.197400487408611e-06,
+      "loss": 1.8413,
+      "step": 1669
+    },
+    {
+      "epoch": 1.83013698630137,
+      "grad_norm": 2.392995834350586,
+      "learning_rate": 2.195369618196588e-06,
+      "loss": 1.7663,
+      "step": 1670
+    },
+    {
+      "epoch": 1.83013698630137,
+      "eval_loss": 1.9936721324920654,
+      "eval_runtime": 37.0717,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 1670
+    },
+    {
+      "epoch": 1.831232876712329,
+      "grad_norm": 2.323596477508545,
+      "learning_rate": 2.1933387489845655e-06,
+      "loss": 1.7463,
+      "step": 1671
+    },
+    {
+      "epoch": 1.8323287671232875,
+      "grad_norm": 2.3553171157836914,
+      "learning_rate": 2.1913078797725427e-06,
+      "loss": 1.7844,
+      "step": 1672
+    },
+    {
+      "epoch": 1.8334246575342465,
+      "grad_norm": 2.3724591732025146,
+      "learning_rate": 2.18927701056052e-06,
+      "loss": 1.7955,
+      "step": 1673
+    },
+    {
+      "epoch": 1.8345205479452056,
+      "grad_norm": 2.4300284385681152,
+      "learning_rate": 2.1872461413484973e-06,
+      "loss": 1.8697,
+      "step": 1674
+    },
+    {
+      "epoch": 1.8356164383561644,
+      "grad_norm": 2.4300284385681152,
+      "learning_rate": 2.1872461413484973e-06,
+      "loss": 1.8065,
+      "step": 1675
+    },
+    {
+      "epoch": 1.8367123287671232,
+      "grad_norm": 2.523627519607544,
+      "learning_rate": 2.1852152721364746e-06,
+      "loss": 1.801,
+      "step": 1676
+    },
+    {
+      "epoch": 1.8378082191780822,
+      "grad_norm": 2.2884891033172607,
+      "learning_rate": 2.183184402924452e-06,
+      "loss": 1.7512,
+      "step": 1677
+    },
+    {
+      "epoch": 1.838904109589041,
+      "grad_norm": 2.3114986419677734,
+      "learning_rate": 2.181153533712429e-06,
+      "loss": 1.8245,
+      "step": 1678
+    },
+    {
+      "epoch": 1.8399999999999999,
+      "grad_norm": 2.326855421066284,
+      "learning_rate": 2.1791226645004064e-06,
+      "loss": 1.7921,
+      "step": 1679
+    },
+    {
+      "epoch": 1.841095890410959,
+      "grad_norm": 2.3448753356933594,
+      "learning_rate": 2.1770917952883837e-06,
+      "loss": 1.7709,
+      "step": 1680
+    },
+    {
+      "epoch": 1.841095890410959,
+      "eval_loss": 1.996195673942566,
+      "eval_runtime": 37.0481,
+      "eval_samples_per_second": 26.992,
+      "eval_steps_per_second": 1.134,
+      "step": 1680
+    },
+    {
+      "epoch": 1.842191780821918,
+      "grad_norm": 2.223470687866211,
+      "learning_rate": 2.175060926076361e-06,
+      "loss": 1.8468,
+      "step": 1681
+    },
+    {
+      "epoch": 1.8432876712328767,
+      "grad_norm": 2.29726505279541,
+      "learning_rate": 2.1730300568643382e-06,
+      "loss": 1.8753,
+      "step": 1682
+    },
+    {
+      "epoch": 1.8443835616438355,
+      "grad_norm": 2.3869571685791016,
+      "learning_rate": 2.1709991876523155e-06,
+      "loss": 1.7835,
+      "step": 1683
+    },
+    {
+      "epoch": 1.8454794520547946,
+      "grad_norm": 2.192608594894409,
+      "learning_rate": 2.1689683184402928e-06,
+      "loss": 1.7556,
+      "step": 1684
+    },
+    {
+      "epoch": 1.8465753424657534,
+      "grad_norm": 2.355975866317749,
+      "learning_rate": 2.16693744922827e-06,
+      "loss": 1.8323,
+      "step": 1685
+    },
+    {
+      "epoch": 1.8476712328767122,
+      "grad_norm": 2.243938446044922,
+      "learning_rate": 2.1649065800162473e-06,
+      "loss": 1.7942,
+      "step": 1686
+    },
+    {
+      "epoch": 1.8487671232876712,
+      "grad_norm": 2.337069511413574,
+      "learning_rate": 2.1628757108042246e-06,
+      "loss": 1.7709,
+      "step": 1687
+    },
+    {
+      "epoch": 1.8498630136986303,
+      "grad_norm": 2.257504463195801,
+      "learning_rate": 2.160844841592202e-06,
+      "loss": 1.7496,
+      "step": 1688
+    },
+    {
+      "epoch": 1.850958904109589,
+      "grad_norm": 2.287336826324463,
+      "learning_rate": 2.1588139723801787e-06,
+      "loss": 1.7454,
+      "step": 1689
+    },
+    {
+      "epoch": 1.8520547945205479,
+      "grad_norm": 2.2864668369293213,
+      "learning_rate": 2.156783103168156e-06,
+      "loss": 1.7807,
+      "step": 1690
+    },
+    {
+      "epoch": 1.8520547945205479,
+      "eval_loss": 1.9940879344940186,
+      "eval_runtime": 37.0261,
+      "eval_samples_per_second": 27.008,
+      "eval_steps_per_second": 1.134,
+      "step": 1690
+    },
+    {
+      "epoch": 1.853150684931507,
+      "grad_norm": 2.1782145500183105,
+      "learning_rate": 2.1547522339561333e-06,
+      "loss": 1.7567,
+      "step": 1691
+    },
+    {
+      "epoch": 1.8542465753424657,
+      "grad_norm": 2.252023458480835,
+      "learning_rate": 2.1527213647441106e-06,
+      "loss": 1.811,
+      "step": 1692
+    },
+    {
+      "epoch": 1.8553424657534245,
+      "grad_norm": 2.2532849311828613,
+      "learning_rate": 2.150690495532088e-06,
+      "loss": 1.8491,
+      "step": 1693
+    },
+    {
+      "epoch": 1.8564383561643836,
+      "grad_norm": 2.35648250579834,
+      "learning_rate": 2.148659626320065e-06,
+      "loss": 1.8125,
+      "step": 1694
+    },
+    {
+      "epoch": 1.8575342465753426,
+      "grad_norm": 2.3490521907806396,
+      "learning_rate": 2.1466287571080424e-06,
+      "loss": 1.8035,
+      "step": 1695
+    },
+    {
+      "epoch": 1.8586301369863014,
+      "grad_norm": 2.2851879596710205,
+      "learning_rate": 2.1445978878960197e-06,
+      "loss": 1.7755,
+      "step": 1696
+    },
+    {
+      "epoch": 1.8597260273972602,
+      "grad_norm": 2.166754961013794,
+      "learning_rate": 2.142567018683997e-06,
+      "loss": 1.7861,
+      "step": 1697
+    },
+    {
+      "epoch": 1.8608219178082193,
+      "grad_norm": 2.2356410026550293,
+      "learning_rate": 2.140536149471974e-06,
+      "loss": 1.8412,
+      "step": 1698
+    },
+    {
+      "epoch": 1.861917808219178,
+      "grad_norm": 2.215618133544922,
+      "learning_rate": 2.1385052802599515e-06,
+      "loss": 1.7623,
+      "step": 1699
+    },
+    {
+      "epoch": 1.8630136986301369,
+      "grad_norm": 2.283298969268799,
+      "learning_rate": 2.1364744110479288e-06,
+      "loss": 1.7692,
+      "step": 1700
+    },
+    {
+      "epoch": 1.8630136986301369,
+      "eval_loss": 1.9961501359939575,
+      "eval_runtime": 37.0338,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1700
+    },
+    {
+      "epoch": 1.864109589041096,
+      "grad_norm": 2.292818546295166,
+      "learning_rate": 2.134443541835906e-06,
+      "loss": 1.8251,
+      "step": 1701
+    },
+    {
+      "epoch": 1.865205479452055,
+      "grad_norm": 2.235895872116089,
+      "learning_rate": 2.132412672623883e-06,
+      "loss": 1.7726,
+      "step": 1702
+    },
+    {
+      "epoch": 1.8663013698630135,
+      "grad_norm": 2.239413022994995,
+      "learning_rate": 2.13038180341186e-06,
+      "loss": 1.7509,
+      "step": 1703
+    },
+    {
+      "epoch": 1.8673972602739726,
+      "grad_norm": 2.4125585556030273,
+      "learning_rate": 2.1283509341998374e-06,
+      "loss": 1.7973,
+      "step": 1704
+    },
+    {
+      "epoch": 1.8684931506849316,
+      "grad_norm": 2.3774733543395996,
+      "learning_rate": 2.1263200649878147e-06,
+      "loss": 1.8526,
+      "step": 1705
+    },
+    {
+      "epoch": 1.8695890410958904,
+      "grad_norm": 2.344918966293335,
+      "learning_rate": 2.124289195775792e-06,
+      "loss": 1.7708,
+      "step": 1706
+    },
+    {
+      "epoch": 1.8706849315068492,
+      "grad_norm": 2.1933865547180176,
+      "learning_rate": 2.1222583265637693e-06,
+      "loss": 1.7935,
+      "step": 1707
+    },
+    {
+      "epoch": 1.8717808219178083,
+      "grad_norm": 2.4939582347869873,
+      "learning_rate": 2.1202274573517465e-06,
+      "loss": 1.706,
+      "step": 1708
+    },
+    {
+      "epoch": 1.872876712328767,
+      "grad_norm": 2.407532215118408,
+      "learning_rate": 2.118196588139724e-06,
+      "loss": 1.7522,
+      "step": 1709
+    },
+    {
+      "epoch": 1.8739726027397259,
+      "grad_norm": 2.297900915145874,
+      "learning_rate": 2.116165718927701e-06,
+      "loss": 1.7558,
+      "step": 1710
+    },
+    {
+      "epoch": 1.8739726027397259,
+      "eval_loss": 1.9946404695510864,
+      "eval_runtime": 36.7569,
+      "eval_samples_per_second": 27.206,
+      "eval_steps_per_second": 1.143,
+      "step": 1710
+    },
+    {
+      "epoch": 1.875068493150685,
+      "grad_norm": 2.2298660278320312,
+      "learning_rate": 2.1141348497156784e-06,
+      "loss": 1.7836,
+      "step": 1711
+    },
+    {
+      "epoch": 1.876164383561644,
+      "grad_norm": 2.308060884475708,
+      "learning_rate": 2.1121039805036556e-06,
+      "loss": 1.7616,
+      "step": 1712
+    },
+    {
+      "epoch": 1.8772602739726028,
+      "grad_norm": 2.288381338119507,
+      "learning_rate": 2.110073111291633e-06,
+      "loss": 1.7502,
+      "step": 1713
+    },
+    {
+      "epoch": 1.8783561643835616,
+      "grad_norm": 2.2191097736358643,
+      "learning_rate": 2.10804224207961e-06,
+      "loss": 1.7946,
+      "step": 1714
+    },
+    {
+      "epoch": 1.8794520547945206,
+      "grad_norm": 2.211777925491333,
+      "learning_rate": 2.1060113728675875e-06,
+      "loss": 1.8238,
+      "step": 1715
+    },
+    {
+      "epoch": 1.8805479452054794,
+      "grad_norm": 2.3946406841278076,
+      "learning_rate": 2.1039805036555647e-06,
+      "loss": 1.8146,
+      "step": 1716
+    },
+    {
+      "epoch": 1.8816438356164382,
+      "grad_norm": 2.2949914932250977,
+      "learning_rate": 2.101949634443542e-06,
+      "loss": 1.6877,
+      "step": 1717
+    },
+    {
+      "epoch": 1.8827397260273973,
+      "grad_norm": 2.2905256748199463,
+      "learning_rate": 2.0999187652315193e-06,
+      "loss": 1.7868,
+      "step": 1718
+    },
+    {
+      "epoch": 1.8838356164383563,
+      "grad_norm": 2.289280891418457,
+      "learning_rate": 2.0978878960194966e-06,
+      "loss": 1.7387,
+      "step": 1719
+    },
+    {
+      "epoch": 1.884931506849315,
+      "grad_norm": 2.35497784614563,
+      "learning_rate": 2.095857026807474e-06,
+      "loss": 1.8002,
+      "step": 1720
+    },
+    {
+      "epoch": 1.884931506849315,
+      "eval_loss": 1.9928913116455078,
+      "eval_runtime": 36.9289,
+      "eval_samples_per_second": 27.079,
+      "eval_steps_per_second": 1.137,
+      "step": 1720
+    },
+    {
+      "epoch": 1.886027397260274,
+      "grad_norm": 2.4404468536376953,
+      "learning_rate": 2.093826157595451e-06,
+      "loss": 1.7981,
+      "step": 1721
+    },
+    {
+      "epoch": 1.887123287671233,
+      "grad_norm": 2.315718173980713,
+      "learning_rate": 2.0917952883834284e-06,
+      "loss": 1.7581,
+      "step": 1722
+    },
+    {
+      "epoch": 1.8882191780821918,
+      "grad_norm": 2.2984843254089355,
+      "learning_rate": 2.0897644191714057e-06,
+      "loss": 1.8153,
+      "step": 1723
+    },
+    {
+      "epoch": 1.8893150684931506,
+      "grad_norm": 2.3332557678222656,
+      "learning_rate": 2.087733549959383e-06,
+      "loss": 1.831,
+      "step": 1724
+    },
+    {
+      "epoch": 1.8904109589041096,
+      "grad_norm": 2.3394649028778076,
+      "learning_rate": 2.0857026807473602e-06,
+      "loss": 1.7486,
+      "step": 1725
+    },
+    {
+      "epoch": 1.8915068493150686,
+      "grad_norm": 2.2985246181488037,
+      "learning_rate": 2.0836718115353375e-06,
+      "loss": 1.721,
+      "step": 1726
+    },
+    {
+      "epoch": 1.8926027397260274,
+      "grad_norm": 2.318504810333252,
+      "learning_rate": 2.0816409423233144e-06,
+      "loss": 1.8469,
+      "step": 1727
+    },
+    {
+      "epoch": 1.8936986301369862,
+      "grad_norm": 2.4938831329345703,
+      "learning_rate": 2.0796100731112916e-06,
+      "loss": 1.7525,
+      "step": 1728
+    },
+    {
+      "epoch": 1.8947945205479453,
+      "grad_norm": 2.3391780853271484,
+      "learning_rate": 2.077579203899269e-06,
+      "loss": 1.7733,
+      "step": 1729
+    },
+    {
+      "epoch": 1.895890410958904,
+      "grad_norm": 2.437540054321289,
+      "learning_rate": 2.075548334687246e-06,
+      "loss": 1.7344,
+      "step": 1730
+    },
+    {
+      "epoch": 1.895890410958904,
+      "eval_loss": 1.990493655204773,
+      "eval_runtime": 37.0347,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1730
+    },
+    {
+      "epoch": 1.896986301369863,
+      "grad_norm": 2.413628578186035,
+      "learning_rate": 2.0735174654752235e-06,
+      "loss": 1.7978,
+      "step": 1731
+    },
+    {
+      "epoch": 1.898082191780822,
+      "grad_norm": 2.3673973083496094,
+      "learning_rate": 2.0714865962632007e-06,
+      "loss": 1.7414,
+      "step": 1732
+    },
+    {
+      "epoch": 1.899178082191781,
+      "grad_norm": 2.3961968421936035,
+      "learning_rate": 2.069455727051178e-06,
+      "loss": 1.7854,
+      "step": 1733
+    },
+    {
+      "epoch": 1.9002739726027398,
+      "grad_norm": 2.5583736896514893,
+      "learning_rate": 2.0674248578391553e-06,
+      "loss": 1.7629,
+      "step": 1734
+    },
+    {
+      "epoch": 1.9013698630136986,
+      "grad_norm": 2.256418228149414,
+      "learning_rate": 2.0653939886271326e-06,
+      "loss": 1.7845,
+      "step": 1735
+    },
+    {
+      "epoch": 1.9024657534246576,
+      "grad_norm": 2.4235012531280518,
+      "learning_rate": 2.06336311941511e-06,
+      "loss": 1.8151,
+      "step": 1736
+    },
+    {
+      "epoch": 1.9035616438356164,
+      "grad_norm": 2.2945358753204346,
+      "learning_rate": 2.061332250203087e-06,
+      "loss": 1.7665,
+      "step": 1737
+    },
+    {
+      "epoch": 1.9046575342465752,
+      "grad_norm": 2.371809959411621,
+      "learning_rate": 2.0593013809910644e-06,
+      "loss": 1.7355,
+      "step": 1738
+    },
+    {
+      "epoch": 1.9057534246575343,
+      "grad_norm": 2.26947021484375,
+      "learning_rate": 2.0572705117790417e-06,
+      "loss": 1.7702,
+      "step": 1739
+    },
+    {
+      "epoch": 1.9068493150684933,
+      "grad_norm": 2.314237356185913,
+      "learning_rate": 2.055239642567019e-06,
+      "loss": 1.8084,
+      "step": 1740
+    },
+    {
+      "epoch": 1.9068493150684933,
+      "eval_loss": 1.9961000680923462,
+      "eval_runtime": 37.0775,
+      "eval_samples_per_second": 26.971,
+      "eval_steps_per_second": 1.133,
+      "step": 1740
+    },
+    {
+      "epoch": 1.907945205479452,
+      "grad_norm": 2.248680591583252,
+      "learning_rate": 2.0532087733549962e-06,
+      "loss": 1.7898,
+      "step": 1741
+    },
+    {
+      "epoch": 1.909041095890411,
+      "grad_norm": 2.336378812789917,
+      "learning_rate": 2.0511779041429735e-06,
+      "loss": 1.795,
+      "step": 1742
+    },
+    {
+      "epoch": 1.91013698630137,
+      "grad_norm": 2.2420554161071777,
+      "learning_rate": 2.0491470349309508e-06,
+      "loss": 1.76,
+      "step": 1743
+    },
+    {
+      "epoch": 1.9112328767123288,
+      "grad_norm": 2.3305511474609375,
+      "learning_rate": 2.047116165718928e-06,
+      "loss": 1.7959,
+      "step": 1744
+    },
+    {
+      "epoch": 1.9123287671232876,
+      "grad_norm": 2.2677786350250244,
+      "learning_rate": 2.0450852965069053e-06,
+      "loss": 1.7922,
+      "step": 1745
+    },
+    {
+      "epoch": 1.9134246575342466,
+      "grad_norm": 2.224148988723755,
+      "learning_rate": 2.0430544272948826e-06,
+      "loss": 1.841,
+      "step": 1746
+    },
+    {
+      "epoch": 1.9145205479452054,
+      "grad_norm": 2.172091245651245,
+      "learning_rate": 2.04102355808286e-06,
+      "loss": 1.784,
+      "step": 1747
+    },
+    {
+      "epoch": 1.9156164383561642,
+      "grad_norm": 2.319962501525879,
+      "learning_rate": 2.038992688870837e-06,
+      "loss": 1.7238,
+      "step": 1748
+    },
+    {
+      "epoch": 1.9167123287671233,
+      "grad_norm": 2.299760103225708,
+      "learning_rate": 2.0369618196588144e-06,
+      "loss": 1.7501,
+      "step": 1749
+    },
+    {
+      "epoch": 1.9178082191780823,
+      "grad_norm": 2.218407154083252,
+      "learning_rate": 2.0349309504467917e-06,
+      "loss": 1.7628,
+      "step": 1750
+    },
+    {
+      "epoch": 1.9178082191780823,
+      "eval_loss": 1.9934331178665161,
+      "eval_runtime": 37.0986,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 1750
+    },
+    {
+      "epoch": 1.9189041095890411,
+      "grad_norm": 2.4121744632720947,
+      "learning_rate": 2.032900081234769e-06,
+      "loss": 1.8128,
+      "step": 1751
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 2.253225326538086,
+      "learning_rate": 2.030869212022746e-06,
+      "loss": 1.8161,
+      "step": 1752
+    },
+    {
+      "epoch": 1.921095890410959,
+      "grad_norm": 2.28702449798584,
+      "learning_rate": 2.028838342810723e-06,
+      "loss": 1.7769,
+      "step": 1753
+    },
+    {
+      "epoch": 1.9221917808219178,
+      "grad_norm": 2.3004465103149414,
+      "learning_rate": 2.0268074735987004e-06,
+      "loss": 1.7631,
+      "step": 1754
+    },
+    {
+      "epoch": 1.9232876712328766,
+      "grad_norm": 2.2733960151672363,
+      "learning_rate": 2.0247766043866777e-06,
+      "loss": 1.75,
+      "step": 1755
+    },
+    {
+      "epoch": 1.9243835616438356,
+      "grad_norm": 2.274019241333008,
+      "learning_rate": 2.022745735174655e-06,
+      "loss": 1.8317,
+      "step": 1756
+    },
+    {
+      "epoch": 1.9254794520547946,
+      "grad_norm": 2.3094100952148438,
+      "learning_rate": 2.020714865962632e-06,
+      "loss": 1.7841,
+      "step": 1757
+    },
+    {
+      "epoch": 1.9265753424657535,
+      "grad_norm": 2.269321918487549,
+      "learning_rate": 2.0186839967506095e-06,
+      "loss": 1.8052,
+      "step": 1758
+    },
+    {
+      "epoch": 1.9276712328767123,
+      "grad_norm": 2.5324459075927734,
+      "learning_rate": 2.0166531275385868e-06,
+      "loss": 1.7916,
+      "step": 1759
+    },
+    {
+      "epoch": 1.9287671232876713,
+      "grad_norm": 2.302569627761841,
+      "learning_rate": 2.014622258326564e-06,
+      "loss": 1.7499,
+      "step": 1760
+    },
+    {
+      "epoch": 1.9287671232876713,
+      "eval_loss": 1.9879337549209595,
+      "eval_runtime": 37.1186,
+      "eval_samples_per_second": 26.941,
+      "eval_steps_per_second": 1.132,
+      "step": 1760
+    },
+    {
+      "epoch": 1.9298630136986301,
+      "grad_norm": 2.29470157623291,
+      "learning_rate": 2.0125913891145413e-06,
+      "loss": 1.7922,
+      "step": 1761
+    },
+    {
+      "epoch": 1.930958904109589,
+      "grad_norm": 2.4307754039764404,
+      "learning_rate": 2.0105605199025186e-06,
+      "loss": 1.7802,
+      "step": 1762
+    },
+    {
+      "epoch": 1.932054794520548,
+      "grad_norm": 2.179588794708252,
+      "learning_rate": 2.008529650690496e-06,
+      "loss": 1.7391,
+      "step": 1763
+    },
+    {
+      "epoch": 1.933150684931507,
+      "grad_norm": 2.5824451446533203,
+      "learning_rate": 2.0064987814784727e-06,
+      "loss": 1.8106,
+      "step": 1764
+    },
+    {
+      "epoch": 1.9342465753424658,
+      "grad_norm": 2.32308292388916,
+      "learning_rate": 2.00446791226645e-06,
+      "loss": 1.7689,
+      "step": 1765
+    },
+    {
+      "epoch": 1.9353424657534246,
+      "grad_norm": 2.262533664703369,
+      "learning_rate": 2.0024370430544273e-06,
+      "loss": 1.7604,
+      "step": 1766
+    },
+    {
+      "epoch": 1.9364383561643836,
+      "grad_norm": 2.325253963470459,
+      "learning_rate": 2.0004061738424045e-06,
+      "loss": 1.7695,
+      "step": 1767
+    },
+    {
+      "epoch": 1.9375342465753425,
+      "grad_norm": 2.2855470180511475,
+      "learning_rate": 1.998375304630382e-06,
+      "loss": 1.8233,
+      "step": 1768
+    },
+    {
+      "epoch": 1.9386301369863013,
+      "grad_norm": 2.3688673973083496,
+      "learning_rate": 1.996344435418359e-06,
+      "loss": 1.7799,
+      "step": 1769
+    },
+    {
+      "epoch": 1.9397260273972603,
+      "grad_norm": 2.4682323932647705,
+      "learning_rate": 1.9943135662063364e-06,
+      "loss": 1.8036,
+      "step": 1770
+    },
+    {
+      "epoch": 1.9397260273972603,
+      "eval_loss": 1.9906389713287354,
+      "eval_runtime": 37.0959,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 1770
+    },
+    {
+      "epoch": 1.9408219178082193,
+      "grad_norm": 2.369166374206543,
+      "learning_rate": 1.9922826969943136e-06,
+      "loss": 1.7878,
+      "step": 1771
+    },
+    {
+      "epoch": 1.941917808219178,
+      "grad_norm": 2.4669666290283203,
+      "learning_rate": 1.990251827782291e-06,
+      "loss": 1.8228,
+      "step": 1772
+    },
+    {
+      "epoch": 1.943013698630137,
+      "grad_norm": 2.3083250522613525,
+      "learning_rate": 1.988220958570268e-06,
+      "loss": 1.7997,
+      "step": 1773
+    },
+    {
+      "epoch": 1.944109589041096,
+      "grad_norm": 2.4884281158447266,
+      "learning_rate": 1.9861900893582455e-06,
+      "loss": 1.8105,
+      "step": 1774
+    },
+    {
+      "epoch": 1.9452054794520548,
+      "grad_norm": 2.3504858016967773,
+      "learning_rate": 1.9841592201462227e-06,
+      "loss": 1.8159,
+      "step": 1775
+    },
+    {
+      "epoch": 1.9463013698630136,
+      "grad_norm": 2.618431329727173,
+      "learning_rate": 1.9821283509342e-06,
+      "loss": 1.7352,
+      "step": 1776
+    },
+    {
+      "epoch": 1.9473972602739726,
+      "grad_norm": 2.3452653884887695,
+      "learning_rate": 1.9800974817221773e-06,
+      "loss": 1.7451,
+      "step": 1777
+    },
+    {
+      "epoch": 1.9484931506849315,
+      "grad_norm": 2.5966849327087402,
+      "learning_rate": 1.9780666125101546e-06,
+      "loss": 1.7965,
+      "step": 1778
+    },
+    {
+      "epoch": 1.9495890410958903,
+      "grad_norm": 2.3017661571502686,
+      "learning_rate": 1.976035743298132e-06,
+      "loss": 1.7629,
+      "step": 1779
+    },
+    {
+      "epoch": 1.9506849315068493,
+      "grad_norm": 2.3766098022460938,
+      "learning_rate": 1.974004874086109e-06,
+      "loss": 1.8403,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9506849315068493,
+      "eval_loss": 1.9936374425888062,
+      "eval_runtime": 37.1158,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9517808219178083,
+      "grad_norm": 2.2327704429626465,
+      "learning_rate": 1.9719740048740864e-06,
+      "loss": 1.7096,
+      "step": 1781
+    },
+    {
+      "epoch": 1.9528767123287671,
+      "grad_norm": 2.458707809448242,
+      "learning_rate": 1.9699431356620637e-06,
+      "loss": 1.7636,
+      "step": 1782
+    },
+    {
+      "epoch": 1.953972602739726,
+      "grad_norm": 2.4143803119659424,
+      "learning_rate": 1.967912266450041e-06,
+      "loss": 1.757,
+      "step": 1783
+    },
+    {
+      "epoch": 1.955068493150685,
+      "grad_norm": 2.465148448944092,
+      "learning_rate": 1.9658813972380182e-06,
+      "loss": 1.7526,
+      "step": 1784
+    },
+    {
+      "epoch": 1.9561643835616438,
+      "grad_norm": 2.356210231781006,
+      "learning_rate": 1.9638505280259955e-06,
+      "loss": 1.8705,
+      "step": 1785
+    },
+    {
+      "epoch": 1.9572602739726026,
+      "grad_norm": 2.7456002235412598,
+      "learning_rate": 1.9618196588139728e-06,
+      "loss": 1.7574,
+      "step": 1786
+    },
+    {
+      "epoch": 1.9583561643835616,
+      "grad_norm": 2.3724262714385986,
+      "learning_rate": 1.95978878960195e-06,
+      "loss": 1.7854,
+      "step": 1787
+    },
+    {
+      "epoch": 1.9594520547945207,
+      "grad_norm": 2.275827646255493,
+      "learning_rate": 1.9577579203899273e-06,
+      "loss": 1.8447,
+      "step": 1788
+    },
+    {
+      "epoch": 1.9605479452054795,
+      "grad_norm": 2.369992256164551,
+      "learning_rate": 1.955727051177904e-06,
+      "loss": 1.8299,
+      "step": 1789
+    },
+    {
+      "epoch": 1.9616438356164383,
+      "grad_norm": 2.462372064590454,
+      "learning_rate": 1.9536961819658815e-06,
+      "loss": 1.7596,
+      "step": 1790
+    },
+    {
+      "epoch": 1.9616438356164383,
+      "eval_loss": 1.9947243928909302,
+      "eval_runtime": 37.111,
+      "eval_samples_per_second": 26.946,
+      "eval_steps_per_second": 1.132,
+      "step": 1790
+    },
+    {
+      "epoch": 1.9627397260273973,
+      "grad_norm": 2.3040034770965576,
+      "learning_rate": 1.9516653127538587e-06,
+      "loss": 1.7391,
+      "step": 1791
+    },
+    {
+      "epoch": 1.9638356164383561,
+      "grad_norm": 2.449328899383545,
+      "learning_rate": 1.949634443541836e-06,
+      "loss": 1.7428,
+      "step": 1792
+    },
+    {
+      "epoch": 1.964931506849315,
+      "grad_norm": 2.2679781913757324,
+      "learning_rate": 1.9476035743298133e-06,
+      "loss": 1.738,
+      "step": 1793
+    },
+    {
+      "epoch": 1.966027397260274,
+      "grad_norm": 2.3444032669067383,
+      "learning_rate": 1.9455727051177906e-06,
+      "loss": 1.8089,
+      "step": 1794
+    },
+    {
+      "epoch": 1.967123287671233,
+      "grad_norm": 2.3005306720733643,
+      "learning_rate": 1.943541835905768e-06,
+      "loss": 1.7718,
+      "step": 1795
+    },
+    {
+      "epoch": 1.9682191780821918,
+      "grad_norm": 2.6022629737854004,
+      "learning_rate": 1.941510966693745e-06,
+      "loss": 1.8325,
+      "step": 1796
+    },
+    {
+      "epoch": 1.9693150684931506,
+      "grad_norm": 2.407585382461548,
+      "learning_rate": 1.9394800974817224e-06,
+      "loss": 1.737,
+      "step": 1797
+    },
+    {
+      "epoch": 1.9704109589041097,
+      "grad_norm": 2.349339723587036,
+      "learning_rate": 1.9374492282696997e-06,
+      "loss": 1.81,
+      "step": 1798
+    },
+    {
+      "epoch": 1.9715068493150685,
+      "grad_norm": 2.296192169189453,
+      "learning_rate": 1.935418359057677e-06,
+      "loss": 1.8424,
+      "step": 1799
+    },
+    {
+      "epoch": 1.9726027397260273,
+      "grad_norm": 2.175875663757324,
+      "learning_rate": 1.9333874898456542e-06,
+      "loss": 1.7977,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9726027397260273,
+      "eval_loss": 1.9929841756820679,
+      "eval_runtime": 37.1159,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9736986301369863,
+      "grad_norm": 2.37428617477417,
+      "learning_rate": 1.931356620633631e-06,
+      "loss": 1.8248,
+      "step": 1801
+    },
+    {
+      "epoch": 1.9747945205479454,
+      "grad_norm": 2.297621965408325,
+      "learning_rate": 1.9293257514216083e-06,
+      "loss": 1.7215,
+      "step": 1802
+    },
+    {
+      "epoch": 1.9758904109589042,
+      "grad_norm": 2.217275857925415,
+      "learning_rate": 1.9272948822095856e-06,
+      "loss": 1.7828,
+      "step": 1803
+    },
+    {
+      "epoch": 1.976986301369863,
+      "grad_norm": 2.250854015350342,
+      "learning_rate": 1.925264012997563e-06,
+      "loss": 1.7245,
+      "step": 1804
+    },
+    {
+      "epoch": 1.978082191780822,
+      "grad_norm": 2.2556822299957275,
+      "learning_rate": 1.92323314378554e-06,
+      "loss": 1.7715,
+      "step": 1805
+    },
+    {
+      "epoch": 1.9791780821917808,
+      "grad_norm": 2.316389322280884,
+      "learning_rate": 1.9212022745735174e-06,
+      "loss": 1.8063,
+      "step": 1806
+    },
+    {
+      "epoch": 1.9802739726027396,
+      "grad_norm": 2.4358153343200684,
+      "learning_rate": 1.9191714053614947e-06,
+      "loss": 1.747,
+      "step": 1807
+    },
+    {
+      "epoch": 1.9813698630136987,
+      "grad_norm": 2.437631368637085,
+      "learning_rate": 1.917140536149472e-06,
+      "loss": 1.7485,
+      "step": 1808
+    },
+    {
+      "epoch": 1.9824657534246577,
+      "grad_norm": 2.3554601669311523,
+      "learning_rate": 1.9151096669374493e-06,
+      "loss": 1.7195,
+      "step": 1809
+    },
+    {
+      "epoch": 1.9835616438356163,
+      "grad_norm": 2.575559616088867,
+      "learning_rate": 1.9130787977254266e-06,
+      "loss": 1.8368,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9835616438356163,
+      "eval_loss": 1.9903514385223389,
+      "eval_runtime": 36.8082,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9846575342465753,
+      "grad_norm": 2.3074066638946533,
+      "learning_rate": 1.911047928513404e-06,
+      "loss": 1.7952,
+      "step": 1811
+    },
+    {
+      "epoch": 1.9857534246575343,
+      "grad_norm": 2.3162307739257812,
+      "learning_rate": 1.909017059301381e-06,
+      "loss": 1.7455,
+      "step": 1812
+    },
+    {
+      "epoch": 1.9868493150684932,
+      "grad_norm": 2.3085129261016846,
+      "learning_rate": 1.9069861900893586e-06,
+      "loss": 1.7513,
+      "step": 1813
+    },
+    {
+      "epoch": 1.987945205479452,
+      "grad_norm": 2.3454785346984863,
+      "learning_rate": 1.9049553208773354e-06,
+      "loss": 1.803,
+      "step": 1814
+    },
+    {
+      "epoch": 1.989041095890411,
+      "grad_norm": 2.4522876739501953,
+      "learning_rate": 1.9029244516653127e-06,
+      "loss": 1.7567,
+      "step": 1815
+    },
+    {
+      "epoch": 1.9901369863013698,
+      "grad_norm": 2.3876044750213623,
+      "learning_rate": 1.90089358245329e-06,
+      "loss": 1.7126,
+      "step": 1816
+    },
+    {
+      "epoch": 1.9912328767123286,
+      "grad_norm": 2.4061570167541504,
+      "learning_rate": 1.8988627132412673e-06,
+      "loss": 1.7618,
+      "step": 1817
+    },
+    {
+      "epoch": 1.9923287671232877,
+      "grad_norm": 2.307445764541626,
+      "learning_rate": 1.8968318440292445e-06,
+      "loss": 1.8014,
+      "step": 1818
+    },
+    {
+      "epoch": 1.9934246575342467,
+      "grad_norm": 2.4414806365966797,
+      "learning_rate": 1.8948009748172218e-06,
+      "loss": 1.7202,
+      "step": 1819
+    },
+    {
+      "epoch": 1.9945205479452055,
+      "grad_norm": 2.3933329582214355,
+      "learning_rate": 1.892770105605199e-06,
+      "loss": 1.8133,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9945205479452055,
+      "eval_loss": 1.991702675819397,
+      "eval_runtime": 37.0127,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9956164383561643,
+      "grad_norm": 2.237699508666992,
+      "learning_rate": 1.8907392363931764e-06,
+      "loss": 1.8026,
+      "step": 1821
+    },
+    {
+      "epoch": 1.9967123287671233,
+      "grad_norm": 2.290004014968872,
+      "learning_rate": 1.8887083671811536e-06,
+      "loss": 1.7689,
+      "step": 1822
+    },
+    {
+      "epoch": 1.9978082191780822,
+      "grad_norm": 2.2264676094055176,
+      "learning_rate": 1.886677497969131e-06,
+      "loss": 1.7862,
+      "step": 1823
+    },
+    {
+      "epoch": 1.998904109589041,
+      "grad_norm": 2.2764949798583984,
+      "learning_rate": 1.8846466287571082e-06,
+      "loss": 1.6763,
+      "step": 1824
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 2.3024826049804688,
+      "learning_rate": 1.8826157595450855e-06,
+      "loss": 1.8076,
+      "step": 1825
+    },
+    {
+      "epoch": 2.001095890410959,
+      "grad_norm": 2.362882614135742,
+      "learning_rate": 1.8805848903330625e-06,
+      "loss": 1.7009,
+      "step": 1826
+    },
+    {
+      "epoch": 2.0021917808219176,
+      "grad_norm": 2.432370185852051,
+      "learning_rate": 1.8785540211210398e-06,
+      "loss": 1.8712,
+      "step": 1827
+    },
+    {
+      "epoch": 2.0032876712328767,
+      "grad_norm": 2.3584744930267334,
+      "learning_rate": 1.876523151909017e-06,
+      "loss": 1.7869,
+      "step": 1828
+    },
+    {
+      "epoch": 2.0043835616438357,
+      "grad_norm": 2.534205198287964,
+      "learning_rate": 1.8744922826969944e-06,
+      "loss": 1.7376,
+      "step": 1829
+    },
+    {
+      "epoch": 2.0054794520547947,
+      "grad_norm": 2.58241605758667,
+      "learning_rate": 1.8724614134849716e-06,
+      "loss": 1.7204,
+      "step": 1830
+    },
+    {
+      "epoch": 2.0054794520547947,
+      "eval_loss": 1.9912636280059814,
+      "eval_runtime": 37.0969,
+      "eval_samples_per_second": 26.956,
+      "eval_steps_per_second": 1.132,
+      "step": 1830
+    },
+    {
+      "epoch": 2.0065753424657533,
+      "grad_norm": 2.3742504119873047,
+      "learning_rate": 1.870430544272949e-06,
+      "loss": 1.7537,
+      "step": 1831
+    },
+    {
+      "epoch": 2.0076712328767123,
+      "grad_norm": 2.6560916900634766,
+      "learning_rate": 1.8683996750609262e-06,
+      "loss": 1.8524,
+      "step": 1832
+    },
+    {
+      "epoch": 2.0087671232876714,
+      "grad_norm": 2.523041248321533,
+      "learning_rate": 1.8663688058489035e-06,
+      "loss": 1.7405,
+      "step": 1833
+    },
+    {
+      "epoch": 2.00986301369863,
+      "grad_norm": 2.5444304943084717,
+      "learning_rate": 1.8643379366368807e-06,
+      "loss": 1.7588,
+      "step": 1834
+    },
+    {
+      "epoch": 2.010958904109589,
+      "grad_norm": 2.708984613418579,
+      "learning_rate": 1.862307067424858e-06,
+      "loss": 1.7686,
+      "step": 1835
+    },
+    {
+      "epoch": 2.012054794520548,
+      "grad_norm": 2.726691961288452,
+      "learning_rate": 1.8602761982128353e-06,
+      "loss": 1.7733,
+      "step": 1836
+    },
+    {
+      "epoch": 2.013150684931507,
+      "grad_norm": 2.5085904598236084,
+      "learning_rate": 1.8582453290008126e-06,
+      "loss": 1.81,
+      "step": 1837
+    },
+    {
+      "epoch": 2.0142465753424657,
+      "grad_norm": 2.4699466228485107,
+      "learning_rate": 1.8562144597887898e-06,
+      "loss": 1.7369,
+      "step": 1838
+    },
+    {
+      "epoch": 2.0153424657534247,
+      "grad_norm": 2.531027317047119,
+      "learning_rate": 1.854183590576767e-06,
+      "loss": 1.7611,
+      "step": 1839
+    },
+    {
+      "epoch": 2.0164383561643837,
+      "grad_norm": 2.5061161518096924,
+      "learning_rate": 1.8521527213647442e-06,
+      "loss": 1.811,
+      "step": 1840
+    },
+    {
+      "epoch": 2.0164383561643837,
+      "eval_loss": 1.9893144369125366,
+      "eval_runtime": 37.1262,
+      "eval_samples_per_second": 26.935,
+      "eval_steps_per_second": 1.131,
+      "step": 1840
+    },
+    {
+      "epoch": 2.0175342465753423,
+      "grad_norm": 2.5739994049072266,
+      "learning_rate": 1.8501218521527215e-06,
+      "loss": 1.7153,
+      "step": 1841
+    },
+    {
+      "epoch": 2.0186301369863013,
+      "grad_norm": 2.5673015117645264,
+      "learning_rate": 1.8480909829406987e-06,
+      "loss": 1.6879,
+      "step": 1842
+    },
+    {
+      "epoch": 2.0197260273972604,
+      "grad_norm": 2.717430353164673,
+      "learning_rate": 1.846060113728676e-06,
+      "loss": 1.723,
+      "step": 1843
+    },
+    {
+      "epoch": 2.020821917808219,
+      "grad_norm": 2.4905285835266113,
+      "learning_rate": 1.8440292445166533e-06,
+      "loss": 1.7651,
+      "step": 1844
+    },
+    {
+      "epoch": 2.021917808219178,
+      "grad_norm": 2.6798791885375977,
+      "learning_rate": 1.8419983753046306e-06,
+      "loss": 1.7308,
+      "step": 1845
+    },
+    {
+      "epoch": 2.023013698630137,
+      "grad_norm": 2.591614007949829,
+      "learning_rate": 1.8399675060926078e-06,
+      "loss": 1.786,
+      "step": 1846
+    },
+    {
+      "epoch": 2.024109589041096,
+      "grad_norm": 2.5405824184417725,
+      "learning_rate": 1.8379366368805851e-06,
+      "loss": 1.7699,
+      "step": 1847
+    },
+    {
+      "epoch": 2.0252054794520546,
+      "grad_norm": 2.4189934730529785,
+      "learning_rate": 1.8359057676685624e-06,
+      "loss": 1.732,
+      "step": 1848
+    },
+    {
+      "epoch": 2.0263013698630137,
+      "grad_norm": 2.5857982635498047,
+      "learning_rate": 1.8338748984565397e-06,
+      "loss": 1.7123,
+      "step": 1849
+    },
+    {
+      "epoch": 2.0273972602739727,
+      "grad_norm": 2.4710893630981445,
+      "learning_rate": 1.831844029244517e-06,
+      "loss": 1.7419,
+      "step": 1850
+    },
+    {
+      "epoch": 2.0273972602739727,
+      "eval_loss": 1.9874999523162842,
+      "eval_runtime": 37.1588,
+      "eval_samples_per_second": 26.912,
+      "eval_steps_per_second": 1.13,
+      "step": 1850
+    },
+    {
+      "epoch": 2.0284931506849313,
+      "grad_norm": 2.3414645195007324,
+      "learning_rate": 1.829813160032494e-06,
+      "loss": 1.7102,
+      "step": 1851
+    },
+    {
+      "epoch": 2.0295890410958903,
+      "grad_norm": 2.4148900508880615,
+      "learning_rate": 1.8277822908204713e-06,
+      "loss": 1.7287,
+      "step": 1852
+    },
+    {
+      "epoch": 2.0306849315068494,
+      "grad_norm": 2.5501346588134766,
+      "learning_rate": 1.8257514216084486e-06,
+      "loss": 1.7633,
+      "step": 1853
+    },
+    {
+      "epoch": 2.0317808219178084,
+      "grad_norm": 2.52215576171875,
+      "learning_rate": 1.8237205523964258e-06,
+      "loss": 1.7385,
+      "step": 1854
+    },
+    {
+      "epoch": 2.032876712328767,
+      "grad_norm": 2.433335065841675,
+      "learning_rate": 1.8216896831844031e-06,
+      "loss": 1.7566,
+      "step": 1855
+    },
+    {
+      "epoch": 2.033972602739726,
+      "grad_norm": 2.441232442855835,
+      "learning_rate": 1.8196588139723804e-06,
+      "loss": 1.8027,
+      "step": 1856
+    },
+    {
+      "epoch": 2.035068493150685,
+      "grad_norm": 2.4601082801818848,
+      "learning_rate": 1.8176279447603577e-06,
+      "loss": 1.6859,
+      "step": 1857
+    },
+    {
+      "epoch": 2.0361643835616436,
+      "grad_norm": 2.5008108615875244,
+      "learning_rate": 1.815597075548335e-06,
+      "loss": 1.7294,
+      "step": 1858
+    },
+    {
+      "epoch": 2.0372602739726027,
+      "grad_norm": 2.433915615081787,
+      "learning_rate": 1.8135662063363122e-06,
+      "loss": 1.755,
+      "step": 1859
+    },
+    {
+      "epoch": 2.0383561643835617,
+      "grad_norm": 2.4938697814941406,
+      "learning_rate": 1.8115353371242895e-06,
+      "loss": 1.7904,
+      "step": 1860
+    },
+    {
+      "epoch": 2.0383561643835617,
+      "eval_loss": 1.9939266443252563,
+      "eval_runtime": 37.1539,
+      "eval_samples_per_second": 26.915,
+      "eval_steps_per_second": 1.13,
+      "step": 1860
+    },
+    {
+      "epoch": 2.0394520547945207,
+      "grad_norm": 2.382324695587158,
+      "learning_rate": 1.8095044679122668e-06,
+      "loss": 1.6856,
+      "step": 1861
+    },
+    {
+      "epoch": 2.0405479452054793,
+      "grad_norm": 2.5250437259674072,
+      "learning_rate": 1.807473598700244e-06,
+      "loss": 1.7434,
+      "step": 1862
+    },
+    {
+      "epoch": 2.0416438356164384,
+      "grad_norm": 2.63969087600708,
+      "learning_rate": 1.805442729488221e-06,
+      "loss": 1.7343,
+      "step": 1863
+    },
+    {
+      "epoch": 2.0427397260273974,
+      "grad_norm": 2.4852068424224854,
+      "learning_rate": 1.8034118602761982e-06,
+      "loss": 1.7494,
+      "step": 1864
+    },
+    {
+      "epoch": 2.043835616438356,
+      "grad_norm": 2.393967390060425,
+      "learning_rate": 1.8013809910641754e-06,
+      "loss": 1.7503,
+      "step": 1865
+    },
+    {
+      "epoch": 2.044931506849315,
+      "grad_norm": 2.4447648525238037,
+      "learning_rate": 1.7993501218521527e-06,
+      "loss": 1.7539,
+      "step": 1866
+    },
+    {
+      "epoch": 2.046027397260274,
+      "grad_norm": 2.483530044555664,
+      "learning_rate": 1.79731925264013e-06,
+      "loss": 1.7001,
+      "step": 1867
+    },
+    {
+      "epoch": 2.047123287671233,
+      "grad_norm": 2.56669282913208,
+      "learning_rate": 1.7952883834281073e-06,
+      "loss": 1.7828,
+      "step": 1868
+    },
+    {
+      "epoch": 2.0482191780821917,
+      "grad_norm": 2.4894964694976807,
+      "learning_rate": 1.7932575142160846e-06,
+      "loss": 1.7208,
+      "step": 1869
+    },
+    {
+      "epoch": 2.0493150684931507,
+      "grad_norm": 2.4713854789733887,
+      "learning_rate": 1.7912266450040618e-06,
+      "loss": 1.7687,
+      "step": 1870
+    },
+    {
+      "epoch": 2.0493150684931507,
+      "eval_loss": 1.9938212633132935,
+      "eval_runtime": 37.1298,
+      "eval_samples_per_second": 26.933,
+      "eval_steps_per_second": 1.131,
+      "step": 1870
+    },
+    {
+      "epoch": 2.0504109589041097,
+      "grad_norm": 2.6204898357391357,
+      "learning_rate": 1.789195775792039e-06,
+      "loss": 1.7718,
+      "step": 1871
+    },
+    {
+      "epoch": 2.0515068493150683,
+      "grad_norm": 2.5194199085235596,
+      "learning_rate": 1.7871649065800164e-06,
+      "loss": 1.6653,
+      "step": 1872
+    },
+    {
+      "epoch": 2.0526027397260274,
+      "grad_norm": 2.5077965259552,
+      "learning_rate": 1.7851340373679937e-06,
+      "loss": 1.8098,
+      "step": 1873
+    },
+    {
+      "epoch": 2.0536986301369864,
+      "grad_norm": 2.5841965675354004,
+      "learning_rate": 1.783103168155971e-06,
+      "loss": 1.7126,
+      "step": 1874
+    },
+    {
+      "epoch": 2.0547945205479454,
+      "grad_norm": 2.617910623550415,
+      "learning_rate": 1.7810722989439482e-06,
+      "loss": 1.7294,
+      "step": 1875
+    },
+    {
+      "epoch": 2.055890410958904,
+      "grad_norm": 2.519253969192505,
+      "learning_rate": 1.7790414297319253e-06,
+      "loss": 1.7497,
+      "step": 1876
+    },
+    {
+      "epoch": 2.056986301369863,
+      "grad_norm": 2.5518412590026855,
+      "learning_rate": 1.7770105605199025e-06,
+      "loss": 1.7188,
+      "step": 1877
+    },
+    {
+      "epoch": 2.058082191780822,
+      "grad_norm": 2.5372824668884277,
+      "learning_rate": 1.7749796913078798e-06,
+      "loss": 1.6998,
+      "step": 1878
+    },
+    {
+      "epoch": 2.0591780821917807,
+      "grad_norm": 2.6649036407470703,
+      "learning_rate": 1.772948822095857e-06,
+      "loss": 1.7621,
+      "step": 1879
+    },
+    {
+      "epoch": 2.0602739726027397,
+      "grad_norm": 2.5216212272644043,
+      "learning_rate": 1.7709179528838344e-06,
+      "loss": 1.7963,
+      "step": 1880
+    },
+    {
+      "epoch": 2.0602739726027397,
+      "eval_loss": 1.993401050567627,
+      "eval_runtime": 37.0954,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 1880
+    },
+    {
+      "epoch": 2.0613698630136987,
+      "grad_norm": 2.5226478576660156,
+      "learning_rate": 1.7688870836718116e-06,
+      "loss": 1.7841,
+      "step": 1881
+    },
+    {
+      "epoch": 2.0624657534246573,
+      "grad_norm": 2.584319829940796,
+      "learning_rate": 1.766856214459789e-06,
+      "loss": 1.7449,
+      "step": 1882
+    },
+    {
+      "epoch": 2.0635616438356164,
+      "grad_norm": 2.5305304527282715,
+      "learning_rate": 1.7648253452477662e-06,
+      "loss": 1.7373,
+      "step": 1883
+    },
+    {
+      "epoch": 2.0646575342465754,
+      "grad_norm": 2.48797607421875,
+      "learning_rate": 1.7627944760357435e-06,
+      "loss": 1.722,
+      "step": 1884
+    },
+    {
+      "epoch": 2.0657534246575344,
+      "grad_norm": 2.4461028575897217,
+      "learning_rate": 1.7607636068237208e-06,
+      "loss": 1.7983,
+      "step": 1885
+    },
+    {
+      "epoch": 2.066849315068493,
+      "grad_norm": 2.624058961868286,
+      "learning_rate": 1.758732737611698e-06,
+      "loss": 1.7709,
+      "step": 1886
+    },
+    {
+      "epoch": 2.067945205479452,
+      "grad_norm": 2.848062515258789,
+      "learning_rate": 1.7567018683996753e-06,
+      "loss": 1.7513,
+      "step": 1887
+    },
+    {
+      "epoch": 2.069041095890411,
+      "grad_norm": 2.5219593048095703,
+      "learning_rate": 1.7546709991876524e-06,
+      "loss": 1.7387,
+      "step": 1888
+    },
+    {
+      "epoch": 2.0701369863013697,
+      "grad_norm": 2.5296876430511475,
+      "learning_rate": 1.7526401299756296e-06,
+      "loss": 1.7378,
+      "step": 1889
+    },
+    {
+      "epoch": 2.0712328767123287,
+      "grad_norm": 2.5452613830566406,
+      "learning_rate": 1.750609260763607e-06,
+      "loss": 1.7402,
+      "step": 1890
+    },
+    {
+      "epoch": 2.0712328767123287,
+      "eval_loss": 1.9914054870605469,
+      "eval_runtime": 37.0548,
+      "eval_samples_per_second": 26.987,
+      "eval_steps_per_second": 1.133,
+      "step": 1890
+    },
+    {
+      "epoch": 2.0723287671232877,
+      "grad_norm": 2.653596878051758,
+      "learning_rate": 1.7485783915515842e-06,
+      "loss": 1.7369,
+      "step": 1891
+    },
+    {
+      "epoch": 2.0734246575342468,
+      "grad_norm": 2.615581750869751,
+      "learning_rate": 1.7465475223395615e-06,
+      "loss": 1.7317,
+      "step": 1892
+    },
+    {
+      "epoch": 2.0745205479452054,
+      "grad_norm": 2.610304832458496,
+      "learning_rate": 1.7445166531275387e-06,
+      "loss": 1.7121,
+      "step": 1893
+    },
+    {
+      "epoch": 2.0756164383561644,
+      "grad_norm": 2.6463494300842285,
+      "learning_rate": 1.742485783915516e-06,
+      "loss": 1.8523,
+      "step": 1894
+    },
+    {
+      "epoch": 2.0767123287671234,
+      "grad_norm": 2.4814131259918213,
+      "learning_rate": 1.7404549147034933e-06,
+      "loss": 1.8239,
+      "step": 1895
+    },
+    {
+      "epoch": 2.077808219178082,
+      "grad_norm": 2.7416090965270996,
+      "learning_rate": 1.7384240454914706e-06,
+      "loss": 1.669,
+      "step": 1896
+    },
+    {
+      "epoch": 2.078904109589041,
+      "grad_norm": 2.9458305835723877,
+      "learning_rate": 1.7363931762794478e-06,
+      "loss": 1.7504,
+      "step": 1897
+    },
+    {
+      "epoch": 2.08,
+      "grad_norm": 2.5118932723999023,
+      "learning_rate": 1.7343623070674251e-06,
+      "loss": 1.7854,
+      "step": 1898
+    },
+    {
+      "epoch": 2.081095890410959,
+      "grad_norm": 2.479421615600586,
+      "learning_rate": 1.7323314378554024e-06,
+      "loss": 1.7325,
+      "step": 1899
+    },
+    {
+      "epoch": 2.0821917808219177,
+      "grad_norm": 2.7400424480438232,
+      "learning_rate": 1.7303005686433797e-06,
+      "loss": 1.7434,
+      "step": 1900
+    },
+    {
+      "epoch": 2.0821917808219177,
+      "eval_loss": 1.9887213706970215,
+      "eval_runtime": 37.0484,
+      "eval_samples_per_second": 26.992,
+      "eval_steps_per_second": 1.134,
+      "step": 1900
+    },
+    {
+      "epoch": 2.0832876712328767,
+      "grad_norm": 2.587714433670044,
+      "learning_rate": 1.7282696994313567e-06,
+      "loss": 1.7335,
+      "step": 1901
+    },
+    {
+      "epoch": 2.0843835616438358,
+      "grad_norm": 2.526818037033081,
+      "learning_rate": 1.726238830219334e-06,
+      "loss": 1.7538,
+      "step": 1902
+    },
+    {
+      "epoch": 2.0854794520547943,
+      "grad_norm": 2.595221519470215,
+      "learning_rate": 1.7242079610073113e-06,
+      "loss": 1.7331,
+      "step": 1903
+    },
+    {
+      "epoch": 2.0865753424657534,
+      "grad_norm": 2.5018110275268555,
+      "learning_rate": 1.7221770917952886e-06,
+      "loss": 1.8069,
+      "step": 1904
+    },
+    {
+      "epoch": 2.0876712328767124,
+      "grad_norm": 2.4378256797790527,
+      "learning_rate": 1.7201462225832658e-06,
+      "loss": 1.757,
+      "step": 1905
+    },
+    {
+      "epoch": 2.0887671232876714,
+      "grad_norm": 2.651167392730713,
+      "learning_rate": 1.7181153533712431e-06,
+      "loss": 1.7696,
+      "step": 1906
+    },
+    {
+      "epoch": 2.08986301369863,
+      "grad_norm": 2.550239086151123,
+      "learning_rate": 1.7160844841592204e-06,
+      "loss": 1.6551,
+      "step": 1907
+    },
+    {
+      "epoch": 2.090958904109589,
+      "grad_norm": 2.4192144870758057,
+      "learning_rate": 1.7140536149471977e-06,
+      "loss": 1.7031,
+      "step": 1908
+    },
+    {
+      "epoch": 2.092054794520548,
+      "grad_norm": 2.5327398777008057,
+      "learning_rate": 1.712022745735175e-06,
+      "loss": 1.734,
+      "step": 1909
+    },
+    {
+      "epoch": 2.0931506849315067,
+      "grad_norm": 2.5962865352630615,
+      "learning_rate": 1.7099918765231522e-06,
+      "loss": 1.732,
+      "step": 1910
+    },
+    {
+      "epoch": 2.0931506849315067,
+      "eval_loss": 1.9915482997894287,
+      "eval_runtime": 36.7856,
+      "eval_samples_per_second": 27.185,
+      "eval_steps_per_second": 1.142,
+      "step": 1910
+    },
+    {
+      "epoch": 2.0942465753424657,
+      "grad_norm": 2.561943531036377,
+      "learning_rate": 1.7079610073111295e-06,
+      "loss": 1.7556,
+      "step": 1911
+    },
+    {
+      "epoch": 2.0953424657534248,
+      "grad_norm": 2.529829263687134,
+      "learning_rate": 1.7059301380991068e-06,
+      "loss": 1.6839,
+      "step": 1912
+    },
+    {
+      "epoch": 2.0964383561643833,
+      "grad_norm": 2.636559009552002,
+      "learning_rate": 1.7038992688870836e-06,
+      "loss": 1.7393,
+      "step": 1913
+    },
+    {
+      "epoch": 2.0975342465753424,
+      "grad_norm": 2.558166742324829,
+      "learning_rate": 1.701868399675061e-06,
+      "loss": 1.7346,
+      "step": 1914
+    },
+    {
+      "epoch": 2.0986301369863014,
+      "grad_norm": 2.5193750858306885,
+      "learning_rate": 1.6998375304630382e-06,
+      "loss": 1.7666,
+      "step": 1915
+    },
+    {
+      "epoch": 2.0997260273972604,
+      "grad_norm": 2.498244047164917,
+      "learning_rate": 1.6978066612510155e-06,
+      "loss": 1.827,
+      "step": 1916
+    },
+    {
+      "epoch": 2.100821917808219,
+      "grad_norm": 2.799603223800659,
+      "learning_rate": 1.6957757920389927e-06,
+      "loss": 1.788,
+      "step": 1917
+    },
+    {
+      "epoch": 2.101917808219178,
+      "grad_norm": 2.733391761779785,
+      "learning_rate": 1.69374492282697e-06,
+      "loss": 1.7487,
+      "step": 1918
+    },
+    {
+      "epoch": 2.103013698630137,
+      "grad_norm": 2.5135183334350586,
+      "learning_rate": 1.6917140536149473e-06,
+      "loss": 1.7114,
+      "step": 1919
+    },
+    {
+      "epoch": 2.1041095890410957,
+      "grad_norm": 2.5275800228118896,
+      "learning_rate": 1.6896831844029246e-06,
+      "loss": 1.797,
+      "step": 1920
+    },
+    {
+      "epoch": 2.1041095890410957,
+      "eval_loss": 1.9926362037658691,
+      "eval_runtime": 36.9326,
+      "eval_samples_per_second": 27.076,
+      "eval_steps_per_second": 1.137,
+      "step": 1920
+    },
+    {
+      "epoch": 2.1052054794520547,
+      "grad_norm": 2.682774782180786,
+      "learning_rate": 1.6876523151909018e-06,
+      "loss": 1.7581,
+      "step": 1921
+    },
+    {
+      "epoch": 2.1063013698630138,
+      "grad_norm": 2.625620126724243,
+      "learning_rate": 1.6856214459788791e-06,
+      "loss": 1.7606,
+      "step": 1922
+    },
+    {
+      "epoch": 2.107397260273973,
+      "grad_norm": 2.5005173683166504,
+      "learning_rate": 1.6835905767668564e-06,
+      "loss": 1.7139,
+      "step": 1923
+    },
+    {
+      "epoch": 2.1084931506849314,
+      "grad_norm": 2.6731526851654053,
+      "learning_rate": 1.6815597075548337e-06,
+      "loss": 1.7174,
+      "step": 1924
+    },
+    {
+      "epoch": 2.1095890410958904,
+      "grad_norm": 2.5407724380493164,
+      "learning_rate": 1.679528838342811e-06,
+      "loss": 1.8049,
+      "step": 1925
+    },
+    {
+      "epoch": 2.1106849315068494,
+      "grad_norm": 2.502427101135254,
+      "learning_rate": 1.677497969130788e-06,
+      "loss": 1.7068,
+      "step": 1926
+    },
+    {
+      "epoch": 2.111780821917808,
+      "grad_norm": 2.5063414573669434,
+      "learning_rate": 1.6754670999187653e-06,
+      "loss": 1.7174,
+      "step": 1927
+    },
+    {
+      "epoch": 2.112876712328767,
+      "grad_norm": 2.524588108062744,
+      "learning_rate": 1.6734362307067426e-06,
+      "loss": 1.7442,
+      "step": 1928
+    },
+    {
+      "epoch": 2.113972602739726,
+      "grad_norm": 2.531909465789795,
+      "learning_rate": 1.6714053614947198e-06,
+      "loss": 1.692,
+      "step": 1929
+    },
+    {
+      "epoch": 2.115068493150685,
+      "grad_norm": 2.5104269981384277,
+      "learning_rate": 1.669374492282697e-06,
+      "loss": 1.7039,
+      "step": 1930
+    },
+    {
+      "epoch": 2.115068493150685,
+      "eval_loss": 1.9927221536636353,
+      "eval_runtime": 37.004,
+      "eval_samples_per_second": 27.024,
+      "eval_steps_per_second": 1.135,
+      "step": 1930
+    },
+    {
+      "epoch": 2.1161643835616437,
+      "grad_norm": 2.5443670749664307,
+      "learning_rate": 1.6673436230706744e-06,
+      "loss": 1.7896,
+      "step": 1931
+    },
+    {
+      "epoch": 2.1172602739726027,
+      "grad_norm": 2.6495745182037354,
+      "learning_rate": 1.6653127538586517e-06,
+      "loss": 1.7735,
+      "step": 1932
+    },
+    {
+      "epoch": 2.118356164383562,
+      "grad_norm": 2.7233121395111084,
+      "learning_rate": 1.663281884646629e-06,
+      "loss": 1.7186,
+      "step": 1933
+    },
+    {
+      "epoch": 2.1194520547945204,
+      "grad_norm": 2.4411816596984863,
+      "learning_rate": 1.6612510154346062e-06,
+      "loss": 1.7378,
+      "step": 1934
+    },
+    {
+      "epoch": 2.1205479452054794,
+      "grad_norm": 2.6478421688079834,
+      "learning_rate": 1.6592201462225835e-06,
+      "loss": 1.7989,
+      "step": 1935
+    },
+    {
+      "epoch": 2.1216438356164384,
+      "grad_norm": 2.56093168258667,
+      "learning_rate": 1.6571892770105608e-06,
+      "loss": 1.7663,
+      "step": 1936
+    },
+    {
+      "epoch": 2.1227397260273975,
+      "grad_norm": 2.5598859786987305,
+      "learning_rate": 1.655158407798538e-06,
+      "loss": 1.7956,
+      "step": 1937
+    },
+    {
+      "epoch": 2.123835616438356,
+      "grad_norm": 2.4958112239837646,
+      "learning_rate": 1.653127538586515e-06,
+      "loss": 1.7482,
+      "step": 1938
+    },
+    {
+      "epoch": 2.124931506849315,
+      "grad_norm": 2.5133280754089355,
+      "learning_rate": 1.6510966693744924e-06,
+      "loss": 1.7811,
+      "step": 1939
+    },
+    {
+      "epoch": 2.126027397260274,
+      "grad_norm": 2.492892026901245,
+      "learning_rate": 1.6490658001624696e-06,
+      "loss": 1.7883,
+      "step": 1940
+    },
+    {
+      "epoch": 2.126027397260274,
+      "eval_loss": 1.9917834997177124,
+      "eval_runtime": 37.0222,
+      "eval_samples_per_second": 27.011,
+      "eval_steps_per_second": 1.134,
+      "step": 1940
+    },
+    {
+      "epoch": 2.1271232876712327,
+      "grad_norm": 2.560483455657959,
+      "learning_rate": 1.647034930950447e-06,
+      "loss": 1.7871,
+      "step": 1941
+    },
+    {
+      "epoch": 2.1282191780821917,
+      "grad_norm": 2.549100160598755,
+      "learning_rate": 1.6450040617384242e-06,
+      "loss": 1.7097,
+      "step": 1942
+    },
+    {
+      "epoch": 2.1293150684931508,
+      "grad_norm": 2.4832775592803955,
+      "learning_rate": 1.6429731925264015e-06,
+      "loss": 1.736,
+      "step": 1943
+    },
+    {
+      "epoch": 2.1304109589041094,
+      "grad_norm": 2.60006046295166,
+      "learning_rate": 1.6409423233143788e-06,
+      "loss": 1.765,
+      "step": 1944
+    },
+    {
+      "epoch": 2.1315068493150684,
+      "grad_norm": 2.4875144958496094,
+      "learning_rate": 1.638911454102356e-06,
+      "loss": 1.7317,
+      "step": 1945
+    },
+    {
+      "epoch": 2.1326027397260274,
+      "grad_norm": 2.5562970638275146,
+      "learning_rate": 1.6368805848903333e-06,
+      "loss": 1.6982,
+      "step": 1946
+    },
+    {
+      "epoch": 2.1336986301369865,
+      "grad_norm": 2.4928972721099854,
+      "learning_rate": 1.6348497156783106e-06,
+      "loss": 1.7821,
+      "step": 1947
+    },
+    {
+      "epoch": 2.134794520547945,
+      "grad_norm": 2.6159043312072754,
+      "learning_rate": 1.6328188464662879e-06,
+      "loss": 1.7223,
+      "step": 1948
+    },
+    {
+      "epoch": 2.135890410958904,
+      "grad_norm": 2.541520833969116,
+      "learning_rate": 1.6307879772542651e-06,
+      "loss": 1.7321,
+      "step": 1949
+    },
+    {
+      "epoch": 2.136986301369863,
+      "grad_norm": 2.5165696144104004,
+      "learning_rate": 1.628757108042242e-06,
+      "loss": 1.7571,
+      "step": 1950
+    },
+    {
+      "epoch": 2.136986301369863,
+      "eval_loss": 1.9868476390838623,
+      "eval_runtime": 37.0367,
+      "eval_samples_per_second": 27.0,
+      "eval_steps_per_second": 1.134,
+      "step": 1950
+    },
+    {
+      "epoch": 2.138082191780822,
+      "grad_norm": 2.5660274028778076,
+      "learning_rate": 1.6267262388302193e-06,
+      "loss": 1.7855,
+      "step": 1951
+    },
+    {
+      "epoch": 2.1391780821917807,
+      "grad_norm": 2.4606971740722656,
+      "learning_rate": 1.6246953696181965e-06,
+      "loss": 1.7324,
+      "step": 1952
+    },
+    {
+      "epoch": 2.1402739726027398,
+      "grad_norm": 2.6528608798980713,
+      "learning_rate": 1.6226645004061738e-06,
+      "loss": 1.7715,
+      "step": 1953
+    },
+    {
+      "epoch": 2.141369863013699,
+      "grad_norm": 2.7232985496520996,
+      "learning_rate": 1.620633631194151e-06,
+      "loss": 1.837,
+      "step": 1954
+    },
+    {
+      "epoch": 2.1424657534246574,
+      "grad_norm": 2.6264185905456543,
+      "learning_rate": 1.6186027619821284e-06,
+      "loss": 1.7678,
+      "step": 1955
+    },
+    {
+      "epoch": 2.1435616438356164,
+      "grad_norm": 2.6248252391815186,
+      "learning_rate": 1.6165718927701056e-06,
+      "loss": 1.7218,
+      "step": 1956
+    },
+    {
+      "epoch": 2.1446575342465755,
+      "grad_norm": 3.00539493560791,
+      "learning_rate": 1.614541023558083e-06,
+      "loss": 1.7678,
+      "step": 1957
+    },
+    {
+      "epoch": 2.145753424657534,
+      "grad_norm": 2.5368528366088867,
+      "learning_rate": 1.6125101543460602e-06,
+      "loss": 1.7737,
+      "step": 1958
+    },
+    {
+      "epoch": 2.146849315068493,
+      "grad_norm": 2.5805916786193848,
+      "learning_rate": 1.6104792851340375e-06,
+      "loss": 1.7394,
+      "step": 1959
+    },
+    {
+      "epoch": 2.147945205479452,
+      "grad_norm": 2.615856647491455,
+      "learning_rate": 1.6084484159220147e-06,
+      "loss": 1.766,
+      "step": 1960
+    },
+    {
+      "epoch": 2.147945205479452,
+      "eval_loss": 1.987140417098999,
+      "eval_runtime": 37.0593,
+      "eval_samples_per_second": 26.984,
+      "eval_steps_per_second": 1.133,
+      "step": 1960
+    },
+    {
+      "epoch": 2.149041095890411,
+      "grad_norm": 2.6550803184509277,
+      "learning_rate": 1.606417546709992e-06,
+      "loss": 1.7342,
+      "step": 1961
+    },
+    {
+      "epoch": 2.1501369863013697,
+      "grad_norm": 2.5999467372894287,
+      "learning_rate": 1.6043866774979693e-06,
+      "loss": 1.7175,
+      "step": 1962
+    },
+    {
+      "epoch": 2.1512328767123288,
+      "grad_norm": 2.561642646789551,
+      "learning_rate": 1.6023558082859464e-06,
+      "loss": 1.7409,
+      "step": 1963
+    },
+    {
+      "epoch": 2.152328767123288,
+      "grad_norm": 2.6062211990356445,
+      "learning_rate": 1.6003249390739236e-06,
+      "loss": 1.7052,
+      "step": 1964
+    },
+    {
+      "epoch": 2.1534246575342464,
+      "grad_norm": 2.5122230052948,
+      "learning_rate": 1.598294069861901e-06,
+      "loss": 1.7444,
+      "step": 1965
+    },
+    {
+      "epoch": 2.1545205479452054,
+      "grad_norm": 2.5838866233825684,
+      "learning_rate": 1.5962632006498782e-06,
+      "loss": 1.716,
+      "step": 1966
+    },
+    {
+      "epoch": 2.1556164383561645,
+      "grad_norm": 2.5454747676849365,
+      "learning_rate": 1.5942323314378555e-06,
+      "loss": 1.7902,
+      "step": 1967
+    },
+    {
+      "epoch": 2.1567123287671235,
+      "grad_norm": 2.57637619972229,
+      "learning_rate": 1.5922014622258327e-06,
+      "loss": 1.6963,
+      "step": 1968
+    },
+    {
+      "epoch": 2.157808219178082,
+      "grad_norm": 2.5542941093444824,
+      "learning_rate": 1.59017059301381e-06,
+      "loss": 1.7151,
+      "step": 1969
+    },
+    {
+      "epoch": 2.158904109589041,
+      "grad_norm": 2.4912102222442627,
+      "learning_rate": 1.5881397238017873e-06,
+      "loss": 1.7624,
+      "step": 1970
+    },
+    {
+      "epoch": 2.158904109589041,
+      "eval_loss": 1.9855093955993652,
+      "eval_runtime": 37.0796,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 1970
+    },
+    {
+      "epoch": 2.16,
+      "grad_norm": 2.4793057441711426,
+      "learning_rate": 1.5861088545897646e-06,
+      "loss": 1.7617,
+      "step": 1971
+    },
+    {
+      "epoch": 2.1610958904109587,
+      "grad_norm": 2.5231704711914062,
+      "learning_rate": 1.5840779853777418e-06,
+      "loss": 1.6845,
+      "step": 1972
+    },
+    {
+      "epoch": 2.1621917808219178,
+      "grad_norm": 2.5126073360443115,
+      "learning_rate": 1.5820471161657191e-06,
+      "loss": 1.7468,
+      "step": 1973
+    },
+    {
+      "epoch": 2.163287671232877,
+      "grad_norm": 2.5558865070343018,
+      "learning_rate": 1.5800162469536964e-06,
+      "loss": 1.7253,
+      "step": 1974
+    },
+    {
+      "epoch": 2.1643835616438354,
+      "grad_norm": 2.808987855911255,
+      "learning_rate": 1.5779853777416735e-06,
+      "loss": 1.7844,
+      "step": 1975
+    },
+    {
+      "epoch": 2.1654794520547944,
+      "grad_norm": 2.765275001525879,
+      "learning_rate": 1.5759545085296507e-06,
+      "loss": 1.7658,
+      "step": 1976
+    },
+    {
+      "epoch": 2.1665753424657535,
+      "grad_norm": 2.5721709728240967,
+      "learning_rate": 1.573923639317628e-06,
+      "loss": 1.74,
+      "step": 1977
+    },
+    {
+      "epoch": 2.1676712328767125,
+      "grad_norm": 2.4276933670043945,
+      "learning_rate": 1.5718927701056053e-06,
+      "loss": 1.7613,
+      "step": 1978
+    },
+    {
+      "epoch": 2.168767123287671,
+      "grad_norm": 2.5665223598480225,
+      "learning_rate": 1.5698619008935826e-06,
+      "loss": 1.7514,
+      "step": 1979
+    },
+    {
+      "epoch": 2.16986301369863,
+      "grad_norm": 2.5588295459747314,
+      "learning_rate": 1.5678310316815598e-06,
+      "loss": 1.7749,
+      "step": 1980
+    },
+    {
+      "epoch": 2.16986301369863,
+      "eval_loss": 1.9858319759368896,
+      "eval_runtime": 37.0705,
+      "eval_samples_per_second": 26.976,
+      "eval_steps_per_second": 1.133,
+      "step": 1980
+    },
+    {
+      "epoch": 2.170958904109589,
+      "grad_norm": 2.5413172245025635,
+      "learning_rate": 1.5658001624695371e-06,
+      "loss": 1.706,
+      "step": 1981
+    },
+    {
+      "epoch": 2.172054794520548,
+      "grad_norm": 2.5246310234069824,
+      "learning_rate": 1.5637692932575144e-06,
+      "loss": 1.7825,
+      "step": 1982
+    },
+    {
+      "epoch": 2.1731506849315068,
+      "grad_norm": 2.385380506515503,
+      "learning_rate": 1.5617384240454917e-06,
+      "loss": 1.7268,
+      "step": 1983
+    },
+    {
+      "epoch": 2.174246575342466,
+      "grad_norm": 2.4967970848083496,
+      "learning_rate": 1.559707554833469e-06,
+      "loss": 1.7424,
+      "step": 1984
+    },
+    {
+      "epoch": 2.175342465753425,
+      "grad_norm": 2.563683271408081,
+      "learning_rate": 1.5576766856214462e-06,
+      "loss": 1.7904,
+      "step": 1985
+    },
+    {
+      "epoch": 2.1764383561643834,
+      "grad_norm": 2.518751859664917,
+      "learning_rate": 1.5556458164094235e-06,
+      "loss": 1.6616,
+      "step": 1986
+    },
+    {
+      "epoch": 2.1775342465753424,
+      "grad_norm": 2.5789380073547363,
+      "learning_rate": 1.5536149471974008e-06,
+      "loss": 1.7363,
+      "step": 1987
+    },
+    {
+      "epoch": 2.1786301369863015,
+      "grad_norm": 2.567406415939331,
+      "learning_rate": 1.5515840779853778e-06,
+      "loss": 1.6835,
+      "step": 1988
+    },
+    {
+      "epoch": 2.17972602739726,
+      "grad_norm": 2.509427309036255,
+      "learning_rate": 1.549553208773355e-06,
+      "loss": 1.7493,
+      "step": 1989
+    },
+    {
+      "epoch": 2.180821917808219,
+      "grad_norm": 2.6355340480804443,
+      "learning_rate": 1.5475223395613324e-06,
+      "loss": 1.7292,
+      "step": 1990
+    },
+    {
+      "epoch": 2.180821917808219,
+      "eval_loss": 1.9893935918807983,
+      "eval_runtime": 37.0928,
+      "eval_samples_per_second": 26.959,
+      "eval_steps_per_second": 1.132,
+      "step": 1990
+    },
+    {
+      "epoch": 2.181917808219178,
+      "grad_norm": 2.564587116241455,
+      "learning_rate": 1.5454914703493097e-06,
+      "loss": 1.7248,
+      "step": 1991
+    },
+    {
+      "epoch": 2.183013698630137,
+      "grad_norm": 2.6708462238311768,
+      "learning_rate": 1.543460601137287e-06,
+      "loss": 1.8123,
+      "step": 1992
+    },
+    {
+      "epoch": 2.1841095890410958,
+      "grad_norm": 2.6396312713623047,
+      "learning_rate": 1.5414297319252642e-06,
+      "loss": 1.7742,
+      "step": 1993
+    },
+    {
+      "epoch": 2.185205479452055,
+      "grad_norm": 2.553544044494629,
+      "learning_rate": 1.5393988627132415e-06,
+      "loss": 1.7705,
+      "step": 1994
+    },
+    {
+      "epoch": 2.186301369863014,
+      "grad_norm": 2.554997682571411,
+      "learning_rate": 1.5373679935012188e-06,
+      "loss": 1.737,
+      "step": 1995
+    },
+    {
+      "epoch": 2.1873972602739724,
+      "grad_norm": 2.5650475025177,
+      "learning_rate": 1.535337124289196e-06,
+      "loss": 1.6865,
+      "step": 1996
+    },
+    {
+      "epoch": 2.1884931506849314,
+      "grad_norm": 2.8019111156463623,
+      "learning_rate": 1.5333062550771733e-06,
+      "loss": 1.7428,
+      "step": 1997
+    },
+    {
+      "epoch": 2.1895890410958905,
+      "grad_norm": 2.5704092979431152,
+      "learning_rate": 1.5312753858651506e-06,
+      "loss": 1.8291,
+      "step": 1998
+    },
+    {
+      "epoch": 2.1906849315068495,
+      "grad_norm": 2.5332226753234863,
+      "learning_rate": 1.5292445166531279e-06,
+      "loss": 1.7274,
+      "step": 1999
+    },
+    {
+      "epoch": 2.191780821917808,
+      "grad_norm": 2.555288553237915,
+      "learning_rate": 1.5272136474411047e-06,
+      "loss": 1.7497,
+      "step": 2000
+    },
+    {
+      "epoch": 2.191780821917808,
+      "eval_loss": 1.9881768226623535,
+      "eval_runtime": 37.0996,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 2000
+    },
+    {
+      "epoch": 2.192876712328767,
+      "grad_norm": 2.6068379878997803,
+      "learning_rate": 1.525182778229082e-06,
+      "loss": 1.7955,
+      "step": 2001
+    },
+    {
+      "epoch": 2.193972602739726,
+      "grad_norm": 2.479121208190918,
+      "learning_rate": 1.5231519090170593e-06,
+      "loss": 1.7342,
+      "step": 2002
+    },
+    {
+      "epoch": 2.1950684931506848,
+      "grad_norm": 2.495670795440674,
+      "learning_rate": 1.5211210398050365e-06,
+      "loss": 1.7125,
+      "step": 2003
+    },
+    {
+      "epoch": 2.196164383561644,
+      "grad_norm": 2.4988651275634766,
+      "learning_rate": 1.5190901705930138e-06,
+      "loss": 1.7137,
+      "step": 2004
+    },
+    {
+      "epoch": 2.197260273972603,
+      "grad_norm": 2.5191574096679688,
+      "learning_rate": 1.517059301380991e-06,
+      "loss": 1.7672,
+      "step": 2005
+    },
+    {
+      "epoch": 2.198356164383562,
+      "grad_norm": 2.5469892024993896,
+      "learning_rate": 1.5150284321689684e-06,
+      "loss": 1.7127,
+      "step": 2006
+    },
+    {
+      "epoch": 2.1994520547945204,
+      "grad_norm": 2.490896701812744,
+      "learning_rate": 1.5129975629569456e-06,
+      "loss": 1.7397,
+      "step": 2007
+    },
+    {
+      "epoch": 2.2005479452054795,
+      "grad_norm": 2.5142805576324463,
+      "learning_rate": 1.510966693744923e-06,
+      "loss": 1.7431,
+      "step": 2008
+    },
+    {
+      "epoch": 2.2016438356164385,
+      "grad_norm": 2.5397887229919434,
+      "learning_rate": 1.5089358245329002e-06,
+      "loss": 1.7439,
+      "step": 2009
+    },
+    {
+      "epoch": 2.202739726027397,
+      "grad_norm": 2.5063931941986084,
+      "learning_rate": 1.5069049553208775e-06,
+      "loss": 1.7259,
+      "step": 2010
+    },
+    {
+      "epoch": 2.202739726027397,
+      "eval_loss": 1.9853808879852295,
+      "eval_runtime": 36.8066,
+      "eval_samples_per_second": 27.169,
+      "eval_steps_per_second": 1.141,
+      "step": 2010
+    },
+    {
+      "epoch": 2.203835616438356,
+      "grad_norm": 2.6017589569091797,
+      "learning_rate": 1.5048740861088547e-06,
+      "loss": 1.7714,
+      "step": 2011
+    },
+    {
+      "epoch": 2.204931506849315,
+      "grad_norm": 2.5295040607452393,
+      "learning_rate": 1.502843216896832e-06,
+      "loss": 1.7561,
+      "step": 2012
+    },
+    {
+      "epoch": 2.206027397260274,
+      "grad_norm": 2.5187482833862305,
+      "learning_rate": 1.500812347684809e-06,
+      "loss": 1.6761,
+      "step": 2013
+    },
+    {
+      "epoch": 2.207123287671233,
+      "grad_norm": 2.5163650512695312,
+      "learning_rate": 1.4987814784727864e-06,
+      "loss": 1.7469,
+      "step": 2014
+    },
+    {
+      "epoch": 2.208219178082192,
+      "grad_norm": 2.731238603591919,
+      "learning_rate": 1.4967506092607636e-06,
+      "loss": 1.7527,
+      "step": 2015
+    },
+    {
+      "epoch": 2.209315068493151,
+      "grad_norm": 2.5298311710357666,
+      "learning_rate": 1.494719740048741e-06,
+      "loss": 1.6912,
+      "step": 2016
+    },
+    {
+      "epoch": 2.2104109589041094,
+      "grad_norm": 2.6385974884033203,
+      "learning_rate": 1.4926888708367182e-06,
+      "loss": 1.7664,
+      "step": 2017
+    },
+    {
+      "epoch": 2.2115068493150685,
+      "grad_norm": 2.6150033473968506,
+      "learning_rate": 1.4906580016246955e-06,
+      "loss": 1.7708,
+      "step": 2018
+    },
+    {
+      "epoch": 2.2126027397260275,
+      "grad_norm": 2.589329957962036,
+      "learning_rate": 1.4886271324126727e-06,
+      "loss": 1.7875,
+      "step": 2019
+    },
+    {
+      "epoch": 2.213698630136986,
+      "grad_norm": 2.6769933700561523,
+      "learning_rate": 1.48659626320065e-06,
+      "loss": 1.8136,
+      "step": 2020
+    },
+    {
+      "epoch": 2.213698630136986,
+      "eval_loss": 1.9827086925506592,
+      "eval_runtime": 37.0026,
+      "eval_samples_per_second": 27.025,
+      "eval_steps_per_second": 1.135,
+      "step": 2020
+    },
+    {
+      "epoch": 2.214794520547945,
+      "grad_norm": 2.7220497131347656,
+      "learning_rate": 1.4845653939886273e-06,
+      "loss": 1.6627,
+      "step": 2021
+    },
+    {
+      "epoch": 2.215890410958904,
+      "grad_norm": 2.6115968227386475,
+      "learning_rate": 1.4825345247766046e-06,
+      "loss": 1.6676,
+      "step": 2022
+    },
+    {
+      "epoch": 2.216986301369863,
+      "grad_norm": 2.4612581729888916,
+      "learning_rate": 1.4805036555645818e-06,
+      "loss": 1.6644,
+      "step": 2023
+    },
+    {
+      "epoch": 2.2180821917808218,
+      "grad_norm": 2.5708696842193604,
+      "learning_rate": 1.4784727863525591e-06,
+      "loss": 1.7357,
+      "step": 2024
+    },
+    {
+      "epoch": 2.219178082191781,
+      "grad_norm": 2.578158140182495,
+      "learning_rate": 1.4764419171405362e-06,
+      "loss": 1.7011,
+      "step": 2025
+    },
+    {
+      "epoch": 2.22027397260274,
+      "grad_norm": 2.547666072845459,
+      "learning_rate": 1.4744110479285135e-06,
+      "loss": 1.7787,
+      "step": 2026
+    },
+    {
+      "epoch": 2.2213698630136984,
+      "grad_norm": 2.617551326751709,
+      "learning_rate": 1.4723801787164907e-06,
+      "loss": 1.784,
+      "step": 2027
+    },
+    {
+      "epoch": 2.2224657534246575,
+      "grad_norm": 2.4385478496551514,
+      "learning_rate": 1.470349309504468e-06,
+      "loss": 1.7764,
+      "step": 2028
+    },
+    {
+      "epoch": 2.2235616438356165,
+      "grad_norm": 2.6088693141937256,
+      "learning_rate": 1.4683184402924453e-06,
+      "loss": 1.7477,
+      "step": 2029
+    },
+    {
+      "epoch": 2.2246575342465755,
+      "grad_norm": 2.6165895462036133,
+      "learning_rate": 1.4662875710804226e-06,
+      "loss": 1.7531,
+      "step": 2030
+    },
+    {
+      "epoch": 2.2246575342465755,
+      "eval_loss": 1.9826008081436157,
+      "eval_runtime": 37.085,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 2030
+    },
+    {
+      "epoch": 2.225753424657534,
+      "grad_norm": 2.6285006999969482,
+      "learning_rate": 1.4642567018683998e-06,
+      "loss": 1.8403,
+      "step": 2031
+    },
+    {
+      "epoch": 2.226849315068493,
+      "grad_norm": 2.549326181411743,
+      "learning_rate": 1.4622258326563771e-06,
+      "loss": 1.7304,
+      "step": 2032
+    },
+    {
+      "epoch": 2.227945205479452,
+      "grad_norm": 2.6198782920837402,
+      "learning_rate": 1.4601949634443544e-06,
+      "loss": 1.7147,
+      "step": 2033
+    },
+    {
+      "epoch": 2.2290410958904108,
+      "grad_norm": 2.536947727203369,
+      "learning_rate": 1.4581640942323317e-06,
+      "loss": 1.7639,
+      "step": 2034
+    },
+    {
+      "epoch": 2.23013698630137,
+      "grad_norm": 2.6757264137268066,
+      "learning_rate": 1.456133225020309e-06,
+      "loss": 1.8434,
+      "step": 2035
+    },
+    {
+      "epoch": 2.231232876712329,
+      "grad_norm": 2.598095655441284,
+      "learning_rate": 1.4541023558082862e-06,
+      "loss": 1.7593,
+      "step": 2036
+    },
+    {
+      "epoch": 2.232328767123288,
+      "grad_norm": 2.6104440689086914,
+      "learning_rate": 1.4520714865962635e-06,
+      "loss": 1.7473,
+      "step": 2037
+    },
+    {
+      "epoch": 2.2334246575342465,
+      "grad_norm": 2.6350162029266357,
+      "learning_rate": 1.4500406173842406e-06,
+      "loss": 1.777,
+      "step": 2038
+    },
+    {
+      "epoch": 2.2345205479452055,
+      "grad_norm": 2.593874216079712,
+      "learning_rate": 1.4480097481722178e-06,
+      "loss": 1.7506,
+      "step": 2039
+    },
+    {
+      "epoch": 2.2356164383561645,
+      "grad_norm": 2.6991093158721924,
+      "learning_rate": 1.4459788789601951e-06,
+      "loss": 1.7375,
+      "step": 2040
+    },
+    {
+      "epoch": 2.2356164383561645,
+      "eval_loss": 1.9815349578857422,
+      "eval_runtime": 37.1126,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 2040
+    },
+    {
+      "epoch": 2.236712328767123,
+      "grad_norm": 2.6339058876037598,
+      "learning_rate": 1.4439480097481724e-06,
+      "loss": 1.7045,
+      "step": 2041
+    },
+    {
+      "epoch": 2.237808219178082,
+      "grad_norm": 2.691631555557251,
+      "learning_rate": 1.4419171405361497e-06,
+      "loss": 1.7063,
+      "step": 2042
+    },
+    {
+      "epoch": 2.238904109589041,
+      "grad_norm": 3.021667003631592,
+      "learning_rate": 1.439886271324127e-06,
+      "loss": 1.7607,
+      "step": 2043
+    },
+    {
+      "epoch": 2.24,
+      "grad_norm": 2.690802812576294,
+      "learning_rate": 1.4378554021121042e-06,
+      "loss": 1.7594,
+      "step": 2044
+    },
+    {
+      "epoch": 2.241095890410959,
+      "grad_norm": 2.5371177196502686,
+      "learning_rate": 1.4358245329000815e-06,
+      "loss": 1.7177,
+      "step": 2045
+    },
+    {
+      "epoch": 2.242191780821918,
+      "grad_norm": 2.5586729049682617,
+      "learning_rate": 1.4337936636880588e-06,
+      "loss": 1.7864,
+      "step": 2046
+    },
+    {
+      "epoch": 2.243287671232877,
+      "grad_norm": 2.6621484756469727,
+      "learning_rate": 1.431762794476036e-06,
+      "loss": 1.7616,
+      "step": 2047
+    },
+    {
+      "epoch": 2.2443835616438355,
+      "grad_norm": 2.787508010864258,
+      "learning_rate": 1.4297319252640133e-06,
+      "loss": 1.7056,
+      "step": 2048
+    },
+    {
+      "epoch": 2.2454794520547945,
+      "grad_norm": 2.6896064281463623,
+      "learning_rate": 1.4277010560519906e-06,
+      "loss": 1.7413,
+      "step": 2049
+    },
+    {
+      "epoch": 2.2465753424657535,
+      "grad_norm": 2.540524959564209,
+      "learning_rate": 1.4256701868399674e-06,
+      "loss": 1.7096,
+      "step": 2050
+    },
+    {
+      "epoch": 2.2465753424657535,
+      "eval_loss": 1.9832353591918945,
+      "eval_runtime": 37.1202,
+      "eval_samples_per_second": 26.94,
+      "eval_steps_per_second": 1.131,
+      "step": 2050
+    },
+    {
+      "epoch": 2.247671232876712,
+      "grad_norm": 2.582171678543091,
+      "learning_rate": 1.4236393176279447e-06,
+      "loss": 1.8262,
+      "step": 2051
+    },
+    {
+      "epoch": 2.248767123287671,
+      "grad_norm": 2.659044027328491,
+      "learning_rate": 1.421608448415922e-06,
+      "loss": 1.7458,
+      "step": 2052
+    },
+    {
+      "epoch": 2.24986301369863,
+      "grad_norm": 2.7265775203704834,
+      "learning_rate": 1.4195775792038993e-06,
+      "loss": 1.8341,
+      "step": 2053
+    },
+    {
+      "epoch": 2.250958904109589,
+      "grad_norm": 2.922330379486084,
+      "learning_rate": 1.4175467099918765e-06,
+      "loss": 1.7103,
+      "step": 2054
+    },
+    {
+      "epoch": 2.252054794520548,
+      "grad_norm": 2.4665236473083496,
+      "learning_rate": 1.4155158407798538e-06,
+      "loss": 1.7461,
+      "step": 2055
+    },
+    {
+      "epoch": 2.253150684931507,
+      "grad_norm": 2.639538049697876,
+      "learning_rate": 1.413484971567831e-06,
+      "loss": 1.7683,
+      "step": 2056
+    },
+    {
+      "epoch": 2.254246575342466,
+      "grad_norm": 2.558363676071167,
+      "learning_rate": 1.4114541023558084e-06,
+      "loss": 1.6821,
+      "step": 2057
+    },
+    {
+      "epoch": 2.255342465753425,
+      "grad_norm": 2.6308321952819824,
+      "learning_rate": 1.4094232331437856e-06,
+      "loss": 1.7226,
+      "step": 2058
+    },
+    {
+      "epoch": 2.2564383561643835,
+      "grad_norm": 2.6325299739837646,
+      "learning_rate": 1.407392363931763e-06,
+      "loss": 1.783,
+      "step": 2059
+    },
+    {
+      "epoch": 2.2575342465753425,
+      "grad_norm": 2.6024723052978516,
+      "learning_rate": 1.4053614947197402e-06,
+      "loss": 1.8402,
+      "step": 2060
+    },
+    {
+      "epoch": 2.2575342465753425,
+      "eval_loss": 1.983332633972168,
+      "eval_runtime": 37.1133,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 2060
+    },
+    {
+      "epoch": 2.2586301369863016,
+      "grad_norm": 2.548522710800171,
+      "learning_rate": 1.4033306255077175e-06,
+      "loss": 1.7625,
+      "step": 2061
+    },
+    {
+      "epoch": 2.25972602739726,
+      "grad_norm": 2.5953831672668457,
+      "learning_rate": 1.4012997562956945e-06,
+      "loss": 1.6632,
+      "step": 2062
+    },
+    {
+      "epoch": 2.260821917808219,
+      "grad_norm": 2.5880215167999268,
+      "learning_rate": 1.3992688870836718e-06,
+      "loss": 1.7957,
+      "step": 2063
+    },
+    {
+      "epoch": 2.261917808219178,
+      "grad_norm": 2.439326047897339,
+      "learning_rate": 1.397238017871649e-06,
+      "loss": 1.76,
+      "step": 2064
+    },
+    {
+      "epoch": 2.263013698630137,
+      "grad_norm": 2.5185437202453613,
+      "learning_rate": 1.3952071486596264e-06,
+      "loss": 1.7385,
+      "step": 2065
+    },
+    {
+      "epoch": 2.264109589041096,
+      "grad_norm": 2.5943732261657715,
+      "learning_rate": 1.3931762794476036e-06,
+      "loss": 1.7066,
+      "step": 2066
+    },
+    {
+      "epoch": 2.265205479452055,
+      "grad_norm": 2.6375551223754883,
+      "learning_rate": 1.391145410235581e-06,
+      "loss": 1.7884,
+      "step": 2067
+    },
+    {
+      "epoch": 2.266301369863014,
+      "grad_norm": 2.4962515830993652,
+      "learning_rate": 1.3891145410235582e-06,
+      "loss": 1.7295,
+      "step": 2068
+    },
+    {
+      "epoch": 2.2673972602739725,
+      "grad_norm": 2.5873053073883057,
+      "learning_rate": 1.3870836718115355e-06,
+      "loss": 1.6697,
+      "step": 2069
+    },
+    {
+      "epoch": 2.2684931506849315,
+      "grad_norm": 2.573211908340454,
+      "learning_rate": 1.3850528025995127e-06,
+      "loss": 1.7198,
+      "step": 2070
+    },
+    {
+      "epoch": 2.2684931506849315,
+      "eval_loss": 1.9826557636260986,
+      "eval_runtime": 37.1271,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 2070
+    },
+    {
+      "epoch": 2.2695890410958905,
+      "grad_norm": 2.68015718460083,
+      "learning_rate": 1.38302193338749e-06,
+      "loss": 1.7294,
+      "step": 2071
+    },
+    {
+      "epoch": 2.270684931506849,
+      "grad_norm": 2.4882888793945312,
+      "learning_rate": 1.3809910641754673e-06,
+      "loss": 1.6592,
+      "step": 2072
+    },
+    {
+      "epoch": 2.271780821917808,
+      "grad_norm": 2.587048053741455,
+      "learning_rate": 1.3789601949634446e-06,
+      "loss": 1.7605,
+      "step": 2073
+    },
+    {
+      "epoch": 2.272876712328767,
+      "grad_norm": 2.6452062129974365,
+      "learning_rate": 1.3769293257514218e-06,
+      "loss": 1.6664,
+      "step": 2074
+    },
+    {
+      "epoch": 2.2739726027397262,
+      "grad_norm": 2.5663232803344727,
+      "learning_rate": 1.374898456539399e-06,
+      "loss": 1.771,
+      "step": 2075
+    },
+    {
+      "epoch": 2.275068493150685,
+      "grad_norm": 2.652989149093628,
+      "learning_rate": 1.3728675873273762e-06,
+      "loss": 1.7096,
+      "step": 2076
+    },
+    {
+      "epoch": 2.276164383561644,
+      "grad_norm": 2.538435220718384,
+      "learning_rate": 1.3708367181153535e-06,
+      "loss": 1.798,
+      "step": 2077
+    },
+    {
+      "epoch": 2.277260273972603,
+      "grad_norm": 2.4685096740722656,
+      "learning_rate": 1.3688058489033307e-06,
+      "loss": 1.7904,
+      "step": 2078
+    },
+    {
+      "epoch": 2.2783561643835615,
+      "grad_norm": 2.566723108291626,
+      "learning_rate": 1.366774979691308e-06,
+      "loss": 1.7569,
+      "step": 2079
+    },
+    {
+      "epoch": 2.2794520547945205,
+      "grad_norm": 2.560960531234741,
+      "learning_rate": 1.3647441104792853e-06,
+      "loss": 1.7734,
+      "step": 2080
+    },
+    {
+      "epoch": 2.2794520547945205,
+      "eval_loss": 1.9809880256652832,
+      "eval_runtime": 37.1137,
+      "eval_samples_per_second": 26.944,
+      "eval_steps_per_second": 1.132,
+      "step": 2080
+    },
+    {
+      "epoch": 2.2805479452054795,
+      "grad_norm": 2.576608419418335,
+      "learning_rate": 1.3627132412672626e-06,
+      "loss": 1.7431,
+      "step": 2081
+    },
+    {
+      "epoch": 2.281643835616438,
+      "grad_norm": 2.648301601409912,
+      "learning_rate": 1.3606823720552398e-06,
+      "loss": 1.6959,
+      "step": 2082
+    },
+    {
+      "epoch": 2.282739726027397,
+      "grad_norm": 2.516493082046509,
+      "learning_rate": 1.3586515028432171e-06,
+      "loss": 1.7519,
+      "step": 2083
+    },
+    {
+      "epoch": 2.283835616438356,
+      "grad_norm": 2.474616050720215,
+      "learning_rate": 1.3566206336311944e-06,
+      "loss": 1.6796,
+      "step": 2084
+    },
+    {
+      "epoch": 2.2849315068493152,
+      "grad_norm": 2.482044219970703,
+      "learning_rate": 1.3545897644191717e-06,
+      "loss": 1.7443,
+      "step": 2085
+    },
+    {
+      "epoch": 2.286027397260274,
+      "grad_norm": 2.6574456691741943,
+      "learning_rate": 1.352558895207149e-06,
+      "loss": 1.7145,
+      "step": 2086
+    },
+    {
+      "epoch": 2.287123287671233,
+      "grad_norm": 2.5350115299224854,
+      "learning_rate": 1.3505280259951258e-06,
+      "loss": 1.765,
+      "step": 2087
+    },
+    {
+      "epoch": 2.288219178082192,
+      "grad_norm": 2.6488730907440186,
+      "learning_rate": 1.348497156783103e-06,
+      "loss": 1.7407,
+      "step": 2088
+    },
+    {
+      "epoch": 2.289315068493151,
+      "grad_norm": 2.6020500659942627,
+      "learning_rate": 1.3464662875710804e-06,
+      "loss": 1.779,
+      "step": 2089
+    },
+    {
+      "epoch": 2.2904109589041095,
+      "grad_norm": 2.633013963699341,
+      "learning_rate": 1.3444354183590576e-06,
+      "loss": 1.6824,
+      "step": 2090
+    },
+    {
+      "epoch": 2.2904109589041095,
+      "eval_loss": 1.982380986213684,
+      "eval_runtime": 37.1166,
+      "eval_samples_per_second": 26.942,
+      "eval_steps_per_second": 1.132,
+      "step": 2090
+    },
+    {
+      "epoch": 2.2915068493150685,
+      "grad_norm": 2.6399149894714355,
+      "learning_rate": 1.342404549147035e-06,
+      "loss": 1.7634,
+      "step": 2091
+    },
+    {
+      "epoch": 2.2926027397260276,
+      "grad_norm": 2.580634832382202,
+      "learning_rate": 1.3403736799350122e-06,
+      "loss": 1.7099,
+      "step": 2092
+    },
+    {
+      "epoch": 2.293698630136986,
+      "grad_norm": 2.523400068283081,
+      "learning_rate": 1.3383428107229895e-06,
+      "loss": 1.7705,
+      "step": 2093
+    },
+    {
+      "epoch": 2.294794520547945,
+      "grad_norm": 2.5114550590515137,
+      "learning_rate": 1.3363119415109667e-06,
+      "loss": 1.7961,
+      "step": 2094
+    },
+    {
+      "epoch": 2.2958904109589042,
+      "grad_norm": 2.457369804382324,
+      "learning_rate": 1.334281072298944e-06,
+      "loss": 1.7507,
+      "step": 2095
+    },
+    {
+      "epoch": 2.296986301369863,
+      "grad_norm": 2.5671467781066895,
+      "learning_rate": 1.3322502030869213e-06,
+      "loss": 1.7607,
+      "step": 2096
+    },
+    {
+      "epoch": 2.298082191780822,
+      "grad_norm": 2.459540367126465,
+      "learning_rate": 1.3302193338748986e-06,
+      "loss": 1.7193,
+      "step": 2097
+    },
+    {
+      "epoch": 2.299178082191781,
+      "grad_norm": 2.4934308528900146,
+      "learning_rate": 1.3281884646628758e-06,
+      "loss": 1.7775,
+      "step": 2098
+    },
+    {
+      "epoch": 2.30027397260274,
+      "grad_norm": 2.520641565322876,
+      "learning_rate": 1.3261575954508531e-06,
+      "loss": 1.7847,
+      "step": 2099
+    },
+    {
+      "epoch": 2.3013698630136985,
+      "grad_norm": 2.4607160091400146,
+      "learning_rate": 1.3241267262388302e-06,
+      "loss": 1.7812,
+      "step": 2100
+    },
+    {
+      "epoch": 2.3013698630136985,
+      "eval_loss": 1.9836152791976929,
+      "eval_runtime": 37.1239,
+      "eval_samples_per_second": 26.937,
+      "eval_steps_per_second": 1.131,
+      "step": 2100
+    },
+    {
+      "epoch": 2.3024657534246575,
+      "grad_norm": 2.577761650085449,
+      "learning_rate": 1.3220958570268074e-06,
+      "loss": 1.7452,
+      "step": 2101
+    },
+    {
+      "epoch": 2.3035616438356166,
+      "grad_norm": 2.5331356525421143,
+      "learning_rate": 1.3200649878147847e-06,
+      "loss": 1.6989,
+      "step": 2102
+    },
+    {
+      "epoch": 2.304657534246575,
+      "grad_norm": 2.632581949234009,
+      "learning_rate": 1.318034118602762e-06,
+      "loss": 1.7368,
+      "step": 2103
+    },
+    {
+      "epoch": 2.305753424657534,
+      "grad_norm": 2.55940318107605,
+      "learning_rate": 1.3160032493907393e-06,
+      "loss": 1.7806,
+      "step": 2104
+    },
+    {
+      "epoch": 2.3068493150684932,
+      "grad_norm": 2.5680153369903564,
+      "learning_rate": 1.3139723801787166e-06,
+      "loss": 1.7482,
+      "step": 2105
+    },
+    {
+      "epoch": 2.3079452054794523,
+      "grad_norm": 2.7915756702423096,
+      "learning_rate": 1.3119415109666938e-06,
+      "loss": 1.832,
+      "step": 2106
+    },
+    {
+      "epoch": 2.309041095890411,
+      "grad_norm": 2.518031597137451,
+      "learning_rate": 1.309910641754671e-06,
+      "loss": 1.7401,
+      "step": 2107
+    },
+    {
+      "epoch": 2.31013698630137,
+      "grad_norm": 2.7410221099853516,
+      "learning_rate": 1.3078797725426484e-06,
+      "loss": 1.6992,
+      "step": 2108
+    },
+    {
+      "epoch": 2.311232876712329,
+      "grad_norm": 2.611825466156006,
+      "learning_rate": 1.3058489033306257e-06,
+      "loss": 1.7085,
+      "step": 2109
+    },
+    {
+      "epoch": 2.3123287671232875,
+      "grad_norm": 2.5833330154418945,
+      "learning_rate": 1.303818034118603e-06,
+      "loss": 1.7129,
+      "step": 2110
+    },
+    {
+      "epoch": 2.3123287671232875,
+      "eval_loss": 1.9824191331863403,
+      "eval_runtime": 36.773,
+      "eval_samples_per_second": 27.194,
+      "eval_steps_per_second": 1.142,
+      "step": 2110
+    },
+    {
+      "epoch": 2.3134246575342465,
+      "grad_norm": 2.587327003479004,
+      "learning_rate": 1.3017871649065802e-06,
+      "loss": 1.7439,
+      "step": 2111
+    },
+    {
+      "epoch": 2.3145205479452056,
+      "grad_norm": 2.574925422668457,
+      "learning_rate": 1.2997562956945573e-06,
+      "loss": 1.7173,
+      "step": 2112
+    },
+    {
+      "epoch": 2.315616438356164,
+      "grad_norm": 2.6404287815093994,
+      "learning_rate": 1.2977254264825345e-06,
+      "loss": 1.7052,
+      "step": 2113
+    },
+    {
+      "epoch": 2.316712328767123,
+      "grad_norm": 2.634584903717041,
+      "learning_rate": 1.2956945572705118e-06,
+      "loss": 1.741,
+      "step": 2114
+    },
+    {
+      "epoch": 2.317808219178082,
+      "grad_norm": 2.66159987449646,
+      "learning_rate": 1.293663688058489e-06,
+      "loss": 1.76,
+      "step": 2115
+    },
+    {
+      "epoch": 2.3189041095890413,
+      "grad_norm": 2.6337332725524902,
+      "learning_rate": 1.2916328188464664e-06,
+      "loss": 1.7035,
+      "step": 2116
+    },
+    {
+      "epoch": 2.32,
+      "grad_norm": 2.528524398803711,
+      "learning_rate": 1.2896019496344436e-06,
+      "loss": 1.6439,
+      "step": 2117
+    },
+    {
+      "epoch": 2.321095890410959,
+      "grad_norm": 2.5938384532928467,
+      "learning_rate": 1.287571080422421e-06,
+      "loss": 1.7134,
+      "step": 2118
+    },
+    {
+      "epoch": 2.322191780821918,
+      "grad_norm": 2.606271505355835,
+      "learning_rate": 1.2855402112103982e-06,
+      "loss": 1.7758,
+      "step": 2119
+    },
+    {
+      "epoch": 2.323287671232877,
+      "grad_norm": 2.530071258544922,
+      "learning_rate": 1.2835093419983755e-06,
+      "loss": 1.6927,
+      "step": 2120
+    },
+    {
+      "epoch": 2.323287671232877,
+      "eval_loss": 1.9856306314468384,
+      "eval_runtime": 36.9674,
+      "eval_samples_per_second": 27.051,
+      "eval_steps_per_second": 1.136,
+      "step": 2120
+    },
+    {
+      "epoch": 2.3243835616438355,
+      "grad_norm": 2.8125321865081787,
+      "learning_rate": 1.2814784727863528e-06,
+      "loss": 1.7225,
+      "step": 2121
+    },
+    {
+      "epoch": 2.3254794520547946,
+      "grad_norm": 2.57012939453125,
+      "learning_rate": 1.27944760357433e-06,
+      "loss": 1.7177,
+      "step": 2122
+    },
+    {
+      "epoch": 2.3265753424657536,
+      "grad_norm": 2.6595780849456787,
+      "learning_rate": 1.2774167343623073e-06,
+      "loss": 1.7807,
+      "step": 2123
+    },
+    {
+      "epoch": 2.327671232876712,
+      "grad_norm": 2.557115077972412,
+      "learning_rate": 1.2753858651502846e-06,
+      "loss": 1.7762,
+      "step": 2124
+    },
+    {
+      "epoch": 2.328767123287671,
+      "grad_norm": 2.763218879699707,
+      "learning_rate": 1.2733549959382616e-06,
+      "loss": 1.7638,
+      "step": 2125
+    },
+    {
+      "epoch": 2.3298630136986302,
+      "grad_norm": 2.7284505367279053,
+      "learning_rate": 1.271324126726239e-06,
+      "loss": 1.7387,
+      "step": 2126
+    },
+    {
+      "epoch": 2.330958904109589,
+      "grad_norm": 2.5893704891204834,
+      "learning_rate": 1.2692932575142162e-06,
+      "loss": 1.7322,
+      "step": 2127
+    },
+    {
+      "epoch": 2.332054794520548,
+      "grad_norm": 2.625582456588745,
+      "learning_rate": 1.2672623883021935e-06,
+      "loss": 1.6865,
+      "step": 2128
+    },
+    {
+      "epoch": 2.333150684931507,
+      "grad_norm": 2.5458641052246094,
+      "learning_rate": 1.2652315190901707e-06,
+      "loss": 1.6817,
+      "step": 2129
+    },
+    {
+      "epoch": 2.334246575342466,
+      "grad_norm": 2.5718271732330322,
+      "learning_rate": 1.263200649878148e-06,
+      "loss": 1.8076,
+      "step": 2130
+    },
+    {
+      "epoch": 2.334246575342466,
+      "eval_loss": 1.9806523323059082,
+      "eval_runtime": 36.9843,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 2130
+    },
+    {
+      "epoch": 2.3353424657534245,
+      "grad_norm": 2.4864919185638428,
+      "learning_rate": 1.2611697806661253e-06,
+      "loss": 1.7308,
+      "step": 2131
+    },
+    {
+      "epoch": 2.3364383561643836,
+      "grad_norm": 2.6007416248321533,
+      "learning_rate": 1.2591389114541026e-06,
+      "loss": 1.7523,
+      "step": 2132
+    },
+    {
+      "epoch": 2.3375342465753426,
+      "grad_norm": 2.600177049636841,
+      "learning_rate": 1.2571080422420798e-06,
+      "loss": 1.7422,
+      "step": 2133
+    },
+    {
+      "epoch": 2.338630136986301,
+      "grad_norm": 2.489025831222534,
+      "learning_rate": 1.2550771730300571e-06,
+      "loss": 1.7184,
+      "step": 2134
+    },
+    {
+      "epoch": 2.33972602739726,
+      "grad_norm": 2.5677363872528076,
+      "learning_rate": 1.2530463038180344e-06,
+      "loss": 1.7922,
+      "step": 2135
+    },
+    {
+      "epoch": 2.3408219178082192,
+      "grad_norm": 2.616976737976074,
+      "learning_rate": 1.2510154346060117e-06,
+      "loss": 1.7354,
+      "step": 2136
+    },
+    {
+      "epoch": 2.3419178082191783,
+      "grad_norm": 2.646456241607666,
+      "learning_rate": 1.2489845653939887e-06,
+      "loss": 1.7377,
+      "step": 2137
+    },
+    {
+      "epoch": 2.343013698630137,
+      "grad_norm": 2.550020933151245,
+      "learning_rate": 1.246953696181966e-06,
+      "loss": 1.7751,
+      "step": 2138
+    },
+    {
+      "epoch": 2.344109589041096,
+      "grad_norm": 2.634279251098633,
+      "learning_rate": 1.2449228269699433e-06,
+      "loss": 1.8157,
+      "step": 2139
+    },
+    {
+      "epoch": 2.345205479452055,
+      "grad_norm": 2.5621821880340576,
+      "learning_rate": 1.2428919577579204e-06,
+      "loss": 1.7383,
+      "step": 2140
+    },
+    {
+      "epoch": 2.345205479452055,
+      "eval_loss": 1.9809924364089966,
+      "eval_runtime": 37.004,
+      "eval_samples_per_second": 27.024,
+      "eval_steps_per_second": 1.135,
+      "step": 2140
+    },
+    {
+      "epoch": 2.3463013698630135,
+      "grad_norm": 2.638831615447998,
+      "learning_rate": 1.2408610885458976e-06,
+      "loss": 1.6988,
+      "step": 2141
+    },
+    {
+      "epoch": 2.3473972602739726,
+      "grad_norm": 2.634612798690796,
+      "learning_rate": 1.238830219333875e-06,
+      "loss": 1.7331,
+      "step": 2142
+    },
+    {
+      "epoch": 2.3484931506849316,
+      "grad_norm": 2.552955389022827,
+      "learning_rate": 1.2367993501218522e-06,
+      "loss": 1.7148,
+      "step": 2143
+    },
+    {
+      "epoch": 2.34958904109589,
+      "grad_norm": 2.6475753784179688,
+      "learning_rate": 1.2347684809098295e-06,
+      "loss": 1.7815,
+      "step": 2144
+    },
+    {
+      "epoch": 2.350684931506849,
+      "grad_norm": 2.7666079998016357,
+      "learning_rate": 1.2327376116978067e-06,
+      "loss": 1.7634,
+      "step": 2145
+    },
+    {
+      "epoch": 2.3517808219178082,
+      "grad_norm": 2.6810429096221924,
+      "learning_rate": 1.230706742485784e-06,
+      "loss": 1.6769,
+      "step": 2146
+    },
+    {
+      "epoch": 2.3528767123287673,
+      "grad_norm": 2.671382188796997,
+      "learning_rate": 1.2286758732737613e-06,
+      "loss": 1.6884,
+      "step": 2147
+    },
+    {
+      "epoch": 2.353972602739726,
+      "grad_norm": 2.614684820175171,
+      "learning_rate": 1.2266450040617386e-06,
+      "loss": 1.8226,
+      "step": 2148
+    },
+    {
+      "epoch": 2.355068493150685,
+      "grad_norm": 2.47189998626709,
+      "learning_rate": 1.2246141348497158e-06,
+      "loss": 1.819,
+      "step": 2149
+    },
+    {
+      "epoch": 2.356164383561644,
+      "grad_norm": 2.4641449451446533,
+      "learning_rate": 1.2225832656376931e-06,
+      "loss": 1.7494,
+      "step": 2150
+    },
+    {
+      "epoch": 2.356164383561644,
+      "eval_loss": 1.9806426763534546,
+      "eval_runtime": 37.0153,
+      "eval_samples_per_second": 27.016,
+      "eval_steps_per_second": 1.135,
+      "step": 2150
+    },
+    {
+      "epoch": 2.357260273972603,
+      "grad_norm": 2.7310373783111572,
+      "learning_rate": 1.2205523964256704e-06,
+      "loss": 1.7432,
+      "step": 2151
+    },
+    {
+      "epoch": 2.3583561643835615,
+      "grad_norm": 2.5554895401000977,
+      "learning_rate": 1.2185215272136477e-06,
+      "loss": 1.7213,
+      "step": 2152
+    },
+    {
+      "epoch": 2.3594520547945206,
+      "grad_norm": 2.5401952266693115,
+      "learning_rate": 1.2164906580016247e-06,
+      "loss": 1.7469,
+      "step": 2153
+    },
+    {
+      "epoch": 2.3605479452054796,
+      "grad_norm": 2.5213980674743652,
+      "learning_rate": 1.214459788789602e-06,
+      "loss": 1.7459,
+      "step": 2154
+    },
+    {
+      "epoch": 2.361643835616438,
+      "grad_norm": 2.5585155487060547,
+      "learning_rate": 1.2124289195775793e-06,
+      "loss": 1.6587,
+      "step": 2155
+    },
+    {
+      "epoch": 2.3627397260273972,
+      "grad_norm": 2.6329963207244873,
+      "learning_rate": 1.2103980503655566e-06,
+      "loss": 1.6925,
+      "step": 2156
+    },
+    {
+      "epoch": 2.3638356164383563,
+      "grad_norm": 2.678351640701294,
+      "learning_rate": 1.2083671811535338e-06,
+      "loss": 1.7322,
+      "step": 2157
+    },
+    {
+      "epoch": 2.364931506849315,
+      "grad_norm": 2.595313549041748,
+      "learning_rate": 1.2063363119415111e-06,
+      "loss": 1.7142,
+      "step": 2158
+    },
+    {
+      "epoch": 2.366027397260274,
+      "grad_norm": 2.5650861263275146,
+      "learning_rate": 1.2043054427294882e-06,
+      "loss": 1.7024,
+      "step": 2159
+    },
+    {
+      "epoch": 2.367123287671233,
+      "grad_norm": 2.495866537094116,
+      "learning_rate": 1.2022745735174654e-06,
+      "loss": 1.7205,
+      "step": 2160
+    },
+    {
+      "epoch": 2.367123287671233,
+      "eval_loss": 1.976922631263733,
+      "eval_runtime": 37.012,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 2160
+    },
+    {
+      "epoch": 2.368219178082192,
+      "grad_norm": 2.570307493209839,
+      "learning_rate": 1.2002437043054427e-06,
+      "loss": 1.7601,
+      "step": 2161
+    },
+    {
+      "epoch": 2.3693150684931505,
+      "grad_norm": 2.444373846054077,
+      "learning_rate": 1.19821283509342e-06,
+      "loss": 1.7857,
+      "step": 2162
+    },
+    {
+      "epoch": 2.3704109589041096,
+      "grad_norm": 2.525505781173706,
+      "learning_rate": 1.1961819658813973e-06,
+      "loss": 1.7875,
+      "step": 2163
+    },
+    {
+      "epoch": 2.3715068493150686,
+      "grad_norm": 2.622260093688965,
+      "learning_rate": 1.1941510966693746e-06,
+      "loss": 1.7596,
+      "step": 2164
+    },
+    {
+      "epoch": 2.3726027397260276,
+      "grad_norm": 2.542076349258423,
+      "learning_rate": 1.1921202274573518e-06,
+      "loss": 1.7248,
+      "step": 2165
+    },
+    {
+      "epoch": 2.3736986301369862,
+      "grad_norm": 2.510263681411743,
+      "learning_rate": 1.190089358245329e-06,
+      "loss": 1.7168,
+      "step": 2166
+    },
+    {
+      "epoch": 2.3747945205479453,
+      "grad_norm": 2.605023145675659,
+      "learning_rate": 1.1880584890333064e-06,
+      "loss": 1.798,
+      "step": 2167
+    },
+    {
+      "epoch": 2.3758904109589043,
+      "grad_norm": 2.593261480331421,
+      "learning_rate": 1.1860276198212837e-06,
+      "loss": 1.7247,
+      "step": 2168
+    },
+    {
+      "epoch": 2.376986301369863,
+      "grad_norm": 2.6010820865631104,
+      "learning_rate": 1.183996750609261e-06,
+      "loss": 1.7607,
+      "step": 2169
+    },
+    {
+      "epoch": 2.378082191780822,
+      "grad_norm": 2.62886118888855,
+      "learning_rate": 1.1819658813972382e-06,
+      "loss": 1.7641,
+      "step": 2170
+    },
+    {
+      "epoch": 2.378082191780822,
+      "eval_loss": 1.9790476560592651,
+      "eval_runtime": 37.0413,
+      "eval_samples_per_second": 26.997,
+      "eval_steps_per_second": 1.134,
+      "step": 2170
+    },
+    {
+      "epoch": 2.379178082191781,
+      "grad_norm": 2.5506582260131836,
+      "learning_rate": 1.1799350121852153e-06,
+      "loss": 1.7407,
+      "step": 2171
+    },
+    {
+      "epoch": 2.3802739726027395,
+      "grad_norm": 2.547445774078369,
+      "learning_rate": 1.1779041429731925e-06,
+      "loss": 1.7808,
+      "step": 2172
+    },
+    {
+      "epoch": 2.3813698630136986,
+      "grad_norm": 2.5443003177642822,
+      "learning_rate": 1.1758732737611698e-06,
+      "loss": 1.7842,
+      "step": 2173
+    },
+    {
+      "epoch": 2.3824657534246576,
+      "grad_norm": 2.5733375549316406,
+      "learning_rate": 1.173842404549147e-06,
+      "loss": 1.6902,
+      "step": 2174
+    },
+    {
+      "epoch": 2.383561643835616,
+      "grad_norm": 2.63702392578125,
+      "learning_rate": 1.1718115353371244e-06,
+      "loss": 1.8049,
+      "step": 2175
+    },
+    {
+      "epoch": 2.3846575342465752,
+      "grad_norm": 2.5577163696289062,
+      "learning_rate": 1.1697806661251016e-06,
+      "loss": 1.7162,
+      "step": 2176
+    },
+    {
+      "epoch": 2.3857534246575343,
+      "grad_norm": 2.4259274005889893,
+      "learning_rate": 1.167749796913079e-06,
+      "loss": 1.7883,
+      "step": 2177
+    },
+    {
+      "epoch": 2.3868493150684933,
+      "grad_norm": 2.537519931793213,
+      "learning_rate": 1.1657189277010562e-06,
+      "loss": 1.7882,
+      "step": 2178
+    },
+    {
+      "epoch": 2.387945205479452,
+      "grad_norm": 2.5480241775512695,
+      "learning_rate": 1.1636880584890335e-06,
+      "loss": 1.7245,
+      "step": 2179
+    },
+    {
+      "epoch": 2.389041095890411,
+      "grad_norm": 2.630337953567505,
+      "learning_rate": 1.1616571892770108e-06,
+      "loss": 1.7165,
+      "step": 2180
+    },
+    {
+      "epoch": 2.389041095890411,
+      "eval_loss": 1.9772392511367798,
+      "eval_runtime": 37.1283,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 2180
+    },
+    {
+      "epoch": 2.39013698630137,
+      "grad_norm": 2.5741171836853027,
+      "learning_rate": 1.159626320064988e-06,
+      "loss": 1.7806,
+      "step": 2181
+    },
+    {
+      "epoch": 2.391232876712329,
+      "grad_norm": 2.6166305541992188,
+      "learning_rate": 1.1575954508529653e-06,
+      "loss": 1.744,
+      "step": 2182
+    },
+    {
+      "epoch": 2.3923287671232876,
+      "grad_norm": 2.6965837478637695,
+      "learning_rate": 1.1555645816409426e-06,
+      "loss": 1.7736,
+      "step": 2183
+    },
+    {
+      "epoch": 2.3934246575342466,
+      "grad_norm": 2.586827516555786,
+      "learning_rate": 1.1535337124289196e-06,
+      "loss": 1.7455,
+      "step": 2184
+    },
+    {
+      "epoch": 2.3945205479452056,
+      "grad_norm": 2.587235689163208,
+      "learning_rate": 1.151502843216897e-06,
+      "loss": 1.6999,
+      "step": 2185
+    },
+    {
+      "epoch": 2.3956164383561642,
+      "grad_norm": 2.656740665435791,
+      "learning_rate": 1.1494719740048742e-06,
+      "loss": 1.6518,
+      "step": 2186
+    },
+    {
+      "epoch": 2.3967123287671233,
+      "grad_norm": 2.5353455543518066,
+      "learning_rate": 1.1474411047928515e-06,
+      "loss": 1.7389,
+      "step": 2187
+    },
+    {
+      "epoch": 2.3978082191780823,
+      "grad_norm": 2.622314453125,
+      "learning_rate": 1.1454102355808287e-06,
+      "loss": 1.7309,
+      "step": 2188
+    },
+    {
+      "epoch": 2.398904109589041,
+      "grad_norm": 2.6286797523498535,
+      "learning_rate": 1.143379366368806e-06,
+      "loss": 1.7094,
+      "step": 2189
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 2.5203921794891357,
+      "learning_rate": 1.141348497156783e-06,
+      "loss": 1.6905,
+      "step": 2190
+    },
+    {
+      "epoch": 2.4,
+      "eval_loss": 1.9822566509246826,
+      "eval_runtime": 37.0581,
+      "eval_samples_per_second": 26.985,
+      "eval_steps_per_second": 1.133,
+      "step": 2190
+    },
+    {
+      "epoch": 2.401095890410959,
+      "grad_norm": 2.6562373638153076,
+      "learning_rate": 1.1393176279447604e-06,
+      "loss": 1.8293,
+      "step": 2191
+    },
+    {
+      "epoch": 2.402191780821918,
+      "grad_norm": 2.5581510066986084,
+      "learning_rate": 1.1372867587327376e-06,
+      "loss": 1.7936,
+      "step": 2192
+    },
+    {
+      "epoch": 2.4032876712328766,
+      "grad_norm": 2.534675121307373,
+      "learning_rate": 1.135255889520715e-06,
+      "loss": 1.6648,
+      "step": 2193
+    },
+    {
+      "epoch": 2.4043835616438356,
+      "grad_norm": 2.7236168384552,
+      "learning_rate": 1.1332250203086922e-06,
+      "loss": 1.7654,
+      "step": 2194
+    },
+    {
+      "epoch": 2.4054794520547946,
+      "grad_norm": 2.5706138610839844,
+      "learning_rate": 1.1311941510966695e-06,
+      "loss": 1.7427,
+      "step": 2195
+    },
+    {
+      "epoch": 2.4065753424657537,
+      "grad_norm": 2.7030224800109863,
+      "learning_rate": 1.1291632818846467e-06,
+      "loss": 1.7216,
+      "step": 2196
+    },
+    {
+      "epoch": 2.4076712328767123,
+      "grad_norm": 2.4936161041259766,
+      "learning_rate": 1.127132412672624e-06,
+      "loss": 1.7033,
+      "step": 2197
+    },
+    {
+      "epoch": 2.4087671232876713,
+      "grad_norm": 2.5384793281555176,
+      "learning_rate": 1.1251015434606013e-06,
+      "loss": 1.7023,
+      "step": 2198
+    },
+    {
+      "epoch": 2.4098630136986303,
+      "grad_norm": 2.5271613597869873,
+      "learning_rate": 1.1230706742485786e-06,
+      "loss": 1.738,
+      "step": 2199
+    },
+    {
+      "epoch": 2.410958904109589,
+      "grad_norm": 2.557610034942627,
+      "learning_rate": 1.1210398050365558e-06,
+      "loss": 1.7927,
+      "step": 2200
+    },
+    {
+      "epoch": 2.410958904109589,
+      "eval_loss": 1.9795563220977783,
+      "eval_runtime": 37.0832,
+      "eval_samples_per_second": 26.966,
+      "eval_steps_per_second": 1.133,
+      "step": 2200
+    },
+    {
+      "epoch": 2.412054794520548,
+      "grad_norm": 2.5497145652770996,
+      "learning_rate": 1.1190089358245331e-06,
+      "loss": 1.7361,
+      "step": 2201
+    },
+    {
+      "epoch": 2.413150684931507,
+      "grad_norm": 2.622793436050415,
+      "learning_rate": 1.1169780666125102e-06,
+      "loss": 1.6624,
+      "step": 2202
+    },
+    {
+      "epoch": 2.4142465753424656,
+      "grad_norm": 2.694563865661621,
+      "learning_rate": 1.1149471974004875e-06,
+      "loss": 1.7755,
+      "step": 2203
+    },
+    {
+      "epoch": 2.4153424657534246,
+      "grad_norm": 2.68461012840271,
+      "learning_rate": 1.1129163281884647e-06,
+      "loss": 1.7132,
+      "step": 2204
+    },
+    {
+      "epoch": 2.4164383561643836,
+      "grad_norm": 2.65791654586792,
+      "learning_rate": 1.110885458976442e-06,
+      "loss": 1.7728,
+      "step": 2205
+    },
+    {
+      "epoch": 2.417534246575342,
+      "grad_norm": 2.46262788772583,
+      "learning_rate": 1.1088545897644193e-06,
+      "loss": 1.6973,
+      "step": 2206
+    },
+    {
+      "epoch": 2.4186301369863012,
+      "grad_norm": 2.5602169036865234,
+      "learning_rate": 1.1068237205523966e-06,
+      "loss": 1.7651,
+      "step": 2207
+    },
+    {
+      "epoch": 2.4197260273972603,
+      "grad_norm": 2.5492701530456543,
+      "learning_rate": 1.1047928513403736e-06,
+      "loss": 1.7674,
+      "step": 2208
+    },
+    {
+      "epoch": 2.4208219178082193,
+      "grad_norm": 2.798940420150757,
+      "learning_rate": 1.102761982128351e-06,
+      "loss": 1.7719,
+      "step": 2209
+    },
+    {
+      "epoch": 2.421917808219178,
+      "grad_norm": 2.6989545822143555,
+      "learning_rate": 1.1007311129163282e-06,
+      "loss": 1.6821,
+      "step": 2210
+    },
+    {
+      "epoch": 2.421917808219178,
+      "eval_loss": 1.9763859510421753,
+      "eval_runtime": 36.7821,
+      "eval_samples_per_second": 27.187,
+      "eval_steps_per_second": 1.142,
+      "step": 2210
+    },
+    {
+      "epoch": 2.423013698630137,
+      "grad_norm": 2.752720594406128,
+      "learning_rate": 1.0987002437043055e-06,
+      "loss": 1.7432,
+      "step": 2211
+    },
+    {
+      "epoch": 2.424109589041096,
+      "grad_norm": 2.6162850856781006,
+      "learning_rate": 1.0966693744922827e-06,
+      "loss": 1.718,
+      "step": 2212
+    },
+    {
+      "epoch": 2.425205479452055,
+      "grad_norm": 2.938995838165283,
+      "learning_rate": 1.09463850528026e-06,
+      "loss": 1.6761,
+      "step": 2213
+    },
+    {
+      "epoch": 2.4263013698630136,
+      "grad_norm": 2.7038822174072266,
+      "learning_rate": 1.0926076360682373e-06,
+      "loss": 1.7267,
+      "step": 2214
+    },
+    {
+      "epoch": 2.4273972602739726,
+      "grad_norm": 2.6025028228759766,
+      "learning_rate": 1.0905767668562146e-06,
+      "loss": 1.7021,
+      "step": 2215
+    },
+    {
+      "epoch": 2.4284931506849317,
+      "grad_norm": 2.5922977924346924,
+      "learning_rate": 1.0885458976441918e-06,
+      "loss": 1.7221,
+      "step": 2216
+    },
+    {
+      "epoch": 2.4295890410958902,
+      "grad_norm": 2.634716033935547,
+      "learning_rate": 1.0865150284321691e-06,
+      "loss": 1.8047,
+      "step": 2217
+    },
+    {
+      "epoch": 2.4306849315068493,
+      "grad_norm": 2.655155658721924,
+      "learning_rate": 1.0844841592201464e-06,
+      "loss": 1.7426,
+      "step": 2218
+    },
+    {
+      "epoch": 2.4317808219178083,
+      "grad_norm": 2.5151286125183105,
+      "learning_rate": 1.0824532900081237e-06,
+      "loss": 1.7627,
+      "step": 2219
+    },
+    {
+      "epoch": 2.432876712328767,
+      "grad_norm": 2.696597099304199,
+      "learning_rate": 1.080422420796101e-06,
+      "loss": 1.7325,
+      "step": 2220
+    },
+    {
+      "epoch": 2.432876712328767,
+      "eval_loss": 1.9818732738494873,
+      "eval_runtime": 36.9854,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 2220
+    },
+    {
+      "epoch": 2.433972602739726,
+      "grad_norm": 2.7444307804107666,
+      "learning_rate": 1.078391551584078e-06,
+      "loss": 1.743,
+      "step": 2221
+    },
+    {
+      "epoch": 2.435068493150685,
+      "grad_norm": 2.503706455230713,
+      "learning_rate": 1.0763606823720553e-06,
+      "loss": 1.7699,
+      "step": 2222
+    },
+    {
+      "epoch": 2.436164383561644,
+      "grad_norm": 2.546041965484619,
+      "learning_rate": 1.0743298131600326e-06,
+      "loss": 1.8077,
+      "step": 2223
+    },
+    {
+      "epoch": 2.4372602739726026,
+      "grad_norm": 2.5515291690826416,
+      "learning_rate": 1.0722989439480098e-06,
+      "loss": 1.7637,
+      "step": 2224
+    },
+    {
+      "epoch": 2.4383561643835616,
+      "grad_norm": 2.564180850982666,
+      "learning_rate": 1.070268074735987e-06,
+      "loss": 1.7299,
+      "step": 2225
+    },
+    {
+      "epoch": 2.4394520547945207,
+      "grad_norm": 2.532708168029785,
+      "learning_rate": 1.0682372055239644e-06,
+      "loss": 1.713,
+      "step": 2226
+    },
+    {
+      "epoch": 2.4405479452054797,
+      "grad_norm": 2.672358512878418,
+      "learning_rate": 1.0662063363119414e-06,
+      "loss": 1.7103,
+      "step": 2227
+    },
+    {
+      "epoch": 2.4416438356164383,
+      "grad_norm": 2.7836480140686035,
+      "learning_rate": 1.0641754670999187e-06,
+      "loss": 1.7129,
+      "step": 2228
+    },
+    {
+      "epoch": 2.4427397260273973,
+      "grad_norm": 2.6253440380096436,
+      "learning_rate": 1.062144597887896e-06,
+      "loss": 1.7389,
+      "step": 2229
+    },
+    {
+      "epoch": 2.4438356164383563,
+      "grad_norm": 2.5863685607910156,
+      "learning_rate": 1.0601137286758733e-06,
+      "loss": 1.7293,
+      "step": 2230
+    },
+    {
+      "epoch": 2.4438356164383563,
+      "eval_loss": 1.9771078824996948,
+      "eval_runtime": 37.0603,
+      "eval_samples_per_second": 26.983,
+      "eval_steps_per_second": 1.133,
+      "step": 2230
+    },
+    {
+      "epoch": 2.444931506849315,
+      "grad_norm": 2.732042074203491,
+      "learning_rate": 1.0580828594638505e-06,
+      "loss": 1.8012,
+      "step": 2231
+    },
+    {
+      "epoch": 2.446027397260274,
+      "grad_norm": 2.5275564193725586,
+      "learning_rate": 1.0560519902518278e-06,
+      "loss": 1.6482,
+      "step": 2232
+    },
+    {
+      "epoch": 2.447123287671233,
+      "grad_norm": 2.7054405212402344,
+      "learning_rate": 1.054021121039805e-06,
+      "loss": 1.7799,
+      "step": 2233
+    },
+    {
+      "epoch": 2.4482191780821916,
+      "grad_norm": 2.587719678878784,
+      "learning_rate": 1.0519902518277824e-06,
+      "loss": 1.6797,
+      "step": 2234
+    },
+    {
+      "epoch": 2.4493150684931506,
+      "grad_norm": 2.600907564163208,
+      "learning_rate": 1.0499593826157596e-06,
+      "loss": 1.6727,
+      "step": 2235
+    },
+    {
+      "epoch": 2.4504109589041096,
+      "grad_norm": 2.6509485244750977,
+      "learning_rate": 1.047928513403737e-06,
+      "loss": 1.7535,
+      "step": 2236
+    },
+    {
+      "epoch": 2.4515068493150687,
+      "grad_norm": 2.6827805042266846,
+      "learning_rate": 1.0458976441917142e-06,
+      "loss": 1.7285,
+      "step": 2237
+    },
+    {
+      "epoch": 2.4526027397260273,
+      "grad_norm": 2.7095749378204346,
+      "learning_rate": 1.0438667749796915e-06,
+      "loss": 1.6546,
+      "step": 2238
+    },
+    {
+      "epoch": 2.4536986301369863,
+      "grad_norm": 2.7757186889648438,
+      "learning_rate": 1.0418359057676688e-06,
+      "loss": 1.6659,
+      "step": 2239
+    },
+    {
+      "epoch": 2.4547945205479453,
+      "grad_norm": 2.7696826457977295,
+      "learning_rate": 1.0398050365556458e-06,
+      "loss": 1.7108,
+      "step": 2240
+    },
+    {
+      "epoch": 2.4547945205479453,
+      "eval_loss": 1.980337142944336,
+      "eval_runtime": 37.0883,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 2240
+    },
+    {
+      "epoch": 2.455890410958904,
+      "grad_norm": 2.96648907661438,
+      "learning_rate": 1.037774167343623e-06,
+      "loss": 1.77,
+      "step": 2241
+    },
+    {
+      "epoch": 2.456986301369863,
+      "grad_norm": 2.561461925506592,
+      "learning_rate": 1.0357432981316004e-06,
+      "loss": 1.6962,
+      "step": 2242
+    },
+    {
+      "epoch": 2.458082191780822,
+      "grad_norm": 2.5870988368988037,
+      "learning_rate": 1.0337124289195776e-06,
+      "loss": 1.7385,
+      "step": 2243
+    },
+    {
+      "epoch": 2.459178082191781,
+      "grad_norm": 2.607558488845825,
+      "learning_rate": 1.031681559707555e-06,
+      "loss": 1.7771,
+      "step": 2244
+    },
+    {
+      "epoch": 2.4602739726027396,
+      "grad_norm": 2.6373603343963623,
+      "learning_rate": 1.0296506904955322e-06,
+      "loss": 1.7845,
+      "step": 2245
+    },
+    {
+      "epoch": 2.4613698630136986,
+      "grad_norm": 2.5508599281311035,
+      "learning_rate": 1.0276198212835095e-06,
+      "loss": 1.7384,
+      "step": 2246
+    },
+    {
+      "epoch": 2.4624657534246577,
+      "grad_norm": 2.730696201324463,
+      "learning_rate": 1.0255889520714867e-06,
+      "loss": 1.7458,
+      "step": 2247
+    },
+    {
+      "epoch": 2.4635616438356163,
+      "grad_norm": 2.4905753135681152,
+      "learning_rate": 1.023558082859464e-06,
+      "loss": 1.7195,
+      "step": 2248
+    },
+    {
+      "epoch": 2.4646575342465753,
+      "grad_norm": 2.664184331893921,
+      "learning_rate": 1.0215272136474413e-06,
+      "loss": 1.6986,
+      "step": 2249
+    },
+    {
+      "epoch": 2.4657534246575343,
+      "grad_norm": 2.683645009994507,
+      "learning_rate": 1.0194963444354186e-06,
+      "loss": 1.7981,
+      "step": 2250
+    },
+    {
+      "epoch": 2.4657534246575343,
+      "eval_loss": 1.9768283367156982,
+      "eval_runtime": 37.1086,
+      "eval_samples_per_second": 26.948,
+      "eval_steps_per_second": 1.132,
+      "step": 2250
+    },
+    {
+      "epoch": 2.466849315068493,
+      "grad_norm": 2.633633852005005,
+      "learning_rate": 1.0174654752233958e-06,
+      "loss": 1.6807,
+      "step": 2251
+    },
+    {
+      "epoch": 2.467945205479452,
+      "grad_norm": 2.655794382095337,
+      "learning_rate": 1.015434606011373e-06,
+      "loss": 1.7705,
+      "step": 2252
+    },
+    {
+      "epoch": 2.469041095890411,
+      "grad_norm": 2.6457812786102295,
+      "learning_rate": 1.0134037367993502e-06,
+      "loss": 1.7201,
+      "step": 2253
+    },
+    {
+      "epoch": 2.47013698630137,
+      "grad_norm": 2.623656749725342,
+      "learning_rate": 1.0113728675873275e-06,
+      "loss": 1.7402,
+      "step": 2254
+    },
+    {
+      "epoch": 2.4712328767123286,
+      "grad_norm": 2.65552020072937,
+      "learning_rate": 1.0093419983753047e-06,
+      "loss": 1.7372,
+      "step": 2255
+    },
+    {
+      "epoch": 2.4723287671232876,
+      "grad_norm": 2.707531213760376,
+      "learning_rate": 1.007311129163282e-06,
+      "loss": 1.7299,
+      "step": 2256
+    },
+    {
+      "epoch": 2.4734246575342467,
+      "grad_norm": 2.504805564880371,
+      "learning_rate": 1.0052802599512593e-06,
+      "loss": 1.6858,
+      "step": 2257
+    },
+    {
+      "epoch": 2.4745205479452057,
+      "grad_norm": 2.584251642227173,
+      "learning_rate": 1.0032493907392364e-06,
+      "loss": 1.6867,
+      "step": 2258
+    },
+    {
+      "epoch": 2.4756164383561643,
+      "grad_norm": 2.5356390476226807,
+      "learning_rate": 1.0012185215272136e-06,
+      "loss": 1.7442,
+      "step": 2259
+    },
+    {
+      "epoch": 2.4767123287671233,
+      "grad_norm": 2.5929319858551025,
+      "learning_rate": 9.99187652315191e-07,
+      "loss": 1.6777,
+      "step": 2260
+    },
+    {
+      "epoch": 2.4767123287671233,
+      "eval_loss": 1.9772632122039795,
+      "eval_runtime": 37.0951,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 2260
+    },
+    {
+      "epoch": 2.4778082191780824,
+      "grad_norm": 2.6396026611328125,
+      "learning_rate": 9.971567831031682e-07,
+      "loss": 1.7638,
+      "step": 2261
+    },
+    {
+      "epoch": 2.478904109589041,
+      "grad_norm": 2.5710740089416504,
+      "learning_rate": 9.951259138911455e-07,
+      "loss": 1.7566,
+      "step": 2262
+    },
+    {
+      "epoch": 2.48,
+      "grad_norm": 2.5611767768859863,
+      "learning_rate": 9.930950446791227e-07,
+      "loss": 1.6639,
+      "step": 2263
+    },
+    {
+      "epoch": 2.481095890410959,
+      "grad_norm": 2.5390372276306152,
+      "learning_rate": 9.910641754671e-07,
+      "loss": 1.7887,
+      "step": 2264
+    },
+    {
+      "epoch": 2.4821917808219176,
+      "grad_norm": 2.6124560832977295,
+      "learning_rate": 9.890333062550773e-07,
+      "loss": 1.8221,
+      "step": 2265
+    },
+    {
+      "epoch": 2.4832876712328766,
+      "grad_norm": 2.6506896018981934,
+      "learning_rate": 9.870024370430546e-07,
+      "loss": 1.725,
+      "step": 2266
+    },
+    {
+      "epoch": 2.4843835616438357,
+      "grad_norm": 2.617889642715454,
+      "learning_rate": 9.849715678310318e-07,
+      "loss": 1.7566,
+      "step": 2267
+    },
+    {
+      "epoch": 2.4854794520547947,
+      "grad_norm": 2.571110486984253,
+      "learning_rate": 9.829406986190091e-07,
+      "loss": 1.7327,
+      "step": 2268
+    },
+    {
+      "epoch": 2.4865753424657533,
+      "grad_norm": 2.657665252685547,
+      "learning_rate": 9.809098294069864e-07,
+      "loss": 1.7471,
+      "step": 2269
+    },
+    {
+      "epoch": 2.4876712328767123,
+      "grad_norm": 2.5402324199676514,
+      "learning_rate": 9.788789601949637e-07,
+      "loss": 1.6597,
+      "step": 2270
+    },
+    {
+      "epoch": 2.4876712328767123,
+      "eval_loss": 1.975923776626587,
+      "eval_runtime": 37.0909,
+      "eval_samples_per_second": 26.961,
+      "eval_steps_per_second": 1.132,
+      "step": 2270
+    },
+    {
+      "epoch": 2.4887671232876714,
+      "grad_norm": 2.619765281677246,
+      "learning_rate": 9.768480909829407e-07,
+      "loss": 1.7434,
+      "step": 2271
+    },
+    {
+      "epoch": 2.48986301369863,
+      "grad_norm": 2.664238214492798,
+      "learning_rate": 9.74817221770918e-07,
+      "loss": 1.7937,
+      "step": 2272
+    },
+    {
+      "epoch": 2.490958904109589,
+      "grad_norm": 2.6743860244750977,
+      "learning_rate": 9.727863525588953e-07,
+      "loss": 1.7203,
+      "step": 2273
+    },
+    {
+      "epoch": 2.492054794520548,
+      "grad_norm": 2.695051670074463,
+      "learning_rate": 9.707554833468726e-07,
+      "loss": 1.7473,
+      "step": 2274
+    },
+    {
+      "epoch": 2.493150684931507,
+      "grad_norm": 2.6115036010742188,
+      "learning_rate": 9.687246141348498e-07,
+      "loss": 1.7358,
+      "step": 2275
+    },
+    {
+      "epoch": 2.4942465753424656,
+      "grad_norm": 2.7951300144195557,
+      "learning_rate": 9.666937449228271e-07,
+      "loss": 1.723,
+      "step": 2276
+    },
+    {
+      "epoch": 2.4953424657534247,
+      "grad_norm": 2.6242685317993164,
+      "learning_rate": 9.646628757108042e-07,
+      "loss": 1.7389,
+      "step": 2277
+    },
+    {
+      "epoch": 2.4964383561643837,
+      "grad_norm": 2.567150831222534,
+      "learning_rate": 9.626320064987814e-07,
+      "loss": 1.6775,
+      "step": 2278
+    },
+    {
+      "epoch": 2.4975342465753423,
+      "grad_norm": 2.6157355308532715,
+      "learning_rate": 9.606011372867587e-07,
+      "loss": 1.7624,
+      "step": 2279
+    },
+    {
+      "epoch": 2.4986301369863013,
+      "grad_norm": 2.5663843154907227,
+      "learning_rate": 9.58570268074736e-07,
+      "loss": 1.6955,
+      "step": 2280
+    },
+    {
+      "epoch": 2.4986301369863013,
+      "eval_loss": 1.9744558334350586,
+      "eval_runtime": 37.1094,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 2280
+    },
+    {
+      "epoch": 2.4997260273972604,
+      "grad_norm": 2.667611598968506,
+      "learning_rate": 9.565393988627133e-07,
+      "loss": 1.7261,
+      "step": 2281
+    },
+    {
+      "epoch": 2.500821917808219,
+      "grad_norm": 2.5702202320098877,
+      "learning_rate": 9.545085296506906e-07,
+      "loss": 1.7252,
+      "step": 2282
+    },
+    {
+      "epoch": 2.501917808219178,
+      "grad_norm": 2.5271430015563965,
+      "learning_rate": 9.524776604386677e-07,
+      "loss": 1.7407,
+      "step": 2283
+    },
+    {
+      "epoch": 2.503013698630137,
+      "grad_norm": 2.5804171562194824,
+      "learning_rate": 9.50446791226645e-07,
+      "loss": 1.739,
+      "step": 2284
+    },
+    {
+      "epoch": 2.504109589041096,
+      "grad_norm": 2.571345090866089,
+      "learning_rate": 9.484159220146223e-07,
+      "loss": 1.668,
+      "step": 2285
+    },
+    {
+      "epoch": 2.5052054794520546,
+      "grad_norm": 2.5697009563446045,
+      "learning_rate": 9.463850528025995e-07,
+      "loss": 1.7677,
+      "step": 2286
+    },
+    {
+      "epoch": 2.5063013698630137,
+      "grad_norm": 2.5846893787384033,
+      "learning_rate": 9.443541835905768e-07,
+      "loss": 1.7589,
+      "step": 2287
+    },
+    {
+      "epoch": 2.5073972602739727,
+      "grad_norm": 2.803380250930786,
+      "learning_rate": 9.423233143785541e-07,
+      "loss": 1.8104,
+      "step": 2288
+    },
+    {
+      "epoch": 2.5084931506849317,
+      "grad_norm": 2.6451611518859863,
+      "learning_rate": 9.402924451665313e-07,
+      "loss": 1.8061,
+      "step": 2289
+    },
+    {
+      "epoch": 2.5095890410958903,
+      "grad_norm": 2.6164097785949707,
+      "learning_rate": 9.382615759545085e-07,
+      "loss": 1.6899,
+      "step": 2290
+    },
+    {
+      "epoch": 2.5095890410958903,
+      "eval_loss": 1.9739093780517578,
+      "eval_runtime": 37.1222,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 2290
+    },
+    {
+      "epoch": 2.5106849315068493,
+      "grad_norm": 2.5874385833740234,
+      "learning_rate": 9.362307067424858e-07,
+      "loss": 1.7692,
+      "step": 2291
+    },
+    {
+      "epoch": 2.5117808219178084,
+      "grad_norm": 2.5705063343048096,
+      "learning_rate": 9.341998375304631e-07,
+      "loss": 1.7303,
+      "step": 2292
+    },
+    {
+      "epoch": 2.512876712328767,
+      "grad_norm": 2.6130380630493164,
+      "learning_rate": 9.321689683184404e-07,
+      "loss": 1.7892,
+      "step": 2293
+    },
+    {
+      "epoch": 2.513972602739726,
+      "grad_norm": 2.5991110801696777,
+      "learning_rate": 9.301380991064176e-07,
+      "loss": 1.6877,
+      "step": 2294
+    },
+    {
+      "epoch": 2.515068493150685,
+      "grad_norm": 2.6092689037323,
+      "learning_rate": 9.281072298943949e-07,
+      "loss": 1.6729,
+      "step": 2295
+    },
+    {
+      "epoch": 2.5161643835616436,
+      "grad_norm": 2.5563883781433105,
+      "learning_rate": 9.260763606823721e-07,
+      "loss": 1.7181,
+      "step": 2296
+    },
+    {
+      "epoch": 2.5172602739726027,
+      "grad_norm": 2.623948574066162,
+      "learning_rate": 9.240454914703494e-07,
+      "loss": 1.7528,
+      "step": 2297
+    },
+    {
+      "epoch": 2.5183561643835617,
+      "grad_norm": 2.661874294281006,
+      "learning_rate": 9.220146222583266e-07,
+      "loss": 1.7582,
+      "step": 2298
+    },
+    {
+      "epoch": 2.5194520547945203,
+      "grad_norm": 2.5936238765716553,
+      "learning_rate": 9.199837530463039e-07,
+      "loss": 1.6987,
+      "step": 2299
+    },
+    {
+      "epoch": 2.5205479452054793,
+      "grad_norm": 2.5134992599487305,
+      "learning_rate": 9.179528838342812e-07,
+      "loss": 1.7265,
+      "step": 2300
+    },
+    {
+      "epoch": 2.5205479452054793,
+      "eval_loss": 1.9737082719802856,
+      "eval_runtime": 37.1105,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 2300
+    },
+    {
+      "epoch": 2.5216438356164383,
+      "grad_norm": 2.627927541732788,
+      "learning_rate": 9.159220146222585e-07,
+      "loss": 1.7059,
+      "step": 2301
+    },
+    {
+      "epoch": 2.5227397260273974,
+      "grad_norm": 2.6419436931610107,
+      "learning_rate": 9.138911454102356e-07,
+      "loss": 1.7558,
+      "step": 2302
+    },
+    {
+      "epoch": 2.5238356164383564,
+      "grad_norm": 2.6184327602386475,
+      "learning_rate": 9.118602761982129e-07,
+      "loss": 1.6281,
+      "step": 2303
+    },
+    {
+      "epoch": 2.524931506849315,
+      "grad_norm": 2.6111767292022705,
+      "learning_rate": 9.098294069861902e-07,
+      "loss": 1.7408,
+      "step": 2304
+    },
+    {
+      "epoch": 2.526027397260274,
+      "grad_norm": 2.6416513919830322,
+      "learning_rate": 9.077985377741675e-07,
+      "loss": 1.7397,
+      "step": 2305
+    },
+    {
+      "epoch": 2.527123287671233,
+      "grad_norm": 2.5487682819366455,
+      "learning_rate": 9.057676685621447e-07,
+      "loss": 1.6941,
+      "step": 2306
+    },
+    {
+      "epoch": 2.5282191780821917,
+      "grad_norm": 2.6250264644622803,
+      "learning_rate": 9.03736799350122e-07,
+      "loss": 1.66,
+      "step": 2307
+    },
+    {
+      "epoch": 2.5293150684931507,
+      "grad_norm": 2.6054790019989014,
+      "learning_rate": 9.017059301380991e-07,
+      "loss": 1.7829,
+      "step": 2308
+    },
+    {
+      "epoch": 2.5304109589041097,
+      "grad_norm": 2.751016616821289,
+      "learning_rate": 8.996750609260764e-07,
+      "loss": 1.7693,
+      "step": 2309
+    },
+    {
+      "epoch": 2.5315068493150683,
+      "grad_norm": 2.6321675777435303,
+      "learning_rate": 8.976441917140536e-07,
+      "loss": 1.7971,
+      "step": 2310
+    },
+    {
+      "epoch": 2.5315068493150683,
+      "eval_loss": 1.9757944345474243,
+      "eval_runtime": 36.77,
+      "eval_samples_per_second": 27.196,
+      "eval_steps_per_second": 1.142,
+      "step": 2310
+    },
+    {
+      "epoch": 2.5326027397260273,
+      "grad_norm": 2.642702341079712,
+      "learning_rate": 8.956133225020309e-07,
+      "loss": 1.7703,
+      "step": 2311
+    },
+    {
+      "epoch": 2.5336986301369864,
+      "grad_norm": 2.614771604537964,
+      "learning_rate": 8.935824532900082e-07,
+      "loss": 1.8009,
+      "step": 2312
+    },
+    {
+      "epoch": 2.534794520547945,
+      "grad_norm": 2.514615058898926,
+      "learning_rate": 8.915515840779855e-07,
+      "loss": 1.6954,
+      "step": 2313
+    },
+    {
+      "epoch": 2.535890410958904,
+      "grad_norm": 2.644207715988159,
+      "learning_rate": 8.895207148659626e-07,
+      "loss": 1.6779,
+      "step": 2314
+    },
+    {
+      "epoch": 2.536986301369863,
+      "grad_norm": 2.7844860553741455,
+      "learning_rate": 8.874898456539399e-07,
+      "loss": 1.7595,
+      "step": 2315
+    },
+    {
+      "epoch": 2.538082191780822,
+      "grad_norm": 2.714629650115967,
+      "learning_rate": 8.854589764419172e-07,
+      "loss": 1.7448,
+      "step": 2316
+    },
+    {
+      "epoch": 2.539178082191781,
+      "grad_norm": 2.57271409034729,
+      "learning_rate": 8.834281072298945e-07,
+      "loss": 1.7438,
+      "step": 2317
+    },
+    {
+      "epoch": 2.5402739726027397,
+      "grad_norm": 2.6566264629364014,
+      "learning_rate": 8.813972380178717e-07,
+      "loss": 1.6924,
+      "step": 2318
+    },
+    {
+      "epoch": 2.5413698630136987,
+      "grad_norm": 2.619744300842285,
+      "learning_rate": 8.79366368805849e-07,
+      "loss": 1.7317,
+      "step": 2319
+    },
+    {
+      "epoch": 2.5424657534246577,
+      "grad_norm": 2.6249327659606934,
+      "learning_rate": 8.773354995938262e-07,
+      "loss": 1.7467,
+      "step": 2320
+    },
+    {
+      "epoch": 2.5424657534246577,
+      "eval_loss": 1.9757587909698486,
+      "eval_runtime": 36.9112,
+      "eval_samples_per_second": 27.092,
+      "eval_steps_per_second": 1.138,
+      "step": 2320
+    },
+    {
+      "epoch": 2.5435616438356163,
+      "grad_norm": 2.584317445755005,
+      "learning_rate": 8.753046303818035e-07,
+      "loss": 1.6962,
+      "step": 2321
+    },
+    {
+      "epoch": 2.5446575342465754,
+      "grad_norm": 2.7346856594085693,
+      "learning_rate": 8.732737611697807e-07,
+      "loss": 1.6863,
+      "step": 2322
+    },
+    {
+      "epoch": 2.5457534246575344,
+      "grad_norm": 2.574941635131836,
+      "learning_rate": 8.71242891957758e-07,
+      "loss": 1.7271,
+      "step": 2323
+    },
+    {
+      "epoch": 2.546849315068493,
+      "grad_norm": 2.546217441558838,
+      "learning_rate": 8.692120227457353e-07,
+      "loss": 1.751,
+      "step": 2324
+    },
+    {
+      "epoch": 2.547945205479452,
+      "grad_norm": 2.534414768218994,
+      "learning_rate": 8.671811535337126e-07,
+      "loss": 1.7692,
+      "step": 2325
+    },
+    {
+      "epoch": 2.549041095890411,
+      "grad_norm": 2.625565528869629,
+      "learning_rate": 8.651502843216898e-07,
+      "loss": 1.7135,
+      "step": 2326
+    },
+    {
+      "epoch": 2.5501369863013696,
+      "grad_norm": 2.715276002883911,
+      "learning_rate": 8.63119415109667e-07,
+      "loss": 1.8093,
+      "step": 2327
+    },
+    {
+      "epoch": 2.5512328767123287,
+      "grad_norm": 2.6494171619415283,
+      "learning_rate": 8.610885458976443e-07,
+      "loss": 1.7589,
+      "step": 2328
+    },
+    {
+      "epoch": 2.5523287671232877,
+      "grad_norm": 2.6840505599975586,
+      "learning_rate": 8.590576766856216e-07,
+      "loss": 1.7484,
+      "step": 2329
+    },
+    {
+      "epoch": 2.5534246575342463,
+      "grad_norm": 2.510047435760498,
+      "learning_rate": 8.570268074735988e-07,
+      "loss": 1.6985,
+      "step": 2330
+    },
+    {
+      "epoch": 2.5534246575342463,
+      "eval_loss": 1.9751781225204468,
+      "eval_runtime": 36.9998,
+      "eval_samples_per_second": 27.027,
+      "eval_steps_per_second": 1.135,
+      "step": 2330
+    },
+    {
+      "epoch": 2.5545205479452053,
+      "grad_norm": 2.6174721717834473,
+      "learning_rate": 8.549959382615761e-07,
+      "loss": 1.6873,
+      "step": 2331
+    },
+    {
+      "epoch": 2.5556164383561644,
+      "grad_norm": 2.570403814315796,
+      "learning_rate": 8.529650690495534e-07,
+      "loss": 1.7058,
+      "step": 2332
+    },
+    {
+      "epoch": 2.5567123287671234,
+      "grad_norm": 2.8619518280029297,
+      "learning_rate": 8.509341998375305e-07,
+      "loss": 1.7342,
+      "step": 2333
+    },
+    {
+      "epoch": 2.5578082191780824,
+      "grad_norm": 2.6569507122039795,
+      "learning_rate": 8.489033306255077e-07,
+      "loss": 1.6736,
+      "step": 2334
+    },
+    {
+      "epoch": 2.558904109589041,
+      "grad_norm": 2.5437800884246826,
+      "learning_rate": 8.46872461413485e-07,
+      "loss": 1.648,
+      "step": 2335
+    },
+    {
+      "epoch": 2.56,
+      "grad_norm": 2.652275323867798,
+      "learning_rate": 8.448415922014623e-07,
+      "loss": 1.692,
+      "step": 2336
+    },
+    {
+      "epoch": 2.561095890410959,
+      "grad_norm": 2.6103672981262207,
+      "learning_rate": 8.428107229894396e-07,
+      "loss": 1.7033,
+      "step": 2337
+    },
+    {
+      "epoch": 2.5621917808219177,
+      "grad_norm": 2.6765308380126953,
+      "learning_rate": 8.407798537774168e-07,
+      "loss": 1.7473,
+      "step": 2338
+    },
+    {
+      "epoch": 2.5632876712328767,
+      "grad_norm": 2.708935260772705,
+      "learning_rate": 8.38748984565394e-07,
+      "loss": 1.7346,
+      "step": 2339
+    },
+    {
+      "epoch": 2.5643835616438357,
+      "grad_norm": 2.597233772277832,
+      "learning_rate": 8.367181153533713e-07,
+      "loss": 1.6927,
+      "step": 2340
+    },
+    {
+      "epoch": 2.5643835616438357,
+      "eval_loss": 1.9768768548965454,
+      "eval_runtime": 37.0191,
+      "eval_samples_per_second": 27.013,
+      "eval_steps_per_second": 1.135,
+      "step": 2340
+    },
+    {
+      "epoch": 2.5654794520547943,
+      "grad_norm": 2.624343156814575,
+      "learning_rate": 8.346872461413486e-07,
+      "loss": 1.7219,
+      "step": 2341
+    },
+    {
+      "epoch": 2.5665753424657534,
+      "grad_norm": 2.675095796585083,
+      "learning_rate": 8.326563769293258e-07,
+      "loss": 1.7136,
+      "step": 2342
+    },
+    {
+      "epoch": 2.5676712328767124,
+      "grad_norm": 2.575009822845459,
+      "learning_rate": 8.306255077173031e-07,
+      "loss": 1.7187,
+      "step": 2343
+    },
+    {
+      "epoch": 2.568767123287671,
+      "grad_norm": 2.539958953857422,
+      "learning_rate": 8.285946385052804e-07,
+      "loss": 1.8232,
+      "step": 2344
+    },
+    {
+      "epoch": 2.56986301369863,
+      "grad_norm": 2.613976001739502,
+      "learning_rate": 8.265637692932575e-07,
+      "loss": 1.6746,
+      "step": 2345
+    },
+    {
+      "epoch": 2.570958904109589,
+      "grad_norm": 2.5001742839813232,
+      "learning_rate": 8.245329000812348e-07,
+      "loss": 1.7542,
+      "step": 2346
+    },
+    {
+      "epoch": 2.572054794520548,
+      "grad_norm": 2.6975460052490234,
+      "learning_rate": 8.225020308692121e-07,
+      "loss": 1.7434,
+      "step": 2347
+    },
+    {
+      "epoch": 2.573150684931507,
+      "grad_norm": 2.6900603771209717,
+      "learning_rate": 8.204711616571894e-07,
+      "loss": 1.6843,
+      "step": 2348
+    },
+    {
+      "epoch": 2.5742465753424657,
+      "grad_norm": 2.5895581245422363,
+      "learning_rate": 8.184402924451667e-07,
+      "loss": 1.7279,
+      "step": 2349
+    },
+    {
+      "epoch": 2.5753424657534247,
+      "grad_norm": 2.5448262691497803,
+      "learning_rate": 8.164094232331439e-07,
+      "loss": 1.6743,
+      "step": 2350
+    },
+    {
+      "epoch": 2.5753424657534247,
+      "eval_loss": 1.9762015342712402,
+      "eval_runtime": 37.0186,
+      "eval_samples_per_second": 27.013,
+      "eval_steps_per_second": 1.135,
+      "step": 2350
+    },
+    {
+      "epoch": 2.5764383561643838,
+      "grad_norm": 2.6080751419067383,
+      "learning_rate": 8.14378554021121e-07,
+      "loss": 1.7525,
+      "step": 2351
+    },
+    {
+      "epoch": 2.5775342465753424,
+      "grad_norm": 2.6307425498962402,
+      "learning_rate": 8.123476848090983e-07,
+      "loss": 1.7304,
+      "step": 2352
+    },
+    {
+      "epoch": 2.5786301369863014,
+      "grad_norm": 2.7442290782928467,
+      "learning_rate": 8.103168155970755e-07,
+      "loss": 1.7018,
+      "step": 2353
+    },
+    {
+      "epoch": 2.5797260273972604,
+      "grad_norm": 2.5672149658203125,
+      "learning_rate": 8.082859463850528e-07,
+      "loss": 1.7115,
+      "step": 2354
+    },
+    {
+      "epoch": 2.580821917808219,
+      "grad_norm": 2.7201836109161377,
+      "learning_rate": 8.062550771730301e-07,
+      "loss": 1.7461,
+      "step": 2355
+    },
+    {
+      "epoch": 2.581917808219178,
+      "grad_norm": 2.7145047187805176,
+      "learning_rate": 8.042242079610074e-07,
+      "loss": 1.7169,
+      "step": 2356
+    },
+    {
+      "epoch": 2.583013698630137,
+      "grad_norm": 2.7162413597106934,
+      "learning_rate": 8.021933387489846e-07,
+      "loss": 1.71,
+      "step": 2357
+    },
+    {
+      "epoch": 2.5841095890410957,
+      "grad_norm": 2.515169858932495,
+      "learning_rate": 8.001624695369618e-07,
+      "loss": 1.7016,
+      "step": 2358
+    },
+    {
+      "epoch": 2.5852054794520547,
+      "grad_norm": 2.620922803878784,
+      "learning_rate": 7.981316003249391e-07,
+      "loss": 1.7251,
+      "step": 2359
+    },
+    {
+      "epoch": 2.5863013698630137,
+      "grad_norm": 2.8066515922546387,
+      "learning_rate": 7.961007311129164e-07,
+      "loss": 1.787,
+      "step": 2360
+    },
+    {
+      "epoch": 2.5863013698630137,
+      "eval_loss": 1.9778543710708618,
+      "eval_runtime": 37.0094,
+      "eval_samples_per_second": 27.02,
+      "eval_steps_per_second": 1.135,
+      "step": 2360
+    },
+    {
+      "epoch": 2.5873972602739728,
+      "grad_norm": 2.5266482830047607,
+      "learning_rate": 7.940698619008936e-07,
+      "loss": 1.6654,
+      "step": 2361
+    },
+    {
+      "epoch": 2.5884931506849314,
+      "grad_norm": 2.6038002967834473,
+      "learning_rate": 7.920389926888709e-07,
+      "loss": 1.7532,
+      "step": 2362
+    },
+    {
+      "epoch": 2.5895890410958904,
+      "grad_norm": 2.716871976852417,
+      "learning_rate": 7.900081234768482e-07,
+      "loss": 1.8235,
+      "step": 2363
+    },
+    {
+      "epoch": 2.5906849315068494,
+      "grad_norm": 2.6962411403656006,
+      "learning_rate": 7.879772542648254e-07,
+      "loss": 1.6927,
+      "step": 2364
+    },
+    {
+      "epoch": 2.5917808219178085,
+      "grad_norm": 2.670116662979126,
+      "learning_rate": 7.859463850528026e-07,
+      "loss": 1.7448,
+      "step": 2365
+    },
+    {
+      "epoch": 2.592876712328767,
+      "grad_norm": 2.6886184215545654,
+      "learning_rate": 7.839155158407799e-07,
+      "loss": 1.7369,
+      "step": 2366
+    },
+    {
+      "epoch": 2.593972602739726,
+      "grad_norm": 2.6570565700531006,
+      "learning_rate": 7.818846466287572e-07,
+      "loss": 1.7368,
+      "step": 2367
+    },
+    {
+      "epoch": 2.595068493150685,
+      "grad_norm": 2.551548719406128,
+      "learning_rate": 7.798537774167345e-07,
+      "loss": 1.7357,
+      "step": 2368
+    },
+    {
+      "epoch": 2.5961643835616437,
+      "grad_norm": 2.646280527114868,
+      "learning_rate": 7.778229082047117e-07,
+      "loss": 1.7614,
+      "step": 2369
+    },
+    {
+      "epoch": 2.5972602739726027,
+      "grad_norm": 2.626239776611328,
+      "learning_rate": 7.757920389926889e-07,
+      "loss": 1.694,
+      "step": 2370
+    },
+    {
+      "epoch": 2.5972602739726027,
+      "eval_loss": 1.9757529497146606,
+      "eval_runtime": 36.9858,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 2370
+    },
+    {
+      "epoch": 2.5983561643835618,
+      "grad_norm": 2.6001148223876953,
+      "learning_rate": 7.737611697806662e-07,
+      "loss": 1.7342,
+      "step": 2371
+    },
+    {
+      "epoch": 2.5994520547945204,
+      "grad_norm": 2.5435168743133545,
+      "learning_rate": 7.717303005686435e-07,
+      "loss": 1.7031,
+      "step": 2372
+    },
+    {
+      "epoch": 2.6005479452054794,
+      "grad_norm": 2.652031183242798,
+      "learning_rate": 7.696994313566207e-07,
+      "loss": 1.7598,
+      "step": 2373
+    },
+    {
+      "epoch": 2.6016438356164384,
+      "grad_norm": 2.635340690612793,
+      "learning_rate": 7.67668562144598e-07,
+      "loss": 1.8008,
+      "step": 2374
+    },
+    {
+      "epoch": 2.602739726027397,
+      "grad_norm": 2.6935770511627197,
+      "learning_rate": 7.656376929325753e-07,
+      "loss": 1.6821,
+      "step": 2375
+    },
+    {
+      "epoch": 2.603835616438356,
+      "grad_norm": 2.621493339538574,
+      "learning_rate": 7.636068237205524e-07,
+      "loss": 1.6844,
+      "step": 2376
+    },
+    {
+      "epoch": 2.604931506849315,
+      "grad_norm": 2.6231918334960938,
+      "learning_rate": 7.615759545085296e-07,
+      "loss": 1.7622,
+      "step": 2377
+    },
+    {
+      "epoch": 2.606027397260274,
+      "grad_norm": 2.608091115951538,
+      "learning_rate": 7.595450852965069e-07,
+      "loss": 1.6772,
+      "step": 2378
+    },
+    {
+      "epoch": 2.607123287671233,
+      "grad_norm": 2.7701456546783447,
+      "learning_rate": 7.575142160844842e-07,
+      "loss": 1.7139,
+      "step": 2379
+    },
+    {
+      "epoch": 2.6082191780821917,
+      "grad_norm": 2.6648287773132324,
+      "learning_rate": 7.554833468724615e-07,
+      "loss": 1.7525,
+      "step": 2380
+    },
+    {
+      "epoch": 2.6082191780821917,
+      "eval_loss": 1.9792429208755493,
+      "eval_runtime": 37.0103,
+      "eval_samples_per_second": 27.02,
+      "eval_steps_per_second": 1.135,
+      "step": 2380
+    },
+    {
+      "epoch": 2.6093150684931508,
+      "grad_norm": 2.690938949584961,
+      "learning_rate": 7.534524776604387e-07,
+      "loss": 1.679,
+      "step": 2381
+    },
+    {
+      "epoch": 2.61041095890411,
+      "grad_norm": 2.658691167831421,
+      "learning_rate": 7.51421608448416e-07,
+      "loss": 1.7033,
+      "step": 2382
+    },
+    {
+      "epoch": 2.6115068493150684,
+      "grad_norm": 2.576317310333252,
+      "learning_rate": 7.493907392363932e-07,
+      "loss": 1.7527,
+      "step": 2383
+    },
+    {
+      "epoch": 2.6126027397260274,
+      "grad_norm": 2.5913753509521484,
+      "learning_rate": 7.473598700243705e-07,
+      "loss": 1.702,
+      "step": 2384
+    },
+    {
+      "epoch": 2.6136986301369864,
+      "grad_norm": 2.61553692817688,
+      "learning_rate": 7.453290008123477e-07,
+      "loss": 1.6568,
+      "step": 2385
+    },
+    {
+      "epoch": 2.614794520547945,
+      "grad_norm": 2.6631391048431396,
+      "learning_rate": 7.43298131600325e-07,
+      "loss": 1.7216,
+      "step": 2386
+    },
+    {
+      "epoch": 2.615890410958904,
+      "grad_norm": 2.6120619773864746,
+      "learning_rate": 7.412672623883023e-07,
+      "loss": 1.7278,
+      "step": 2387
+    },
+    {
+      "epoch": 2.616986301369863,
+      "grad_norm": 2.6166980266571045,
+      "learning_rate": 7.392363931762796e-07,
+      "loss": 1.7275,
+      "step": 2388
+    },
+    {
+      "epoch": 2.6180821917808217,
+      "grad_norm": 2.7203712463378906,
+      "learning_rate": 7.372055239642567e-07,
+      "loss": 1.6657,
+      "step": 2389
+    },
+    {
+      "epoch": 2.6191780821917807,
+      "grad_norm": 2.6785244941711426,
+      "learning_rate": 7.35174654752234e-07,
+      "loss": 1.7206,
+      "step": 2390
+    },
+    {
+      "epoch": 2.6191780821917807,
+      "eval_loss": 1.9763447046279907,
+      "eval_runtime": 37.0352,
+      "eval_samples_per_second": 27.001,
+      "eval_steps_per_second": 1.134,
+      "step": 2390
+    },
+    {
+      "epoch": 2.6202739726027398,
+      "grad_norm": 2.72511887550354,
+      "learning_rate": 7.331437855402113e-07,
+      "loss": 1.7474,
+      "step": 2391
+    },
+    {
+      "epoch": 2.621369863013699,
+      "grad_norm": 2.6428632736206055,
+      "learning_rate": 7.311129163281886e-07,
+      "loss": 1.7694,
+      "step": 2392
+    },
+    {
+      "epoch": 2.6224657534246574,
+      "grad_norm": 2.6593737602233887,
+      "learning_rate": 7.290820471161658e-07,
+      "loss": 1.6883,
+      "step": 2393
+    },
+    {
+      "epoch": 2.6235616438356164,
+      "grad_norm": 2.6048357486724854,
+      "learning_rate": 7.270511779041431e-07,
+      "loss": 1.7457,
+      "step": 2394
+    },
+    {
+      "epoch": 2.6246575342465754,
+      "grad_norm": 2.6435704231262207,
+      "learning_rate": 7.250203086921203e-07,
+      "loss": 1.7508,
+      "step": 2395
+    },
+    {
+      "epoch": 2.6257534246575345,
+      "grad_norm": 2.609815835952759,
+      "learning_rate": 7.229894394800976e-07,
+      "loss": 1.6603,
+      "step": 2396
+    },
+    {
+      "epoch": 2.626849315068493,
+      "grad_norm": 2.5431087017059326,
+      "learning_rate": 7.209585702680748e-07,
+      "loss": 1.725,
+      "step": 2397
+    },
+    {
+      "epoch": 2.627945205479452,
+      "grad_norm": 2.6712050437927246,
+      "learning_rate": 7.189277010560521e-07,
+      "loss": 1.7428,
+      "step": 2398
+    },
+    {
+      "epoch": 2.629041095890411,
+      "grad_norm": 2.6398820877075195,
+      "learning_rate": 7.168968318440294e-07,
+      "loss": 1.6834,
+      "step": 2399
+    },
+    {
+      "epoch": 2.6301369863013697,
+      "grad_norm": 2.5986127853393555,
+      "learning_rate": 7.148659626320067e-07,
+      "loss": 1.6985,
+      "step": 2400
+    },
+    {
+      "epoch": 2.6301369863013697,
+      "eval_loss": 1.9727513790130615,
+      "eval_runtime": 37.0552,
+      "eval_samples_per_second": 26.987,
+      "eval_steps_per_second": 1.133,
+      "step": 2400
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2736,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.481542582468608e+17,
+  "train_batch_size": 3,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/training_args.bin b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..b7adbe5a2dbf27f05befec6cc65fbc93e90eefd0
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aaa882279512a8c278376c0c601206933f3c5662ddcd3c13c53bf32ae3680d14
+size 6011
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2400/zero_to_fp32.py
@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/config.json
@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "meta-llama/Llama-3.2-3B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "torch_dtype": "float16",
+  "transformers_version": "4.45.1",
+  "use_cache": true,
+  "vocab_size": 128256
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/generation_config.json
@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": 128001,
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.45.1"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/latest b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/latest
new file mode 100644
index 0000000000000000000000000000000000000000..98f8bed9a5485ee900d9931cc06950de69499848
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/latest
@@ -0,0 +1 @@
+global_step2500
\ No newline at end of file
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/model.safetensors.index.json
new file mode 100644
index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/model.safetensors.index.json
@@ -0,0 +1,262 @@
+{
+  "metadata": {
+    "total_size": 7213504512
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00002-of-00002.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/rng_state_0.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..f084998befba4c7b08cfeba516748af5fcfa6071
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8bfdf5ccc88de7823e7dc6081ffb5cb38ed71058da20638fbf7f61d94c58874c
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/rng_state_1.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..a9f12d82a8fc9d14b8b35ce910d38bc891b2dafa
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa288bebadeaf65eb824b833596a55b793d13b5c957085cf3d7dc8c0e6de2ddf
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/rng_state_2.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..cf71113371921ea2e7fb737a20484b1a62f38481
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a08b70b84bf331849cc389e31e88e5bf289f53f5f48e4ac516e4a788f9bfae7d
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/scheduler.pt b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..bec29e5be2144ffb82ae168b0fa37e3576366d60
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20714c35ae35f55976c8c397055b166efcbc327d5c3799903c9ed73908591c4b
+size 627
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/tokenizer.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb
+size 17210374
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128256": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|end_of_text|>",
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "[PAD]",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..02ccc3586457cdbbc1e3c6a944206708979ed48f
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/trainer_state.json
@@ -0,0 +1,19533 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.73972602739726,
+  "eval_steps": 10,
+  "global_step": 2500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001095890410958904,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0636,
+      "step": 1
+    },
+    {
+      "epoch": 0.002191780821917808,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0501,
+      "step": 2
+    },
+    {
+      "epoch": 0.0032876712328767125,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0238,
+      "step": 3
+    },
+    {
+      "epoch": 0.004383561643835616,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0346,
+      "step": 4
+    },
+    {
+      "epoch": 0.005479452054794521,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0108,
+      "step": 5
+    },
+    {
+      "epoch": 0.006575342465753425,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.131,
+      "step": 6
+    },
+    {
+      "epoch": 0.007671232876712329,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0488,
+      "step": 7
+    },
+    {
+      "epoch": 0.008767123287671232,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0905,
+      "step": 8
+    },
+    {
+      "epoch": 0.009863013698630137,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0493,
+      "step": 9
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.1038,
+      "step": 10
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "eval_loss": 3.100078821182251,
+      "eval_runtime": 37.5838,
+      "eval_samples_per_second": 26.607,
+      "eval_steps_per_second": 1.118,
+      "step": 10
+    },
+    {
+      "epoch": 0.012054794520547946,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0804,
+      "step": 11
+    },
+    {
+      "epoch": 0.01315068493150685,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0053,
+      "step": 12
+    },
+    {
+      "epoch": 0.014246575342465753,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0607,
+      "step": 13
+    },
+    {
+      "epoch": 0.015342465753424657,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0539,
+      "step": 14
+    },
+    {
+      "epoch": 0.01643835616438356,
+      "grad_norm": 5.420984745025635,
+      "learning_rate": 1.8248175182481753e-08,
+      "loss": 3.0941,
+      "step": 15
+    },
+    {
+      "epoch": 0.017534246575342465,
+      "grad_norm": 5.76840877532959,
+      "learning_rate": 3.6496350364963505e-08,
+      "loss": 3.036,
+      "step": 16
+    },
+    {
+      "epoch": 0.01863013698630137,
+      "grad_norm": 5.905357360839844,
+      "learning_rate": 5.4744525547445264e-08,
+      "loss": 3.0491,
+      "step": 17
+    },
+    {
+      "epoch": 0.019726027397260273,
+      "grad_norm": 5.981553554534912,
+      "learning_rate": 7.299270072992701e-08,
+      "loss": 3.0775,
+      "step": 18
+    },
+    {
+      "epoch": 0.020821917808219178,
+      "grad_norm": 6.240172863006592,
+      "learning_rate": 9.124087591240876e-08,
+      "loss": 3.0345,
+      "step": 19
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "grad_norm": 5.59375,
+      "learning_rate": 1.0948905109489053e-07,
+      "loss": 3.1063,
+      "step": 20
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "eval_loss": 3.098820686340332,
+      "eval_runtime": 37.2695,
+      "eval_samples_per_second": 26.832,
+      "eval_steps_per_second": 1.127,
+      "step": 20
+    },
+    {
+      "epoch": 0.023013698630136987,
+      "grad_norm": 5.819795608520508,
+      "learning_rate": 1.2773722627737227e-07,
+      "loss": 3.0808,
+      "step": 21
+    },
+    {
+      "epoch": 0.02410958904109589,
+      "grad_norm": 5.7714762687683105,
+      "learning_rate": 1.4598540145985402e-07,
+      "loss": 3.0055,
+      "step": 22
+    },
+    {
+      "epoch": 0.025205479452054796,
+      "grad_norm": 5.51552152633667,
+      "learning_rate": 1.642335766423358e-07,
+      "loss": 3.0396,
+      "step": 23
+    },
+    {
+      "epoch": 0.0263013698630137,
+      "grad_norm": 5.594012260437012,
+      "learning_rate": 1.8248175182481753e-07,
+      "loss": 3.0824,
+      "step": 24
+    },
+    {
+      "epoch": 0.0273972602739726,
+      "grad_norm": 5.1974029541015625,
+      "learning_rate": 2.0072992700729928e-07,
+      "loss": 3.08,
+      "step": 25
+    },
+    {
+      "epoch": 0.028493150684931506,
+      "grad_norm": 4.733968257904053,
+      "learning_rate": 2.1897810218978106e-07,
+      "loss": 3.0597,
+      "step": 26
+    },
+    {
+      "epoch": 0.02958904109589041,
+      "grad_norm": 4.717319011688232,
+      "learning_rate": 2.3722627737226278e-07,
+      "loss": 3.0696,
+      "step": 27
+    },
+    {
+      "epoch": 0.030684931506849315,
+      "grad_norm": 4.6303534507751465,
+      "learning_rate": 2.5547445255474454e-07,
+      "loss": 3.0273,
+      "step": 28
+    },
+    {
+      "epoch": 0.03178082191780822,
+      "grad_norm": 4.413999557495117,
+      "learning_rate": 2.737226277372263e-07,
+      "loss": 3.0202,
+      "step": 29
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "grad_norm": 4.516043186187744,
+      "learning_rate": 2.9197080291970804e-07,
+      "loss": 2.9926,
+      "step": 30
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "eval_loss": 3.0561797618865967,
+      "eval_runtime": 37.0351,
+      "eval_samples_per_second": 27.001,
+      "eval_steps_per_second": 1.134,
+      "step": 30
+    },
+    {
+      "epoch": 0.03397260273972603,
+      "grad_norm": 3.7179207801818848,
+      "learning_rate": 3.102189781021898e-07,
+      "loss": 3.0086,
+      "step": 31
+    },
+    {
+      "epoch": 0.03506849315068493,
+      "grad_norm": 3.7840425968170166,
+      "learning_rate": 3.284671532846716e-07,
+      "loss": 2.9986,
+      "step": 32
+    },
+    {
+      "epoch": 0.03616438356164384,
+      "grad_norm": 3.843116521835327,
+      "learning_rate": 3.467153284671533e-07,
+      "loss": 2.9677,
+      "step": 33
+    },
+    {
+      "epoch": 0.03726027397260274,
+      "grad_norm": 3.8528714179992676,
+      "learning_rate": 3.6496350364963505e-07,
+      "loss": 2.9732,
+      "step": 34
+    },
+    {
+      "epoch": 0.038356164383561646,
+      "grad_norm": 3.76963210105896,
+      "learning_rate": 3.8321167883211686e-07,
+      "loss": 2.9855,
+      "step": 35
+    },
+    {
+      "epoch": 0.03945205479452055,
+      "grad_norm": 3.6240015029907227,
+      "learning_rate": 4.0145985401459856e-07,
+      "loss": 3.0117,
+      "step": 36
+    },
+    {
+      "epoch": 0.040547945205479455,
+      "grad_norm": 3.1380624771118164,
+      "learning_rate": 4.197080291970803e-07,
+      "loss": 2.9481,
+      "step": 37
+    },
+    {
+      "epoch": 0.041643835616438356,
+      "grad_norm": 3.565329074859619,
+      "learning_rate": 4.379562043795621e-07,
+      "loss": 2.9175,
+      "step": 38
+    },
+    {
+      "epoch": 0.042739726027397264,
+      "grad_norm": 3.8062448501586914,
+      "learning_rate": 4.562043795620438e-07,
+      "loss": 3.0074,
+      "step": 39
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "grad_norm": 3.412046432495117,
+      "learning_rate": 4.7445255474452557e-07,
+      "loss": 2.9119,
+      "step": 40
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "eval_loss": 2.9874370098114014,
+      "eval_runtime": 37.0514,
+      "eval_samples_per_second": 26.99,
+      "eval_steps_per_second": 1.134,
+      "step": 40
+    },
+    {
+      "epoch": 0.044931506849315066,
+      "grad_norm": 3.227437973022461,
+      "learning_rate": 4.927007299270073e-07,
+      "loss": 2.9436,
+      "step": 41
+    },
+    {
+      "epoch": 0.046027397260273974,
+      "grad_norm": 2.9633443355560303,
+      "learning_rate": 5.109489051094891e-07,
+      "loss": 2.967,
+      "step": 42
+    },
+    {
+      "epoch": 0.047123287671232875,
+      "grad_norm": 2.942678928375244,
+      "learning_rate": 5.291970802919708e-07,
+      "loss": 2.9375,
+      "step": 43
+    },
+    {
+      "epoch": 0.04821917808219178,
+      "grad_norm": 2.9806888103485107,
+      "learning_rate": 5.474452554744526e-07,
+      "loss": 2.9474,
+      "step": 44
+    },
+    {
+      "epoch": 0.049315068493150684,
+      "grad_norm": 3.000063419342041,
+      "learning_rate": 5.656934306569343e-07,
+      "loss": 2.9289,
+      "step": 45
+    },
+    {
+      "epoch": 0.05041095890410959,
+      "grad_norm": 3.1481194496154785,
+      "learning_rate": 5.839416058394161e-07,
+      "loss": 2.8869,
+      "step": 46
+    },
+    {
+      "epoch": 0.05150684931506849,
+      "grad_norm": 2.949202060699463,
+      "learning_rate": 6.021897810218979e-07,
+      "loss": 2.8783,
+      "step": 47
+    },
+    {
+      "epoch": 0.0526027397260274,
+      "grad_norm": 2.706526279449463,
+      "learning_rate": 6.204379562043796e-07,
+      "loss": 2.9024,
+      "step": 48
+    },
+    {
+      "epoch": 0.0536986301369863,
+      "grad_norm": 2.7317655086517334,
+      "learning_rate": 6.386861313868614e-07,
+      "loss": 2.8325,
+      "step": 49
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "grad_norm": 2.25644588470459,
+      "learning_rate": 6.569343065693432e-07,
+      "loss": 2.9409,
+      "step": 50
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "eval_loss": 2.909320116043091,
+      "eval_runtime": 37.0714,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 50
+    },
+    {
+      "epoch": 0.05589041095890411,
+      "grad_norm": 2.8789613246917725,
+      "learning_rate": 6.751824817518248e-07,
+      "loss": 2.8459,
+      "step": 51
+    },
+    {
+      "epoch": 0.05698630136986301,
+      "grad_norm": 2.686873435974121,
+      "learning_rate": 6.934306569343066e-07,
+      "loss": 2.8653,
+      "step": 52
+    },
+    {
+      "epoch": 0.05808219178082192,
+      "grad_norm": 2.54382586479187,
+      "learning_rate": 7.116788321167883e-07,
+      "loss": 2.7973,
+      "step": 53
+    },
+    {
+      "epoch": 0.05917808219178082,
+      "grad_norm": 2.1871724128723145,
+      "learning_rate": 7.299270072992701e-07,
+      "loss": 2.7982,
+      "step": 54
+    },
+    {
+      "epoch": 0.06027397260273973,
+      "grad_norm": 2.145303249359131,
+      "learning_rate": 7.48175182481752e-07,
+      "loss": 2.8195,
+      "step": 55
+    },
+    {
+      "epoch": 0.06136986301369863,
+      "grad_norm": 2.1635069847106934,
+      "learning_rate": 7.664233576642337e-07,
+      "loss": 2.825,
+      "step": 56
+    },
+    {
+      "epoch": 0.06246575342465754,
+      "grad_norm": 2.3530752658843994,
+      "learning_rate": 7.846715328467154e-07,
+      "loss": 2.8407,
+      "step": 57
+    },
+    {
+      "epoch": 0.06356164383561644,
+      "grad_norm": 2.2455942630767822,
+      "learning_rate": 8.029197080291971e-07,
+      "loss": 2.8182,
+      "step": 58
+    },
+    {
+      "epoch": 0.06465753424657535,
+      "grad_norm": 1.8985673189163208,
+      "learning_rate": 8.211678832116789e-07,
+      "loss": 2.772,
+      "step": 59
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "grad_norm": 1.9699499607086182,
+      "learning_rate": 8.394160583941606e-07,
+      "loss": 2.7367,
+      "step": 60
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "eval_loss": 2.8451859951019287,
+      "eval_runtime": 37.0793,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 60
+    },
+    {
+      "epoch": 0.06684931506849315,
+      "grad_norm": 2.159547805786133,
+      "learning_rate": 8.576642335766425e-07,
+      "loss": 2.7635,
+      "step": 61
+    },
+    {
+      "epoch": 0.06794520547945206,
+      "grad_norm": 2.12705659866333,
+      "learning_rate": 8.759124087591242e-07,
+      "loss": 2.7624,
+      "step": 62
+    },
+    {
+      "epoch": 0.06904109589041096,
+      "grad_norm": 1.8393241167068481,
+      "learning_rate": 8.94160583941606e-07,
+      "loss": 2.7763,
+      "step": 63
+    },
+    {
+      "epoch": 0.07013698630136986,
+      "grad_norm": 1.683522343635559,
+      "learning_rate": 9.124087591240876e-07,
+      "loss": 2.7762,
+      "step": 64
+    },
+    {
+      "epoch": 0.07123287671232877,
+      "grad_norm": 1.9260376691818237,
+      "learning_rate": 9.306569343065694e-07,
+      "loss": 2.7418,
+      "step": 65
+    },
+    {
+      "epoch": 0.07232876712328767,
+      "grad_norm": 2.09322190284729,
+      "learning_rate": 9.489051094890511e-07,
+      "loss": 2.7625,
+      "step": 66
+    },
+    {
+      "epoch": 0.07342465753424658,
+      "grad_norm": 1.9183429479599,
+      "learning_rate": 9.67153284671533e-07,
+      "loss": 2.7453,
+      "step": 67
+    },
+    {
+      "epoch": 0.07452054794520548,
+      "grad_norm": 1.9792778491973877,
+      "learning_rate": 9.854014598540146e-07,
+      "loss": 2.7858,
+      "step": 68
+    },
+    {
+      "epoch": 0.07561643835616438,
+      "grad_norm": 2.1397311687469482,
+      "learning_rate": 1.0036496350364965e-06,
+      "loss": 2.7242,
+      "step": 69
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "grad_norm": 1.8556463718414307,
+      "learning_rate": 1.0218978102189781e-06,
+      "loss": 2.6853,
+      "step": 70
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "eval_loss": 2.7787046432495117,
+      "eval_runtime": 37.1017,
+      "eval_samples_per_second": 26.953,
+      "eval_steps_per_second": 1.132,
+      "step": 70
+    },
+    {
+      "epoch": 0.07780821917808219,
+      "grad_norm": 1.9776800870895386,
+      "learning_rate": 1.04014598540146e-06,
+      "loss": 2.749,
+      "step": 71
+    },
+    {
+      "epoch": 0.0789041095890411,
+      "grad_norm": 2.3972630500793457,
+      "learning_rate": 1.0583941605839416e-06,
+      "loss": 2.6457,
+      "step": 72
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 2.090210199356079,
+      "learning_rate": 1.0766423357664233e-06,
+      "loss": 2.6401,
+      "step": 73
+    },
+    {
+      "epoch": 0.08109589041095891,
+      "grad_norm": 2.4117701053619385,
+      "learning_rate": 1.0948905109489052e-06,
+      "loss": 2.6649,
+      "step": 74
+    },
+    {
+      "epoch": 0.0821917808219178,
+      "grad_norm": 2.4845173358917236,
+      "learning_rate": 1.113138686131387e-06,
+      "loss": 2.6855,
+      "step": 75
+    },
+    {
+      "epoch": 0.08328767123287671,
+      "grad_norm": 1.7779242992401123,
+      "learning_rate": 1.1313868613138687e-06,
+      "loss": 2.6926,
+      "step": 76
+    },
+    {
+      "epoch": 0.08438356164383562,
+      "grad_norm": 2.3163535594940186,
+      "learning_rate": 1.1496350364963505e-06,
+      "loss": 2.6302,
+      "step": 77
+    },
+    {
+      "epoch": 0.08547945205479453,
+      "grad_norm": 2.2296528816223145,
+      "learning_rate": 1.1678832116788322e-06,
+      "loss": 2.6054,
+      "step": 78
+    },
+    {
+      "epoch": 0.08657534246575342,
+      "grad_norm": 2.253431797027588,
+      "learning_rate": 1.1861313868613138e-06,
+      "loss": 2.6104,
+      "step": 79
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "grad_norm": 2.490609645843506,
+      "learning_rate": 1.2043795620437959e-06,
+      "loss": 2.6078,
+      "step": 80
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "eval_loss": 2.706103563308716,
+      "eval_runtime": 37.0892,
+      "eval_samples_per_second": 26.962,
+      "eval_steps_per_second": 1.132,
+      "step": 80
+    },
+    {
+      "epoch": 0.08876712328767124,
+      "grad_norm": 2.3048248291015625,
+      "learning_rate": 1.2226277372262775e-06,
+      "loss": 2.6096,
+      "step": 81
+    },
+    {
+      "epoch": 0.08986301369863013,
+      "grad_norm": 2.440666913986206,
+      "learning_rate": 1.2408759124087592e-06,
+      "loss": 2.6412,
+      "step": 82
+    },
+    {
+      "epoch": 0.09095890410958904,
+      "grad_norm": 1.9456374645233154,
+      "learning_rate": 1.259124087591241e-06,
+      "loss": 2.591,
+      "step": 83
+    },
+    {
+      "epoch": 0.09205479452054795,
+      "grad_norm": 2.598172664642334,
+      "learning_rate": 1.2773722627737229e-06,
+      "loss": 2.6093,
+      "step": 84
+    },
+    {
+      "epoch": 0.09315068493150686,
+      "grad_norm": 2.2587032318115234,
+      "learning_rate": 1.2956204379562045e-06,
+      "loss": 2.5949,
+      "step": 85
+    },
+    {
+      "epoch": 0.09424657534246575,
+      "grad_norm": 2.2927448749542236,
+      "learning_rate": 1.3138686131386864e-06,
+      "loss": 2.5464,
+      "step": 86
+    },
+    {
+      "epoch": 0.09534246575342466,
+      "grad_norm": 2.5940468311309814,
+      "learning_rate": 1.332116788321168e-06,
+      "loss": 2.5613,
+      "step": 87
+    },
+    {
+      "epoch": 0.09643835616438357,
+      "grad_norm": 2.073207378387451,
+      "learning_rate": 1.3503649635036497e-06,
+      "loss": 2.4955,
+      "step": 88
+    },
+    {
+      "epoch": 0.09753424657534246,
+      "grad_norm": 3.211766242980957,
+      "learning_rate": 1.3686131386861315e-06,
+      "loss": 2.5315,
+      "step": 89
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "grad_norm": 2.2590038776397705,
+      "learning_rate": 1.3868613138686132e-06,
+      "loss": 2.4641,
+      "step": 90
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "eval_loss": 2.625087261199951,
+      "eval_runtime": 37.0852,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 90
+    },
+    {
+      "epoch": 0.09972602739726028,
+      "grad_norm": 2.3959836959838867,
+      "learning_rate": 1.405109489051095e-06,
+      "loss": 2.5537,
+      "step": 91
+    },
+    {
+      "epoch": 0.10082191780821918,
+      "grad_norm": 2.596761465072632,
+      "learning_rate": 1.4233576642335767e-06,
+      "loss": 2.5135,
+      "step": 92
+    },
+    {
+      "epoch": 0.10191780821917808,
+      "grad_norm": 2.475759267807007,
+      "learning_rate": 1.4416058394160583e-06,
+      "loss": 2.5436,
+      "step": 93
+    },
+    {
+      "epoch": 0.10301369863013699,
+      "grad_norm": 3.0613298416137695,
+      "learning_rate": 1.4598540145985402e-06,
+      "loss": 2.5066,
+      "step": 94
+    },
+    {
+      "epoch": 0.10410958904109589,
+      "grad_norm": 2.406097650527954,
+      "learning_rate": 1.478102189781022e-06,
+      "loss": 2.4646,
+      "step": 95
+    },
+    {
+      "epoch": 0.1052054794520548,
+      "grad_norm": 2.5872504711151123,
+      "learning_rate": 1.496350364963504e-06,
+      "loss": 2.5351,
+      "step": 96
+    },
+    {
+      "epoch": 0.1063013698630137,
+      "grad_norm": 2.4294471740722656,
+      "learning_rate": 1.5145985401459856e-06,
+      "loss": 2.4953,
+      "step": 97
+    },
+    {
+      "epoch": 0.1073972602739726,
+      "grad_norm": 2.674884080886841,
+      "learning_rate": 1.5328467153284674e-06,
+      "loss": 2.5133,
+      "step": 98
+    },
+    {
+      "epoch": 0.10849315068493151,
+      "grad_norm": 2.5096852779388428,
+      "learning_rate": 1.551094890510949e-06,
+      "loss": 2.4693,
+      "step": 99
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "grad_norm": 2.7390334606170654,
+      "learning_rate": 1.5693430656934307e-06,
+      "loss": 2.4195,
+      "step": 100
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "eval_loss": 2.55997371673584,
+      "eval_runtime": 37.0732,
+      "eval_samples_per_second": 26.974,
+      "eval_steps_per_second": 1.133,
+      "step": 100
+    },
+    {
+      "epoch": 0.11068493150684931,
+      "grad_norm": 2.8887312412261963,
+      "learning_rate": 1.5875912408759126e-06,
+      "loss": 2.4193,
+      "step": 101
+    },
+    {
+      "epoch": 0.11178082191780822,
+      "grad_norm": 3.0721421241760254,
+      "learning_rate": 1.6058394160583942e-06,
+      "loss": 2.4709,
+      "step": 102
+    },
+    {
+      "epoch": 0.11287671232876713,
+      "grad_norm": 2.728156566619873,
+      "learning_rate": 1.624087591240876e-06,
+      "loss": 2.5105,
+      "step": 103
+    },
+    {
+      "epoch": 0.11397260273972602,
+      "grad_norm": 3.1214346885681152,
+      "learning_rate": 1.6423357664233577e-06,
+      "loss": 2.4316,
+      "step": 104
+    },
+    {
+      "epoch": 0.11506849315068493,
+      "grad_norm": 2.9249517917633057,
+      "learning_rate": 1.6605839416058394e-06,
+      "loss": 2.3737,
+      "step": 105
+    },
+    {
+      "epoch": 0.11616438356164384,
+      "grad_norm": 2.6646173000335693,
+      "learning_rate": 1.6788321167883212e-06,
+      "loss": 2.4165,
+      "step": 106
+    },
+    {
+      "epoch": 0.11726027397260275,
+      "grad_norm": 3.108067750930786,
+      "learning_rate": 1.6970802919708029e-06,
+      "loss": 2.4217,
+      "step": 107
+    },
+    {
+      "epoch": 0.11835616438356164,
+      "grad_norm": 3.142338991165161,
+      "learning_rate": 1.715328467153285e-06,
+      "loss": 2.3527,
+      "step": 108
+    },
+    {
+      "epoch": 0.11945205479452055,
+      "grad_norm": 2.9078011512756348,
+      "learning_rate": 1.7335766423357666e-06,
+      "loss": 2.3927,
+      "step": 109
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "grad_norm": 3.40766978263855,
+      "learning_rate": 1.7518248175182485e-06,
+      "loss": 2.4021,
+      "step": 110
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "eval_loss": 2.5031485557556152,
+      "eval_runtime": 36.8147,
+      "eval_samples_per_second": 27.163,
+      "eval_steps_per_second": 1.141,
+      "step": 110
+    },
+    {
+      "epoch": 0.12164383561643835,
+      "grad_norm": 4.396062850952148,
+      "learning_rate": 1.77007299270073e-06,
+      "loss": 2.4373,
+      "step": 111
+    },
+    {
+      "epoch": 0.12273972602739726,
+      "grad_norm": 4.0441365242004395,
+      "learning_rate": 1.788321167883212e-06,
+      "loss": 2.4162,
+      "step": 112
+    },
+    {
+      "epoch": 0.12383561643835617,
+      "grad_norm": 2.8869969844818115,
+      "learning_rate": 1.8065693430656936e-06,
+      "loss": 2.4515,
+      "step": 113
+    },
+    {
+      "epoch": 0.12493150684931507,
+      "grad_norm": 3.4833669662475586,
+      "learning_rate": 1.8248175182481753e-06,
+      "loss": 2.3965,
+      "step": 114
+    },
+    {
+      "epoch": 0.12602739726027398,
+      "grad_norm": 2.9996068477630615,
+      "learning_rate": 1.8430656934306571e-06,
+      "loss": 2.3374,
+      "step": 115
+    },
+    {
+      "epoch": 0.12712328767123288,
+      "grad_norm": 2.701371192932129,
+      "learning_rate": 1.8613138686131388e-06,
+      "loss": 2.3608,
+      "step": 116
+    },
+    {
+      "epoch": 0.12821917808219177,
+      "grad_norm": 2.9491288661956787,
+      "learning_rate": 1.8795620437956206e-06,
+      "loss": 2.355,
+      "step": 117
+    },
+    {
+      "epoch": 0.1293150684931507,
+      "grad_norm": 2.9266560077667236,
+      "learning_rate": 1.8978102189781023e-06,
+      "loss": 2.3475,
+      "step": 118
+    },
+    {
+      "epoch": 0.1304109589041096,
+      "grad_norm": 3.0419797897338867,
+      "learning_rate": 1.916058394160584e-06,
+      "loss": 2.3475,
+      "step": 119
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "grad_norm": 2.4531214237213135,
+      "learning_rate": 1.934306569343066e-06,
+      "loss": 2.3724,
+      "step": 120
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "eval_loss": 2.455423593521118,
+      "eval_runtime": 37.0304,
+      "eval_samples_per_second": 27.005,
+      "eval_steps_per_second": 1.134,
+      "step": 120
+    },
+    {
+      "epoch": 0.1326027397260274,
+      "grad_norm": 3.572502851486206,
+      "learning_rate": 1.952554744525548e-06,
+      "loss": 2.2671,
+      "step": 121
+    },
+    {
+      "epoch": 0.1336986301369863,
+      "grad_norm": 3.5408504009246826,
+      "learning_rate": 1.9708029197080293e-06,
+      "loss": 2.3133,
+      "step": 122
+    },
+    {
+      "epoch": 0.13479452054794522,
+      "grad_norm": 2.488725423812866,
+      "learning_rate": 1.989051094890511e-06,
+      "loss": 2.3141,
+      "step": 123
+    },
+    {
+      "epoch": 0.1358904109589041,
+      "grad_norm": 3.241276264190674,
+      "learning_rate": 2.007299270072993e-06,
+      "loss": 2.318,
+      "step": 124
+    },
+    {
+      "epoch": 0.136986301369863,
+      "grad_norm": 2.853959798812866,
+      "learning_rate": 2.025547445255475e-06,
+      "loss": 2.3712,
+      "step": 125
+    },
+    {
+      "epoch": 0.13808219178082193,
+      "grad_norm": 2.9798169136047363,
+      "learning_rate": 2.0437956204379563e-06,
+      "loss": 2.3072,
+      "step": 126
+    },
+    {
+      "epoch": 0.13917808219178082,
+      "grad_norm": 2.8529601097106934,
+      "learning_rate": 2.062043795620438e-06,
+      "loss": 2.318,
+      "step": 127
+    },
+    {
+      "epoch": 0.14027397260273972,
+      "grad_norm": 3.475337028503418,
+      "learning_rate": 2.08029197080292e-06,
+      "loss": 2.31,
+      "step": 128
+    },
+    {
+      "epoch": 0.14136986301369864,
+      "grad_norm": 4.213955879211426,
+      "learning_rate": 2.0985401459854014e-06,
+      "loss": 2.3734,
+      "step": 129
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "grad_norm": 3.0833330154418945,
+      "learning_rate": 2.1167883211678833e-06,
+      "loss": 2.2602,
+      "step": 130
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "eval_loss": 2.4205682277679443,
+      "eval_runtime": 37.0798,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 130
+    },
+    {
+      "epoch": 0.14356164383561643,
+      "grad_norm": 3.411235809326172,
+      "learning_rate": 2.135036496350365e-06,
+      "loss": 2.3255,
+      "step": 131
+    },
+    {
+      "epoch": 0.14465753424657535,
+      "grad_norm": 2.882053852081299,
+      "learning_rate": 2.1532846715328466e-06,
+      "loss": 2.3483,
+      "step": 132
+    },
+    {
+      "epoch": 0.14575342465753424,
+      "grad_norm": 3.2436420917510986,
+      "learning_rate": 2.171532846715329e-06,
+      "loss": 2.2492,
+      "step": 133
+    },
+    {
+      "epoch": 0.14684931506849316,
+      "grad_norm": 2.929466962814331,
+      "learning_rate": 2.1897810218978103e-06,
+      "loss": 2.2647,
+      "step": 134
+    },
+    {
+      "epoch": 0.14794520547945206,
+      "grad_norm": 2.9344966411590576,
+      "learning_rate": 2.208029197080292e-06,
+      "loss": 2.3061,
+      "step": 135
+    },
+    {
+      "epoch": 0.14904109589041095,
+      "grad_norm": 3.2567203044891357,
+      "learning_rate": 2.226277372262774e-06,
+      "loss": 2.2746,
+      "step": 136
+    },
+    {
+      "epoch": 0.15013698630136987,
+      "grad_norm": 3.1494054794311523,
+      "learning_rate": 2.244525547445256e-06,
+      "loss": 2.2866,
+      "step": 137
+    },
+    {
+      "epoch": 0.15123287671232877,
+      "grad_norm": 2.723390817642212,
+      "learning_rate": 2.2627737226277373e-06,
+      "loss": 2.3109,
+      "step": 138
+    },
+    {
+      "epoch": 0.15232876712328766,
+      "grad_norm": 2.8904967308044434,
+      "learning_rate": 2.281021897810219e-06,
+      "loss": 2.301,
+      "step": 139
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "grad_norm": 2.8899483680725098,
+      "learning_rate": 2.299270072992701e-06,
+      "loss": 2.2637,
+      "step": 140
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "eval_loss": 2.3952724933624268,
+      "eval_runtime": 37.1272,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 140
+    },
+    {
+      "epoch": 0.15452054794520548,
+      "grad_norm": 3.5309066772460938,
+      "learning_rate": 2.3175182481751825e-06,
+      "loss": 2.2572,
+      "step": 141
+    },
+    {
+      "epoch": 0.15561643835616437,
+      "grad_norm": 2.8369667530059814,
+      "learning_rate": 2.3357664233576643e-06,
+      "loss": 2.2712,
+      "step": 142
+    },
+    {
+      "epoch": 0.1567123287671233,
+      "grad_norm": 3.629161834716797,
+      "learning_rate": 2.354014598540146e-06,
+      "loss": 2.2301,
+      "step": 143
+    },
+    {
+      "epoch": 0.1578082191780822,
+      "grad_norm": 3.5313000679016113,
+      "learning_rate": 2.3722627737226276e-06,
+      "loss": 2.2712,
+      "step": 144
+    },
+    {
+      "epoch": 0.1589041095890411,
+      "grad_norm": 3.1328771114349365,
+      "learning_rate": 2.39051094890511e-06,
+      "loss": 2.2706,
+      "step": 145
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 4.6992950439453125,
+      "learning_rate": 2.4087591240875918e-06,
+      "loss": 2.219,
+      "step": 146
+    },
+    {
+      "epoch": 0.1610958904109589,
+      "grad_norm": 5.230221748352051,
+      "learning_rate": 2.427007299270073e-06,
+      "loss": 2.2396,
+      "step": 147
+    },
+    {
+      "epoch": 0.16219178082191782,
+      "grad_norm": 3.5680761337280273,
+      "learning_rate": 2.445255474452555e-06,
+      "loss": 2.2434,
+      "step": 148
+    },
+    {
+      "epoch": 0.1632876712328767,
+      "grad_norm": 3.1642441749572754,
+      "learning_rate": 2.463503649635037e-06,
+      "loss": 2.2856,
+      "step": 149
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "grad_norm": 2.694085121154785,
+      "learning_rate": 2.4817518248175183e-06,
+      "loss": 2.2006,
+      "step": 150
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "eval_loss": 2.37199068069458,
+      "eval_runtime": 37.1342,
+      "eval_samples_per_second": 26.929,
+      "eval_steps_per_second": 1.131,
+      "step": 150
+    },
+    {
+      "epoch": 0.16547945205479453,
+      "grad_norm": 2.9868853092193604,
+      "learning_rate": 2.5e-06,
+      "loss": 2.1694,
+      "step": 151
+    },
+    {
+      "epoch": 0.16657534246575342,
+      "grad_norm": 2.793532371520996,
+      "learning_rate": 2.518248175182482e-06,
+      "loss": 2.2965,
+      "step": 152
+    },
+    {
+      "epoch": 0.16767123287671232,
+      "grad_norm": 2.60152530670166,
+      "learning_rate": 2.5364963503649635e-06,
+      "loss": 2.1954,
+      "step": 153
+    },
+    {
+      "epoch": 0.16876712328767124,
+      "grad_norm": 2.6192219257354736,
+      "learning_rate": 2.5547445255474458e-06,
+      "loss": 2.1909,
+      "step": 154
+    },
+    {
+      "epoch": 0.16986301369863013,
+      "grad_norm": 2.467344045639038,
+      "learning_rate": 2.5729927007299272e-06,
+      "loss": 2.2438,
+      "step": 155
+    },
+    {
+      "epoch": 0.17095890410958905,
+      "grad_norm": 2.772012948989868,
+      "learning_rate": 2.591240875912409e-06,
+      "loss": 2.2213,
+      "step": 156
+    },
+    {
+      "epoch": 0.17205479452054795,
+      "grad_norm": 2.613675355911255,
+      "learning_rate": 2.6094890510948905e-06,
+      "loss": 2.2213,
+      "step": 157
+    },
+    {
+      "epoch": 0.17315068493150684,
+      "grad_norm": 2.849245071411133,
+      "learning_rate": 2.627737226277373e-06,
+      "loss": 2.255,
+      "step": 158
+    },
+    {
+      "epoch": 0.17424657534246576,
+      "grad_norm": 2.8388969898223877,
+      "learning_rate": 2.6459854014598542e-06,
+      "loss": 2.2572,
+      "step": 159
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "grad_norm": 2.5219321250915527,
+      "learning_rate": 2.664233576642336e-06,
+      "loss": 2.2253,
+      "step": 160
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "eval_loss": 2.3509061336517334,
+      "eval_runtime": 37.0953,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 160
+    },
+    {
+      "epoch": 0.17643835616438355,
+      "grad_norm": 2.8460891246795654,
+      "learning_rate": 2.6824817518248175e-06,
+      "loss": 2.2009,
+      "step": 161
+    },
+    {
+      "epoch": 0.17753424657534247,
+      "grad_norm": 2.673377275466919,
+      "learning_rate": 2.7007299270072994e-06,
+      "loss": 2.2211,
+      "step": 162
+    },
+    {
+      "epoch": 0.17863013698630137,
+      "grad_norm": 2.8514182567596436,
+      "learning_rate": 2.718978102189781e-06,
+      "loss": 2.1955,
+      "step": 163
+    },
+    {
+      "epoch": 0.17972602739726026,
+      "grad_norm": 3.530385971069336,
+      "learning_rate": 2.737226277372263e-06,
+      "loss": 2.2536,
+      "step": 164
+    },
+    {
+      "epoch": 0.18082191780821918,
+      "grad_norm": 2.733394145965576,
+      "learning_rate": 2.755474452554745e-06,
+      "loss": 2.2367,
+      "step": 165
+    },
+    {
+      "epoch": 0.18191780821917808,
+      "grad_norm": 3.871150016784668,
+      "learning_rate": 2.7737226277372264e-06,
+      "loss": 2.1949,
+      "step": 166
+    },
+    {
+      "epoch": 0.18301369863013697,
+      "grad_norm": 3.7782530784606934,
+      "learning_rate": 2.7919708029197082e-06,
+      "loss": 2.2979,
+      "step": 167
+    },
+    {
+      "epoch": 0.1841095890410959,
+      "grad_norm": 3.094599485397339,
+      "learning_rate": 2.81021897810219e-06,
+      "loss": 2.2512,
+      "step": 168
+    },
+    {
+      "epoch": 0.1852054794520548,
+      "grad_norm": 3.2166848182678223,
+      "learning_rate": 2.828467153284672e-06,
+      "loss": 2.1784,
+      "step": 169
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "grad_norm": 3.4127073287963867,
+      "learning_rate": 2.8467153284671534e-06,
+      "loss": 2.2241,
+      "step": 170
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "eval_loss": 2.333245038986206,
+      "eval_runtime": 37.0402,
+      "eval_samples_per_second": 26.998,
+      "eval_steps_per_second": 1.134,
+      "step": 170
+    },
+    {
+      "epoch": 0.1873972602739726,
+      "grad_norm": 2.677579879760742,
+      "learning_rate": 2.8649635036496353e-06,
+      "loss": 2.135,
+      "step": 171
+    },
+    {
+      "epoch": 0.1884931506849315,
+      "grad_norm": 2.9602150917053223,
+      "learning_rate": 2.8832116788321167e-06,
+      "loss": 2.1916,
+      "step": 172
+    },
+    {
+      "epoch": 0.18958904109589042,
+      "grad_norm": 2.580122232437134,
+      "learning_rate": 2.901459854014599e-06,
+      "loss": 2.1477,
+      "step": 173
+    },
+    {
+      "epoch": 0.19068493150684931,
+      "grad_norm": 2.652217388153076,
+      "learning_rate": 2.9197080291970804e-06,
+      "loss": 2.2394,
+      "step": 174
+    },
+    {
+      "epoch": 0.1917808219178082,
+      "grad_norm": 2.6887567043304443,
+      "learning_rate": 2.9379562043795623e-06,
+      "loss": 2.2245,
+      "step": 175
+    },
+    {
+      "epoch": 0.19287671232876713,
+      "grad_norm": 2.768165111541748,
+      "learning_rate": 2.956204379562044e-06,
+      "loss": 2.2017,
+      "step": 176
+    },
+    {
+      "epoch": 0.19397260273972602,
+      "grad_norm": 2.9375691413879395,
+      "learning_rate": 2.9744525547445256e-06,
+      "loss": 2.2464,
+      "step": 177
+    },
+    {
+      "epoch": 0.19506849315068492,
+      "grad_norm": 2.8364248275756836,
+      "learning_rate": 2.992700729927008e-06,
+      "loss": 2.2155,
+      "step": 178
+    },
+    {
+      "epoch": 0.19616438356164384,
+      "grad_norm": 2.843487024307251,
+      "learning_rate": 3.0109489051094893e-06,
+      "loss": 2.1577,
+      "step": 179
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "grad_norm": 3.005230665206909,
+      "learning_rate": 3.029197080291971e-06,
+      "loss": 2.1549,
+      "step": 180
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "eval_loss": 2.3209409713745117,
+      "eval_runtime": 37.022,
+      "eval_samples_per_second": 27.011,
+      "eval_steps_per_second": 1.134,
+      "step": 180
+    },
+    {
+      "epoch": 0.19835616438356166,
+      "grad_norm": 2.490260601043701,
+      "learning_rate": 3.0474452554744526e-06,
+      "loss": 2.2313,
+      "step": 181
+    },
+    {
+      "epoch": 0.19945205479452055,
+      "grad_norm": 2.973118543624878,
+      "learning_rate": 3.065693430656935e-06,
+      "loss": 2.2316,
+      "step": 182
+    },
+    {
+      "epoch": 0.20054794520547944,
+      "grad_norm": 3.3308680057525635,
+      "learning_rate": 3.0839416058394163e-06,
+      "loss": 2.2087,
+      "step": 183
+    },
+    {
+      "epoch": 0.20164383561643837,
+      "grad_norm": 3.1440727710723877,
+      "learning_rate": 3.102189781021898e-06,
+      "loss": 2.2256,
+      "step": 184
+    },
+    {
+      "epoch": 0.20273972602739726,
+      "grad_norm": 2.5696325302124023,
+      "learning_rate": 3.1204379562043796e-06,
+      "loss": 2.2093,
+      "step": 185
+    },
+    {
+      "epoch": 0.20383561643835615,
+      "grad_norm": 3.1225037574768066,
+      "learning_rate": 3.1386861313868614e-06,
+      "loss": 2.1425,
+      "step": 186
+    },
+    {
+      "epoch": 0.20493150684931508,
+      "grad_norm": 3.650743246078491,
+      "learning_rate": 3.156934306569343e-06,
+      "loss": 2.1339,
+      "step": 187
+    },
+    {
+      "epoch": 0.20602739726027397,
+      "grad_norm": 2.924072027206421,
+      "learning_rate": 3.175182481751825e-06,
+      "loss": 2.2363,
+      "step": 188
+    },
+    {
+      "epoch": 0.20712328767123286,
+      "grad_norm": 2.9561989307403564,
+      "learning_rate": 3.193430656934307e-06,
+      "loss": 2.1934,
+      "step": 189
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "grad_norm": 2.731393814086914,
+      "learning_rate": 3.2116788321167884e-06,
+      "loss": 2.1718,
+      "step": 190
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "eval_loss": 2.3087782859802246,
+      "eval_runtime": 37.0161,
+      "eval_samples_per_second": 27.015,
+      "eval_steps_per_second": 1.135,
+      "step": 190
+    },
+    {
+      "epoch": 0.20931506849315068,
+      "grad_norm": 2.774442434310913,
+      "learning_rate": 3.2299270072992707e-06,
+      "loss": 2.1841,
+      "step": 191
+    },
+    {
+      "epoch": 0.2104109589041096,
+      "grad_norm": 3.0579135417938232,
+      "learning_rate": 3.248175182481752e-06,
+      "loss": 2.1677,
+      "step": 192
+    },
+    {
+      "epoch": 0.2115068493150685,
+      "grad_norm": 2.732984781265259,
+      "learning_rate": 3.266423357664234e-06,
+      "loss": 2.0975,
+      "step": 193
+    },
+    {
+      "epoch": 0.2126027397260274,
+      "grad_norm": 3.442659854888916,
+      "learning_rate": 3.2846715328467155e-06,
+      "loss": 2.1923,
+      "step": 194
+    },
+    {
+      "epoch": 0.2136986301369863,
+      "grad_norm": 3.064223527908325,
+      "learning_rate": 3.3029197080291973e-06,
+      "loss": 2.1707,
+      "step": 195
+    },
+    {
+      "epoch": 0.2147945205479452,
+      "grad_norm": 2.5138468742370605,
+      "learning_rate": 3.3211678832116788e-06,
+      "loss": 2.1933,
+      "step": 196
+    },
+    {
+      "epoch": 0.2158904109589041,
+      "grad_norm": 3.1076464653015137,
+      "learning_rate": 3.339416058394161e-06,
+      "loss": 2.1725,
+      "step": 197
+    },
+    {
+      "epoch": 0.21698630136986302,
+      "grad_norm": 2.7291388511657715,
+      "learning_rate": 3.3576642335766425e-06,
+      "loss": 2.1722,
+      "step": 198
+    },
+    {
+      "epoch": 0.21808219178082192,
+      "grad_norm": 2.805506467819214,
+      "learning_rate": 3.3759124087591243e-06,
+      "loss": 2.1688,
+      "step": 199
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "grad_norm": 2.6190085411071777,
+      "learning_rate": 3.3941605839416058e-06,
+      "loss": 2.1708,
+      "step": 200
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "eval_loss": 2.295546054840088,
+      "eval_runtime": 37.0133,
+      "eval_samples_per_second": 27.017,
+      "eval_steps_per_second": 1.135,
+      "step": 200
+    },
+    {
+      "epoch": 0.22027397260273973,
+      "grad_norm": 3.1270673274993896,
+      "learning_rate": 3.412408759124088e-06,
+      "loss": 2.1759,
+      "step": 201
+    },
+    {
+      "epoch": 0.22136986301369863,
+      "grad_norm": 3.494178533554077,
+      "learning_rate": 3.43065693430657e-06,
+      "loss": 2.1795,
+      "step": 202
+    },
+    {
+      "epoch": 0.22246575342465755,
+      "grad_norm": 2.7398736476898193,
+      "learning_rate": 3.4489051094890513e-06,
+      "loss": 2.1201,
+      "step": 203
+    },
+    {
+      "epoch": 0.22356164383561644,
+      "grad_norm": 2.981506824493408,
+      "learning_rate": 3.467153284671533e-06,
+      "loss": 2.1693,
+      "step": 204
+    },
+    {
+      "epoch": 0.22465753424657534,
+      "grad_norm": 2.4129178524017334,
+      "learning_rate": 3.4854014598540146e-06,
+      "loss": 2.1125,
+      "step": 205
+    },
+    {
+      "epoch": 0.22575342465753426,
+      "grad_norm": 2.701197385787964,
+      "learning_rate": 3.503649635036497e-06,
+      "loss": 2.1565,
+      "step": 206
+    },
+    {
+      "epoch": 0.22684931506849315,
+      "grad_norm": 2.533306837081909,
+      "learning_rate": 3.5218978102189783e-06,
+      "loss": 2.1487,
+      "step": 207
+    },
+    {
+      "epoch": 0.22794520547945205,
+      "grad_norm": 2.8030788898468018,
+      "learning_rate": 3.54014598540146e-06,
+      "loss": 2.1801,
+      "step": 208
+    },
+    {
+      "epoch": 0.22904109589041097,
+      "grad_norm": 2.895414352416992,
+      "learning_rate": 3.5583941605839416e-06,
+      "loss": 2.189,
+      "step": 209
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "grad_norm": 2.508390426635742,
+      "learning_rate": 3.576642335766424e-06,
+      "loss": 2.2376,
+      "step": 210
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "eval_loss": 2.2900516986846924,
+      "eval_runtime": 36.7645,
+      "eval_samples_per_second": 27.2,
+      "eval_steps_per_second": 1.142,
+      "step": 210
+    },
+    {
+      "epoch": 0.23123287671232876,
+      "grad_norm": 2.6393816471099854,
+      "learning_rate": 3.5948905109489054e-06,
+      "loss": 2.2267,
+      "step": 211
+    },
+    {
+      "epoch": 0.23232876712328768,
+      "grad_norm": 3.0687549114227295,
+      "learning_rate": 3.6131386861313872e-06,
+      "loss": 2.0744,
+      "step": 212
+    },
+    {
+      "epoch": 0.23342465753424657,
+      "grad_norm": 2.6342105865478516,
+      "learning_rate": 3.6313868613138687e-06,
+      "loss": 2.0737,
+      "step": 213
+    },
+    {
+      "epoch": 0.2345205479452055,
+      "grad_norm": 3.5662875175476074,
+      "learning_rate": 3.6496350364963505e-06,
+      "loss": 2.1147,
+      "step": 214
+    },
+    {
+      "epoch": 0.2356164383561644,
+      "grad_norm": 2.5357561111450195,
+      "learning_rate": 3.667883211678833e-06,
+      "loss": 2.1268,
+      "step": 215
+    },
+    {
+      "epoch": 0.23671232876712328,
+      "grad_norm": 3.439711093902588,
+      "learning_rate": 3.6861313868613142e-06,
+      "loss": 2.1118,
+      "step": 216
+    },
+    {
+      "epoch": 0.2378082191780822,
+      "grad_norm": 2.9410204887390137,
+      "learning_rate": 3.704379562043796e-06,
+      "loss": 2.1087,
+      "step": 217
+    },
+    {
+      "epoch": 0.2389041095890411,
+      "grad_norm": 3.3182294368743896,
+      "learning_rate": 3.7226277372262775e-06,
+      "loss": 2.196,
+      "step": 218
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 3.3834011554718018,
+      "learning_rate": 3.7408759124087594e-06,
+      "loss": 2.1633,
+      "step": 219
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "grad_norm": 3.114952325820923,
+      "learning_rate": 3.7591240875912412e-06,
+      "loss": 2.1107,
+      "step": 220
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "eval_loss": 2.277580976486206,
+      "eval_runtime": 36.9846,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 220
+    },
+    {
+      "epoch": 0.2421917808219178,
+      "grad_norm": 3.1151392459869385,
+      "learning_rate": 3.777372262773723e-06,
+      "loss": 2.1471,
+      "step": 221
+    },
+    {
+      "epoch": 0.2432876712328767,
+      "grad_norm": 3.798450469970703,
+      "learning_rate": 3.7956204379562045e-06,
+      "loss": 2.1324,
+      "step": 222
+    },
+    {
+      "epoch": 0.24438356164383562,
+      "grad_norm": 2.6187098026275635,
+      "learning_rate": 3.8138686131386864e-06,
+      "loss": 2.1061,
+      "step": 223
+    },
+    {
+      "epoch": 0.24547945205479452,
+      "grad_norm": 4.014018535614014,
+      "learning_rate": 3.832116788321168e-06,
+      "loss": 2.166,
+      "step": 224
+    },
+    {
+      "epoch": 0.2465753424657534,
+      "grad_norm": 3.0043938159942627,
+      "learning_rate": 3.85036496350365e-06,
+      "loss": 2.1197,
+      "step": 225
+    },
+    {
+      "epoch": 0.24767123287671233,
+      "grad_norm": 3.507066249847412,
+      "learning_rate": 3.868613138686132e-06,
+      "loss": 2.0848,
+      "step": 226
+    },
+    {
+      "epoch": 0.24876712328767123,
+      "grad_norm": 3.5713281631469727,
+      "learning_rate": 3.886861313868613e-06,
+      "loss": 2.1239,
+      "step": 227
+    },
+    {
+      "epoch": 0.24986301369863015,
+      "grad_norm": 3.961082696914673,
+      "learning_rate": 3.905109489051096e-06,
+      "loss": 2.1406,
+      "step": 228
+    },
+    {
+      "epoch": 0.25095890410958904,
+      "grad_norm": 3.1514859199523926,
+      "learning_rate": 3.923357664233577e-06,
+      "loss": 2.1689,
+      "step": 229
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "grad_norm": 3.8486833572387695,
+      "learning_rate": 3.9416058394160585e-06,
+      "loss": 2.1321,
+      "step": 230
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "eval_loss": 2.2620601654052734,
+      "eval_runtime": 37.058,
+      "eval_samples_per_second": 26.985,
+      "eval_steps_per_second": 1.133,
+      "step": 230
+    },
+    {
+      "epoch": 0.25315068493150683,
+      "grad_norm": 2.5626907348632812,
+      "learning_rate": 3.95985401459854e-06,
+      "loss": 2.1777,
+      "step": 231
+    },
+    {
+      "epoch": 0.25424657534246575,
+      "grad_norm": 4.105278015136719,
+      "learning_rate": 3.978102189781022e-06,
+      "loss": 2.2099,
+      "step": 232
+    },
+    {
+      "epoch": 0.2553424657534247,
+      "grad_norm": 2.9818739891052246,
+      "learning_rate": 3.996350364963504e-06,
+      "loss": 2.0705,
+      "step": 233
+    },
+    {
+      "epoch": 0.25643835616438354,
+      "grad_norm": 2.87030291557312,
+      "learning_rate": 4.014598540145986e-06,
+      "loss": 2.1453,
+      "step": 234
+    },
+    {
+      "epoch": 0.25753424657534246,
+      "grad_norm": 2.4847915172576904,
+      "learning_rate": 4.032846715328467e-06,
+      "loss": 2.0461,
+      "step": 235
+    },
+    {
+      "epoch": 0.2586301369863014,
+      "grad_norm": 3.451420783996582,
+      "learning_rate": 4.05109489051095e-06,
+      "loss": 2.1133,
+      "step": 236
+    },
+    {
+      "epoch": 0.25972602739726025,
+      "grad_norm": 2.954723358154297,
+      "learning_rate": 4.069343065693431e-06,
+      "loss": 2.1316,
+      "step": 237
+    },
+    {
+      "epoch": 0.2608219178082192,
+      "grad_norm": 2.5295896530151367,
+      "learning_rate": 4.0875912408759126e-06,
+      "loss": 2.0901,
+      "step": 238
+    },
+    {
+      "epoch": 0.2619178082191781,
+      "grad_norm": 2.6359453201293945,
+      "learning_rate": 4.105839416058395e-06,
+      "loss": 2.117,
+      "step": 239
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "grad_norm": 2.5236105918884277,
+      "learning_rate": 4.124087591240876e-06,
+      "loss": 2.1958,
+      "step": 240
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "eval_loss": 2.250295400619507,
+      "eval_runtime": 37.0811,
+      "eval_samples_per_second": 26.968,
+      "eval_steps_per_second": 1.133,
+      "step": 240
+    },
+    {
+      "epoch": 0.2641095890410959,
+      "grad_norm": 2.581918954849243,
+      "learning_rate": 4.1423357664233586e-06,
+      "loss": 2.1054,
+      "step": 241
+    },
+    {
+      "epoch": 0.2652054794520548,
+      "grad_norm": 2.501332998275757,
+      "learning_rate": 4.16058394160584e-06,
+      "loss": 2.0878,
+      "step": 242
+    },
+    {
+      "epoch": 0.2663013698630137,
+      "grad_norm": 2.6477770805358887,
+      "learning_rate": 4.1788321167883214e-06,
+      "loss": 2.1656,
+      "step": 243
+    },
+    {
+      "epoch": 0.2673972602739726,
+      "grad_norm": 2.7252438068389893,
+      "learning_rate": 4.197080291970803e-06,
+      "loss": 2.0497,
+      "step": 244
+    },
+    {
+      "epoch": 0.2684931506849315,
+      "grad_norm": 2.6056301593780518,
+      "learning_rate": 4.215328467153285e-06,
+      "loss": 2.1411,
+      "step": 245
+    },
+    {
+      "epoch": 0.26958904109589044,
+      "grad_norm": 2.6581931114196777,
+      "learning_rate": 4.233576642335767e-06,
+      "loss": 2.0829,
+      "step": 246
+    },
+    {
+      "epoch": 0.2706849315068493,
+      "grad_norm": 2.6316752433776855,
+      "learning_rate": 4.251824817518249e-06,
+      "loss": 2.0882,
+      "step": 247
+    },
+    {
+      "epoch": 0.2717808219178082,
+      "grad_norm": 2.803823947906494,
+      "learning_rate": 4.27007299270073e-06,
+      "loss": 2.1203,
+      "step": 248
+    },
+    {
+      "epoch": 0.27287671232876715,
+      "grad_norm": 2.9551174640655518,
+      "learning_rate": 4.288321167883212e-06,
+      "loss": 2.0897,
+      "step": 249
+    },
+    {
+      "epoch": 0.273972602739726,
+      "grad_norm": 2.4847161769866943,
+      "learning_rate": 4.306569343065693e-06,
+      "loss": 2.0845,
+      "step": 250
+    },
+    {
+      "epoch": 0.273972602739726,
+      "eval_loss": 2.2491259574890137,
+      "eval_runtime": 37.1024,
+      "eval_samples_per_second": 26.952,
+      "eval_steps_per_second": 1.132,
+      "step": 250
+    },
+    {
+      "epoch": 0.27506849315068493,
+      "grad_norm": 2.6398351192474365,
+      "learning_rate": 4.3248175182481755e-06,
+      "loss": 2.0479,
+      "step": 251
+    },
+    {
+      "epoch": 0.27616438356164386,
+      "grad_norm": 3.0948562622070312,
+      "learning_rate": 4.343065693430658e-06,
+      "loss": 2.1184,
+      "step": 252
+    },
+    {
+      "epoch": 0.2772602739726027,
+      "grad_norm": 2.5378012657165527,
+      "learning_rate": 4.361313868613139e-06,
+      "loss": 2.1099,
+      "step": 253
+    },
+    {
+      "epoch": 0.27835616438356164,
+      "grad_norm": 3.0282225608825684,
+      "learning_rate": 4.379562043795621e-06,
+      "loss": 2.0982,
+      "step": 254
+    },
+    {
+      "epoch": 0.27945205479452057,
+      "grad_norm": 2.6205334663391113,
+      "learning_rate": 4.397810218978102e-06,
+      "loss": 2.1285,
+      "step": 255
+    },
+    {
+      "epoch": 0.28054794520547943,
+      "grad_norm": 3.2056095600128174,
+      "learning_rate": 4.416058394160584e-06,
+      "loss": 2.1221,
+      "step": 256
+    },
+    {
+      "epoch": 0.28164383561643835,
+      "grad_norm": 3.1370127201080322,
+      "learning_rate": 4.434306569343066e-06,
+      "loss": 2.1016,
+      "step": 257
+    },
+    {
+      "epoch": 0.2827397260273973,
+      "grad_norm": 2.494168519973755,
+      "learning_rate": 4.452554744525548e-06,
+      "loss": 2.0754,
+      "step": 258
+    },
+    {
+      "epoch": 0.28383561643835614,
+      "grad_norm": 2.7979183197021484,
+      "learning_rate": 4.4708029197080295e-06,
+      "loss": 2.1218,
+      "step": 259
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "grad_norm": 2.755682945251465,
+      "learning_rate": 4.489051094890512e-06,
+      "loss": 2.1136,
+      "step": 260
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "eval_loss": 2.2433817386627197,
+      "eval_runtime": 37.1121,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 260
+    },
+    {
+      "epoch": 0.286027397260274,
+      "grad_norm": 2.3641891479492188,
+      "learning_rate": 4.507299270072993e-06,
+      "loss": 2.1488,
+      "step": 261
+    },
+    {
+      "epoch": 0.28712328767123285,
+      "grad_norm": 2.741292953491211,
+      "learning_rate": 4.525547445255475e-06,
+      "loss": 2.1048,
+      "step": 262
+    },
+    {
+      "epoch": 0.2882191780821918,
+      "grad_norm": 2.811483383178711,
+      "learning_rate": 4.543795620437956e-06,
+      "loss": 2.0996,
+      "step": 263
+    },
+    {
+      "epoch": 0.2893150684931507,
+      "grad_norm": 2.6705479621887207,
+      "learning_rate": 4.562043795620438e-06,
+      "loss": 2.0652,
+      "step": 264
+    },
+    {
+      "epoch": 0.29041095890410956,
+      "grad_norm": 2.9383747577667236,
+      "learning_rate": 4.580291970802921e-06,
+      "loss": 2.1574,
+      "step": 265
+    },
+    {
+      "epoch": 0.2915068493150685,
+      "grad_norm": 2.591373920440674,
+      "learning_rate": 4.598540145985402e-06,
+      "loss": 2.1655,
+      "step": 266
+    },
+    {
+      "epoch": 0.2926027397260274,
+      "grad_norm": 3.16955828666687,
+      "learning_rate": 4.6167883211678835e-06,
+      "loss": 2.1285,
+      "step": 267
+    },
+    {
+      "epoch": 0.2936986301369863,
+      "grad_norm": 2.66941499710083,
+      "learning_rate": 4.635036496350365e-06,
+      "loss": 2.0777,
+      "step": 268
+    },
+    {
+      "epoch": 0.2947945205479452,
+      "grad_norm": 2.768139600753784,
+      "learning_rate": 4.653284671532847e-06,
+      "loss": 2.0748,
+      "step": 269
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "grad_norm": 2.5951924324035645,
+      "learning_rate": 4.671532846715329e-06,
+      "loss": 2.1289,
+      "step": 270
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "eval_loss": 2.2393550872802734,
+      "eval_runtime": 37.0993,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 270
+    },
+    {
+      "epoch": 0.29698630136986304,
+      "grad_norm": 2.5422160625457764,
+      "learning_rate": 4.689781021897811e-06,
+      "loss": 2.1342,
+      "step": 271
+    },
+    {
+      "epoch": 0.2980821917808219,
+      "grad_norm": 2.540555477142334,
+      "learning_rate": 4.708029197080292e-06,
+      "loss": 2.097,
+      "step": 272
+    },
+    {
+      "epoch": 0.2991780821917808,
+      "grad_norm": 2.995366334915161,
+      "learning_rate": 4.726277372262774e-06,
+      "loss": 2.1452,
+      "step": 273
+    },
+    {
+      "epoch": 0.30027397260273975,
+      "grad_norm": 2.2561020851135254,
+      "learning_rate": 4.744525547445255e-06,
+      "loss": 2.1069,
+      "step": 274
+    },
+    {
+      "epoch": 0.3013698630136986,
+      "grad_norm": 2.53476619720459,
+      "learning_rate": 4.7627737226277375e-06,
+      "loss": 2.0381,
+      "step": 275
+    },
+    {
+      "epoch": 0.30246575342465754,
+      "grad_norm": 2.3415749073028564,
+      "learning_rate": 4.78102189781022e-06,
+      "loss": 2.1027,
+      "step": 276
+    },
+    {
+      "epoch": 0.30356164383561646,
+      "grad_norm": 2.9756431579589844,
+      "learning_rate": 4.799270072992701e-06,
+      "loss": 2.1136,
+      "step": 277
+    },
+    {
+      "epoch": 0.3046575342465753,
+      "grad_norm": 2.606208324432373,
+      "learning_rate": 4.8175182481751835e-06,
+      "loss": 2.0797,
+      "step": 278
+    },
+    {
+      "epoch": 0.30575342465753425,
+      "grad_norm": 2.5089313983917236,
+      "learning_rate": 4.835766423357665e-06,
+      "loss": 2.1026,
+      "step": 279
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "grad_norm": 2.5003838539123535,
+      "learning_rate": 4.854014598540146e-06,
+      "loss": 2.0923,
+      "step": 280
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "eval_loss": 2.2351036071777344,
+      "eval_runtime": 37.111,
+      "eval_samples_per_second": 26.946,
+      "eval_steps_per_second": 1.132,
+      "step": 280
+    },
+    {
+      "epoch": 0.30794520547945203,
+      "grad_norm": 2.403602361679077,
+      "learning_rate": 4.872262773722628e-06,
+      "loss": 2.0388,
+      "step": 281
+    },
+    {
+      "epoch": 0.30904109589041096,
+      "grad_norm": 2.7775959968566895,
+      "learning_rate": 4.89051094890511e-06,
+      "loss": 2.0561,
+      "step": 282
+    },
+    {
+      "epoch": 0.3101369863013699,
+      "grad_norm": 2.7128653526306152,
+      "learning_rate": 4.9087591240875915e-06,
+      "loss": 2.042,
+      "step": 283
+    },
+    {
+      "epoch": 0.31123287671232874,
+      "grad_norm": 2.7029991149902344,
+      "learning_rate": 4.927007299270074e-06,
+      "loss": 2.0814,
+      "step": 284
+    },
+    {
+      "epoch": 0.31232876712328766,
+      "grad_norm": 2.3942933082580566,
+      "learning_rate": 4.945255474452555e-06,
+      "loss": 2.0906,
+      "step": 285
+    },
+    {
+      "epoch": 0.3134246575342466,
+      "grad_norm": 2.5086541175842285,
+      "learning_rate": 4.963503649635037e-06,
+      "loss": 2.0625,
+      "step": 286
+    },
+    {
+      "epoch": 0.31452054794520545,
+      "grad_norm": 2.6668760776519775,
+      "learning_rate": 4.981751824817518e-06,
+      "loss": 2.1538,
+      "step": 287
+    },
+    {
+      "epoch": 0.3156164383561644,
+      "grad_norm": 2.5066750049591064,
+      "learning_rate": 5e-06,
+      "loss": 2.1423,
+      "step": 288
+    },
+    {
+      "epoch": 0.3167123287671233,
+      "grad_norm": 2.84967041015625,
+      "learning_rate": 4.997969130787978e-06,
+      "loss": 2.0943,
+      "step": 289
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "grad_norm": 2.151946544647217,
+      "learning_rate": 4.995938261575955e-06,
+      "loss": 2.1098,
+      "step": 290
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "eval_loss": 2.223247528076172,
+      "eval_runtime": 37.1755,
+      "eval_samples_per_second": 26.899,
+      "eval_steps_per_second": 1.13,
+      "step": 290
+    },
+    {
+      "epoch": 0.3189041095890411,
+      "grad_norm": 2.887166976928711,
+      "learning_rate": 4.993907392363932e-06,
+      "loss": 2.0281,
+      "step": 291
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 2.5930211544036865,
+      "learning_rate": 4.9918765231519095e-06,
+      "loss": 2.1063,
+      "step": 292
+    },
+    {
+      "epoch": 0.32109589041095893,
+      "grad_norm": 2.5568413734436035,
+      "learning_rate": 4.989845653939887e-06,
+      "loss": 2.0313,
+      "step": 293
+    },
+    {
+      "epoch": 0.3221917808219178,
+      "grad_norm": 2.536095380783081,
+      "learning_rate": 4.987814784727864e-06,
+      "loss": 2.0749,
+      "step": 294
+    },
+    {
+      "epoch": 0.3232876712328767,
+      "grad_norm": 2.3385651111602783,
+      "learning_rate": 4.985783915515841e-06,
+      "loss": 2.0666,
+      "step": 295
+    },
+    {
+      "epoch": 0.32438356164383564,
+      "grad_norm": 2.504826068878174,
+      "learning_rate": 4.983753046303819e-06,
+      "loss": 2.0677,
+      "step": 296
+    },
+    {
+      "epoch": 0.3254794520547945,
+      "grad_norm": 2.5289409160614014,
+      "learning_rate": 4.981722177091796e-06,
+      "loss": 2.0828,
+      "step": 297
+    },
+    {
+      "epoch": 0.3265753424657534,
+      "grad_norm": 2.803795099258423,
+      "learning_rate": 4.979691307879773e-06,
+      "loss": 2.0733,
+      "step": 298
+    },
+    {
+      "epoch": 0.32767123287671235,
+      "grad_norm": 2.651273012161255,
+      "learning_rate": 4.9776604386677504e-06,
+      "loss": 2.0987,
+      "step": 299
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "grad_norm": 2.4715864658355713,
+      "learning_rate": 4.975629569455728e-06,
+      "loss": 2.0733,
+      "step": 300
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "eval_loss": 2.214858293533325,
+      "eval_runtime": 37.0559,
+      "eval_samples_per_second": 26.986,
+      "eval_steps_per_second": 1.133,
+      "step": 300
+    },
+    {
+      "epoch": 0.32986301369863014,
+      "grad_norm": 2.2573442459106445,
+      "learning_rate": 4.973598700243705e-06,
+      "loss": 2.1291,
+      "step": 301
+    },
+    {
+      "epoch": 0.33095890410958906,
+      "grad_norm": 2.645966053009033,
+      "learning_rate": 4.971567831031681e-06,
+      "loss": 2.046,
+      "step": 302
+    },
+    {
+      "epoch": 0.3320547945205479,
+      "grad_norm": 2.361392021179199,
+      "learning_rate": 4.9695369618196595e-06,
+      "loss": 2.0447,
+      "step": 303
+    },
+    {
+      "epoch": 0.33315068493150685,
+      "grad_norm": 2.272714376449585,
+      "learning_rate": 4.967506092607636e-06,
+      "loss": 1.9272,
+      "step": 304
+    },
+    {
+      "epoch": 0.33424657534246577,
+      "grad_norm": 2.5846943855285645,
+      "learning_rate": 4.965475223395614e-06,
+      "loss": 2.0199,
+      "step": 305
+    },
+    {
+      "epoch": 0.33534246575342463,
+      "grad_norm": 2.38985013961792,
+      "learning_rate": 4.9634443541835905e-06,
+      "loss": 2.0394,
+      "step": 306
+    },
+    {
+      "epoch": 0.33643835616438356,
+      "grad_norm": 2.3526251316070557,
+      "learning_rate": 4.961413484971569e-06,
+      "loss": 2.0403,
+      "step": 307
+    },
+    {
+      "epoch": 0.3375342465753425,
+      "grad_norm": 2.399531602859497,
+      "learning_rate": 4.959382615759545e-06,
+      "loss": 2.0607,
+      "step": 308
+    },
+    {
+      "epoch": 0.33863013698630134,
+      "grad_norm": 2.50803804397583,
+      "learning_rate": 4.957351746547523e-06,
+      "loss": 2.0688,
+      "step": 309
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "grad_norm": 2.278088092803955,
+      "learning_rate": 4.9553208773355e-06,
+      "loss": 2.0453,
+      "step": 310
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "eval_loss": 2.2185235023498535,
+      "eval_runtime": 36.9712,
+      "eval_samples_per_second": 27.048,
+      "eval_steps_per_second": 1.136,
+      "step": 310
+    },
+    {
+      "epoch": 0.3408219178082192,
+      "grad_norm": 2.555300712585449,
+      "learning_rate": 4.953290008123478e-06,
+      "loss": 2.1726,
+      "step": 311
+    },
+    {
+      "epoch": 0.3419178082191781,
+      "grad_norm": 2.614870548248291,
+      "learning_rate": 4.951259138911454e-06,
+      "loss": 2.0659,
+      "step": 312
+    },
+    {
+      "epoch": 0.343013698630137,
+      "grad_norm": 2.1932661533355713,
+      "learning_rate": 4.9492282696994315e-06,
+      "loss": 1.9981,
+      "step": 313
+    },
+    {
+      "epoch": 0.3441095890410959,
+      "grad_norm": 2.707451581954956,
+      "learning_rate": 4.947197400487409e-06,
+      "loss": 2.1321,
+      "step": 314
+    },
+    {
+      "epoch": 0.3452054794520548,
+      "grad_norm": 2.4317946434020996,
+      "learning_rate": 4.945166531275386e-06,
+      "loss": 2.0548,
+      "step": 315
+    },
+    {
+      "epoch": 0.3463013698630137,
+      "grad_norm": 2.639883279800415,
+      "learning_rate": 4.943135662063363e-06,
+      "loss": 2.0512,
+      "step": 316
+    },
+    {
+      "epoch": 0.3473972602739726,
+      "grad_norm": 3.1968822479248047,
+      "learning_rate": 4.9411047928513406e-06,
+      "loss": 2.0611,
+      "step": 317
+    },
+    {
+      "epoch": 0.34849315068493153,
+      "grad_norm": 2.4361510276794434,
+      "learning_rate": 4.939073923639318e-06,
+      "loss": 2.0974,
+      "step": 318
+    },
+    {
+      "epoch": 0.3495890410958904,
+      "grad_norm": 2.8278262615203857,
+      "learning_rate": 4.937043054427295e-06,
+      "loss": 2.0992,
+      "step": 319
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "grad_norm": 3.01247239112854,
+      "learning_rate": 4.935012185215272e-06,
+      "loss": 2.0372,
+      "step": 320
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "eval_loss": 2.206442356109619,
+      "eval_runtime": 36.8797,
+      "eval_samples_per_second": 27.115,
+      "eval_steps_per_second": 1.139,
+      "step": 320
+    },
+    {
+      "epoch": 0.35178082191780824,
+      "grad_norm": 2.3372156620025635,
+      "learning_rate": 4.93298131600325e-06,
+      "loss": 2.0211,
+      "step": 321
+    },
+    {
+      "epoch": 0.3528767123287671,
+      "grad_norm": 2.91377329826355,
+      "learning_rate": 4.930950446791227e-06,
+      "loss": 2.0879,
+      "step": 322
+    },
+    {
+      "epoch": 0.353972602739726,
+      "grad_norm": 2.3207666873931885,
+      "learning_rate": 4.928919577579204e-06,
+      "loss": 2.056,
+      "step": 323
+    },
+    {
+      "epoch": 0.35506849315068495,
+      "grad_norm": 2.3260514736175537,
+      "learning_rate": 4.9268887083671815e-06,
+      "loss": 2.0684,
+      "step": 324
+    },
+    {
+      "epoch": 0.3561643835616438,
+      "grad_norm": 2.3070554733276367,
+      "learning_rate": 4.924857839155159e-06,
+      "loss": 2.0494,
+      "step": 325
+    },
+    {
+      "epoch": 0.35726027397260274,
+      "grad_norm": 2.189871072769165,
+      "learning_rate": 4.922826969943136e-06,
+      "loss": 2.0531,
+      "step": 326
+    },
+    {
+      "epoch": 0.35835616438356166,
+      "grad_norm": 2.553520917892456,
+      "learning_rate": 4.920796100731113e-06,
+      "loss": 2.0749,
+      "step": 327
+    },
+    {
+      "epoch": 0.3594520547945205,
+      "grad_norm": 2.62296199798584,
+      "learning_rate": 4.918765231519091e-06,
+      "loss": 1.9983,
+      "step": 328
+    },
+    {
+      "epoch": 0.36054794520547945,
+      "grad_norm": 2.4612369537353516,
+      "learning_rate": 4.916734362307068e-06,
+      "loss": 2.0641,
+      "step": 329
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "grad_norm": 2.656157970428467,
+      "learning_rate": 4.914703493095045e-06,
+      "loss": 2.0148,
+      "step": 330
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "eval_loss": 2.2054214477539062,
+      "eval_runtime": 36.886,
+      "eval_samples_per_second": 27.111,
+      "eval_steps_per_second": 1.139,
+      "step": 330
+    },
+    {
+      "epoch": 0.36273972602739724,
+      "grad_norm": 2.2823736667633057,
+      "learning_rate": 4.912672623883022e-06,
+      "loss": 2.0221,
+      "step": 331
+    },
+    {
+      "epoch": 0.36383561643835616,
+      "grad_norm": 2.8403289318084717,
+      "learning_rate": 4.910641754671e-06,
+      "loss": 2.0149,
+      "step": 332
+    },
+    {
+      "epoch": 0.3649315068493151,
+      "grad_norm": 2.453887462615967,
+      "learning_rate": 4.908610885458977e-06,
+      "loss": 2.0411,
+      "step": 333
+    },
+    {
+      "epoch": 0.36602739726027395,
+      "grad_norm": 2.6094300746917725,
+      "learning_rate": 4.906580016246954e-06,
+      "loss": 2.0293,
+      "step": 334
+    },
+    {
+      "epoch": 0.36712328767123287,
+      "grad_norm": 2.4412384033203125,
+      "learning_rate": 4.9045491470349315e-06,
+      "loss": 2.0778,
+      "step": 335
+    },
+    {
+      "epoch": 0.3682191780821918,
+      "grad_norm": 2.4435057640075684,
+      "learning_rate": 4.902518277822909e-06,
+      "loss": 2.0597,
+      "step": 336
+    },
+    {
+      "epoch": 0.3693150684931507,
+      "grad_norm": 2.2390244007110596,
+      "learning_rate": 4.900487408610886e-06,
+      "loss": 2.0654,
+      "step": 337
+    },
+    {
+      "epoch": 0.3704109589041096,
+      "grad_norm": 2.4243953227996826,
+      "learning_rate": 4.898456539398863e-06,
+      "loss": 2.0696,
+      "step": 338
+    },
+    {
+      "epoch": 0.3715068493150685,
+      "grad_norm": 2.375720262527466,
+      "learning_rate": 4.89642567018684e-06,
+      "loss": 2.0866,
+      "step": 339
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "grad_norm": 2.6379668712615967,
+      "learning_rate": 4.894394800974818e-06,
+      "loss": 2.1552,
+      "step": 340
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "eval_loss": 2.196237802505493,
+      "eval_runtime": 36.8953,
+      "eval_samples_per_second": 27.104,
+      "eval_steps_per_second": 1.138,
+      "step": 340
+    },
+    {
+      "epoch": 0.3736986301369863,
+      "grad_norm": 2.417145252227783,
+      "learning_rate": 4.892363931762794e-06,
+      "loss": 2.0024,
+      "step": 341
+    },
+    {
+      "epoch": 0.3747945205479452,
+      "grad_norm": 2.234590768814087,
+      "learning_rate": 4.8903330625507725e-06,
+      "loss": 2.0837,
+      "step": 342
+    },
+    {
+      "epoch": 0.37589041095890413,
+      "grad_norm": 2.404266595840454,
+      "learning_rate": 4.888302193338749e-06,
+      "loss": 2.018,
+      "step": 343
+    },
+    {
+      "epoch": 0.376986301369863,
+      "grad_norm": 2.364802837371826,
+      "learning_rate": 4.886271324126727e-06,
+      "loss": 1.994,
+      "step": 344
+    },
+    {
+      "epoch": 0.3780821917808219,
+      "grad_norm": 2.2262251377105713,
+      "learning_rate": 4.8842404549147034e-06,
+      "loss": 2.0929,
+      "step": 345
+    },
+    {
+      "epoch": 0.37917808219178084,
+      "grad_norm": 2.4854862689971924,
+      "learning_rate": 4.8822095857026816e-06,
+      "loss": 2.0801,
+      "step": 346
+    },
+    {
+      "epoch": 0.3802739726027397,
+      "grad_norm": 2.3861470222473145,
+      "learning_rate": 4.880178716490658e-06,
+      "loss": 2.0496,
+      "step": 347
+    },
+    {
+      "epoch": 0.38136986301369863,
+      "grad_norm": 2.372326612472534,
+      "learning_rate": 4.878147847278636e-06,
+      "loss": 1.9939,
+      "step": 348
+    },
+    {
+      "epoch": 0.38246575342465755,
+      "grad_norm": 2.3593149185180664,
+      "learning_rate": 4.8761169780666125e-06,
+      "loss": 2.0108,
+      "step": 349
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "grad_norm": 2.1802141666412354,
+      "learning_rate": 4.874086108854591e-06,
+      "loss": 1.983,
+      "step": 350
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "eval_loss": 2.1947567462921143,
+      "eval_runtime": 36.8798,
+      "eval_samples_per_second": 27.115,
+      "eval_steps_per_second": 1.139,
+      "step": 350
+    },
+    {
+      "epoch": 0.38465753424657534,
+      "grad_norm": 2.47067928314209,
+      "learning_rate": 4.872055239642567e-06,
+      "loss": 2.0002,
+      "step": 351
+    },
+    {
+      "epoch": 0.38575342465753426,
+      "grad_norm": 2.359609365463257,
+      "learning_rate": 4.870024370430544e-06,
+      "loss": 1.9799,
+      "step": 352
+    },
+    {
+      "epoch": 0.3868493150684931,
+      "grad_norm": 2.3711702823638916,
+      "learning_rate": 4.867993501218522e-06,
+      "loss": 2.0341,
+      "step": 353
+    },
+    {
+      "epoch": 0.38794520547945205,
+      "grad_norm": 2.246647834777832,
+      "learning_rate": 4.865962632006499e-06,
+      "loss": 2.0687,
+      "step": 354
+    },
+    {
+      "epoch": 0.38904109589041097,
+      "grad_norm": 2.3031938076019287,
+      "learning_rate": 4.863931762794476e-06,
+      "loss": 1.9994,
+      "step": 355
+    },
+    {
+      "epoch": 0.39013698630136984,
+      "grad_norm": 2.492159605026245,
+      "learning_rate": 4.8619008935824535e-06,
+      "loss": 2.0195,
+      "step": 356
+    },
+    {
+      "epoch": 0.39123287671232876,
+      "grad_norm": 2.2878570556640625,
+      "learning_rate": 4.859870024370431e-06,
+      "loss": 2.0589,
+      "step": 357
+    },
+    {
+      "epoch": 0.3923287671232877,
+      "grad_norm": 2.444096088409424,
+      "learning_rate": 4.857839155158408e-06,
+      "loss": 2.0796,
+      "step": 358
+    },
+    {
+      "epoch": 0.3934246575342466,
+      "grad_norm": 2.463857650756836,
+      "learning_rate": 4.855808285946385e-06,
+      "loss": 2.0824,
+      "step": 359
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "grad_norm": 2.0893168449401855,
+      "learning_rate": 4.8537774167343626e-06,
+      "loss": 2.0651,
+      "step": 360
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "eval_loss": 2.188523054122925,
+      "eval_runtime": 36.8679,
+      "eval_samples_per_second": 27.124,
+      "eval_steps_per_second": 1.139,
+      "step": 360
+    },
+    {
+      "epoch": 0.3956164383561644,
+      "grad_norm": 2.602850914001465,
+      "learning_rate": 4.85174654752234e-06,
+      "loss": 2.0507,
+      "step": 361
+    },
+    {
+      "epoch": 0.3967123287671233,
+      "grad_norm": 2.4185140132904053,
+      "learning_rate": 4.849715678310317e-06,
+      "loss": 2.0386,
+      "step": 362
+    },
+    {
+      "epoch": 0.3978082191780822,
+      "grad_norm": 2.1492245197296143,
+      "learning_rate": 4.847684809098294e-06,
+      "loss": 2.0376,
+      "step": 363
+    },
+    {
+      "epoch": 0.3989041095890411,
+      "grad_norm": 2.5710153579711914,
+      "learning_rate": 4.845653939886272e-06,
+      "loss": 2.0383,
+      "step": 364
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 2.4986608028411865,
+      "learning_rate": 4.843623070674249e-06,
+      "loss": 2.0464,
+      "step": 365
+    },
+    {
+      "epoch": 0.4010958904109589,
+      "grad_norm": 2.260388135910034,
+      "learning_rate": 4.841592201462226e-06,
+      "loss": 2.0217,
+      "step": 366
+    },
+    {
+      "epoch": 0.4021917808219178,
+      "grad_norm": 2.6500394344329834,
+      "learning_rate": 4.8395613322502035e-06,
+      "loss": 2.0777,
+      "step": 367
+    },
+    {
+      "epoch": 0.40328767123287673,
+      "grad_norm": 2.159274101257324,
+      "learning_rate": 4.837530463038181e-06,
+      "loss": 1.9747,
+      "step": 368
+    },
+    {
+      "epoch": 0.4043835616438356,
+      "grad_norm": 2.4623043537139893,
+      "learning_rate": 4.835499593826158e-06,
+      "loss": 2.039,
+      "step": 369
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "grad_norm": 2.2265007495880127,
+      "learning_rate": 4.833468724614135e-06,
+      "loss": 2.0618,
+      "step": 370
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "eval_loss": 2.184051275253296,
+      "eval_runtime": 36.9284,
+      "eval_samples_per_second": 27.079,
+      "eval_steps_per_second": 1.137,
+      "step": 370
+    },
+    {
+      "epoch": 0.40657534246575344,
+      "grad_norm": 2.300276517868042,
+      "learning_rate": 4.831437855402113e-06,
+      "loss": 2.0382,
+      "step": 371
+    },
+    {
+      "epoch": 0.4076712328767123,
+      "grad_norm": 2.341156482696533,
+      "learning_rate": 4.82940698619009e-06,
+      "loss": 1.9361,
+      "step": 372
+    },
+    {
+      "epoch": 0.40876712328767123,
+      "grad_norm": 2.4196999073028564,
+      "learning_rate": 4.827376116978067e-06,
+      "loss": 2.0087,
+      "step": 373
+    },
+    {
+      "epoch": 0.40986301369863015,
+      "grad_norm": 2.4650514125823975,
+      "learning_rate": 4.8253452477660444e-06,
+      "loss": 2.0687,
+      "step": 374
+    },
+    {
+      "epoch": 0.410958904109589,
+      "grad_norm": 2.393739700317383,
+      "learning_rate": 4.823314378554022e-06,
+      "loss": 2.0421,
+      "step": 375
+    },
+    {
+      "epoch": 0.41205479452054794,
+      "grad_norm": 2.2436327934265137,
+      "learning_rate": 4.821283509341998e-06,
+      "loss": 2.0358,
+      "step": 376
+    },
+    {
+      "epoch": 0.41315068493150686,
+      "grad_norm": 2.1801650524139404,
+      "learning_rate": 4.819252640129976e-06,
+      "loss": 2.0081,
+      "step": 377
+    },
+    {
+      "epoch": 0.41424657534246573,
+      "grad_norm": 2.300882339477539,
+      "learning_rate": 4.817221770917953e-06,
+      "loss": 2.0321,
+      "step": 378
+    },
+    {
+      "epoch": 0.41534246575342465,
+      "grad_norm": 2.2136058807373047,
+      "learning_rate": 4.815190901705931e-06,
+      "loss": 1.9999,
+      "step": 379
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "grad_norm": 2.168034076690674,
+      "learning_rate": 4.813160032493907e-06,
+      "loss": 2.0336,
+      "step": 380
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "eval_loss": 2.1770179271698,
+      "eval_runtime": 37.0015,
+      "eval_samples_per_second": 27.026,
+      "eval_steps_per_second": 1.135,
+      "step": 380
+    },
+    {
+      "epoch": 0.41753424657534244,
+      "grad_norm": 2.1604163646698,
+      "learning_rate": 4.811129163281885e-06,
+      "loss": 2.0212,
+      "step": 381
+    },
+    {
+      "epoch": 0.41863013698630136,
+      "grad_norm": 2.23923659324646,
+      "learning_rate": 4.809098294069862e-06,
+      "loss": 2.0495,
+      "step": 382
+    },
+    {
+      "epoch": 0.4197260273972603,
+      "grad_norm": 2.149949312210083,
+      "learning_rate": 4.80706742485784e-06,
+      "loss": 2.041,
+      "step": 383
+    },
+    {
+      "epoch": 0.4208219178082192,
+      "grad_norm": 2.49120831489563,
+      "learning_rate": 4.805036555645816e-06,
+      "loss": 2.0931,
+      "step": 384
+    },
+    {
+      "epoch": 0.42191780821917807,
+      "grad_norm": 2.2123563289642334,
+      "learning_rate": 4.8030056864337945e-06,
+      "loss": 2.0087,
+      "step": 385
+    },
+    {
+      "epoch": 0.423013698630137,
+      "grad_norm": 2.4235594272613525,
+      "learning_rate": 4.800974817221771e-06,
+      "loss": 2.0656,
+      "step": 386
+    },
+    {
+      "epoch": 0.4241095890410959,
+      "grad_norm": 2.158252239227295,
+      "learning_rate": 4.798943948009749e-06,
+      "loss": 2.0181,
+      "step": 387
+    },
+    {
+      "epoch": 0.4252054794520548,
+      "grad_norm": 2.2412970066070557,
+      "learning_rate": 4.7969130787977254e-06,
+      "loss": 2.0323,
+      "step": 388
+    },
+    {
+      "epoch": 0.4263013698630137,
+      "grad_norm": 2.2520956993103027,
+      "learning_rate": 4.794882209585703e-06,
+      "loss": 2.026,
+      "step": 389
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "grad_norm": 2.168954849243164,
+      "learning_rate": 4.79285134037368e-06,
+      "loss": 2.0459,
+      "step": 390
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "eval_loss": 2.1705472469329834,
+      "eval_runtime": 37.0913,
+      "eval_samples_per_second": 26.96,
+      "eval_steps_per_second": 1.132,
+      "step": 390
+    },
+    {
+      "epoch": 0.4284931506849315,
+      "grad_norm": 2.1397407054901123,
+      "learning_rate": 4.790820471161657e-06,
+      "loss": 1.9995,
+      "step": 391
+    },
+    {
+      "epoch": 0.4295890410958904,
+      "grad_norm": 2.1747751235961914,
+      "learning_rate": 4.7887896019496345e-06,
+      "loss": 2.0088,
+      "step": 392
+    },
+    {
+      "epoch": 0.43068493150684933,
+      "grad_norm": 2.2860991954803467,
+      "learning_rate": 4.786758732737612e-06,
+      "loss": 2.0134,
+      "step": 393
+    },
+    {
+      "epoch": 0.4317808219178082,
+      "grad_norm": 2.2133519649505615,
+      "learning_rate": 4.784727863525589e-06,
+      "loss": 2.0586,
+      "step": 394
+    },
+    {
+      "epoch": 0.4328767123287671,
+      "grad_norm": 2.2147040367126465,
+      "learning_rate": 4.782696994313566e-06,
+      "loss": 1.9748,
+      "step": 395
+    },
+    {
+      "epoch": 0.43397260273972604,
+      "grad_norm": 2.343418836593628,
+      "learning_rate": 4.780666125101544e-06,
+      "loss": 2.0223,
+      "step": 396
+    },
+    {
+      "epoch": 0.4350684931506849,
+      "grad_norm": 2.198935031890869,
+      "learning_rate": 4.778635255889521e-06,
+      "loss": 1.9627,
+      "step": 397
+    },
+    {
+      "epoch": 0.43616438356164383,
+      "grad_norm": 2.1314704418182373,
+      "learning_rate": 4.776604386677498e-06,
+      "loss": 1.9938,
+      "step": 398
+    },
+    {
+      "epoch": 0.43726027397260275,
+      "grad_norm": 2.3005123138427734,
+      "learning_rate": 4.7745735174654755e-06,
+      "loss": 2.0029,
+      "step": 399
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "grad_norm": 2.2826664447784424,
+      "learning_rate": 4.772542648253453e-06,
+      "loss": 2.0585,
+      "step": 400
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "eval_loss": 2.1691415309906006,
+      "eval_runtime": 37.0852,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 400
+    },
+    {
+      "epoch": 0.43945205479452054,
+      "grad_norm": 2.0547878742218018,
+      "learning_rate": 4.77051177904143e-06,
+      "loss": 2.0582,
+      "step": 401
+    },
+    {
+      "epoch": 0.44054794520547946,
+      "grad_norm": 2.200042486190796,
+      "learning_rate": 4.768480909829407e-06,
+      "loss": 1.9936,
+      "step": 402
+    },
+    {
+      "epoch": 0.44164383561643833,
+      "grad_norm": 2.354637861251831,
+      "learning_rate": 4.766450040617385e-06,
+      "loss": 2.0327,
+      "step": 403
+    },
+    {
+      "epoch": 0.44273972602739725,
+      "grad_norm": 2.2332348823547363,
+      "learning_rate": 4.764419171405362e-06,
+      "loss": 2.0052,
+      "step": 404
+    },
+    {
+      "epoch": 0.4438356164383562,
+      "grad_norm": 2.288060426712036,
+      "learning_rate": 4.762388302193339e-06,
+      "loss": 2.0472,
+      "step": 405
+    },
+    {
+      "epoch": 0.4449315068493151,
+      "grad_norm": 2.1368887424468994,
+      "learning_rate": 4.760357432981316e-06,
+      "loss": 2.0305,
+      "step": 406
+    },
+    {
+      "epoch": 0.44602739726027396,
+      "grad_norm": 2.362630844116211,
+      "learning_rate": 4.758326563769294e-06,
+      "loss": 2.002,
+      "step": 407
+    },
+    {
+      "epoch": 0.4471232876712329,
+      "grad_norm": 2.283419370651245,
+      "learning_rate": 4.756295694557271e-06,
+      "loss": 2.0271,
+      "step": 408
+    },
+    {
+      "epoch": 0.4482191780821918,
+      "grad_norm": 2.0400819778442383,
+      "learning_rate": 4.754264825345248e-06,
+      "loss": 2.0256,
+      "step": 409
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "grad_norm": 2.588416814804077,
+      "learning_rate": 4.7522339561332255e-06,
+      "loss": 1.9756,
+      "step": 410
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "eval_loss": 2.17041015625,
+      "eval_runtime": 36.958,
+      "eval_samples_per_second": 27.058,
+      "eval_steps_per_second": 1.136,
+      "step": 410
+    },
+    {
+      "epoch": 0.4504109589041096,
+      "grad_norm": 2.2397494316101074,
+      "learning_rate": 4.750203086921203e-06,
+      "loss": 1.9876,
+      "step": 411
+    },
+    {
+      "epoch": 0.4515068493150685,
+      "grad_norm": 2.2085509300231934,
+      "learning_rate": 4.74817221770918e-06,
+      "loss": 1.9464,
+      "step": 412
+    },
+    {
+      "epoch": 0.4526027397260274,
+      "grad_norm": 2.140974760055542,
+      "learning_rate": 4.7461413484971565e-06,
+      "loss": 2.0179,
+      "step": 413
+    },
+    {
+      "epoch": 0.4536986301369863,
+      "grad_norm": 2.124366044998169,
+      "learning_rate": 4.744110479285135e-06,
+      "loss": 2.0732,
+      "step": 414
+    },
+    {
+      "epoch": 0.4547945205479452,
+      "grad_norm": 2.2463996410369873,
+      "learning_rate": 4.742079610073111e-06,
+      "loss": 1.9616,
+      "step": 415
+    },
+    {
+      "epoch": 0.4558904109589041,
+      "grad_norm": 2.1249961853027344,
+      "learning_rate": 4.740048740861089e-06,
+      "loss": 1.9889,
+      "step": 416
+    },
+    {
+      "epoch": 0.456986301369863,
+      "grad_norm": 2.25842022895813,
+      "learning_rate": 4.738017871649066e-06,
+      "loss": 2.0033,
+      "step": 417
+    },
+    {
+      "epoch": 0.45808219178082193,
+      "grad_norm": 2.271935224533081,
+      "learning_rate": 4.735987002437044e-06,
+      "loss": 2.0148,
+      "step": 418
+    },
+    {
+      "epoch": 0.4591780821917808,
+      "grad_norm": 2.4797163009643555,
+      "learning_rate": 4.73395613322502e-06,
+      "loss": 2.0736,
+      "step": 419
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "grad_norm": 2.1901395320892334,
+      "learning_rate": 4.731925264012998e-06,
+      "loss": 2.0763,
+      "step": 420
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "eval_loss": 2.161660671234131,
+      "eval_runtime": 36.9099,
+      "eval_samples_per_second": 27.093,
+      "eval_steps_per_second": 1.138,
+      "step": 420
+    },
+    {
+      "epoch": 0.46136986301369864,
+      "grad_norm": 2.083536386489868,
+      "learning_rate": 4.729894394800975e-06,
+      "loss": 2.0299,
+      "step": 421
+    },
+    {
+      "epoch": 0.4624657534246575,
+      "grad_norm": 2.1118173599243164,
+      "learning_rate": 4.727863525588953e-06,
+      "loss": 1.9742,
+      "step": 422
+    },
+    {
+      "epoch": 0.46356164383561643,
+      "grad_norm": 2.195802927017212,
+      "learning_rate": 4.725832656376929e-06,
+      "loss": 1.9318,
+      "step": 423
+    },
+    {
+      "epoch": 0.46465753424657535,
+      "grad_norm": 2.318098545074463,
+      "learning_rate": 4.723801787164907e-06,
+      "loss": 1.9875,
+      "step": 424
+    },
+    {
+      "epoch": 0.4657534246575342,
+      "grad_norm": 2.1698496341705322,
+      "learning_rate": 4.721770917952885e-06,
+      "loss": 2.031,
+      "step": 425
+    },
+    {
+      "epoch": 0.46684931506849314,
+      "grad_norm": 2.283935785293579,
+      "learning_rate": 4.719740048740861e-06,
+      "loss": 1.9766,
+      "step": 426
+    },
+    {
+      "epoch": 0.46794520547945206,
+      "grad_norm": 2.4180757999420166,
+      "learning_rate": 4.717709179528839e-06,
+      "loss": 1.9976,
+      "step": 427
+    },
+    {
+      "epoch": 0.469041095890411,
+      "grad_norm": 2.2689030170440674,
+      "learning_rate": 4.715678310316816e-06,
+      "loss": 2.03,
+      "step": 428
+    },
+    {
+      "epoch": 0.47013698630136985,
+      "grad_norm": 2.423708200454712,
+      "learning_rate": 4.713647441104794e-06,
+      "loss": 2.0202,
+      "step": 429
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "grad_norm": 2.184705972671509,
+      "learning_rate": 4.71161657189277e-06,
+      "loss": 1.9655,
+      "step": 430
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "eval_loss": 2.1548638343811035,
+      "eval_runtime": 36.9426,
+      "eval_samples_per_second": 27.069,
+      "eval_steps_per_second": 1.137,
+      "step": 430
+    },
+    {
+      "epoch": 0.4723287671232877,
+      "grad_norm": 2.295851945877075,
+      "learning_rate": 4.709585702680748e-06,
+      "loss": 1.9561,
+      "step": 431
+    },
+    {
+      "epoch": 0.47342465753424656,
+      "grad_norm": 2.1762046813964844,
+      "learning_rate": 4.707554833468725e-06,
+      "loss": 2.0341,
+      "step": 432
+    },
+    {
+      "epoch": 0.4745205479452055,
+      "grad_norm": 2.4684343338012695,
+      "learning_rate": 4.705523964256703e-06,
+      "loss": 2.0898,
+      "step": 433
+    },
+    {
+      "epoch": 0.4756164383561644,
+      "grad_norm": 2.293318271636963,
+      "learning_rate": 4.703493095044679e-06,
+      "loss": 1.9794,
+      "step": 434
+    },
+    {
+      "epoch": 0.4767123287671233,
+      "grad_norm": 2.1884210109710693,
+      "learning_rate": 4.701462225832657e-06,
+      "loss": 1.9979,
+      "step": 435
+    },
+    {
+      "epoch": 0.4778082191780822,
+      "grad_norm": 2.412504196166992,
+      "learning_rate": 4.699431356620634e-06,
+      "loss": 1.9237,
+      "step": 436
+    },
+    {
+      "epoch": 0.4789041095890411,
+      "grad_norm": 2.192824602127075,
+      "learning_rate": 4.697400487408612e-06,
+      "loss": 2.0756,
+      "step": 437
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 2.368100643157959,
+      "learning_rate": 4.695369618196588e-06,
+      "loss": 2.028,
+      "step": 438
+    },
+    {
+      "epoch": 0.4810958904109589,
+      "grad_norm": 2.218454122543335,
+      "learning_rate": 4.693338748984566e-06,
+      "loss": 2.0049,
+      "step": 439
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "grad_norm": 2.356168031692505,
+      "learning_rate": 4.691307879772543e-06,
+      "loss": 1.9958,
+      "step": 440
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "eval_loss": 2.1540727615356445,
+      "eval_runtime": 36.9557,
+      "eval_samples_per_second": 27.059,
+      "eval_steps_per_second": 1.136,
+      "step": 440
+    },
+    {
+      "epoch": 0.4832876712328767,
+      "grad_norm": 2.09956693649292,
+      "learning_rate": 4.68927701056052e-06,
+      "loss": 1.9388,
+      "step": 441
+    },
+    {
+      "epoch": 0.4843835616438356,
+      "grad_norm": 2.5746614933013916,
+      "learning_rate": 4.6872461413484975e-06,
+      "loss": 2.0425,
+      "step": 442
+    },
+    {
+      "epoch": 0.48547945205479454,
+      "grad_norm": 2.2064783573150635,
+      "learning_rate": 4.685215272136475e-06,
+      "loss": 2.0082,
+      "step": 443
+    },
+    {
+      "epoch": 0.4865753424657534,
+      "grad_norm": 2.159029006958008,
+      "learning_rate": 4.683184402924452e-06,
+      "loss": 2.0055,
+      "step": 444
+    },
+    {
+      "epoch": 0.4876712328767123,
+      "grad_norm": 2.5222008228302,
+      "learning_rate": 4.681153533712429e-06,
+      "loss": 2.0007,
+      "step": 445
+    },
+    {
+      "epoch": 0.48876712328767125,
+      "grad_norm": 2.163487672805786,
+      "learning_rate": 4.679122664500407e-06,
+      "loss": 2.0719,
+      "step": 446
+    },
+    {
+      "epoch": 0.4898630136986301,
+      "grad_norm": 2.3383803367614746,
+      "learning_rate": 4.677091795288384e-06,
+      "loss": 1.9623,
+      "step": 447
+    },
+    {
+      "epoch": 0.49095890410958903,
+      "grad_norm": 2.1056675910949707,
+      "learning_rate": 4.675060926076361e-06,
+      "loss": 1.9443,
+      "step": 448
+    },
+    {
+      "epoch": 0.49205479452054796,
+      "grad_norm": 2.145270824432373,
+      "learning_rate": 4.673030056864338e-06,
+      "loss": 2.0469,
+      "step": 449
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "grad_norm": 2.578706741333008,
+      "learning_rate": 4.670999187652316e-06,
+      "loss": 1.9569,
+      "step": 450
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "eval_loss": 2.1527743339538574,
+      "eval_runtime": 36.8961,
+      "eval_samples_per_second": 27.103,
+      "eval_steps_per_second": 1.138,
+      "step": 450
+    },
+    {
+      "epoch": 0.49424657534246574,
+      "grad_norm": 2.1328542232513428,
+      "learning_rate": 4.668968318440293e-06,
+      "loss": 1.9559,
+      "step": 451
+    },
+    {
+      "epoch": 0.49534246575342467,
+      "grad_norm": 2.148820638656616,
+      "learning_rate": 4.66693744922827e-06,
+      "loss": 1.9545,
+      "step": 452
+    },
+    {
+      "epoch": 0.4964383561643836,
+      "grad_norm": 2.250340461730957,
+      "learning_rate": 4.6649065800162475e-06,
+      "loss": 2.0358,
+      "step": 453
+    },
+    {
+      "epoch": 0.49753424657534245,
+      "grad_norm": 2.452397584915161,
+      "learning_rate": 4.662875710804225e-06,
+      "loss": 1.9294,
+      "step": 454
+    },
+    {
+      "epoch": 0.4986301369863014,
+      "grad_norm": 2.140615701675415,
+      "learning_rate": 4.660844841592202e-06,
+      "loss": 2.0084,
+      "step": 455
+    },
+    {
+      "epoch": 0.4997260273972603,
+      "grad_norm": 2.176806926727295,
+      "learning_rate": 4.658813972380179e-06,
+      "loss": 1.9174,
+      "step": 456
+    },
+    {
+      "epoch": 0.5008219178082192,
+      "grad_norm": 2.349031925201416,
+      "learning_rate": 4.656783103168157e-06,
+      "loss": 2.0377,
+      "step": 457
+    },
+    {
+      "epoch": 0.5019178082191781,
+      "grad_norm": 2.436889171600342,
+      "learning_rate": 4.654752233956134e-06,
+      "loss": 1.9133,
+      "step": 458
+    },
+    {
+      "epoch": 0.503013698630137,
+      "grad_norm": 2.116610288619995,
+      "learning_rate": 4.652721364744111e-06,
+      "loss": 1.9993,
+      "step": 459
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "grad_norm": 2.186915636062622,
+      "learning_rate": 4.6506904955320885e-06,
+      "loss": 1.9744,
+      "step": 460
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "eval_loss": 2.1580278873443604,
+      "eval_runtime": 36.9097,
+      "eval_samples_per_second": 27.093,
+      "eval_steps_per_second": 1.138,
+      "step": 460
+    },
+    {
+      "epoch": 0.5052054794520547,
+      "grad_norm": 2.2009212970733643,
+      "learning_rate": 4.648659626320066e-06,
+      "loss": 2.0173,
+      "step": 461
+    },
+    {
+      "epoch": 0.5063013698630137,
+      "grad_norm": 2.117286443710327,
+      "learning_rate": 4.646628757108043e-06,
+      "loss": 2.0238,
+      "step": 462
+    },
+    {
+      "epoch": 0.5073972602739726,
+      "grad_norm": 2.238790273666382,
+      "learning_rate": 4.6445978878960194e-06,
+      "loss": 1.9885,
+      "step": 463
+    },
+    {
+      "epoch": 0.5084931506849315,
+      "grad_norm": 2.3493096828460693,
+      "learning_rate": 4.6425670186839976e-06,
+      "loss": 1.9593,
+      "step": 464
+    },
+    {
+      "epoch": 0.5095890410958904,
+      "grad_norm": 2.3276610374450684,
+      "learning_rate": 4.640536149471974e-06,
+      "loss": 1.9993,
+      "step": 465
+    },
+    {
+      "epoch": 0.5106849315068493,
+      "grad_norm": 2.3671715259552,
+      "learning_rate": 4.638505280259952e-06,
+      "loss": 2.0256,
+      "step": 466
+    },
+    {
+      "epoch": 0.5117808219178083,
+      "grad_norm": 2.251803159713745,
+      "learning_rate": 4.6364744110479285e-06,
+      "loss": 2.0436,
+      "step": 467
+    },
+    {
+      "epoch": 0.5128767123287671,
+      "grad_norm": 2.3023312091827393,
+      "learning_rate": 4.634443541835907e-06,
+      "loss": 1.9968,
+      "step": 468
+    },
+    {
+      "epoch": 0.513972602739726,
+      "grad_norm": 2.3660333156585693,
+      "learning_rate": 4.632412672623883e-06,
+      "loss": 2.0492,
+      "step": 469
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "grad_norm": 2.262179136276245,
+      "learning_rate": 4.630381803411861e-06,
+      "loss": 1.9698,
+      "step": 470
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "eval_loss": 2.1516528129577637,
+      "eval_runtime": 36.7956,
+      "eval_samples_per_second": 27.177,
+      "eval_steps_per_second": 1.141,
+      "step": 470
+    },
+    {
+      "epoch": 0.5161643835616438,
+      "grad_norm": 2.450770616531372,
+      "learning_rate": 4.628350934199838e-06,
+      "loss": 1.8799,
+      "step": 471
+    },
+    {
+      "epoch": 0.5172602739726028,
+      "grad_norm": 2.187458038330078,
+      "learning_rate": 4.626320064987816e-06,
+      "loss": 1.9829,
+      "step": 472
+    },
+    {
+      "epoch": 0.5183561643835617,
+      "grad_norm": 2.1726722717285156,
+      "learning_rate": 4.624289195775792e-06,
+      "loss": 2.012,
+      "step": 473
+    },
+    {
+      "epoch": 0.5194520547945205,
+      "grad_norm": 2.4177136421203613,
+      "learning_rate": 4.62225832656377e-06,
+      "loss": 1.9796,
+      "step": 474
+    },
+    {
+      "epoch": 0.5205479452054794,
+      "grad_norm": 2.2353951930999756,
+      "learning_rate": 4.620227457351747e-06,
+      "loss": 1.99,
+      "step": 475
+    },
+    {
+      "epoch": 0.5216438356164383,
+      "grad_norm": 2.369591236114502,
+      "learning_rate": 4.618196588139724e-06,
+      "loss": 1.9847,
+      "step": 476
+    },
+    {
+      "epoch": 0.5227397260273973,
+      "grad_norm": 2.546703577041626,
+      "learning_rate": 4.616165718927701e-06,
+      "loss": 2.0047,
+      "step": 477
+    },
+    {
+      "epoch": 0.5238356164383562,
+      "grad_norm": 2.3509185314178467,
+      "learning_rate": 4.6141348497156786e-06,
+      "loss": 1.9589,
+      "step": 478
+    },
+    {
+      "epoch": 0.5249315068493151,
+      "grad_norm": 2.3063488006591797,
+      "learning_rate": 4.612103980503656e-06,
+      "loss": 2.045,
+      "step": 479
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "grad_norm": 2.285572052001953,
+      "learning_rate": 4.610073111291633e-06,
+      "loss": 2.0218,
+      "step": 480
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "eval_loss": 2.1472301483154297,
+      "eval_runtime": 36.7689,
+      "eval_samples_per_second": 27.197,
+      "eval_steps_per_second": 1.142,
+      "step": 480
+    },
+    {
+      "epoch": 0.5271232876712328,
+      "grad_norm": 2.22987961769104,
+      "learning_rate": 4.60804224207961e-06,
+      "loss": 1.9955,
+      "step": 481
+    },
+    {
+      "epoch": 0.5282191780821918,
+      "grad_norm": 2.200918197631836,
+      "learning_rate": 4.606011372867588e-06,
+      "loss": 1.9768,
+      "step": 482
+    },
+    {
+      "epoch": 0.5293150684931507,
+      "grad_norm": 2.2441930770874023,
+      "learning_rate": 4.603980503655565e-06,
+      "loss": 2.0088,
+      "step": 483
+    },
+    {
+      "epoch": 0.5304109589041096,
+      "grad_norm": 2.380383014678955,
+      "learning_rate": 4.601949634443542e-06,
+      "loss": 1.9853,
+      "step": 484
+    },
+    {
+      "epoch": 0.5315068493150685,
+      "grad_norm": 2.317995071411133,
+      "learning_rate": 4.5999187652315195e-06,
+      "loss": 1.9183,
+      "step": 485
+    },
+    {
+      "epoch": 0.5326027397260275,
+      "grad_norm": 2.2462668418884277,
+      "learning_rate": 4.597887896019497e-06,
+      "loss": 1.9201,
+      "step": 486
+    },
+    {
+      "epoch": 0.5336986301369863,
+      "grad_norm": 2.1482582092285156,
+      "learning_rate": 4.595857026807474e-06,
+      "loss": 1.9722,
+      "step": 487
+    },
+    {
+      "epoch": 0.5347945205479452,
+      "grad_norm": 2.079831123352051,
+      "learning_rate": 4.593826157595451e-06,
+      "loss": 1.977,
+      "step": 488
+    },
+    {
+      "epoch": 0.5358904109589041,
+      "grad_norm": 2.2169578075408936,
+      "learning_rate": 4.591795288383429e-06,
+      "loss": 1.9615,
+      "step": 489
+    },
+    {
+      "epoch": 0.536986301369863,
+      "grad_norm": 2.087846279144287,
+      "learning_rate": 4.589764419171406e-06,
+      "loss": 2.0402,
+      "step": 490
+    },
+    {
+      "epoch": 0.536986301369863,
+      "eval_loss": 2.1416702270507812,
+      "eval_runtime": 37.1191,
+      "eval_samples_per_second": 26.94,
+      "eval_steps_per_second": 1.131,
+      "step": 490
+    },
+    {
+      "epoch": 0.538082191780822,
+      "grad_norm": 2.2161977291107178,
+      "learning_rate": 4.587733549959383e-06,
+      "loss": 1.9959,
+      "step": 491
+    },
+    {
+      "epoch": 0.5391780821917809,
+      "grad_norm": 2.19530987739563,
+      "learning_rate": 4.5857026807473604e-06,
+      "loss": 1.961,
+      "step": 492
+    },
+    {
+      "epoch": 0.5402739726027397,
+      "grad_norm": 2.1150615215301514,
+      "learning_rate": 4.583671811535338e-06,
+      "loss": 1.946,
+      "step": 493
+    },
+    {
+      "epoch": 0.5413698630136986,
+      "grad_norm": 2.186352491378784,
+      "learning_rate": 4.581640942323315e-06,
+      "loss": 1.9137,
+      "step": 494
+    },
+    {
+      "epoch": 0.5424657534246575,
+      "grad_norm": 2.233370780944824,
+      "learning_rate": 4.579610073111292e-06,
+      "loss": 2.021,
+      "step": 495
+    },
+    {
+      "epoch": 0.5435616438356164,
+      "grad_norm": 2.1326382160186768,
+      "learning_rate": 4.5775792038992695e-06,
+      "loss": 2.0242,
+      "step": 496
+    },
+    {
+      "epoch": 0.5446575342465754,
+      "grad_norm": 2.536837577819824,
+      "learning_rate": 4.575548334687247e-06,
+      "loss": 2.0185,
+      "step": 497
+    },
+    {
+      "epoch": 0.5457534246575343,
+      "grad_norm": 2.193598747253418,
+      "learning_rate": 4.573517465475224e-06,
+      "loss": 2.0533,
+      "step": 498
+    },
+    {
+      "epoch": 0.5468493150684931,
+      "grad_norm": 2.1346311569213867,
+      "learning_rate": 4.571486596263201e-06,
+      "loss": 1.9564,
+      "step": 499
+    },
+    {
+      "epoch": 0.547945205479452,
+      "grad_norm": 2.4294722080230713,
+      "learning_rate": 4.569455727051178e-06,
+      "loss": 2.0003,
+      "step": 500
+    },
+    {
+      "epoch": 0.547945205479452,
+      "eval_loss": 2.1430764198303223,
+      "eval_runtime": 37.0591,
+      "eval_samples_per_second": 26.984,
+      "eval_steps_per_second": 1.133,
+      "step": 500
+    },
+    {
+      "epoch": 0.549041095890411,
+      "grad_norm": 2.2394182682037354,
+      "learning_rate": 4.567424857839156e-06,
+      "loss": 2.0055,
+      "step": 501
+    },
+    {
+      "epoch": 0.5501369863013699,
+      "grad_norm": 2.2543067932128906,
+      "learning_rate": 4.565393988627132e-06,
+      "loss": 1.9837,
+      "step": 502
+    },
+    {
+      "epoch": 0.5512328767123288,
+      "grad_norm": 2.2805464267730713,
+      "learning_rate": 4.5633631194151105e-06,
+      "loss": 2.0495,
+      "step": 503
+    },
+    {
+      "epoch": 0.5523287671232877,
+      "grad_norm": 2.2838146686553955,
+      "learning_rate": 4.561332250203087e-06,
+      "loss": 1.9866,
+      "step": 504
+    },
+    {
+      "epoch": 0.5534246575342465,
+      "grad_norm": 2.286147356033325,
+      "learning_rate": 4.559301380991065e-06,
+      "loss": 1.9805,
+      "step": 505
+    },
+    {
+      "epoch": 0.5545205479452054,
+      "grad_norm": 2.064378023147583,
+      "learning_rate": 4.5572705117790414e-06,
+      "loss": 1.9259,
+      "step": 506
+    },
+    {
+      "epoch": 0.5556164383561644,
+      "grad_norm": 2.4020493030548096,
+      "learning_rate": 4.5552396425670196e-06,
+      "loss": 1.9731,
+      "step": 507
+    },
+    {
+      "epoch": 0.5567123287671233,
+      "grad_norm": 2.11594557762146,
+      "learning_rate": 4.553208773354996e-06,
+      "loss": 2.0126,
+      "step": 508
+    },
+    {
+      "epoch": 0.5578082191780822,
+      "grad_norm": 2.4520673751831055,
+      "learning_rate": 4.551177904142974e-06,
+      "loss": 1.9396,
+      "step": 509
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "grad_norm": 2.302771806716919,
+      "learning_rate": 4.5491470349309505e-06,
+      "loss": 2.0361,
+      "step": 510
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "eval_loss": 2.1408755779266357,
+      "eval_runtime": 36.9881,
+      "eval_samples_per_second": 27.036,
+      "eval_steps_per_second": 1.136,
+      "step": 510
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 2.250927686691284,
+      "learning_rate": 4.547116165718929e-06,
+      "loss": 1.9616,
+      "step": 511
+    },
+    {
+      "epoch": 0.5610958904109589,
+      "grad_norm": 2.4257938861846924,
+      "learning_rate": 4.545085296506905e-06,
+      "loss": 1.9836,
+      "step": 512
+    },
+    {
+      "epoch": 0.5621917808219178,
+      "grad_norm": 2.08693265914917,
+      "learning_rate": 4.543054427294882e-06,
+      "loss": 1.9541,
+      "step": 513
+    },
+    {
+      "epoch": 0.5632876712328767,
+      "grad_norm": 2.133875846862793,
+      "learning_rate": 4.54102355808286e-06,
+      "loss": 1.9377,
+      "step": 514
+    },
+    {
+      "epoch": 0.5643835616438356,
+      "grad_norm": 2.1217594146728516,
+      "learning_rate": 4.538992688870837e-06,
+      "loss": 1.9369,
+      "step": 515
+    },
+    {
+      "epoch": 0.5654794520547946,
+      "grad_norm": 2.117994546890259,
+      "learning_rate": 4.536961819658814e-06,
+      "loss": 1.9666,
+      "step": 516
+    },
+    {
+      "epoch": 0.5665753424657535,
+      "grad_norm": 2.1514523029327393,
+      "learning_rate": 4.5349309504467915e-06,
+      "loss": 1.9704,
+      "step": 517
+    },
+    {
+      "epoch": 0.5676712328767123,
+      "grad_norm": 2.2309377193450928,
+      "learning_rate": 4.532900081234769e-06,
+      "loss": 1.9858,
+      "step": 518
+    },
+    {
+      "epoch": 0.5687671232876712,
+      "grad_norm": 2.100398540496826,
+      "learning_rate": 4.530869212022746e-06,
+      "loss": 1.9573,
+      "step": 519
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "grad_norm": 2.122840166091919,
+      "learning_rate": 4.528838342810723e-06,
+      "loss": 1.9813,
+      "step": 520
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "eval_loss": 2.135481357574463,
+      "eval_runtime": 36.893,
+      "eval_samples_per_second": 27.105,
+      "eval_steps_per_second": 1.138,
+      "step": 520
+    },
+    {
+      "epoch": 0.570958904109589,
+      "grad_norm": 2.1103999614715576,
+      "learning_rate": 4.526807473598701e-06,
+      "loss": 1.9691,
+      "step": 521
+    },
+    {
+      "epoch": 0.572054794520548,
+      "grad_norm": 2.059232473373413,
+      "learning_rate": 4.524776604386678e-06,
+      "loss": 1.9687,
+      "step": 522
+    },
+    {
+      "epoch": 0.5731506849315069,
+      "grad_norm": 2.207000732421875,
+      "learning_rate": 4.522745735174655e-06,
+      "loss": 1.949,
+      "step": 523
+    },
+    {
+      "epoch": 0.5742465753424657,
+      "grad_norm": 2.264000177383423,
+      "learning_rate": 4.520714865962632e-06,
+      "loss": 2.0066,
+      "step": 524
+    },
+    {
+      "epoch": 0.5753424657534246,
+      "grad_norm": 2.121701717376709,
+      "learning_rate": 4.51868399675061e-06,
+      "loss": 1.9879,
+      "step": 525
+    },
+    {
+      "epoch": 0.5764383561643835,
+      "grad_norm": 2.0745644569396973,
+      "learning_rate": 4.516653127538587e-06,
+      "loss": 1.9924,
+      "step": 526
+    },
+    {
+      "epoch": 0.5775342465753425,
+      "grad_norm": 2.2257349491119385,
+      "learning_rate": 4.514622258326564e-06,
+      "loss": 1.9525,
+      "step": 527
+    },
+    {
+      "epoch": 0.5786301369863014,
+      "grad_norm": 2.1678507328033447,
+      "learning_rate": 4.5125913891145415e-06,
+      "loss": 1.9378,
+      "step": 528
+    },
+    {
+      "epoch": 0.5797260273972603,
+      "grad_norm": 2.2800841331481934,
+      "learning_rate": 4.510560519902519e-06,
+      "loss": 1.9685,
+      "step": 529
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "grad_norm": 2.1471869945526123,
+      "learning_rate": 4.508529650690496e-06,
+      "loss": 2.0027,
+      "step": 530
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "eval_loss": 2.1354892253875732,
+      "eval_runtime": 36.8434,
+      "eval_samples_per_second": 27.142,
+      "eval_steps_per_second": 1.14,
+      "step": 530
+    },
+    {
+      "epoch": 0.581917808219178,
+      "grad_norm": 2.2238621711730957,
+      "learning_rate": 4.506498781478473e-06,
+      "loss": 1.9258,
+      "step": 531
+    },
+    {
+      "epoch": 0.583013698630137,
+      "grad_norm": 2.2224624156951904,
+      "learning_rate": 4.504467912266451e-06,
+      "loss": 1.9168,
+      "step": 532
+    },
+    {
+      "epoch": 0.5841095890410959,
+      "grad_norm": 1.9850294589996338,
+      "learning_rate": 4.502437043054428e-06,
+      "loss": 1.9656,
+      "step": 533
+    },
+    {
+      "epoch": 0.5852054794520548,
+      "grad_norm": 2.1303517818450928,
+      "learning_rate": 4.500406173842405e-06,
+      "loss": 2.0349,
+      "step": 534
+    },
+    {
+      "epoch": 0.5863013698630137,
+      "grad_norm": 2.2114832401275635,
+      "learning_rate": 4.4983753046303824e-06,
+      "loss": 1.9878,
+      "step": 535
+    },
+    {
+      "epoch": 0.5873972602739727,
+      "grad_norm": 1.9375886917114258,
+      "learning_rate": 4.49634443541836e-06,
+      "loss": 1.9724,
+      "step": 536
+    },
+    {
+      "epoch": 0.5884931506849315,
+      "grad_norm": 2.185518741607666,
+      "learning_rate": 4.494313566206336e-06,
+      "loss": 1.9584,
+      "step": 537
+    },
+    {
+      "epoch": 0.5895890410958904,
+      "grad_norm": 2.125638484954834,
+      "learning_rate": 4.492282696994314e-06,
+      "loss": 2.0203,
+      "step": 538
+    },
+    {
+      "epoch": 0.5906849315068493,
+      "grad_norm": 1.9456262588500977,
+      "learning_rate": 4.490251827782291e-06,
+      "loss": 1.9523,
+      "step": 539
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "grad_norm": 2.088125705718994,
+      "learning_rate": 4.488220958570269e-06,
+      "loss": 1.9271,
+      "step": 540
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "eval_loss": 2.134958505630493,
+      "eval_runtime": 36.8342,
+      "eval_samples_per_second": 27.149,
+      "eval_steps_per_second": 1.14,
+      "step": 540
+    },
+    {
+      "epoch": 0.5928767123287672,
+      "grad_norm": 2.1423559188842773,
+      "learning_rate": 4.486190089358245e-06,
+      "loss": 1.9467,
+      "step": 541
+    },
+    {
+      "epoch": 0.5939726027397261,
+      "grad_norm": 2.2169737815856934,
+      "learning_rate": 4.484159220146223e-06,
+      "loss": 1.8901,
+      "step": 542
+    },
+    {
+      "epoch": 0.5950684931506849,
+      "grad_norm": 2.419947385787964,
+      "learning_rate": 4.4821283509342e-06,
+      "loss": 2.0064,
+      "step": 543
+    },
+    {
+      "epoch": 0.5961643835616438,
+      "grad_norm": 2.0396435260772705,
+      "learning_rate": 4.480097481722178e-06,
+      "loss": 1.9819,
+      "step": 544
+    },
+    {
+      "epoch": 0.5972602739726027,
+      "grad_norm": 2.210700511932373,
+      "learning_rate": 4.478066612510154e-06,
+      "loss": 1.9578,
+      "step": 545
+    },
+    {
+      "epoch": 0.5983561643835617,
+      "grad_norm": 2.159698009490967,
+      "learning_rate": 4.4760357432981325e-06,
+      "loss": 1.9313,
+      "step": 546
+    },
+    {
+      "epoch": 0.5994520547945206,
+      "grad_norm": 2.125058650970459,
+      "learning_rate": 4.474004874086109e-06,
+      "loss": 1.9549,
+      "step": 547
+    },
+    {
+      "epoch": 0.6005479452054795,
+      "grad_norm": 2.477320432662964,
+      "learning_rate": 4.471974004874087e-06,
+      "loss": 2.0883,
+      "step": 548
+    },
+    {
+      "epoch": 0.6016438356164383,
+      "grad_norm": 2.214901924133301,
+      "learning_rate": 4.4699431356620635e-06,
+      "loss": 1.9184,
+      "step": 549
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "grad_norm": 2.1001312732696533,
+      "learning_rate": 4.467912266450041e-06,
+      "loss": 2.0231,
+      "step": 550
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "eval_loss": 2.1315157413482666,
+      "eval_runtime": 36.8131,
+      "eval_samples_per_second": 27.164,
+      "eval_steps_per_second": 1.141,
+      "step": 550
+    },
+    {
+      "epoch": 0.6038356164383561,
+      "grad_norm": 2.1030538082122803,
+      "learning_rate": 4.465881397238018e-06,
+      "loss": 1.967,
+      "step": 551
+    },
+    {
+      "epoch": 0.6049315068493151,
+      "grad_norm": 2.0390465259552,
+      "learning_rate": 4.463850528025995e-06,
+      "loss": 1.9615,
+      "step": 552
+    },
+    {
+      "epoch": 0.606027397260274,
+      "grad_norm": 2.1331627368927,
+      "learning_rate": 4.4618196588139726e-06,
+      "loss": 1.9667,
+      "step": 553
+    },
+    {
+      "epoch": 0.6071232876712329,
+      "grad_norm": 2.1987671852111816,
+      "learning_rate": 4.45978878960195e-06,
+      "loss": 1.9922,
+      "step": 554
+    },
+    {
+      "epoch": 0.6082191780821918,
+      "grad_norm": 2.127434253692627,
+      "learning_rate": 4.457757920389927e-06,
+      "loss": 1.9913,
+      "step": 555
+    },
+    {
+      "epoch": 0.6093150684931506,
+      "grad_norm": 2.2486324310302734,
+      "learning_rate": 4.455727051177904e-06,
+      "loss": 1.9239,
+      "step": 556
+    },
+    {
+      "epoch": 0.6104109589041096,
+      "grad_norm": 2.127575397491455,
+      "learning_rate": 4.453696181965882e-06,
+      "loss": 1.9521,
+      "step": 557
+    },
+    {
+      "epoch": 0.6115068493150685,
+      "grad_norm": 2.0798254013061523,
+      "learning_rate": 4.451665312753859e-06,
+      "loss": 2.0085,
+      "step": 558
+    },
+    {
+      "epoch": 0.6126027397260274,
+      "grad_norm": 2.287574291229248,
+      "learning_rate": 4.449634443541836e-06,
+      "loss": 1.98,
+      "step": 559
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "grad_norm": 2.2056541442871094,
+      "learning_rate": 4.4476035743298135e-06,
+      "loss": 2.0282,
+      "step": 560
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "eval_loss": 2.1287026405334473,
+      "eval_runtime": 36.7788,
+      "eval_samples_per_second": 27.19,
+      "eval_steps_per_second": 1.142,
+      "step": 560
+    },
+    {
+      "epoch": 0.6147945205479453,
+      "grad_norm": 2.169157028198242,
+      "learning_rate": 4.445572705117791e-06,
+      "loss": 1.9308,
+      "step": 561
+    },
+    {
+      "epoch": 0.6158904109589041,
+      "grad_norm": 2.1035149097442627,
+      "learning_rate": 4.443541835905768e-06,
+      "loss": 1.9507,
+      "step": 562
+    },
+    {
+      "epoch": 0.616986301369863,
+      "grad_norm": 2.1174328327178955,
+      "learning_rate": 4.441510966693745e-06,
+      "loss": 2.0162,
+      "step": 563
+    },
+    {
+      "epoch": 0.6180821917808219,
+      "grad_norm": 2.201641321182251,
+      "learning_rate": 4.439480097481723e-06,
+      "loss": 1.9573,
+      "step": 564
+    },
+    {
+      "epoch": 0.6191780821917808,
+      "grad_norm": 2.1763603687286377,
+      "learning_rate": 4.4374492282697e-06,
+      "loss": 2.0149,
+      "step": 565
+    },
+    {
+      "epoch": 0.6202739726027398,
+      "grad_norm": 2.2866439819335938,
+      "learning_rate": 4.435418359057677e-06,
+      "loss": 1.8836,
+      "step": 566
+    },
+    {
+      "epoch": 0.6213698630136987,
+      "grad_norm": 2.2060160636901855,
+      "learning_rate": 4.433387489845654e-06,
+      "loss": 1.9806,
+      "step": 567
+    },
+    {
+      "epoch": 0.6224657534246575,
+      "grad_norm": 2.0614662170410156,
+      "learning_rate": 4.431356620633632e-06,
+      "loss": 1.9821,
+      "step": 568
+    },
+    {
+      "epoch": 0.6235616438356164,
+      "grad_norm": 2.1055402755737305,
+      "learning_rate": 4.429325751421609e-06,
+      "loss": 1.9365,
+      "step": 569
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "grad_norm": 1.9907679557800293,
+      "learning_rate": 4.427294882209586e-06,
+      "loss": 1.9429,
+      "step": 570
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "eval_loss": 2.1286206245422363,
+      "eval_runtime": 36.7748,
+      "eval_samples_per_second": 27.193,
+      "eval_steps_per_second": 1.142,
+      "step": 570
+    },
+    {
+      "epoch": 0.6257534246575343,
+      "grad_norm": 2.0169243812561035,
+      "learning_rate": 4.4252640129975635e-06,
+      "loss": 1.9643,
+      "step": 571
+    },
+    {
+      "epoch": 0.6268493150684932,
+      "grad_norm": 2.1082401275634766,
+      "learning_rate": 4.423233143785541e-06,
+      "loss": 1.9009,
+      "step": 572
+    },
+    {
+      "epoch": 0.6279452054794521,
+      "grad_norm": 2.202542543411255,
+      "learning_rate": 4.421202274573518e-06,
+      "loss": 1.9694,
+      "step": 573
+    },
+    {
+      "epoch": 0.6290410958904109,
+      "grad_norm": 2.028489828109741,
+      "learning_rate": 4.4191714053614945e-06,
+      "loss": 1.9697,
+      "step": 574
+    },
+    {
+      "epoch": 0.6301369863013698,
+      "grad_norm": 2.134274482727051,
+      "learning_rate": 4.417140536149473e-06,
+      "loss": 1.8672,
+      "step": 575
+    },
+    {
+      "epoch": 0.6312328767123287,
+      "grad_norm": 2.063542366027832,
+      "learning_rate": 4.415109666937449e-06,
+      "loss": 1.909,
+      "step": 576
+    },
+    {
+      "epoch": 0.6323287671232877,
+      "grad_norm": 1.9747039079666138,
+      "learning_rate": 4.413078797725427e-06,
+      "loss": 1.9795,
+      "step": 577
+    },
+    {
+      "epoch": 0.6334246575342466,
+      "grad_norm": 2.063673496246338,
+      "learning_rate": 4.411047928513404e-06,
+      "loss": 1.9283,
+      "step": 578
+    },
+    {
+      "epoch": 0.6345205479452055,
+      "grad_norm": 2.024611711502075,
+      "learning_rate": 4.409017059301382e-06,
+      "loss": 1.9682,
+      "step": 579
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "grad_norm": 2.0471599102020264,
+      "learning_rate": 4.406986190089358e-06,
+      "loss": 1.969,
+      "step": 580
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "eval_loss": 2.1233692169189453,
+      "eval_runtime": 36.8185,
+      "eval_samples_per_second": 27.16,
+      "eval_steps_per_second": 1.141,
+      "step": 580
+    },
+    {
+      "epoch": 0.6367123287671232,
+      "grad_norm": 2.1038553714752197,
+      "learning_rate": 4.404955320877336e-06,
+      "loss": 1.9915,
+      "step": 581
+    },
+    {
+      "epoch": 0.6378082191780822,
+      "grad_norm": 2.1222152709960938,
+      "learning_rate": 4.402924451665313e-06,
+      "loss": 1.9892,
+      "step": 582
+    },
+    {
+      "epoch": 0.6389041095890411,
+      "grad_norm": 2.09137225151062,
+      "learning_rate": 4.400893582453291e-06,
+      "loss": 1.934,
+      "step": 583
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 2.2158281803131104,
+      "learning_rate": 4.398862713241267e-06,
+      "loss": 1.9188,
+      "step": 584
+    },
+    {
+      "epoch": 0.6410958904109589,
+      "grad_norm": 2.34619140625,
+      "learning_rate": 4.396831844029245e-06,
+      "loss": 1.9495,
+      "step": 585
+    },
+    {
+      "epoch": 0.6421917808219179,
+      "grad_norm": 2.2732315063476562,
+      "learning_rate": 4.394800974817222e-06,
+      "loss": 1.9632,
+      "step": 586
+    },
+    {
+      "epoch": 0.6432876712328767,
+      "grad_norm": 2.153338670730591,
+      "learning_rate": 4.392770105605199e-06,
+      "loss": 1.9551,
+      "step": 587
+    },
+    {
+      "epoch": 0.6443835616438356,
+      "grad_norm": 2.2201521396636963,
+      "learning_rate": 4.390739236393176e-06,
+      "loss": 1.9677,
+      "step": 588
+    },
+    {
+      "epoch": 0.6454794520547945,
+      "grad_norm": 1.9894570112228394,
+      "learning_rate": 4.388708367181154e-06,
+      "loss": 1.9826,
+      "step": 589
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "grad_norm": 2.145061492919922,
+      "learning_rate": 4.386677497969131e-06,
+      "loss": 1.9826,
+      "step": 590
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "eval_loss": 2.122312545776367,
+      "eval_runtime": 37.1153,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 590
+    },
+    {
+      "epoch": 0.6476712328767124,
+      "grad_norm": 2.257237672805786,
+      "learning_rate": 4.384646628757108e-06,
+      "loss": 1.9722,
+      "step": 591
+    },
+    {
+      "epoch": 0.6487671232876713,
+      "grad_norm": 2.2406375408172607,
+      "learning_rate": 4.3826157595450855e-06,
+      "loss": 2.0141,
+      "step": 592
+    },
+    {
+      "epoch": 0.6498630136986301,
+      "grad_norm": 1.9656853675842285,
+      "learning_rate": 4.380584890333063e-06,
+      "loss": 1.9435,
+      "step": 593
+    },
+    {
+      "epoch": 0.650958904109589,
+      "grad_norm": 2.1063120365142822,
+      "learning_rate": 4.37855402112104e-06,
+      "loss": 1.9001,
+      "step": 594
+    },
+    {
+      "epoch": 0.6520547945205479,
+      "grad_norm": 2.1594269275665283,
+      "learning_rate": 4.376523151909017e-06,
+      "loss": 1.8596,
+      "step": 595
+    },
+    {
+      "epoch": 0.6531506849315069,
+      "grad_norm": 2.2261908054351807,
+      "learning_rate": 4.3744922826969946e-06,
+      "loss": 2.016,
+      "step": 596
+    },
+    {
+      "epoch": 0.6542465753424658,
+      "grad_norm": 1.9940093755722046,
+      "learning_rate": 4.372461413484972e-06,
+      "loss": 1.9293,
+      "step": 597
+    },
+    {
+      "epoch": 0.6553424657534247,
+      "grad_norm": 2.62186861038208,
+      "learning_rate": 4.370430544272949e-06,
+      "loss": 1.9656,
+      "step": 598
+    },
+    {
+      "epoch": 0.6564383561643835,
+      "grad_norm": 2.3531494140625,
+      "learning_rate": 4.368399675060926e-06,
+      "loss": 1.9984,
+      "step": 599
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "grad_norm": 2.362769365310669,
+      "learning_rate": 4.366368805848904e-06,
+      "loss": 1.9971,
+      "step": 600
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "eval_loss": 2.1172940731048584,
+      "eval_runtime": 37.0498,
+      "eval_samples_per_second": 26.991,
+      "eval_steps_per_second": 1.134,
+      "step": 600
+    },
+    {
+      "epoch": 0.6586301369863014,
+      "grad_norm": 2.3806607723236084,
+      "learning_rate": 4.364337936636881e-06,
+      "loss": 1.9566,
+      "step": 601
+    },
+    {
+      "epoch": 0.6597260273972603,
+      "grad_norm": 2.5531814098358154,
+      "learning_rate": 4.362307067424858e-06,
+      "loss": 1.9002,
+      "step": 602
+    },
+    {
+      "epoch": 0.6608219178082192,
+      "grad_norm": 2.283364772796631,
+      "learning_rate": 4.3602761982128355e-06,
+      "loss": 1.9633,
+      "step": 603
+    },
+    {
+      "epoch": 0.6619178082191781,
+      "grad_norm": 2.180171489715576,
+      "learning_rate": 4.358245329000813e-06,
+      "loss": 1.9277,
+      "step": 604
+    },
+    {
+      "epoch": 0.663013698630137,
+      "grad_norm": 2.0771350860595703,
+      "learning_rate": 4.35621445978879e-06,
+      "loss": 1.9484,
+      "step": 605
+    },
+    {
+      "epoch": 0.6641095890410958,
+      "grad_norm": 2.3772811889648438,
+      "learning_rate": 4.354183590576767e-06,
+      "loss": 1.9564,
+      "step": 606
+    },
+    {
+      "epoch": 0.6652054794520548,
+      "grad_norm": 2.205486297607422,
+      "learning_rate": 4.352152721364745e-06,
+      "loss": 1.9389,
+      "step": 607
+    },
+    {
+      "epoch": 0.6663013698630137,
+      "grad_norm": 2.085190773010254,
+      "learning_rate": 4.350121852152722e-06,
+      "loss": 1.8948,
+      "step": 608
+    },
+    {
+      "epoch": 0.6673972602739726,
+      "grad_norm": 2.4398982524871826,
+      "learning_rate": 4.348090982940699e-06,
+      "loss": 1.9736,
+      "step": 609
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "grad_norm": 2.194087028503418,
+      "learning_rate": 4.3460601137286764e-06,
+      "loss": 1.9607,
+      "step": 610
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "eval_loss": 2.115041971206665,
+      "eval_runtime": 36.9738,
+      "eval_samples_per_second": 27.046,
+      "eval_steps_per_second": 1.136,
+      "step": 610
+    },
+    {
+      "epoch": 0.6695890410958905,
+      "grad_norm": 2.143054485321045,
+      "learning_rate": 4.344029244516654e-06,
+      "loss": 1.9733,
+      "step": 611
+    },
+    {
+      "epoch": 0.6706849315068493,
+      "grad_norm": 2.2484617233276367,
+      "learning_rate": 4.341998375304631e-06,
+      "loss": 1.9417,
+      "step": 612
+    },
+    {
+      "epoch": 0.6717808219178082,
+      "grad_norm": 2.0430805683135986,
+      "learning_rate": 4.339967506092607e-06,
+      "loss": 2.0192,
+      "step": 613
+    },
+    {
+      "epoch": 0.6728767123287671,
+      "grad_norm": 2.061089277267456,
+      "learning_rate": 4.3379366368805855e-06,
+      "loss": 1.9543,
+      "step": 614
+    },
+    {
+      "epoch": 0.673972602739726,
+      "grad_norm": 2.298022747039795,
+      "learning_rate": 4.335905767668562e-06,
+      "loss": 1.9579,
+      "step": 615
+    },
+    {
+      "epoch": 0.675068493150685,
+      "grad_norm": 2.1418962478637695,
+      "learning_rate": 4.33387489845654e-06,
+      "loss": 1.938,
+      "step": 616
+    },
+    {
+      "epoch": 0.6761643835616439,
+      "grad_norm": 2.1005501747131348,
+      "learning_rate": 4.3318440292445165e-06,
+      "loss": 1.972,
+      "step": 617
+    },
+    {
+      "epoch": 0.6772602739726027,
+      "grad_norm": 2.1665523052215576,
+      "learning_rate": 4.329813160032495e-06,
+      "loss": 1.9535,
+      "step": 618
+    },
+    {
+      "epoch": 0.6783561643835616,
+      "grad_norm": 2.144580841064453,
+      "learning_rate": 4.327782290820471e-06,
+      "loss": 1.9188,
+      "step": 619
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "grad_norm": 1.9855908155441284,
+      "learning_rate": 4.325751421608449e-06,
+      "loss": 1.912,
+      "step": 620
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "eval_loss": 2.118319272994995,
+      "eval_runtime": 36.9677,
+      "eval_samples_per_second": 27.051,
+      "eval_steps_per_second": 1.136,
+      "step": 620
+    },
+    {
+      "epoch": 0.6805479452054795,
+      "grad_norm": 2.395538091659546,
+      "learning_rate": 4.323720552396426e-06,
+      "loss": 1.9886,
+      "step": 621
+    },
+    {
+      "epoch": 0.6816438356164384,
+      "grad_norm": 2.190574884414673,
+      "learning_rate": 4.321689683184404e-06,
+      "loss": 1.9326,
+      "step": 622
+    },
+    {
+      "epoch": 0.6827397260273973,
+      "grad_norm": 2.1826560497283936,
+      "learning_rate": 4.31965881397238e-06,
+      "loss": 1.9278,
+      "step": 623
+    },
+    {
+      "epoch": 0.6838356164383562,
+      "grad_norm": 2.151097536087036,
+      "learning_rate": 4.3176279447603574e-06,
+      "loss": 1.9359,
+      "step": 624
+    },
+    {
+      "epoch": 0.684931506849315,
+      "grad_norm": 2.1830122470855713,
+      "learning_rate": 4.315597075548335e-06,
+      "loss": 1.9728,
+      "step": 625
+    },
+    {
+      "epoch": 0.686027397260274,
+      "grad_norm": 2.203325033187866,
+      "learning_rate": 4.313566206336312e-06,
+      "loss": 1.9638,
+      "step": 626
+    },
+    {
+      "epoch": 0.6871232876712329,
+      "grad_norm": 2.7240309715270996,
+      "learning_rate": 4.311535337124289e-06,
+      "loss": 1.9907,
+      "step": 627
+    },
+    {
+      "epoch": 0.6882191780821918,
+      "grad_norm": 2.133599281311035,
+      "learning_rate": 4.3095044679122665e-06,
+      "loss": 1.9352,
+      "step": 628
+    },
+    {
+      "epoch": 0.6893150684931507,
+      "grad_norm": 2.0318551063537598,
+      "learning_rate": 4.307473598700244e-06,
+      "loss": 1.9432,
+      "step": 629
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "grad_norm": 2.502101182937622,
+      "learning_rate": 4.305442729488221e-06,
+      "loss": 1.9641,
+      "step": 630
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "eval_loss": 2.1158926486968994,
+      "eval_runtime": 36.8818,
+      "eval_samples_per_second": 27.114,
+      "eval_steps_per_second": 1.139,
+      "step": 630
+    },
+    {
+      "epoch": 0.6915068493150685,
+      "grad_norm": 2.3620944023132324,
+      "learning_rate": 4.303411860276198e-06,
+      "loss": 1.8841,
+      "step": 631
+    },
+    {
+      "epoch": 0.6926027397260274,
+      "grad_norm": 2.3139240741729736,
+      "learning_rate": 4.301380991064176e-06,
+      "loss": 1.9896,
+      "step": 632
+    },
+    {
+      "epoch": 0.6936986301369863,
+      "grad_norm": 2.314877510070801,
+      "learning_rate": 4.299350121852153e-06,
+      "loss": 1.9655,
+      "step": 633
+    },
+    {
+      "epoch": 0.6947945205479452,
+      "grad_norm": 2.036371946334839,
+      "learning_rate": 4.29731925264013e-06,
+      "loss": 1.9917,
+      "step": 634
+    },
+    {
+      "epoch": 0.6958904109589041,
+      "grad_norm": 2.1859662532806396,
+      "learning_rate": 4.2952883834281075e-06,
+      "loss": 1.9805,
+      "step": 635
+    },
+    {
+      "epoch": 0.6969863013698631,
+      "grad_norm": 2.2317097187042236,
+      "learning_rate": 4.293257514216085e-06,
+      "loss": 1.9402,
+      "step": 636
+    },
+    {
+      "epoch": 0.6980821917808219,
+      "grad_norm": 2.3249905109405518,
+      "learning_rate": 4.291226645004062e-06,
+      "loss": 1.9013,
+      "step": 637
+    },
+    {
+      "epoch": 0.6991780821917808,
+      "grad_norm": 2.199758768081665,
+      "learning_rate": 4.289195775792039e-06,
+      "loss": 1.9783,
+      "step": 638
+    },
+    {
+      "epoch": 0.7002739726027397,
+      "grad_norm": 2.3598549365997314,
+      "learning_rate": 4.287164906580017e-06,
+      "loss": 1.9936,
+      "step": 639
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "grad_norm": 2.1705522537231445,
+      "learning_rate": 4.285134037367994e-06,
+      "loss": 1.9304,
+      "step": 640
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "eval_loss": 2.1055190563201904,
+      "eval_runtime": 36.8673,
+      "eval_samples_per_second": 27.124,
+      "eval_steps_per_second": 1.139,
+      "step": 640
+    },
+    {
+      "epoch": 0.7024657534246576,
+      "grad_norm": 2.3253612518310547,
+      "learning_rate": 4.283103168155971e-06,
+      "loss": 1.9196,
+      "step": 641
+    },
+    {
+      "epoch": 0.7035616438356165,
+      "grad_norm": 2.3054087162017822,
+      "learning_rate": 4.281072298943948e-06,
+      "loss": 1.9053,
+      "step": 642
+    },
+    {
+      "epoch": 0.7046575342465753,
+      "grad_norm": 2.262057304382324,
+      "learning_rate": 4.279041429731926e-06,
+      "loss": 1.8999,
+      "step": 643
+    },
+    {
+      "epoch": 0.7057534246575342,
+      "grad_norm": 2.421391248703003,
+      "learning_rate": 4.277010560519903e-06,
+      "loss": 1.9438,
+      "step": 644
+    },
+    {
+      "epoch": 0.7068493150684931,
+      "grad_norm": 2.2841498851776123,
+      "learning_rate": 4.27497969130788e-06,
+      "loss": 1.9187,
+      "step": 645
+    },
+    {
+      "epoch": 0.707945205479452,
+      "grad_norm": 2.3226938247680664,
+      "learning_rate": 4.2729488220958575e-06,
+      "loss": 1.9275,
+      "step": 646
+    },
+    {
+      "epoch": 0.709041095890411,
+      "grad_norm": 2.2306301593780518,
+      "learning_rate": 4.270917952883835e-06,
+      "loss": 1.9684,
+      "step": 647
+    },
+    {
+      "epoch": 0.7101369863013699,
+      "grad_norm": 2.0551652908325195,
+      "learning_rate": 4.268887083671812e-06,
+      "loss": 1.9849,
+      "step": 648
+    },
+    {
+      "epoch": 0.7112328767123288,
+      "grad_norm": 2.2952210903167725,
+      "learning_rate": 4.266856214459789e-06,
+      "loss": 1.9026,
+      "step": 649
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "grad_norm": 2.1630797386169434,
+      "learning_rate": 4.264825345247766e-06,
+      "loss": 2.0184,
+      "step": 650
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "eval_loss": 2.1037871837615967,
+      "eval_runtime": 36.8221,
+      "eval_samples_per_second": 27.158,
+      "eval_steps_per_second": 1.141,
+      "step": 650
+    },
+    {
+      "epoch": 0.7134246575342466,
+      "grad_norm": 1.964150309562683,
+      "learning_rate": 4.262794476035744e-06,
+      "loss": 1.8932,
+      "step": 651
+    },
+    {
+      "epoch": 0.7145205479452055,
+      "grad_norm": 2.1474685668945312,
+      "learning_rate": 4.26076360682372e-06,
+      "loss": 1.9383,
+      "step": 652
+    },
+    {
+      "epoch": 0.7156164383561644,
+      "grad_norm": 2.199824094772339,
+      "learning_rate": 4.2587327376116984e-06,
+      "loss": 1.8599,
+      "step": 653
+    },
+    {
+      "epoch": 0.7167123287671233,
+      "grad_norm": 2.12156343460083,
+      "learning_rate": 4.256701868399675e-06,
+      "loss": 1.9162,
+      "step": 654
+    },
+    {
+      "epoch": 0.7178082191780822,
+      "grad_norm": 2.3832154273986816,
+      "learning_rate": 4.254670999187653e-06,
+      "loss": 1.9613,
+      "step": 655
+    },
+    {
+      "epoch": 0.718904109589041,
+      "grad_norm": 2.063906192779541,
+      "learning_rate": 4.252640129975629e-06,
+      "loss": 1.9421,
+      "step": 656
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 2.184842586517334,
+      "learning_rate": 4.2506092607636075e-06,
+      "loss": 1.9471,
+      "step": 657
+    },
+    {
+      "epoch": 0.7210958904109589,
+      "grad_norm": 2.1340036392211914,
+      "learning_rate": 4.248578391551584e-06,
+      "loss": 1.931,
+      "step": 658
+    },
+    {
+      "epoch": 0.7221917808219178,
+      "grad_norm": 2.117544651031494,
+      "learning_rate": 4.246547522339562e-06,
+      "loss": 1.8698,
+      "step": 659
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "grad_norm": 2.308344841003418,
+      "learning_rate": 4.2445166531275385e-06,
+      "loss": 1.9217,
+      "step": 660
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "eval_loss": 2.1086323261260986,
+      "eval_runtime": 36.8082,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 660
+    },
+    {
+      "epoch": 0.7243835616438357,
+      "grad_norm": 2.2239017486572266,
+      "learning_rate": 4.242485783915517e-06,
+      "loss": 1.9286,
+      "step": 661
+    },
+    {
+      "epoch": 0.7254794520547945,
+      "grad_norm": 2.263697624206543,
+      "learning_rate": 4.240454914703493e-06,
+      "loss": 1.8821,
+      "step": 662
+    },
+    {
+      "epoch": 0.7265753424657534,
+      "grad_norm": 2.4305059909820557,
+      "learning_rate": 4.23842404549147e-06,
+      "loss": 1.8996,
+      "step": 663
+    },
+    {
+      "epoch": 0.7276712328767123,
+      "grad_norm": 2.2333695888519287,
+      "learning_rate": 4.236393176279448e-06,
+      "loss": 1.8906,
+      "step": 664
+    },
+    {
+      "epoch": 0.7287671232876712,
+      "grad_norm": 2.160982608795166,
+      "learning_rate": 4.234362307067425e-06,
+      "loss": 1.879,
+      "step": 665
+    },
+    {
+      "epoch": 0.7298630136986302,
+      "grad_norm": 2.3707094192504883,
+      "learning_rate": 4.232331437855402e-06,
+      "loss": 1.9577,
+      "step": 666
+    },
+    {
+      "epoch": 0.7309589041095891,
+      "grad_norm": 2.0502827167510986,
+      "learning_rate": 4.2303005686433795e-06,
+      "loss": 1.9801,
+      "step": 667
+    },
+    {
+      "epoch": 0.7320547945205479,
+      "grad_norm": 2.0648176670074463,
+      "learning_rate": 4.228269699431357e-06,
+      "loss": 1.978,
+      "step": 668
+    },
+    {
+      "epoch": 0.7331506849315068,
+      "grad_norm": 2.346316337585449,
+      "learning_rate": 4.226238830219334e-06,
+      "loss": 1.9387,
+      "step": 669
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "grad_norm": 1.9734236001968384,
+      "learning_rate": 4.224207961007311e-06,
+      "loss": 1.9396,
+      "step": 670
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "eval_loss": 2.1104705333709717,
+      "eval_runtime": 36.8144,
+      "eval_samples_per_second": 27.163,
+      "eval_steps_per_second": 1.141,
+      "step": 670
+    },
+    {
+      "epoch": 0.7353424657534247,
+      "grad_norm": 2.3170549869537354,
+      "learning_rate": 4.2221770917952886e-06,
+      "loss": 1.9398,
+      "step": 671
+    },
+    {
+      "epoch": 0.7364383561643836,
+      "grad_norm": 2.0437264442443848,
+      "learning_rate": 4.220146222583266e-06,
+      "loss": 1.9284,
+      "step": 672
+    },
+    {
+      "epoch": 0.7375342465753425,
+      "grad_norm": 2.0808510780334473,
+      "learning_rate": 4.218115353371243e-06,
+      "loss": 1.9425,
+      "step": 673
+    },
+    {
+      "epoch": 0.7386301369863014,
+      "grad_norm": 2.0854742527008057,
+      "learning_rate": 4.21608448415922e-06,
+      "loss": 1.9358,
+      "step": 674
+    },
+    {
+      "epoch": 0.7397260273972602,
+      "grad_norm": 2.131622552871704,
+      "learning_rate": 4.214053614947198e-06,
+      "loss": 1.9048,
+      "step": 675
+    },
+    {
+      "epoch": 0.7408219178082192,
+      "grad_norm": 2.276359796524048,
+      "learning_rate": 4.212022745735175e-06,
+      "loss": 1.9871,
+      "step": 676
+    },
+    {
+      "epoch": 0.7419178082191781,
+      "grad_norm": 2.2194812297821045,
+      "learning_rate": 4.209991876523152e-06,
+      "loss": 1.882,
+      "step": 677
+    },
+    {
+      "epoch": 0.743013698630137,
+      "grad_norm": 2.181891679763794,
+      "learning_rate": 4.2079610073111295e-06,
+      "loss": 1.9546,
+      "step": 678
+    },
+    {
+      "epoch": 0.7441095890410959,
+      "grad_norm": 2.0956673622131348,
+      "learning_rate": 4.205930138099107e-06,
+      "loss": 1.8836,
+      "step": 679
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "grad_norm": 2.057894229888916,
+      "learning_rate": 4.203899268887084e-06,
+      "loss": 1.9842,
+      "step": 680
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "eval_loss": 2.104182720184326,
+      "eval_runtime": 36.8189,
+      "eval_samples_per_second": 27.16,
+      "eval_steps_per_second": 1.141,
+      "step": 680
+    },
+    {
+      "epoch": 0.7463013698630137,
+      "grad_norm": 2.0965662002563477,
+      "learning_rate": 4.201868399675061e-06,
+      "loss": 1.9378,
+      "step": 681
+    },
+    {
+      "epoch": 0.7473972602739726,
+      "grad_norm": 2.1843457221984863,
+      "learning_rate": 4.199837530463039e-06,
+      "loss": 1.9784,
+      "step": 682
+    },
+    {
+      "epoch": 0.7484931506849315,
+      "grad_norm": 2.2044246196746826,
+      "learning_rate": 4.197806661251016e-06,
+      "loss": 1.8642,
+      "step": 683
+    },
+    {
+      "epoch": 0.7495890410958904,
+      "grad_norm": 2.202115058898926,
+      "learning_rate": 4.195775792038993e-06,
+      "loss": 1.929,
+      "step": 684
+    },
+    {
+      "epoch": 0.7506849315068493,
+      "grad_norm": 2.3671462535858154,
+      "learning_rate": 4.19374492282697e-06,
+      "loss": 1.9172,
+      "step": 685
+    },
+    {
+      "epoch": 0.7517808219178083,
+      "grad_norm": 2.0602803230285645,
+      "learning_rate": 4.191714053614948e-06,
+      "loss": 1.8503,
+      "step": 686
+    },
+    {
+      "epoch": 0.7528767123287671,
+      "grad_norm": 2.12145733833313,
+      "learning_rate": 4.189683184402924e-06,
+      "loss": 1.9511,
+      "step": 687
+    },
+    {
+      "epoch": 0.753972602739726,
+      "grad_norm": 2.1049914360046387,
+      "learning_rate": 4.187652315190902e-06,
+      "loss": 1.9043,
+      "step": 688
+    },
+    {
+      "epoch": 0.7550684931506849,
+      "grad_norm": 2.235497236251831,
+      "learning_rate": 4.185621445978879e-06,
+      "loss": 1.9784,
+      "step": 689
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "grad_norm": 2.173222303390503,
+      "learning_rate": 4.183590576766857e-06,
+      "loss": 1.9052,
+      "step": 690
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "eval_loss": 2.0982251167297363,
+      "eval_runtime": 37.1446,
+      "eval_samples_per_second": 26.922,
+      "eval_steps_per_second": 1.131,
+      "step": 690
+    },
+    {
+      "epoch": 0.7572602739726028,
+      "grad_norm": 2.0693562030792236,
+      "learning_rate": 4.181559707554833e-06,
+      "loss": 1.9223,
+      "step": 691
+    },
+    {
+      "epoch": 0.7583561643835617,
+      "grad_norm": 2.209419012069702,
+      "learning_rate": 4.179528838342811e-06,
+      "loss": 1.9709,
+      "step": 692
+    },
+    {
+      "epoch": 0.7594520547945205,
+      "grad_norm": 2.066041946411133,
+      "learning_rate": 4.177497969130788e-06,
+      "loss": 1.889,
+      "step": 693
+    },
+    {
+      "epoch": 0.7605479452054794,
+      "grad_norm": 2.4114325046539307,
+      "learning_rate": 4.175467099918766e-06,
+      "loss": 1.9059,
+      "step": 694
+    },
+    {
+      "epoch": 0.7616438356164383,
+      "grad_norm": 2.0185043811798096,
+      "learning_rate": 4.173436230706742e-06,
+      "loss": 1.9877,
+      "step": 695
+    },
+    {
+      "epoch": 0.7627397260273973,
+      "grad_norm": 2.23392915725708,
+      "learning_rate": 4.1714053614947205e-06,
+      "loss": 1.9199,
+      "step": 696
+    },
+    {
+      "epoch": 0.7638356164383562,
+      "grad_norm": 2.2158737182617188,
+      "learning_rate": 4.169374492282697e-06,
+      "loss": 1.9001,
+      "step": 697
+    },
+    {
+      "epoch": 0.7649315068493151,
+      "grad_norm": 2.1511690616607666,
+      "learning_rate": 4.167343623070675e-06,
+      "loss": 1.9166,
+      "step": 698
+    },
+    {
+      "epoch": 0.766027397260274,
+      "grad_norm": 2.374741554260254,
+      "learning_rate": 4.165312753858652e-06,
+      "loss": 1.9684,
+      "step": 699
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "grad_norm": 2.0409998893737793,
+      "learning_rate": 4.163281884646629e-06,
+      "loss": 1.9177,
+      "step": 700
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "eval_loss": 2.096895933151245,
+      "eval_runtime": 37.066,
+      "eval_samples_per_second": 26.979,
+      "eval_steps_per_second": 1.133,
+      "step": 700
+    },
+    {
+      "epoch": 0.7682191780821918,
+      "grad_norm": 2.3173422813415527,
+      "learning_rate": 4.161251015434607e-06,
+      "loss": 1.9071,
+      "step": 701
+    },
+    {
+      "epoch": 0.7693150684931507,
+      "grad_norm": 2.077842950820923,
+      "learning_rate": 4.159220146222583e-06,
+      "loss": 1.935,
+      "step": 702
+    },
+    {
+      "epoch": 0.7704109589041096,
+      "grad_norm": 2.0518617630004883,
+      "learning_rate": 4.157189277010561e-06,
+      "loss": 1.8942,
+      "step": 703
+    },
+    {
+      "epoch": 0.7715068493150685,
+      "grad_norm": 1.9950882196426392,
+      "learning_rate": 4.155158407798538e-06,
+      "loss": 1.917,
+      "step": 704
+    },
+    {
+      "epoch": 0.7726027397260274,
+      "grad_norm": 2.3222219944000244,
+      "learning_rate": 4.153127538586516e-06,
+      "loss": 1.9551,
+      "step": 705
+    },
+    {
+      "epoch": 0.7736986301369863,
+      "grad_norm": 2.129333972930908,
+      "learning_rate": 4.151096669374492e-06,
+      "loss": 1.9204,
+      "step": 706
+    },
+    {
+      "epoch": 0.7747945205479452,
+      "grad_norm": 2.0357553958892822,
+      "learning_rate": 4.1490658001624705e-06,
+      "loss": 1.9399,
+      "step": 707
+    },
+    {
+      "epoch": 0.7758904109589041,
+      "grad_norm": 2.233694553375244,
+      "learning_rate": 4.147034930950447e-06,
+      "loss": 1.9182,
+      "step": 708
+    },
+    {
+      "epoch": 0.776986301369863,
+      "grad_norm": 2.1391239166259766,
+      "learning_rate": 4.145004061738425e-06,
+      "loss": 1.9157,
+      "step": 709
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "grad_norm": 2.162320613861084,
+      "learning_rate": 4.1429731925264015e-06,
+      "loss": 1.9115,
+      "step": 710
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "eval_loss": 2.0954039096832275,
+      "eval_runtime": 36.9862,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 710
+    },
+    {
+      "epoch": 0.7791780821917809,
+      "grad_norm": 2.2382194995880127,
+      "learning_rate": 4.140942323314379e-06,
+      "loss": 1.8428,
+      "step": 711
+    },
+    {
+      "epoch": 0.7802739726027397,
+      "grad_norm": 2.3286185264587402,
+      "learning_rate": 4.138911454102356e-06,
+      "loss": 1.8906,
+      "step": 712
+    },
+    {
+      "epoch": 0.7813698630136986,
+      "grad_norm": 2.182708501815796,
+      "learning_rate": 4.136880584890333e-06,
+      "loss": 1.9132,
+      "step": 713
+    },
+    {
+      "epoch": 0.7824657534246575,
+      "grad_norm": 1.9978337287902832,
+      "learning_rate": 4.1348497156783106e-06,
+      "loss": 1.9433,
+      "step": 714
+    },
+    {
+      "epoch": 0.7835616438356164,
+      "grad_norm": 2.3092565536499023,
+      "learning_rate": 4.132818846466288e-06,
+      "loss": 1.9374,
+      "step": 715
+    },
+    {
+      "epoch": 0.7846575342465754,
+      "grad_norm": 2.0351054668426514,
+      "learning_rate": 4.130787977254265e-06,
+      "loss": 1.9361,
+      "step": 716
+    },
+    {
+      "epoch": 0.7857534246575343,
+      "grad_norm": 2.14770770072937,
+      "learning_rate": 4.128757108042242e-06,
+      "loss": 1.9553,
+      "step": 717
+    },
+    {
+      "epoch": 0.7868493150684932,
+      "grad_norm": 2.0613555908203125,
+      "learning_rate": 4.12672623883022e-06,
+      "loss": 1.9506,
+      "step": 718
+    },
+    {
+      "epoch": 0.787945205479452,
+      "grad_norm": 2.1312081813812256,
+      "learning_rate": 4.124695369618197e-06,
+      "loss": 2.0079,
+      "step": 719
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "grad_norm": 2.088552474975586,
+      "learning_rate": 4.122664500406174e-06,
+      "loss": 1.9166,
+      "step": 720
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "eval_loss": 2.093032121658325,
+      "eval_runtime": 36.8821,
+      "eval_samples_per_second": 27.113,
+      "eval_steps_per_second": 1.139,
+      "step": 720
+    },
+    {
+      "epoch": 0.7901369863013699,
+      "grad_norm": 2.1748929023742676,
+      "learning_rate": 4.1206336311941515e-06,
+      "loss": 1.9212,
+      "step": 721
+    },
+    {
+      "epoch": 0.7912328767123288,
+      "grad_norm": 2.1492016315460205,
+      "learning_rate": 4.118602761982129e-06,
+      "loss": 1.9452,
+      "step": 722
+    },
+    {
+      "epoch": 0.7923287671232877,
+      "grad_norm": 2.3364970684051514,
+      "learning_rate": 4.116571892770106e-06,
+      "loss": 1.8821,
+      "step": 723
+    },
+    {
+      "epoch": 0.7934246575342466,
+      "grad_norm": 2.318488597869873,
+      "learning_rate": 4.114541023558083e-06,
+      "loss": 1.8597,
+      "step": 724
+    },
+    {
+      "epoch": 0.7945205479452054,
+      "grad_norm": 2.153841257095337,
+      "learning_rate": 4.112510154346061e-06,
+      "loss": 1.9843,
+      "step": 725
+    },
+    {
+      "epoch": 0.7956164383561644,
+      "grad_norm": 2.3464651107788086,
+      "learning_rate": 4.110479285134038e-06,
+      "loss": 1.9105,
+      "step": 726
+    },
+    {
+      "epoch": 0.7967123287671233,
+      "grad_norm": 2.34413743019104,
+      "learning_rate": 4.108448415922015e-06,
+      "loss": 1.9285,
+      "step": 727
+    },
+    {
+      "epoch": 0.7978082191780822,
+      "grad_norm": 2.0822196006774902,
+      "learning_rate": 4.1064175467099924e-06,
+      "loss": 1.9536,
+      "step": 728
+    },
+    {
+      "epoch": 0.7989041095890411,
+      "grad_norm": 2.1978065967559814,
+      "learning_rate": 4.10438667749797e-06,
+      "loss": 1.9487,
+      "step": 729
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 2.4218783378601074,
+      "learning_rate": 4.102355808285947e-06,
+      "loss": 1.8113,
+      "step": 730
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 2.0963101387023926,
+      "eval_runtime": 36.8853,
+      "eval_samples_per_second": 27.111,
+      "eval_steps_per_second": 1.139,
+      "step": 730
+    },
+    {
+      "epoch": 0.8010958904109589,
+      "grad_norm": 2.3887739181518555,
+      "learning_rate": 4.100324939073924e-06,
+      "loss": 2.0024,
+      "step": 731
+    },
+    {
+      "epoch": 0.8021917808219178,
+      "grad_norm": 2.296579360961914,
+      "learning_rate": 4.0982940698619015e-06,
+      "loss": 1.9092,
+      "step": 732
+    },
+    {
+      "epoch": 0.8032876712328767,
+      "grad_norm": 2.3881773948669434,
+      "learning_rate": 4.096263200649879e-06,
+      "loss": 1.9112,
+      "step": 733
+    },
+    {
+      "epoch": 0.8043835616438356,
+      "grad_norm": 2.40120530128479,
+      "learning_rate": 4.094232331437856e-06,
+      "loss": 1.9191,
+      "step": 734
+    },
+    {
+      "epoch": 0.8054794520547945,
+      "grad_norm": 2.184706211090088,
+      "learning_rate": 4.092201462225833e-06,
+      "loss": 1.8857,
+      "step": 735
+    },
+    {
+      "epoch": 0.8065753424657535,
+      "grad_norm": 2.3338732719421387,
+      "learning_rate": 4.090170593013811e-06,
+      "loss": 1.8889,
+      "step": 736
+    },
+    {
+      "epoch": 0.8076712328767123,
+      "grad_norm": 2.4065468311309814,
+      "learning_rate": 4.088139723801787e-06,
+      "loss": 1.9309,
+      "step": 737
+    },
+    {
+      "epoch": 0.8087671232876712,
+      "grad_norm": 2.044992446899414,
+      "learning_rate": 4.086108854589765e-06,
+      "loss": 1.8727,
+      "step": 738
+    },
+    {
+      "epoch": 0.8098630136986301,
+      "grad_norm": 2.23860239982605,
+      "learning_rate": 4.084077985377742e-06,
+      "loss": 1.8936,
+      "step": 739
+    },
+    {
+      "epoch": 0.810958904109589,
+      "grad_norm": 2.1324410438537598,
+      "learning_rate": 4.08204711616572e-06,
+      "loss": 1.9035,
+      "step": 740
+    },
+    {
+      "epoch": 0.810958904109589,
+      "eval_loss": 2.092885732650757,
+      "eval_runtime": 36.7991,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 740
+    },
+    {
+      "epoch": 0.812054794520548,
+      "grad_norm": 2.337340831756592,
+      "learning_rate": 4.080016246953696e-06,
+      "loss": 1.9747,
+      "step": 741
+    },
+    {
+      "epoch": 0.8131506849315069,
+      "grad_norm": 2.262801170349121,
+      "learning_rate": 4.077985377741674e-06,
+      "loss": 1.8817,
+      "step": 742
+    },
+    {
+      "epoch": 0.8142465753424658,
+      "grad_norm": 2.0494258403778076,
+      "learning_rate": 4.075954508529651e-06,
+      "loss": 1.9493,
+      "step": 743
+    },
+    {
+      "epoch": 0.8153424657534246,
+      "grad_norm": 2.085829019546509,
+      "learning_rate": 4.073923639317629e-06,
+      "loss": 1.9611,
+      "step": 744
+    },
+    {
+      "epoch": 0.8164383561643835,
+      "grad_norm": 2.105806827545166,
+      "learning_rate": 4.071892770105605e-06,
+      "loss": 1.9779,
+      "step": 745
+    },
+    {
+      "epoch": 0.8175342465753425,
+      "grad_norm": 2.2664732933044434,
+      "learning_rate": 4.069861900893583e-06,
+      "loss": 1.9009,
+      "step": 746
+    },
+    {
+      "epoch": 0.8186301369863014,
+      "grad_norm": 1.9899768829345703,
+      "learning_rate": 4.06783103168156e-06,
+      "loss": 1.8585,
+      "step": 747
+    },
+    {
+      "epoch": 0.8197260273972603,
+      "grad_norm": 2.242236614227295,
+      "learning_rate": 4.065800162469538e-06,
+      "loss": 1.8904,
+      "step": 748
+    },
+    {
+      "epoch": 0.8208219178082192,
+      "grad_norm": 2.4481139183044434,
+      "learning_rate": 4.063769293257514e-06,
+      "loss": 1.8922,
+      "step": 749
+    },
+    {
+      "epoch": 0.821917808219178,
+      "grad_norm": 2.231722354888916,
+      "learning_rate": 4.061738424045492e-06,
+      "loss": 1.9292,
+      "step": 750
+    },
+    {
+      "epoch": 0.821917808219178,
+      "eval_loss": 2.0909852981567383,
+      "eval_runtime": 36.8255,
+      "eval_samples_per_second": 27.155,
+      "eval_steps_per_second": 1.141,
+      "step": 750
+    },
+    {
+      "epoch": 0.823013698630137,
+      "grad_norm": 2.323070764541626,
+      "learning_rate": 4.059707554833469e-06,
+      "loss": 1.9053,
+      "step": 751
+    },
+    {
+      "epoch": 0.8241095890410959,
+      "grad_norm": 2.137535572052002,
+      "learning_rate": 4.057676685621446e-06,
+      "loss": 1.9272,
+      "step": 752
+    },
+    {
+      "epoch": 0.8252054794520548,
+      "grad_norm": 2.2087643146514893,
+      "learning_rate": 4.0556458164094235e-06,
+      "loss": 1.9541,
+      "step": 753
+    },
+    {
+      "epoch": 0.8263013698630137,
+      "grad_norm": 2.04811692237854,
+      "learning_rate": 4.053614947197401e-06,
+      "loss": 1.922,
+      "step": 754
+    },
+    {
+      "epoch": 0.8273972602739726,
+      "grad_norm": 2.062622308731079,
+      "learning_rate": 4.051584077985378e-06,
+      "loss": 1.8388,
+      "step": 755
+    },
+    {
+      "epoch": 0.8284931506849315,
+      "grad_norm": 2.14631986618042,
+      "learning_rate": 4.049553208773355e-06,
+      "loss": 1.9153,
+      "step": 756
+    },
+    {
+      "epoch": 0.8295890410958904,
+      "grad_norm": 2.353529214859009,
+      "learning_rate": 4.047522339561333e-06,
+      "loss": 1.9623,
+      "step": 757
+    },
+    {
+      "epoch": 0.8306849315068493,
+      "grad_norm": 2.122342348098755,
+      "learning_rate": 4.04549147034931e-06,
+      "loss": 1.9095,
+      "step": 758
+    },
+    {
+      "epoch": 0.8317808219178082,
+      "grad_norm": 2.2540230751037598,
+      "learning_rate": 4.043460601137287e-06,
+      "loss": 1.9068,
+      "step": 759
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "grad_norm": 2.164384365081787,
+      "learning_rate": 4.041429731925264e-06,
+      "loss": 2.0239,
+      "step": 760
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "eval_loss": 2.0904648303985596,
+      "eval_runtime": 36.7617,
+      "eval_samples_per_second": 27.202,
+      "eval_steps_per_second": 1.142,
+      "step": 760
+    },
+    {
+      "epoch": 0.8339726027397261,
+      "grad_norm": 2.2487905025482178,
+      "learning_rate": 4.039398862713242e-06,
+      "loss": 1.9109,
+      "step": 761
+    },
+    {
+      "epoch": 0.8350684931506849,
+      "grad_norm": 2.0947930812835693,
+      "learning_rate": 4.037367993501219e-06,
+      "loss": 1.9133,
+      "step": 762
+    },
+    {
+      "epoch": 0.8361643835616438,
+      "grad_norm": 2.2259862422943115,
+      "learning_rate": 4.035337124289196e-06,
+      "loss": 1.9821,
+      "step": 763
+    },
+    {
+      "epoch": 0.8372602739726027,
+      "grad_norm": 2.0616455078125,
+      "learning_rate": 4.0333062550771735e-06,
+      "loss": 1.9037,
+      "step": 764
+    },
+    {
+      "epoch": 0.8383561643835616,
+      "grad_norm": 2.2704474925994873,
+      "learning_rate": 4.031275385865151e-06,
+      "loss": 1.8664,
+      "step": 765
+    },
+    {
+      "epoch": 0.8394520547945206,
+      "grad_norm": 2.0059332847595215,
+      "learning_rate": 4.029244516653128e-06,
+      "loss": 1.908,
+      "step": 766
+    },
+    {
+      "epoch": 0.8405479452054795,
+      "grad_norm": 2.0462095737457275,
+      "learning_rate": 4.027213647441105e-06,
+      "loss": 1.9397,
+      "step": 767
+    },
+    {
+      "epoch": 0.8416438356164384,
+      "grad_norm": 2.150207996368408,
+      "learning_rate": 4.025182778229083e-06,
+      "loss": 1.8505,
+      "step": 768
+    },
+    {
+      "epoch": 0.8427397260273972,
+      "grad_norm": 2.174863815307617,
+      "learning_rate": 4.02315190901706e-06,
+      "loss": 1.8895,
+      "step": 769
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "grad_norm": 2.1907355785369873,
+      "learning_rate": 4.021121039805037e-06,
+      "loss": 1.9156,
+      "step": 770
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "eval_loss": 2.0900955200195312,
+      "eval_runtime": 36.8004,
+      "eval_samples_per_second": 27.174,
+      "eval_steps_per_second": 1.141,
+      "step": 770
+    },
+    {
+      "epoch": 0.8449315068493151,
+      "grad_norm": 2.2831835746765137,
+      "learning_rate": 4.0190901705930144e-06,
+      "loss": 1.9115,
+      "step": 771
+    },
+    {
+      "epoch": 0.846027397260274,
+      "grad_norm": 2.039520263671875,
+      "learning_rate": 4.017059301380992e-06,
+      "loss": 2.0047,
+      "step": 772
+    },
+    {
+      "epoch": 0.8471232876712329,
+      "grad_norm": 2.147209405899048,
+      "learning_rate": 4.015028432168969e-06,
+      "loss": 1.8845,
+      "step": 773
+    },
+    {
+      "epoch": 0.8482191780821918,
+      "grad_norm": 2.168997287750244,
+      "learning_rate": 4.012997562956945e-06,
+      "loss": 1.9606,
+      "step": 774
+    },
+    {
+      "epoch": 0.8493150684931506,
+      "grad_norm": 2.038966417312622,
+      "learning_rate": 4.0109666937449235e-06,
+      "loss": 1.8734,
+      "step": 775
+    },
+    {
+      "epoch": 0.8504109589041096,
+      "grad_norm": 2.0531768798828125,
+      "learning_rate": 4.0089358245329e-06,
+      "loss": 1.889,
+      "step": 776
+    },
+    {
+      "epoch": 0.8515068493150685,
+      "grad_norm": 2.348008155822754,
+      "learning_rate": 4.006904955320878e-06,
+      "loss": 1.8972,
+      "step": 777
+    },
+    {
+      "epoch": 0.8526027397260274,
+      "grad_norm": 2.1647253036499023,
+      "learning_rate": 4.0048740861088545e-06,
+      "loss": 1.8652,
+      "step": 778
+    },
+    {
+      "epoch": 0.8536986301369863,
+      "grad_norm": 2.1971640586853027,
+      "learning_rate": 4.002843216896833e-06,
+      "loss": 1.9495,
+      "step": 779
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "grad_norm": 2.169632911682129,
+      "learning_rate": 4.000812347684809e-06,
+      "loss": 1.9307,
+      "step": 780
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "eval_loss": 2.087860107421875,
+      "eval_runtime": 36.7982,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 780
+    },
+    {
+      "epoch": 0.8558904109589041,
+      "grad_norm": 2.4424400329589844,
+      "learning_rate": 3.998781478472787e-06,
+      "loss": 1.8849,
+      "step": 781
+    },
+    {
+      "epoch": 0.856986301369863,
+      "grad_norm": 2.2226672172546387,
+      "learning_rate": 3.996750609260764e-06,
+      "loss": 1.8975,
+      "step": 782
+    },
+    {
+      "epoch": 0.8580821917808219,
+      "grad_norm": 2.1898491382598877,
+      "learning_rate": 3.994719740048742e-06,
+      "loss": 1.8751,
+      "step": 783
+    },
+    {
+      "epoch": 0.8591780821917808,
+      "grad_norm": 2.466963052749634,
+      "learning_rate": 3.992688870836718e-06,
+      "loss": 2.0155,
+      "step": 784
+    },
+    {
+      "epoch": 0.8602739726027397,
+      "grad_norm": 2.184039831161499,
+      "learning_rate": 3.990658001624696e-06,
+      "loss": 1.9133,
+      "step": 785
+    },
+    {
+      "epoch": 0.8613698630136987,
+      "grad_norm": 2.089228868484497,
+      "learning_rate": 3.988627132412673e-06,
+      "loss": 1.8707,
+      "step": 786
+    },
+    {
+      "epoch": 0.8624657534246576,
+      "grad_norm": 2.0700666904449463,
+      "learning_rate": 3.98659626320065e-06,
+      "loss": 1.9487,
+      "step": 787
+    },
+    {
+      "epoch": 0.8635616438356164,
+      "grad_norm": 2.1495554447174072,
+      "learning_rate": 3.984565393988627e-06,
+      "loss": 1.9037,
+      "step": 788
+    },
+    {
+      "epoch": 0.8646575342465753,
+      "grad_norm": 2.0916988849639893,
+      "learning_rate": 3.9825345247766046e-06,
+      "loss": 1.9096,
+      "step": 789
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "grad_norm": 2.072387933731079,
+      "learning_rate": 3.980503655564582e-06,
+      "loss": 1.9139,
+      "step": 790
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "eval_loss": 2.0876142978668213,
+      "eval_runtime": 37.1068,
+      "eval_samples_per_second": 26.949,
+      "eval_steps_per_second": 1.132,
+      "step": 790
+    },
+    {
+      "epoch": 0.8668493150684932,
+      "grad_norm": 2.1206343173980713,
+      "learning_rate": 3.978472786352559e-06,
+      "loss": 1.9331,
+      "step": 791
+    },
+    {
+      "epoch": 0.8679452054794521,
+      "grad_norm": 2.20263934135437,
+      "learning_rate": 3.976441917140536e-06,
+      "loss": 1.9297,
+      "step": 792
+    },
+    {
+      "epoch": 0.869041095890411,
+      "grad_norm": 2.027554750442505,
+      "learning_rate": 3.974411047928514e-06,
+      "loss": 1.919,
+      "step": 793
+    },
+    {
+      "epoch": 0.8701369863013698,
+      "grad_norm": 2.0291616916656494,
+      "learning_rate": 3.972380178716491e-06,
+      "loss": 1.9682,
+      "step": 794
+    },
+    {
+      "epoch": 0.8712328767123287,
+      "grad_norm": 2.2719762325286865,
+      "learning_rate": 3.970349309504468e-06,
+      "loss": 1.9407,
+      "step": 795
+    },
+    {
+      "epoch": 0.8723287671232877,
+      "grad_norm": 2.0452375411987305,
+      "learning_rate": 3.9683184402924455e-06,
+      "loss": 1.9487,
+      "step": 796
+    },
+    {
+      "epoch": 0.8734246575342466,
+      "grad_norm": 2.170441150665283,
+      "learning_rate": 3.966287571080423e-06,
+      "loss": 1.8651,
+      "step": 797
+    },
+    {
+      "epoch": 0.8745205479452055,
+      "grad_norm": 2.3370516300201416,
+      "learning_rate": 3.9642567018684e-06,
+      "loss": 1.8599,
+      "step": 798
+    },
+    {
+      "epoch": 0.8756164383561644,
+      "grad_norm": 2.336737632751465,
+      "learning_rate": 3.962225832656377e-06,
+      "loss": 1.9304,
+      "step": 799
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "grad_norm": 1.9708369970321655,
+      "learning_rate": 3.960194963444355e-06,
+      "loss": 1.9125,
+      "step": 800
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "eval_loss": 2.0812273025512695,
+      "eval_runtime": 37.0752,
+      "eval_samples_per_second": 26.972,
+      "eval_steps_per_second": 1.133,
+      "step": 800
+    },
+    {
+      "epoch": 0.8778082191780822,
+      "grad_norm": 2.154219627380371,
+      "learning_rate": 3.958164094232332e-06,
+      "loss": 1.8304,
+      "step": 801
+    },
+    {
+      "epoch": 0.8789041095890411,
+      "grad_norm": 2.237786054611206,
+      "learning_rate": 3.956133225020309e-06,
+      "loss": 1.9147,
+      "step": 802
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 2.07143235206604,
+      "learning_rate": 3.954102355808286e-06,
+      "loss": 1.9128,
+      "step": 803
+    },
+    {
+      "epoch": 0.8810958904109589,
+      "grad_norm": 2.2780392169952393,
+      "learning_rate": 3.952071486596264e-06,
+      "loss": 1.9145,
+      "step": 804
+    },
+    {
+      "epoch": 0.8821917808219178,
+      "grad_norm": 2.2396292686462402,
+      "learning_rate": 3.950040617384241e-06,
+      "loss": 1.9049,
+      "step": 805
+    },
+    {
+      "epoch": 0.8832876712328767,
+      "grad_norm": 2.426511287689209,
+      "learning_rate": 3.948009748172218e-06,
+      "loss": 1.8984,
+      "step": 806
+    },
+    {
+      "epoch": 0.8843835616438356,
+      "grad_norm": 2.1874325275421143,
+      "learning_rate": 3.9459788789601955e-06,
+      "loss": 1.8888,
+      "step": 807
+    },
+    {
+      "epoch": 0.8854794520547945,
+      "grad_norm": 2.038098096847534,
+      "learning_rate": 3.943948009748173e-06,
+      "loss": 1.9582,
+      "step": 808
+    },
+    {
+      "epoch": 0.8865753424657534,
+      "grad_norm": 2.3748130798339844,
+      "learning_rate": 3.94191714053615e-06,
+      "loss": 1.9091,
+      "step": 809
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "grad_norm": 2.2457494735717773,
+      "learning_rate": 3.939886271324127e-06,
+      "loss": 1.9067,
+      "step": 810
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "eval_loss": 2.0766079425811768,
+      "eval_runtime": 36.9896,
+      "eval_samples_per_second": 27.035,
+      "eval_steps_per_second": 1.135,
+      "step": 810
+    },
+    {
+      "epoch": 0.8887671232876713,
+      "grad_norm": 2.156644821166992,
+      "learning_rate": 3.937855402112104e-06,
+      "loss": 1.9347,
+      "step": 811
+    },
+    {
+      "epoch": 0.8898630136986302,
+      "grad_norm": 2.3991472721099854,
+      "learning_rate": 3.935824532900082e-06,
+      "loss": 1.8645,
+      "step": 812
+    },
+    {
+      "epoch": 0.890958904109589,
+      "grad_norm": 2.252729654312134,
+      "learning_rate": 3.933793663688058e-06,
+      "loss": 1.8537,
+      "step": 813
+    },
+    {
+      "epoch": 0.8920547945205479,
+      "grad_norm": 2.2917964458465576,
+      "learning_rate": 3.9317627944760365e-06,
+      "loss": 1.9271,
+      "step": 814
+    },
+    {
+      "epoch": 0.8931506849315068,
+      "grad_norm": 2.217310667037964,
+      "learning_rate": 3.929731925264013e-06,
+      "loss": 1.9406,
+      "step": 815
+    },
+    {
+      "epoch": 0.8942465753424658,
+      "grad_norm": 2.167931079864502,
+      "learning_rate": 3.927701056051991e-06,
+      "loss": 1.8554,
+      "step": 816
+    },
+    {
+      "epoch": 0.8953424657534247,
+      "grad_norm": 2.18337082862854,
+      "learning_rate": 3.9256701868399674e-06,
+      "loss": 1.8795,
+      "step": 817
+    },
+    {
+      "epoch": 0.8964383561643836,
+      "grad_norm": 2.2348973751068115,
+      "learning_rate": 3.9236393176279456e-06,
+      "loss": 1.8831,
+      "step": 818
+    },
+    {
+      "epoch": 0.8975342465753424,
+      "grad_norm": 2.1822657585144043,
+      "learning_rate": 3.921608448415922e-06,
+      "loss": 1.8728,
+      "step": 819
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "grad_norm": 2.5253243446350098,
+      "learning_rate": 3.9195775792039e-06,
+      "loss": 1.8633,
+      "step": 820
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "eval_loss": 2.0769834518432617,
+      "eval_runtime": 36.838,
+      "eval_samples_per_second": 27.146,
+      "eval_steps_per_second": 1.14,
+      "step": 820
+    },
+    {
+      "epoch": 0.8997260273972603,
+      "grad_norm": 2.2109973430633545,
+      "learning_rate": 3.9175467099918765e-06,
+      "loss": 1.912,
+      "step": 821
+    },
+    {
+      "epoch": 0.9008219178082192,
+      "grad_norm": 2.053692102432251,
+      "learning_rate": 3.915515840779855e-06,
+      "loss": 1.8928,
+      "step": 822
+    },
+    {
+      "epoch": 0.9019178082191781,
+      "grad_norm": 2.1418213844299316,
+      "learning_rate": 3.913484971567831e-06,
+      "loss": 1.9399,
+      "step": 823
+    },
+    {
+      "epoch": 0.903013698630137,
+      "grad_norm": 2.3615458011627197,
+      "learning_rate": 3.911454102355808e-06,
+      "loss": 1.9293,
+      "step": 824
+    },
+    {
+      "epoch": 0.9041095890410958,
+      "grad_norm": 2.098254680633545,
+      "learning_rate": 3.909423233143786e-06,
+      "loss": 1.8301,
+      "step": 825
+    },
+    {
+      "epoch": 0.9052054794520548,
+      "grad_norm": 2.4473679065704346,
+      "learning_rate": 3.907392363931763e-06,
+      "loss": 1.9254,
+      "step": 826
+    },
+    {
+      "epoch": 0.9063013698630137,
+      "grad_norm": 2.252843141555786,
+      "learning_rate": 3.90536149471974e-06,
+      "loss": 1.9155,
+      "step": 827
+    },
+    {
+      "epoch": 0.9073972602739726,
+      "grad_norm": 2.1574337482452393,
+      "learning_rate": 3.9033306255077175e-06,
+      "loss": 1.9809,
+      "step": 828
+    },
+    {
+      "epoch": 0.9084931506849315,
+      "grad_norm": 2.2899720668792725,
+      "learning_rate": 3.901299756295695e-06,
+      "loss": 2.0045,
+      "step": 829
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "grad_norm": 2.227285146713257,
+      "learning_rate": 3.899268887083672e-06,
+      "loss": 2.026,
+      "step": 830
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "eval_loss": 2.0764760971069336,
+      "eval_runtime": 36.9195,
+      "eval_samples_per_second": 27.086,
+      "eval_steps_per_second": 1.138,
+      "step": 830
+    },
+    {
+      "epoch": 0.9106849315068493,
+      "grad_norm": 2.1935312747955322,
+      "learning_rate": 3.897238017871649e-06,
+      "loss": 1.8922,
+      "step": 831
+    },
+    {
+      "epoch": 0.9117808219178082,
+      "grad_norm": 2.129692316055298,
+      "learning_rate": 3.8952071486596266e-06,
+      "loss": 1.8981,
+      "step": 832
+    },
+    {
+      "epoch": 0.9128767123287671,
+      "grad_norm": 2.4435646533966064,
+      "learning_rate": 3.893176279447604e-06,
+      "loss": 1.91,
+      "step": 833
+    },
+    {
+      "epoch": 0.913972602739726,
+      "grad_norm": 2.1556732654571533,
+      "learning_rate": 3.891145410235581e-06,
+      "loss": 1.8688,
+      "step": 834
+    },
+    {
+      "epoch": 0.915068493150685,
+      "grad_norm": 2.175891160964966,
+      "learning_rate": 3.889114541023558e-06,
+      "loss": 1.9777,
+      "step": 835
+    },
+    {
+      "epoch": 0.9161643835616439,
+      "grad_norm": 2.220466375350952,
+      "learning_rate": 3.887083671811536e-06,
+      "loss": 1.9152,
+      "step": 836
+    },
+    {
+      "epoch": 0.9172602739726028,
+      "grad_norm": 2.331397294998169,
+      "learning_rate": 3.885052802599513e-06,
+      "loss": 1.892,
+      "step": 837
+    },
+    {
+      "epoch": 0.9183561643835616,
+      "grad_norm": 2.2434773445129395,
+      "learning_rate": 3.88302193338749e-06,
+      "loss": 1.841,
+      "step": 838
+    },
+    {
+      "epoch": 0.9194520547945205,
+      "grad_norm": 2.3248002529144287,
+      "learning_rate": 3.8809910641754675e-06,
+      "loss": 1.9657,
+      "step": 839
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "grad_norm": 2.090252637863159,
+      "learning_rate": 3.878960194963445e-06,
+      "loss": 1.9121,
+      "step": 840
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "eval_loss": 2.0798065662384033,
+      "eval_runtime": 36.8557,
+      "eval_samples_per_second": 27.133,
+      "eval_steps_per_second": 1.14,
+      "step": 840
+    },
+    {
+      "epoch": 0.9216438356164384,
+      "grad_norm": 2.6968672275543213,
+      "learning_rate": 3.876929325751422e-06,
+      "loss": 1.9003,
+      "step": 841
+    },
+    {
+      "epoch": 0.9227397260273973,
+      "grad_norm": 2.356321096420288,
+      "learning_rate": 3.874898456539399e-06,
+      "loss": 1.9471,
+      "step": 842
+    },
+    {
+      "epoch": 0.9238356164383562,
+      "grad_norm": 2.2678606510162354,
+      "learning_rate": 3.872867587327377e-06,
+      "loss": 1.8488,
+      "step": 843
+    },
+    {
+      "epoch": 0.924931506849315,
+      "grad_norm": 2.1122310161590576,
+      "learning_rate": 3.870836718115354e-06,
+      "loss": 1.9247,
+      "step": 844
+    },
+    {
+      "epoch": 0.9260273972602739,
+      "grad_norm": 2.1219136714935303,
+      "learning_rate": 3.868805848903331e-06,
+      "loss": 1.8829,
+      "step": 845
+    },
+    {
+      "epoch": 0.9271232876712329,
+      "grad_norm": 2.175607204437256,
+      "learning_rate": 3.8667749796913084e-06,
+      "loss": 1.887,
+      "step": 846
+    },
+    {
+      "epoch": 0.9282191780821918,
+      "grad_norm": 2.156214714050293,
+      "learning_rate": 3.864744110479286e-06,
+      "loss": 1.9119,
+      "step": 847
+    },
+    {
+      "epoch": 0.9293150684931507,
+      "grad_norm": 2.158461570739746,
+      "learning_rate": 3.862713241267262e-06,
+      "loss": 1.9092,
+      "step": 848
+    },
+    {
+      "epoch": 0.9304109589041096,
+      "grad_norm": 2.180551767349243,
+      "learning_rate": 3.86068237205524e-06,
+      "loss": 1.946,
+      "step": 849
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "grad_norm": 2.1030542850494385,
+      "learning_rate": 3.858651502843217e-06,
+      "loss": 1.9484,
+      "step": 850
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "eval_loss": 2.0767791271209717,
+      "eval_runtime": 36.8047,
+      "eval_samples_per_second": 27.17,
+      "eval_steps_per_second": 1.141,
+      "step": 850
+    },
+    {
+      "epoch": 0.9326027397260274,
+      "grad_norm": 2.240769386291504,
+      "learning_rate": 3.856620633631195e-06,
+      "loss": 1.9114,
+      "step": 851
+    },
+    {
+      "epoch": 0.9336986301369863,
+      "grad_norm": 2.1964619159698486,
+      "learning_rate": 3.854589764419171e-06,
+      "loss": 1.8339,
+      "step": 852
+    },
+    {
+      "epoch": 0.9347945205479452,
+      "grad_norm": 2.2875845432281494,
+      "learning_rate": 3.852558895207149e-06,
+      "loss": 1.8942,
+      "step": 853
+    },
+    {
+      "epoch": 0.9358904109589041,
+      "grad_norm": 2.1462507247924805,
+      "learning_rate": 3.850528025995126e-06,
+      "loss": 1.9838,
+      "step": 854
+    },
+    {
+      "epoch": 0.936986301369863,
+      "grad_norm": 2.0982913970947266,
+      "learning_rate": 3.848497156783104e-06,
+      "loss": 1.9351,
+      "step": 855
+    },
+    {
+      "epoch": 0.938082191780822,
+      "grad_norm": 2.260263442993164,
+      "learning_rate": 3.84646628757108e-06,
+      "loss": 1.8745,
+      "step": 856
+    },
+    {
+      "epoch": 0.9391780821917808,
+      "grad_norm": 2.091778039932251,
+      "learning_rate": 3.8444354183590585e-06,
+      "loss": 1.8888,
+      "step": 857
+    },
+    {
+      "epoch": 0.9402739726027397,
+      "grad_norm": 2.054974317550659,
+      "learning_rate": 3.842404549147035e-06,
+      "loss": 1.8801,
+      "step": 858
+    },
+    {
+      "epoch": 0.9413698630136986,
+      "grad_norm": 2.079641580581665,
+      "learning_rate": 3.840373679935013e-06,
+      "loss": 1.8879,
+      "step": 859
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "grad_norm": 2.3441810607910156,
+      "learning_rate": 3.8383428107229894e-06,
+      "loss": 1.8571,
+      "step": 860
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "eval_loss": 2.0766854286193848,
+      "eval_runtime": 36.8414,
+      "eval_samples_per_second": 27.143,
+      "eval_steps_per_second": 1.14,
+      "step": 860
+    },
+    {
+      "epoch": 0.9435616438356165,
+      "grad_norm": 2.123061180114746,
+      "learning_rate": 3.836311941510967e-06,
+      "loss": 1.8922,
+      "step": 861
+    },
+    {
+      "epoch": 0.9446575342465754,
+      "grad_norm": 2.1573386192321777,
+      "learning_rate": 3.834281072298944e-06,
+      "loss": 1.8713,
+      "step": 862
+    },
+    {
+      "epoch": 0.9457534246575342,
+      "grad_norm": 2.225186824798584,
+      "learning_rate": 3.832250203086921e-06,
+      "loss": 1.925,
+      "step": 863
+    },
+    {
+      "epoch": 0.9468493150684931,
+      "grad_norm": 2.1539998054504395,
+      "learning_rate": 3.8302193338748985e-06,
+      "loss": 1.8721,
+      "step": 864
+    },
+    {
+      "epoch": 0.947945205479452,
+      "grad_norm": 2.075580358505249,
+      "learning_rate": 3.828188464662876e-06,
+      "loss": 1.9489,
+      "step": 865
+    },
+    {
+      "epoch": 0.949041095890411,
+      "grad_norm": 2.1789908409118652,
+      "learning_rate": 3.826157595450853e-06,
+      "loss": 1.9883,
+      "step": 866
+    },
+    {
+      "epoch": 0.9501369863013699,
+      "grad_norm": 2.0838615894317627,
+      "learning_rate": 3.82412672623883e-06,
+      "loss": 1.9149,
+      "step": 867
+    },
+    {
+      "epoch": 0.9512328767123288,
+      "grad_norm": 2.1321213245391846,
+      "learning_rate": 3.822095857026808e-06,
+      "loss": 1.8505,
+      "step": 868
+    },
+    {
+      "epoch": 0.9523287671232876,
+      "grad_norm": 2.1419479846954346,
+      "learning_rate": 3.820064987814785e-06,
+      "loss": 1.9226,
+      "step": 869
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "grad_norm": 2.290163040161133,
+      "learning_rate": 3.818034118602762e-06,
+      "loss": 1.8737,
+      "step": 870
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "eval_loss": 2.073598623275757,
+      "eval_runtime": 36.7896,
+      "eval_samples_per_second": 27.182,
+      "eval_steps_per_second": 1.142,
+      "step": 870
+    },
+    {
+      "epoch": 0.9545205479452055,
+      "grad_norm": 2.1951074600219727,
+      "learning_rate": 3.8160032493907395e-06,
+      "loss": 1.8699,
+      "step": 871
+    },
+    {
+      "epoch": 0.9556164383561644,
+      "grad_norm": 2.25960111618042,
+      "learning_rate": 3.813972380178717e-06,
+      "loss": 1.8813,
+      "step": 872
+    },
+    {
+      "epoch": 0.9567123287671233,
+      "grad_norm": 2.4376907348632812,
+      "learning_rate": 3.811941510966694e-06,
+      "loss": 1.8782,
+      "step": 873
+    },
+    {
+      "epoch": 0.9578082191780822,
+      "grad_norm": 1.980772614479065,
+      "learning_rate": 3.809910641754671e-06,
+      "loss": 1.9205,
+      "step": 874
+    },
+    {
+      "epoch": 0.958904109589041,
+      "grad_norm": 2.0699028968811035,
+      "learning_rate": 3.8078797725426486e-06,
+      "loss": 1.8674,
+      "step": 875
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 2.2765867710113525,
+      "learning_rate": 3.8058489033306254e-06,
+      "loss": 1.81,
+      "step": 876
+    },
+    {
+      "epoch": 0.9610958904109589,
+      "grad_norm": 2.0998799800872803,
+      "learning_rate": 3.803818034118603e-06,
+      "loss": 1.9061,
+      "step": 877
+    },
+    {
+      "epoch": 0.9621917808219178,
+      "grad_norm": 2.2362914085388184,
+      "learning_rate": 3.80178716490658e-06,
+      "loss": 1.8798,
+      "step": 878
+    },
+    {
+      "epoch": 0.9632876712328767,
+      "grad_norm": 2.0789835453033447,
+      "learning_rate": 3.7997562956945577e-06,
+      "loss": 1.9491,
+      "step": 879
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "grad_norm": 2.0458688735961914,
+      "learning_rate": 3.7977254264825345e-06,
+      "loss": 1.931,
+      "step": 880
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "eval_loss": 2.0672221183776855,
+      "eval_runtime": 36.8218,
+      "eval_samples_per_second": 27.158,
+      "eval_steps_per_second": 1.141,
+      "step": 880
+    },
+    {
+      "epoch": 0.9654794520547946,
+      "grad_norm": 2.361626625061035,
+      "learning_rate": 3.7956945572705122e-06,
+      "loss": 1.8815,
+      "step": 881
+    },
+    {
+      "epoch": 0.9665753424657534,
+      "grad_norm": 2.294160842895508,
+      "learning_rate": 3.793663688058489e-06,
+      "loss": 1.8649,
+      "step": 882
+    },
+    {
+      "epoch": 0.9676712328767123,
+      "grad_norm": 2.2242612838745117,
+      "learning_rate": 3.791632818846467e-06,
+      "loss": 1.9535,
+      "step": 883
+    },
+    {
+      "epoch": 0.9687671232876712,
+      "grad_norm": 2.220269203186035,
+      "learning_rate": 3.7896019496344436e-06,
+      "loss": 1.9506,
+      "step": 884
+    },
+    {
+      "epoch": 0.9698630136986301,
+      "grad_norm": 2.112518072128296,
+      "learning_rate": 3.787571080422421e-06,
+      "loss": 1.9487,
+      "step": 885
+    },
+    {
+      "epoch": 0.9709589041095891,
+      "grad_norm": 2.17695951461792,
+      "learning_rate": 3.785540211210398e-06,
+      "loss": 1.9467,
+      "step": 886
+    },
+    {
+      "epoch": 0.972054794520548,
+      "grad_norm": 2.0990567207336426,
+      "learning_rate": 3.7835093419983755e-06,
+      "loss": 1.9216,
+      "step": 887
+    },
+    {
+      "epoch": 0.9731506849315068,
+      "grad_norm": 2.222566604614258,
+      "learning_rate": 3.7814784727863527e-06,
+      "loss": 1.9043,
+      "step": 888
+    },
+    {
+      "epoch": 0.9742465753424657,
+      "grad_norm": 2.09890079498291,
+      "learning_rate": 3.77944760357433e-06,
+      "loss": 1.884,
+      "step": 889
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "grad_norm": 2.136141061782837,
+      "learning_rate": 3.7774167343623073e-06,
+      "loss": 2.0032,
+      "step": 890
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "eval_loss": 2.0665438175201416,
+      "eval_runtime": 37.1207,
+      "eval_samples_per_second": 26.939,
+      "eval_steps_per_second": 1.131,
+      "step": 890
+    },
+    {
+      "epoch": 0.9764383561643836,
+      "grad_norm": 2.267444372177124,
+      "learning_rate": 3.7753858651502846e-06,
+      "loss": 1.947,
+      "step": 891
+    },
+    {
+      "epoch": 0.9775342465753425,
+      "grad_norm": 2.1141815185546875,
+      "learning_rate": 3.773354995938262e-06,
+      "loss": 1.8705,
+      "step": 892
+    },
+    {
+      "epoch": 0.9786301369863014,
+      "grad_norm": 2.1036782264709473,
+      "learning_rate": 3.771324126726239e-06,
+      "loss": 1.9381,
+      "step": 893
+    },
+    {
+      "epoch": 0.9797260273972602,
+      "grad_norm": 2.081982135772705,
+      "learning_rate": 3.7692932575142164e-06,
+      "loss": 1.9079,
+      "step": 894
+    },
+    {
+      "epoch": 0.9808219178082191,
+      "grad_norm": 2.0815446376800537,
+      "learning_rate": 3.7672623883021937e-06,
+      "loss": 1.9417,
+      "step": 895
+    },
+    {
+      "epoch": 0.9819178082191781,
+      "grad_norm": 2.0531511306762695,
+      "learning_rate": 3.765231519090171e-06,
+      "loss": 1.876,
+      "step": 896
+    },
+    {
+      "epoch": 0.983013698630137,
+      "grad_norm": 2.212940216064453,
+      "learning_rate": 3.7632006498781482e-06,
+      "loss": 1.8008,
+      "step": 897
+    },
+    {
+      "epoch": 0.9841095890410959,
+      "grad_norm": 2.1541531085968018,
+      "learning_rate": 3.761169780666125e-06,
+      "loss": 1.9039,
+      "step": 898
+    },
+    {
+      "epoch": 0.9852054794520548,
+      "grad_norm": 2.113884925842285,
+      "learning_rate": 3.7591389114541028e-06,
+      "loss": 1.8562,
+      "step": 899
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "grad_norm": 2.1503701210021973,
+      "learning_rate": 3.7571080422420796e-06,
+      "loss": 1.9052,
+      "step": 900
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "eval_loss": 2.0664408206939697,
+      "eval_runtime": 37.0876,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 900
+    },
+    {
+      "epoch": 0.9873972602739726,
+      "grad_norm": 2.1584348678588867,
+      "learning_rate": 3.7550771730300573e-06,
+      "loss": 1.9296,
+      "step": 901
+    },
+    {
+      "epoch": 0.9884931506849315,
+      "grad_norm": 2.164318084716797,
+      "learning_rate": 3.753046303818034e-06,
+      "loss": 1.8834,
+      "step": 902
+    },
+    {
+      "epoch": 0.9895890410958904,
+      "grad_norm": 2.0222678184509277,
+      "learning_rate": 3.751015434606012e-06,
+      "loss": 1.9426,
+      "step": 903
+    },
+    {
+      "epoch": 0.9906849315068493,
+      "grad_norm": 2.116835832595825,
+      "learning_rate": 3.7489845653939887e-06,
+      "loss": 1.9195,
+      "step": 904
+    },
+    {
+      "epoch": 0.9917808219178083,
+      "grad_norm": 2.2708873748779297,
+      "learning_rate": 3.7469536961819664e-06,
+      "loss": 1.9631,
+      "step": 905
+    },
+    {
+      "epoch": 0.9928767123287672,
+      "grad_norm": 2.1174001693725586,
+      "learning_rate": 3.7449228269699433e-06,
+      "loss": 1.8602,
+      "step": 906
+    },
+    {
+      "epoch": 0.993972602739726,
+      "grad_norm": 2.091830015182495,
+      "learning_rate": 3.742891957757921e-06,
+      "loss": 1.8906,
+      "step": 907
+    },
+    {
+      "epoch": 0.9950684931506849,
+      "grad_norm": 2.007119655609131,
+      "learning_rate": 3.740861088545898e-06,
+      "loss": 1.9122,
+      "step": 908
+    },
+    {
+      "epoch": 0.9961643835616438,
+      "grad_norm": 2.070396661758423,
+      "learning_rate": 3.7388302193338755e-06,
+      "loss": 1.9106,
+      "step": 909
+    },
+    {
+      "epoch": 0.9972602739726028,
+      "grad_norm": 2.080331325531006,
+      "learning_rate": 3.7367993501218524e-06,
+      "loss": 1.934,
+      "step": 910
+    },
+    {
+      "epoch": 0.9972602739726028,
+      "eval_loss": 2.0694408416748047,
+      "eval_runtime": 36.9953,
+      "eval_samples_per_second": 27.03,
+      "eval_steps_per_second": 1.135,
+      "step": 910
+    },
+    {
+      "epoch": 0.9983561643835617,
+      "grad_norm": 2.0920615196228027,
+      "learning_rate": 3.7347684809098297e-06,
+      "loss": 1.8679,
+      "step": 911
+    },
+    {
+      "epoch": 0.9994520547945206,
+      "grad_norm": 2.112732172012329,
+      "learning_rate": 3.732737611697807e-06,
+      "loss": 1.8723,
+      "step": 912
+    },
+    {
+      "epoch": 1.0005479452054795,
+      "grad_norm": 2.1449220180511475,
+      "learning_rate": 3.7307067424857842e-06,
+      "loss": 1.8509,
+      "step": 913
+    },
+    {
+      "epoch": 1.0016438356164383,
+      "grad_norm": 2.2179951667785645,
+      "learning_rate": 3.7286758732737615e-06,
+      "loss": 1.7911,
+      "step": 914
+    },
+    {
+      "epoch": 1.0027397260273974,
+      "grad_norm": 2.1392273902893066,
+      "learning_rate": 3.7266450040617388e-06,
+      "loss": 1.8475,
+      "step": 915
+    },
+    {
+      "epoch": 1.0038356164383562,
+      "grad_norm": 2.190265417098999,
+      "learning_rate": 3.724614134849716e-06,
+      "loss": 1.8394,
+      "step": 916
+    },
+    {
+      "epoch": 1.004931506849315,
+      "grad_norm": 2.1833760738372803,
+      "learning_rate": 3.7225832656376933e-06,
+      "loss": 1.9696,
+      "step": 917
+    },
+    {
+      "epoch": 1.006027397260274,
+      "grad_norm": 2.2504308223724365,
+      "learning_rate": 3.7205523964256706e-06,
+      "loss": 1.8135,
+      "step": 918
+    },
+    {
+      "epoch": 1.0071232876712328,
+      "grad_norm": 2.3909389972686768,
+      "learning_rate": 3.718521527213648e-06,
+      "loss": 1.9331,
+      "step": 919
+    },
+    {
+      "epoch": 1.0082191780821919,
+      "grad_norm": 2.3367602825164795,
+      "learning_rate": 3.716490658001625e-06,
+      "loss": 1.9026,
+      "step": 920
+    },
+    {
+      "epoch": 1.0082191780821919,
+      "eval_loss": 2.066087484359741,
+      "eval_runtime": 36.8701,
+      "eval_samples_per_second": 27.122,
+      "eval_steps_per_second": 1.139,
+      "step": 920
+    },
+    {
+      "epoch": 1.0093150684931507,
+      "grad_norm": 2.1732144355773926,
+      "learning_rate": 3.7144597887896024e-06,
+      "loss": 1.8888,
+      "step": 921
+    },
+    {
+      "epoch": 1.0104109589041095,
+      "grad_norm": 2.4632201194763184,
+      "learning_rate": 3.7124289195775797e-06,
+      "loss": 1.7725,
+      "step": 922
+    },
+    {
+      "epoch": 1.0115068493150685,
+      "grad_norm": 2.4887173175811768,
+      "learning_rate": 3.710398050365557e-06,
+      "loss": 1.9125,
+      "step": 923
+    },
+    {
+      "epoch": 1.0126027397260273,
+      "grad_norm": 2.4684619903564453,
+      "learning_rate": 3.708367181153534e-06,
+      "loss": 1.8829,
+      "step": 924
+    },
+    {
+      "epoch": 1.0136986301369864,
+      "grad_norm": 2.348788022994995,
+      "learning_rate": 3.7063363119415115e-06,
+      "loss": 1.8542,
+      "step": 925
+    },
+    {
+      "epoch": 1.0147945205479452,
+      "grad_norm": 2.4432766437530518,
+      "learning_rate": 3.7043054427294884e-06,
+      "loss": 1.903,
+      "step": 926
+    },
+    {
+      "epoch": 1.0158904109589042,
+      "grad_norm": 2.425910234451294,
+      "learning_rate": 3.702274573517466e-06,
+      "loss": 1.8532,
+      "step": 927
+    },
+    {
+      "epoch": 1.016986301369863,
+      "grad_norm": 2.447777032852173,
+      "learning_rate": 3.700243704305443e-06,
+      "loss": 1.8277,
+      "step": 928
+    },
+    {
+      "epoch": 1.0180821917808218,
+      "grad_norm": 2.1830408573150635,
+      "learning_rate": 3.6982128350934206e-06,
+      "loss": 1.8546,
+      "step": 929
+    },
+    {
+      "epoch": 1.0191780821917809,
+      "grad_norm": 2.2692818641662598,
+      "learning_rate": 3.6961819658813975e-06,
+      "loss": 1.8356,
+      "step": 930
+    },
+    {
+      "epoch": 1.0191780821917809,
+      "eval_loss": 2.0714614391326904,
+      "eval_runtime": 36.8661,
+      "eval_samples_per_second": 27.125,
+      "eval_steps_per_second": 1.139,
+      "step": 930
+    },
+    {
+      "epoch": 1.0202739726027397,
+      "grad_norm": 2.5044169425964355,
+      "learning_rate": 3.694151096669375e-06,
+      "loss": 1.8173,
+      "step": 931
+    },
+    {
+      "epoch": 1.0213698630136987,
+      "grad_norm": 2.419637680053711,
+      "learning_rate": 3.692120227457352e-06,
+      "loss": 1.9418,
+      "step": 932
+    },
+    {
+      "epoch": 1.0224657534246575,
+      "grad_norm": 2.282288074493408,
+      "learning_rate": 3.6900893582453297e-06,
+      "loss": 1.8698,
+      "step": 933
+    },
+    {
+      "epoch": 1.0235616438356165,
+      "grad_norm": 2.3215551376342773,
+      "learning_rate": 3.6880584890333066e-06,
+      "loss": 1.8935,
+      "step": 934
+    },
+    {
+      "epoch": 1.0246575342465754,
+      "grad_norm": 2.5938258171081543,
+      "learning_rate": 3.6860276198212834e-06,
+      "loss": 1.79,
+      "step": 935
+    },
+    {
+      "epoch": 1.0257534246575342,
+      "grad_norm": 2.215024471282959,
+      "learning_rate": 3.683996750609261e-06,
+      "loss": 1.8768,
+      "step": 936
+    },
+    {
+      "epoch": 1.0268493150684932,
+      "grad_norm": 2.239609956741333,
+      "learning_rate": 3.681965881397238e-06,
+      "loss": 1.8887,
+      "step": 937
+    },
+    {
+      "epoch": 1.027945205479452,
+      "grad_norm": 2.3134827613830566,
+      "learning_rate": 3.6799350121852157e-06,
+      "loss": 1.8897,
+      "step": 938
+    },
+    {
+      "epoch": 1.029041095890411,
+      "grad_norm": 2.4100260734558105,
+      "learning_rate": 3.6779041429731925e-06,
+      "loss": 1.909,
+      "step": 939
+    },
+    {
+      "epoch": 1.0301369863013699,
+      "grad_norm": 2.26204776763916,
+      "learning_rate": 3.6758732737611702e-06,
+      "loss": 1.8514,
+      "step": 940
+    },
+    {
+      "epoch": 1.0301369863013699,
+      "eval_loss": 2.0666027069091797,
+      "eval_runtime": 36.8336,
+      "eval_samples_per_second": 27.149,
+      "eval_steps_per_second": 1.14,
+      "step": 940
+    },
+    {
+      "epoch": 1.0312328767123287,
+      "grad_norm": 2.22210955619812,
+      "learning_rate": 3.673842404549147e-06,
+      "loss": 1.8743,
+      "step": 941
+    },
+    {
+      "epoch": 1.0323287671232877,
+      "grad_norm": 2.2371468544006348,
+      "learning_rate": 3.671811535337125e-06,
+      "loss": 1.8573,
+      "step": 942
+    },
+    {
+      "epoch": 1.0334246575342465,
+      "grad_norm": 2.346745014190674,
+      "learning_rate": 3.6697806661251016e-06,
+      "loss": 1.8614,
+      "step": 943
+    },
+    {
+      "epoch": 1.0345205479452055,
+      "grad_norm": 2.4765074253082275,
+      "learning_rate": 3.6677497969130793e-06,
+      "loss": 1.8654,
+      "step": 944
+    },
+    {
+      "epoch": 1.0356164383561643,
+      "grad_norm": 2.3023204803466797,
+      "learning_rate": 3.665718927701056e-06,
+      "loss": 1.9901,
+      "step": 945
+    },
+    {
+      "epoch": 1.0367123287671234,
+      "grad_norm": 2.322421073913574,
+      "learning_rate": 3.663688058489034e-06,
+      "loss": 1.8775,
+      "step": 946
+    },
+    {
+      "epoch": 1.0378082191780822,
+      "grad_norm": 2.2324275970458984,
+      "learning_rate": 3.6616571892770107e-06,
+      "loss": 1.9303,
+      "step": 947
+    },
+    {
+      "epoch": 1.038904109589041,
+      "grad_norm": 2.372067928314209,
+      "learning_rate": 3.659626320064988e-06,
+      "loss": 1.8673,
+      "step": 948
+    },
+    {
+      "epoch": 1.04,
+      "grad_norm": 2.2404983043670654,
+      "learning_rate": 3.6575954508529653e-06,
+      "loss": 1.8025,
+      "step": 949
+    },
+    {
+      "epoch": 1.0410958904109588,
+      "grad_norm": 2.180673599243164,
+      "learning_rate": 3.6555645816409426e-06,
+      "loss": 1.8563,
+      "step": 950
+    },
+    {
+      "epoch": 1.0410958904109588,
+      "eval_loss": 2.061262845993042,
+      "eval_runtime": 36.7727,
+      "eval_samples_per_second": 27.194,
+      "eval_steps_per_second": 1.142,
+      "step": 950
+    },
+    {
+      "epoch": 1.0421917808219179,
+      "grad_norm": 2.299043893814087,
+      "learning_rate": 3.65353371242892e-06,
+      "loss": 1.8453,
+      "step": 951
+    },
+    {
+      "epoch": 1.0432876712328767,
+      "grad_norm": 2.4878089427948,
+      "learning_rate": 3.651502843216897e-06,
+      "loss": 1.8901,
+      "step": 952
+    },
+    {
+      "epoch": 1.0443835616438357,
+      "grad_norm": 2.150404930114746,
+      "learning_rate": 3.6494719740048744e-06,
+      "loss": 1.7371,
+      "step": 953
+    },
+    {
+      "epoch": 1.0454794520547945,
+      "grad_norm": 2.177750587463379,
+      "learning_rate": 3.6474411047928517e-06,
+      "loss": 1.8814,
+      "step": 954
+    },
+    {
+      "epoch": 1.0465753424657533,
+      "grad_norm": 2.230797052383423,
+      "learning_rate": 3.645410235580829e-06,
+      "loss": 1.8323,
+      "step": 955
+    },
+    {
+      "epoch": 1.0476712328767124,
+      "grad_norm": 2.2484378814697266,
+      "learning_rate": 3.6433793663688062e-06,
+      "loss": 1.8749,
+      "step": 956
+    },
+    {
+      "epoch": 1.0487671232876712,
+      "grad_norm": 2.201279640197754,
+      "learning_rate": 3.6413484971567835e-06,
+      "loss": 1.8518,
+      "step": 957
+    },
+    {
+      "epoch": 1.0498630136986302,
+      "grad_norm": 2.1125736236572266,
+      "learning_rate": 3.6393176279447608e-06,
+      "loss": 1.846,
+      "step": 958
+    },
+    {
+      "epoch": 1.050958904109589,
+      "grad_norm": 2.3078789710998535,
+      "learning_rate": 3.637286758732738e-06,
+      "loss": 1.8791,
+      "step": 959
+    },
+    {
+      "epoch": 1.0520547945205478,
+      "grad_norm": 2.2911148071289062,
+      "learning_rate": 3.6352558895207153e-06,
+      "loss": 1.8901,
+      "step": 960
+    },
+    {
+      "epoch": 1.0520547945205478,
+      "eval_loss": 2.0583066940307617,
+      "eval_runtime": 36.7535,
+      "eval_samples_per_second": 27.208,
+      "eval_steps_per_second": 1.143,
+      "step": 960
+    },
+    {
+      "epoch": 1.0531506849315069,
+      "grad_norm": 2.3990273475646973,
+      "learning_rate": 3.633225020308692e-06,
+      "loss": 1.9071,
+      "step": 961
+    },
+    {
+      "epoch": 1.0542465753424657,
+      "grad_norm": 2.358685255050659,
+      "learning_rate": 3.63119415109667e-06,
+      "loss": 1.8982,
+      "step": 962
+    },
+    {
+      "epoch": 1.0553424657534247,
+      "grad_norm": 2.1145992279052734,
+      "learning_rate": 3.6291632818846467e-06,
+      "loss": 1.8649,
+      "step": 963
+    },
+    {
+      "epoch": 1.0564383561643835,
+      "grad_norm": 2.2302935123443604,
+      "learning_rate": 3.6271324126726244e-06,
+      "loss": 1.8911,
+      "step": 964
+    },
+    {
+      "epoch": 1.0575342465753426,
+      "grad_norm": 2.344797372817993,
+      "learning_rate": 3.6251015434606013e-06,
+      "loss": 1.8,
+      "step": 965
+    },
+    {
+      "epoch": 1.0586301369863014,
+      "grad_norm": 2.2514073848724365,
+      "learning_rate": 3.623070674248579e-06,
+      "loss": 1.8128,
+      "step": 966
+    },
+    {
+      "epoch": 1.0597260273972602,
+      "grad_norm": 2.541221857070923,
+      "learning_rate": 3.621039805036556e-06,
+      "loss": 1.8504,
+      "step": 967
+    },
+    {
+      "epoch": 1.0608219178082192,
+      "grad_norm": 2.3211770057678223,
+      "learning_rate": 3.6190089358245335e-06,
+      "loss": 1.8329,
+      "step": 968
+    },
+    {
+      "epoch": 1.061917808219178,
+      "grad_norm": 2.1008756160736084,
+      "learning_rate": 3.6169780666125104e-06,
+      "loss": 1.8972,
+      "step": 969
+    },
+    {
+      "epoch": 1.063013698630137,
+      "grad_norm": 2.4148898124694824,
+      "learning_rate": 3.614947197400488e-06,
+      "loss": 1.8748,
+      "step": 970
+    },
+    {
+      "epoch": 1.063013698630137,
+      "eval_loss": 2.0592916011810303,
+      "eval_runtime": 36.7663,
+      "eval_samples_per_second": 27.199,
+      "eval_steps_per_second": 1.142,
+      "step": 970
+    },
+    {
+      "epoch": 1.0641095890410959,
+      "grad_norm": 2.4116220474243164,
+      "learning_rate": 3.612916328188465e-06,
+      "loss": 1.8915,
+      "step": 971
+    },
+    {
+      "epoch": 1.0652054794520547,
+      "grad_norm": 2.257411241531372,
+      "learning_rate": 3.610885458976442e-06,
+      "loss": 1.7901,
+      "step": 972
+    },
+    {
+      "epoch": 1.0663013698630137,
+      "grad_norm": 2.166875123977661,
+      "learning_rate": 3.6088545897644195e-06,
+      "loss": 1.8969,
+      "step": 973
+    },
+    {
+      "epoch": 1.0673972602739725,
+      "grad_norm": 2.464716911315918,
+      "learning_rate": 3.6068237205523963e-06,
+      "loss": 1.8844,
+      "step": 974
+    },
+    {
+      "epoch": 1.0684931506849316,
+      "grad_norm": 2.337334156036377,
+      "learning_rate": 3.604792851340374e-06,
+      "loss": 1.8192,
+      "step": 975
+    },
+    {
+      "epoch": 1.0695890410958904,
+      "grad_norm": 2.117663621902466,
+      "learning_rate": 3.602761982128351e-06,
+      "loss": 1.8233,
+      "step": 976
+    },
+    {
+      "epoch": 1.0706849315068494,
+      "grad_norm": 2.3353586196899414,
+      "learning_rate": 3.6007311129163286e-06,
+      "loss": 1.9133,
+      "step": 977
+    },
+    {
+      "epoch": 1.0717808219178082,
+      "grad_norm": 2.444279909133911,
+      "learning_rate": 3.5987002437043054e-06,
+      "loss": 1.8803,
+      "step": 978
+    },
+    {
+      "epoch": 1.072876712328767,
+      "grad_norm": 2.3458383083343506,
+      "learning_rate": 3.596669374492283e-06,
+      "loss": 1.8108,
+      "step": 979
+    },
+    {
+      "epoch": 1.073972602739726,
+      "grad_norm": 2.207977771759033,
+      "learning_rate": 3.59463850528026e-06,
+      "loss": 1.8423,
+      "step": 980
+    },
+    {
+      "epoch": 1.073972602739726,
+      "eval_loss": 2.0578773021698,
+      "eval_runtime": 36.8123,
+      "eval_samples_per_second": 27.165,
+      "eval_steps_per_second": 1.141,
+      "step": 980
+    },
+    {
+      "epoch": 1.0750684931506849,
+      "grad_norm": 2.378127336502075,
+      "learning_rate": 3.5926076360682377e-06,
+      "loss": 1.9268,
+      "step": 981
+    },
+    {
+      "epoch": 1.076164383561644,
+      "grad_norm": 2.248105525970459,
+      "learning_rate": 3.5905767668562145e-06,
+      "loss": 1.8111,
+      "step": 982
+    },
+    {
+      "epoch": 1.0772602739726027,
+      "grad_norm": 2.275592088699341,
+      "learning_rate": 3.5885458976441922e-06,
+      "loss": 1.9099,
+      "step": 983
+    },
+    {
+      "epoch": 1.0783561643835617,
+      "grad_norm": 3.108377456665039,
+      "learning_rate": 3.586515028432169e-06,
+      "loss": 1.8728,
+      "step": 984
+    },
+    {
+      "epoch": 1.0794520547945206,
+      "grad_norm": 2.2702999114990234,
+      "learning_rate": 3.5844841592201464e-06,
+      "loss": 1.9007,
+      "step": 985
+    },
+    {
+      "epoch": 1.0805479452054794,
+      "grad_norm": 3.218647003173828,
+      "learning_rate": 3.5824532900081237e-06,
+      "loss": 1.9157,
+      "step": 986
+    },
+    {
+      "epoch": 1.0816438356164384,
+      "grad_norm": 2.3393027782440186,
+      "learning_rate": 3.580422420796101e-06,
+      "loss": 1.7936,
+      "step": 987
+    },
+    {
+      "epoch": 1.0827397260273972,
+      "grad_norm": 2.420541286468506,
+      "learning_rate": 3.578391551584078e-06,
+      "loss": 1.8608,
+      "step": 988
+    },
+    {
+      "epoch": 1.0838356164383562,
+      "grad_norm": 2.3354556560516357,
+      "learning_rate": 3.5763606823720555e-06,
+      "loss": 1.8684,
+      "step": 989
+    },
+    {
+      "epoch": 1.084931506849315,
+      "grad_norm": 2.200282573699951,
+      "learning_rate": 3.5743298131600328e-06,
+      "loss": 1.7814,
+      "step": 990
+    },
+    {
+      "epoch": 1.084931506849315,
+      "eval_loss": 2.0565249919891357,
+      "eval_runtime": 36.952,
+      "eval_samples_per_second": 27.062,
+      "eval_steps_per_second": 1.137,
+      "step": 990
+    },
+    {
+      "epoch": 1.086027397260274,
+      "grad_norm": 2.149033308029175,
+      "learning_rate": 3.57229894394801e-06,
+      "loss": 1.8708,
+      "step": 991
+    },
+    {
+      "epoch": 1.087123287671233,
+      "grad_norm": 2.4663679599761963,
+      "learning_rate": 3.5702680747359873e-06,
+      "loss": 1.8947,
+      "step": 992
+    },
+    {
+      "epoch": 1.0882191780821917,
+      "grad_norm": 2.3142261505126953,
+      "learning_rate": 3.5682372055239646e-06,
+      "loss": 1.8727,
+      "step": 993
+    },
+    {
+      "epoch": 1.0893150684931507,
+      "grad_norm": 2.2223033905029297,
+      "learning_rate": 3.566206336311942e-06,
+      "loss": 1.8203,
+      "step": 994
+    },
+    {
+      "epoch": 1.0904109589041096,
+      "grad_norm": 2.423855781555176,
+      "learning_rate": 3.564175467099919e-06,
+      "loss": 1.8687,
+      "step": 995
+    },
+    {
+      "epoch": 1.0915068493150686,
+      "grad_norm": 2.4022302627563477,
+      "learning_rate": 3.5621445978878964e-06,
+      "loss": 1.8215,
+      "step": 996
+    },
+    {
+      "epoch": 1.0926027397260274,
+      "grad_norm": 2.378236770629883,
+      "learning_rate": 3.5601137286758737e-06,
+      "loss": 1.7994,
+      "step": 997
+    },
+    {
+      "epoch": 1.0936986301369862,
+      "grad_norm": 2.4499096870422363,
+      "learning_rate": 3.5580828594638505e-06,
+      "loss": 1.8779,
+      "step": 998
+    },
+    {
+      "epoch": 1.0947945205479452,
+      "grad_norm": 2.3181755542755127,
+      "learning_rate": 3.5560519902518282e-06,
+      "loss": 1.8341,
+      "step": 999
+    },
+    {
+      "epoch": 1.095890410958904,
+      "grad_norm": 2.2406623363494873,
+      "learning_rate": 3.554021121039805e-06,
+      "loss": 1.874,
+      "step": 1000
+    },
+    {
+      "epoch": 1.095890410958904,
+      "eval_loss": 2.0523672103881836,
+      "eval_runtime": 37.0454,
+      "eval_samples_per_second": 26.994,
+      "eval_steps_per_second": 1.134,
+      "step": 1000
+    },
+    {
+      "epoch": 1.096986301369863,
+      "grad_norm": 2.254667043685913,
+      "learning_rate": 3.551990251827783e-06,
+      "loss": 1.9017,
+      "step": 1001
+    },
+    {
+      "epoch": 1.098082191780822,
+      "grad_norm": 2.2846896648406982,
+      "learning_rate": 3.5499593826157596e-06,
+      "loss": 1.8505,
+      "step": 1002
+    },
+    {
+      "epoch": 1.099178082191781,
+      "grad_norm": 2.183469772338867,
+      "learning_rate": 3.5479285134037373e-06,
+      "loss": 1.8266,
+      "step": 1003
+    },
+    {
+      "epoch": 1.1002739726027397,
+      "grad_norm": 2.151798963546753,
+      "learning_rate": 3.545897644191714e-06,
+      "loss": 1.8487,
+      "step": 1004
+    },
+    {
+      "epoch": 1.1013698630136985,
+      "grad_norm": 2.2039146423339844,
+      "learning_rate": 3.543866774979692e-06,
+      "loss": 1.8327,
+      "step": 1005
+    },
+    {
+      "epoch": 1.1024657534246576,
+      "grad_norm": 2.2091665267944336,
+      "learning_rate": 3.5418359057676687e-06,
+      "loss": 1.8214,
+      "step": 1006
+    },
+    {
+      "epoch": 1.1035616438356164,
+      "grad_norm": 2.2821404933929443,
+      "learning_rate": 3.5398050365556464e-06,
+      "loss": 1.8591,
+      "step": 1007
+    },
+    {
+      "epoch": 1.1046575342465754,
+      "grad_norm": 2.1726973056793213,
+      "learning_rate": 3.5377741673436233e-06,
+      "loss": 1.8265,
+      "step": 1008
+    },
+    {
+      "epoch": 1.1057534246575342,
+      "grad_norm": 2.2870233058929443,
+      "learning_rate": 3.535743298131601e-06,
+      "loss": 1.8783,
+      "step": 1009
+    },
+    {
+      "epoch": 1.106849315068493,
+      "grad_norm": 2.148080587387085,
+      "learning_rate": 3.533712428919578e-06,
+      "loss": 1.8272,
+      "step": 1010
+    },
+    {
+      "epoch": 1.106849315068493,
+      "eval_loss": 2.049504041671753,
+      "eval_runtime": 36.8717,
+      "eval_samples_per_second": 27.121,
+      "eval_steps_per_second": 1.139,
+      "step": 1010
+    },
+    {
+      "epoch": 1.107945205479452,
+      "grad_norm": 2.182035446166992,
+      "learning_rate": 3.5316815597075547e-06,
+      "loss": 1.8771,
+      "step": 1011
+    },
+    {
+      "epoch": 1.1090410958904109,
+      "grad_norm": 2.135085344314575,
+      "learning_rate": 3.5296506904955324e-06,
+      "loss": 1.9031,
+      "step": 1012
+    },
+    {
+      "epoch": 1.11013698630137,
+      "grad_norm": 2.1017255783081055,
+      "learning_rate": 3.5276198212835093e-06,
+      "loss": 1.8341,
+      "step": 1013
+    },
+    {
+      "epoch": 1.1112328767123287,
+      "grad_norm": 2.2773115634918213,
+      "learning_rate": 3.525588952071487e-06,
+      "loss": 1.8564,
+      "step": 1014
+    },
+    {
+      "epoch": 1.1123287671232878,
+      "grad_norm": 2.2773115634918213,
+      "learning_rate": 3.525588952071487e-06,
+      "loss": 1.8971,
+      "step": 1015
+    },
+    {
+      "epoch": 1.1134246575342466,
+      "grad_norm": 2.273375988006592,
+      "learning_rate": 3.523558082859464e-06,
+      "loss": 1.9074,
+      "step": 1016
+    },
+    {
+      "epoch": 1.1145205479452054,
+      "grad_norm": 2.2055182456970215,
+      "learning_rate": 3.5215272136474415e-06,
+      "loss": 1.8724,
+      "step": 1017
+    },
+    {
+      "epoch": 1.1156164383561644,
+      "grad_norm": 2.246126651763916,
+      "learning_rate": 3.5194963444354184e-06,
+      "loss": 1.8387,
+      "step": 1018
+    },
+    {
+      "epoch": 1.1167123287671232,
+      "grad_norm": 2.3076183795928955,
+      "learning_rate": 3.517465475223396e-06,
+      "loss": 1.8387,
+      "step": 1019
+    },
+    {
+      "epoch": 1.1178082191780823,
+      "grad_norm": 2.3365488052368164,
+      "learning_rate": 3.515434606011373e-06,
+      "loss": 1.8063,
+      "step": 1020
+    },
+    {
+      "epoch": 1.1178082191780823,
+      "eval_loss": 2.0528781414031982,
+      "eval_runtime": 36.8885,
+      "eval_samples_per_second": 27.109,
+      "eval_steps_per_second": 1.139,
+      "step": 1020
+    },
+    {
+      "epoch": 1.118904109589041,
+      "grad_norm": 2.319324493408203,
+      "learning_rate": 3.5134037367993506e-06,
+      "loss": 1.8843,
+      "step": 1021
+    },
+    {
+      "epoch": 1.12,
+      "grad_norm": 2.238720655441284,
+      "learning_rate": 3.5113728675873275e-06,
+      "loss": 1.9375,
+      "step": 1022
+    },
+    {
+      "epoch": 1.121095890410959,
+      "grad_norm": 2.2824833393096924,
+      "learning_rate": 3.5093419983753047e-06,
+      "loss": 1.8749,
+      "step": 1023
+    },
+    {
+      "epoch": 1.1221917808219177,
+      "grad_norm": 2.5120770931243896,
+      "learning_rate": 3.507311129163282e-06,
+      "loss": 1.8774,
+      "step": 1024
+    },
+    {
+      "epoch": 1.1232876712328768,
+      "grad_norm": 2.1762094497680664,
+      "learning_rate": 3.5052802599512593e-06,
+      "loss": 1.8018,
+      "step": 1025
+    },
+    {
+      "epoch": 1.1243835616438356,
+      "grad_norm": 2.2530343532562256,
+      "learning_rate": 3.5032493907392366e-06,
+      "loss": 1.8451,
+      "step": 1026
+    },
+    {
+      "epoch": 1.1254794520547946,
+      "grad_norm": 2.21649169921875,
+      "learning_rate": 3.501218521527214e-06,
+      "loss": 1.8613,
+      "step": 1027
+    },
+    {
+      "epoch": 1.1265753424657534,
+      "grad_norm": 2.306431293487549,
+      "learning_rate": 3.499187652315191e-06,
+      "loss": 1.7947,
+      "step": 1028
+    },
+    {
+      "epoch": 1.1276712328767124,
+      "grad_norm": 2.203005313873291,
+      "learning_rate": 3.4971567831031684e-06,
+      "loss": 1.8483,
+      "step": 1029
+    },
+    {
+      "epoch": 1.1287671232876713,
+      "grad_norm": 2.1645009517669678,
+      "learning_rate": 3.4951259138911457e-06,
+      "loss": 1.8915,
+      "step": 1030
+    },
+    {
+      "epoch": 1.1287671232876713,
+      "eval_loss": 2.053393602371216,
+      "eval_runtime": 36.855,
+      "eval_samples_per_second": 27.133,
+      "eval_steps_per_second": 1.14,
+      "step": 1030
+    },
+    {
+      "epoch": 1.12986301369863,
+      "grad_norm": 2.2863385677337646,
+      "learning_rate": 3.493095044679123e-06,
+      "loss": 1.8073,
+      "step": 1031
+    },
+    {
+      "epoch": 1.130958904109589,
+      "grad_norm": 2.1439812183380127,
+      "learning_rate": 3.4910641754671002e-06,
+      "loss": 1.9191,
+      "step": 1032
+    },
+    {
+      "epoch": 1.132054794520548,
+      "grad_norm": 2.224228858947754,
+      "learning_rate": 3.4890333062550775e-06,
+      "loss": 1.7478,
+      "step": 1033
+    },
+    {
+      "epoch": 1.133150684931507,
+      "grad_norm": 2.319059133529663,
+      "learning_rate": 3.4870024370430548e-06,
+      "loss": 1.8779,
+      "step": 1034
+    },
+    {
+      "epoch": 1.1342465753424658,
+      "grad_norm": 2.5003139972686768,
+      "learning_rate": 3.484971567831032e-06,
+      "loss": 1.776,
+      "step": 1035
+    },
+    {
+      "epoch": 1.1353424657534246,
+      "grad_norm": 2.2032580375671387,
+      "learning_rate": 3.482940698619009e-06,
+      "loss": 1.8221,
+      "step": 1036
+    },
+    {
+      "epoch": 1.1364383561643836,
+      "grad_norm": 2.5253560543060303,
+      "learning_rate": 3.4809098294069866e-06,
+      "loss": 1.8952,
+      "step": 1037
+    },
+    {
+      "epoch": 1.1375342465753424,
+      "grad_norm": 2.3215854167938232,
+      "learning_rate": 3.4788789601949634e-06,
+      "loss": 1.847,
+      "step": 1038
+    },
+    {
+      "epoch": 1.1386301369863014,
+      "grad_norm": 2.2543280124664307,
+      "learning_rate": 3.476848090982941e-06,
+      "loss": 1.8689,
+      "step": 1039
+    },
+    {
+      "epoch": 1.1397260273972603,
+      "grad_norm": 2.1592042446136475,
+      "learning_rate": 3.474817221770918e-06,
+      "loss": 1.8375,
+      "step": 1040
+    },
+    {
+      "epoch": 1.1397260273972603,
+      "eval_loss": 2.0524044036865234,
+      "eval_runtime": 36.8163,
+      "eval_samples_per_second": 27.162,
+      "eval_steps_per_second": 1.141,
+      "step": 1040
+    },
+    {
+      "epoch": 1.140821917808219,
+      "grad_norm": 2.2874810695648193,
+      "learning_rate": 3.4727863525588957e-06,
+      "loss": 1.9442,
+      "step": 1041
+    },
+    {
+      "epoch": 1.141917808219178,
+      "grad_norm": 2.288395404815674,
+      "learning_rate": 3.4707554833468725e-06,
+      "loss": 1.8881,
+      "step": 1042
+    },
+    {
+      "epoch": 1.143013698630137,
+      "grad_norm": 2.2264153957366943,
+      "learning_rate": 3.4687246141348502e-06,
+      "loss": 1.8518,
+      "step": 1043
+    },
+    {
+      "epoch": 1.144109589041096,
+      "grad_norm": 2.4413652420043945,
+      "learning_rate": 3.466693744922827e-06,
+      "loss": 1.8146,
+      "step": 1044
+    },
+    {
+      "epoch": 1.1452054794520548,
+      "grad_norm": 2.3597898483276367,
+      "learning_rate": 3.464662875710805e-06,
+      "loss": 1.8421,
+      "step": 1045
+    },
+    {
+      "epoch": 1.1463013698630138,
+      "grad_norm": 2.2489402294158936,
+      "learning_rate": 3.4626320064987817e-06,
+      "loss": 1.8643,
+      "step": 1046
+    },
+    {
+      "epoch": 1.1473972602739726,
+      "grad_norm": 2.3958940505981445,
+      "learning_rate": 3.4606011372867594e-06,
+      "loss": 1.8522,
+      "step": 1047
+    },
+    {
+      "epoch": 1.1484931506849314,
+      "grad_norm": 2.295349597930908,
+      "learning_rate": 3.458570268074736e-06,
+      "loss": 1.8677,
+      "step": 1048
+    },
+    {
+      "epoch": 1.1495890410958904,
+      "grad_norm": 2.2571280002593994,
+      "learning_rate": 3.4565393988627135e-06,
+      "loss": 1.8393,
+      "step": 1049
+    },
+    {
+      "epoch": 1.1506849315068493,
+      "grad_norm": 2.173983335494995,
+      "learning_rate": 3.4545085296506908e-06,
+      "loss": 1.8267,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1506849315068493,
+      "eval_loss": 2.0489604473114014,
+      "eval_runtime": 36.7678,
+      "eval_samples_per_second": 27.198,
+      "eval_steps_per_second": 1.142,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1517808219178083,
+      "grad_norm": 2.3395209312438965,
+      "learning_rate": 3.452477660438668e-06,
+      "loss": 1.8032,
+      "step": 1051
+    },
+    {
+      "epoch": 1.152876712328767,
+      "grad_norm": 2.1894757747650146,
+      "learning_rate": 3.4504467912266453e-06,
+      "loss": 1.8834,
+      "step": 1052
+    },
+    {
+      "epoch": 1.1539726027397261,
+      "grad_norm": 2.1978437900543213,
+      "learning_rate": 3.4484159220146226e-06,
+      "loss": 1.7897,
+      "step": 1053
+    },
+    {
+      "epoch": 1.155068493150685,
+      "grad_norm": 2.275792360305786,
+      "learning_rate": 3.4463850528026e-06,
+      "loss": 1.8566,
+      "step": 1054
+    },
+    {
+      "epoch": 1.1561643835616437,
+      "grad_norm": 2.2659800052642822,
+      "learning_rate": 3.444354183590577e-06,
+      "loss": 1.8307,
+      "step": 1055
+    },
+    {
+      "epoch": 1.1572602739726028,
+      "grad_norm": 2.232149600982666,
+      "learning_rate": 3.4423233143785544e-06,
+      "loss": 1.8701,
+      "step": 1056
+    },
+    {
+      "epoch": 1.1583561643835616,
+      "grad_norm": 2.3636281490325928,
+      "learning_rate": 3.4402924451665317e-06,
+      "loss": 1.8008,
+      "step": 1057
+    },
+    {
+      "epoch": 1.1594520547945206,
+      "grad_norm": 2.271533489227295,
+      "learning_rate": 3.438261575954509e-06,
+      "loss": 1.7751,
+      "step": 1058
+    },
+    {
+      "epoch": 1.1605479452054794,
+      "grad_norm": 2.3286566734313965,
+      "learning_rate": 3.4362307067424862e-06,
+      "loss": 1.865,
+      "step": 1059
+    },
+    {
+      "epoch": 1.1616438356164385,
+      "grad_norm": 2.46665096282959,
+      "learning_rate": 3.434199837530463e-06,
+      "loss": 1.9001,
+      "step": 1060
+    },
+    {
+      "epoch": 1.1616438356164385,
+      "eval_loss": 2.0477516651153564,
+      "eval_runtime": 36.7818,
+      "eval_samples_per_second": 27.187,
+      "eval_steps_per_second": 1.142,
+      "step": 1060
+    },
+    {
+      "epoch": 1.1627397260273973,
+      "grad_norm": 2.307744264602661,
+      "learning_rate": 3.432168968318441e-06,
+      "loss": 1.8637,
+      "step": 1061
+    },
+    {
+      "epoch": 1.163835616438356,
+      "grad_norm": 2.3935985565185547,
+      "learning_rate": 3.4301380991064176e-06,
+      "loss": 1.7867,
+      "step": 1062
+    },
+    {
+      "epoch": 1.1649315068493151,
+      "grad_norm": 2.290971040725708,
+      "learning_rate": 3.4281072298943953e-06,
+      "loss": 1.8708,
+      "step": 1063
+    },
+    {
+      "epoch": 1.166027397260274,
+      "grad_norm": 2.191629409790039,
+      "learning_rate": 3.426076360682372e-06,
+      "loss": 1.9031,
+      "step": 1064
+    },
+    {
+      "epoch": 1.167123287671233,
+      "grad_norm": 2.284426689147949,
+      "learning_rate": 3.42404549147035e-06,
+      "loss": 1.9026,
+      "step": 1065
+    },
+    {
+      "epoch": 1.1682191780821918,
+      "grad_norm": 2.2567849159240723,
+      "learning_rate": 3.4220146222583267e-06,
+      "loss": 1.8507,
+      "step": 1066
+    },
+    {
+      "epoch": 1.1693150684931506,
+      "grad_norm": 2.2093160152435303,
+      "learning_rate": 3.4199837530463044e-06,
+      "loss": 1.8001,
+      "step": 1067
+    },
+    {
+      "epoch": 1.1704109589041096,
+      "grad_norm": 2.092560291290283,
+      "learning_rate": 3.4179528838342813e-06,
+      "loss": 1.8393,
+      "step": 1068
+    },
+    {
+      "epoch": 1.1715068493150684,
+      "grad_norm": 2.313749074935913,
+      "learning_rate": 3.415922014622259e-06,
+      "loss": 1.9385,
+      "step": 1069
+    },
+    {
+      "epoch": 1.1726027397260275,
+      "grad_norm": 2.488673210144043,
+      "learning_rate": 3.413891145410236e-06,
+      "loss": 1.8924,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1726027397260275,
+      "eval_loss": 2.0482401847839355,
+      "eval_runtime": 36.7919,
+      "eval_samples_per_second": 27.18,
+      "eval_steps_per_second": 1.142,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1736986301369863,
+      "grad_norm": 2.0804412364959717,
+      "learning_rate": 3.4118602761982135e-06,
+      "loss": 1.881,
+      "step": 1071
+    },
+    {
+      "epoch": 1.174794520547945,
+      "grad_norm": 2.165426254272461,
+      "learning_rate": 3.4098294069861904e-06,
+      "loss": 1.9035,
+      "step": 1072
+    },
+    {
+      "epoch": 1.1758904109589041,
+      "grad_norm": 2.3034675121307373,
+      "learning_rate": 3.4077985377741673e-06,
+      "loss": 1.8554,
+      "step": 1073
+    },
+    {
+      "epoch": 1.176986301369863,
+      "grad_norm": 2.303159713745117,
+      "learning_rate": 3.405767668562145e-06,
+      "loss": 1.8255,
+      "step": 1074
+    },
+    {
+      "epoch": 1.178082191780822,
+      "grad_norm": 2.2402753829956055,
+      "learning_rate": 3.403736799350122e-06,
+      "loss": 1.9087,
+      "step": 1075
+    },
+    {
+      "epoch": 1.1791780821917808,
+      "grad_norm": 2.251638650894165,
+      "learning_rate": 3.4017059301380995e-06,
+      "loss": 1.8878,
+      "step": 1076
+    },
+    {
+      "epoch": 1.1802739726027398,
+      "grad_norm": 2.2661657333374023,
+      "learning_rate": 3.3996750609260764e-06,
+      "loss": 1.8244,
+      "step": 1077
+    },
+    {
+      "epoch": 1.1813698630136986,
+      "grad_norm": 2.3280904293060303,
+      "learning_rate": 3.397644191714054e-06,
+      "loss": 1.9328,
+      "step": 1078
+    },
+    {
+      "epoch": 1.1824657534246574,
+      "grad_norm": 2.315779447555542,
+      "learning_rate": 3.395613322502031e-06,
+      "loss": 1.8289,
+      "step": 1079
+    },
+    {
+      "epoch": 1.1835616438356165,
+      "grad_norm": 2.136967182159424,
+      "learning_rate": 3.3935824532900086e-06,
+      "loss": 1.8307,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1835616438356165,
+      "eval_loss": 2.0473902225494385,
+      "eval_runtime": 36.8094,
+      "eval_samples_per_second": 27.167,
+      "eval_steps_per_second": 1.141,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1846575342465753,
+      "grad_norm": 2.3701720237731934,
+      "learning_rate": 3.3915515840779855e-06,
+      "loss": 1.8123,
+      "step": 1081
+    },
+    {
+      "epoch": 1.1857534246575343,
+      "grad_norm": 2.2662320137023926,
+      "learning_rate": 3.389520714865963e-06,
+      "loss": 1.85,
+      "step": 1082
+    },
+    {
+      "epoch": 1.1868493150684931,
+      "grad_norm": 2.2151782512664795,
+      "learning_rate": 3.38748984565394e-06,
+      "loss": 1.845,
+      "step": 1083
+    },
+    {
+      "epoch": 1.1879452054794521,
+      "grad_norm": 2.2209115028381348,
+      "learning_rate": 3.3854589764419177e-06,
+      "loss": 1.8787,
+      "step": 1084
+    },
+    {
+      "epoch": 1.189041095890411,
+      "grad_norm": 2.4760477542877197,
+      "learning_rate": 3.3834281072298946e-06,
+      "loss": 1.818,
+      "step": 1085
+    },
+    {
+      "epoch": 1.1901369863013698,
+      "grad_norm": 2.309034585952759,
+      "learning_rate": 3.381397238017872e-06,
+      "loss": 1.8387,
+      "step": 1086
+    },
+    {
+      "epoch": 1.1912328767123288,
+      "grad_norm": 2.2174596786499023,
+      "learning_rate": 3.379366368805849e-06,
+      "loss": 1.8135,
+      "step": 1087
+    },
+    {
+      "epoch": 1.1923287671232876,
+      "grad_norm": 2.3043556213378906,
+      "learning_rate": 3.3773354995938264e-06,
+      "loss": 1.9156,
+      "step": 1088
+    },
+    {
+      "epoch": 1.1934246575342466,
+      "grad_norm": 2.2240524291992188,
+      "learning_rate": 3.3753046303818037e-06,
+      "loss": 1.8636,
+      "step": 1089
+    },
+    {
+      "epoch": 1.1945205479452055,
+      "grad_norm": 2.1108691692352295,
+      "learning_rate": 3.373273761169781e-06,
+      "loss": 1.8458,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1945205479452055,
+      "eval_loss": 2.04378604888916,
+      "eval_runtime": 36.9058,
+      "eval_samples_per_second": 27.096,
+      "eval_steps_per_second": 1.138,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1956164383561645,
+      "grad_norm": 2.1893398761749268,
+      "learning_rate": 3.3712428919577582e-06,
+      "loss": 1.8446,
+      "step": 1091
+    },
+    {
+      "epoch": 1.1967123287671233,
+      "grad_norm": 2.305298089981079,
+      "learning_rate": 3.3692120227457355e-06,
+      "loss": 1.8283,
+      "step": 1092
+    },
+    {
+      "epoch": 1.1978082191780821,
+      "grad_norm": 2.364433765411377,
+      "learning_rate": 3.3671811535337128e-06,
+      "loss": 1.8708,
+      "step": 1093
+    },
+    {
+      "epoch": 1.1989041095890411,
+      "grad_norm": 2.259892225265503,
+      "learning_rate": 3.36515028432169e-06,
+      "loss": 1.7966,
+      "step": 1094
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 2.2487263679504395,
+      "learning_rate": 3.3631194151096673e-06,
+      "loss": 1.8238,
+      "step": 1095
+    },
+    {
+      "epoch": 1.201095890410959,
+      "grad_norm": 2.1592557430267334,
+      "learning_rate": 3.3610885458976446e-06,
+      "loss": 1.8651,
+      "step": 1096
+    },
+    {
+      "epoch": 1.2021917808219178,
+      "grad_norm": 2.257737159729004,
+      "learning_rate": 3.359057676685622e-06,
+      "loss": 1.8926,
+      "step": 1097
+    },
+    {
+      "epoch": 1.2032876712328768,
+      "grad_norm": 2.0702707767486572,
+      "learning_rate": 3.357026807473599e-06,
+      "loss": 1.8833,
+      "step": 1098
+    },
+    {
+      "epoch": 1.2043835616438356,
+      "grad_norm": 2.1568455696105957,
+      "learning_rate": 3.354995938261576e-06,
+      "loss": 1.7732,
+      "step": 1099
+    },
+    {
+      "epoch": 1.2054794520547945,
+      "grad_norm": 2.294715404510498,
+      "learning_rate": 3.3529650690495537e-06,
+      "loss": 1.8205,
+      "step": 1100
+    },
+    {
+      "epoch": 1.2054794520547945,
+      "eval_loss": 2.044201612472534,
+      "eval_runtime": 36.9861,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 1100
+    },
+    {
+      "epoch": 1.2065753424657535,
+      "grad_norm": 2.3234570026397705,
+      "learning_rate": 3.3509341998375305e-06,
+      "loss": 1.7935,
+      "step": 1101
+    },
+    {
+      "epoch": 1.2076712328767123,
+      "grad_norm": 2.307224750518799,
+      "learning_rate": 3.3489033306255082e-06,
+      "loss": 1.8254,
+      "step": 1102
+    },
+    {
+      "epoch": 1.208767123287671,
+      "grad_norm": 2.3725457191467285,
+      "learning_rate": 3.346872461413485e-06,
+      "loss": 1.8493,
+      "step": 1103
+    },
+    {
+      "epoch": 1.2098630136986301,
+      "grad_norm": 2.2057435512542725,
+      "learning_rate": 3.344841592201463e-06,
+      "loss": 1.8241,
+      "step": 1104
+    },
+    {
+      "epoch": 1.210958904109589,
+      "grad_norm": 2.232947826385498,
+      "learning_rate": 3.3428107229894397e-06,
+      "loss": 1.8057,
+      "step": 1105
+    },
+    {
+      "epoch": 1.212054794520548,
+      "grad_norm": 2.4268925189971924,
+      "learning_rate": 3.3407798537774174e-06,
+      "loss": 1.8334,
+      "step": 1106
+    },
+    {
+      "epoch": 1.2131506849315068,
+      "grad_norm": 2.3134701251983643,
+      "learning_rate": 3.338748984565394e-06,
+      "loss": 1.8516,
+      "step": 1107
+    },
+    {
+      "epoch": 1.2142465753424658,
+      "grad_norm": 2.375873327255249,
+      "learning_rate": 3.336718115353372e-06,
+      "loss": 1.8942,
+      "step": 1108
+    },
+    {
+      "epoch": 1.2153424657534246,
+      "grad_norm": 2.622528076171875,
+      "learning_rate": 3.3346872461413488e-06,
+      "loss": 1.9027,
+      "step": 1109
+    },
+    {
+      "epoch": 1.2164383561643834,
+      "grad_norm": 2.290865898132324,
+      "learning_rate": 3.3326563769293256e-06,
+      "loss": 1.8258,
+      "step": 1110
+    },
+    {
+      "epoch": 1.2164383561643834,
+      "eval_loss": 2.0408973693847656,
+      "eval_runtime": 36.8596,
+      "eval_samples_per_second": 27.13,
+      "eval_steps_per_second": 1.139,
+      "step": 1110
+    },
+    {
+      "epoch": 1.2175342465753425,
+      "grad_norm": 2.1596908569335938,
+      "learning_rate": 3.3306255077173033e-06,
+      "loss": 1.8066,
+      "step": 1111
+    },
+    {
+      "epoch": 1.2186301369863013,
+      "grad_norm": 2.4652187824249268,
+      "learning_rate": 3.32859463850528e-06,
+      "loss": 1.8372,
+      "step": 1112
+    },
+    {
+      "epoch": 1.2197260273972603,
+      "grad_norm": 2.282440185546875,
+      "learning_rate": 3.326563769293258e-06,
+      "loss": 1.8535,
+      "step": 1113
+    },
+    {
+      "epoch": 1.2208219178082191,
+      "grad_norm": 2.2390308380126953,
+      "learning_rate": 3.3245329000812347e-06,
+      "loss": 1.7719,
+      "step": 1114
+    },
+    {
+      "epoch": 1.2219178082191782,
+      "grad_norm": 2.368565320968628,
+      "learning_rate": 3.3225020308692124e-06,
+      "loss": 1.8207,
+      "step": 1115
+    },
+    {
+      "epoch": 1.223013698630137,
+      "grad_norm": 2.463942766189575,
+      "learning_rate": 3.3204711616571893e-06,
+      "loss": 1.8402,
+      "step": 1116
+    },
+    {
+      "epoch": 1.2241095890410958,
+      "grad_norm": 2.266096830368042,
+      "learning_rate": 3.318440292445167e-06,
+      "loss": 1.7854,
+      "step": 1117
+    },
+    {
+      "epoch": 1.2252054794520548,
+      "grad_norm": 2.3463494777679443,
+      "learning_rate": 3.316409423233144e-06,
+      "loss": 1.8449,
+      "step": 1118
+    },
+    {
+      "epoch": 1.2263013698630136,
+      "grad_norm": 2.381727933883667,
+      "learning_rate": 3.3143785540211215e-06,
+      "loss": 1.8085,
+      "step": 1119
+    },
+    {
+      "epoch": 1.2273972602739727,
+      "grad_norm": 2.3251113891601562,
+      "learning_rate": 3.3123476848090984e-06,
+      "loss": 1.8922,
+      "step": 1120
+    },
+    {
+      "epoch": 1.2273972602739727,
+      "eval_loss": 2.0425429344177246,
+      "eval_runtime": 36.851,
+      "eval_samples_per_second": 27.136,
+      "eval_steps_per_second": 1.14,
+      "step": 1120
+    },
+    {
+      "epoch": 1.2284931506849315,
+      "grad_norm": 2.2314298152923584,
+      "learning_rate": 3.310316815597076e-06,
+      "loss": 1.8474,
+      "step": 1121
+    },
+    {
+      "epoch": 1.2295890410958905,
+      "grad_norm": 2.1847472190856934,
+      "learning_rate": 3.308285946385053e-06,
+      "loss": 1.8454,
+      "step": 1122
+    },
+    {
+      "epoch": 1.2306849315068493,
+      "grad_norm": 2.249422788619995,
+      "learning_rate": 3.30625507717303e-06,
+      "loss": 1.9257,
+      "step": 1123
+    },
+    {
+      "epoch": 1.2317808219178081,
+      "grad_norm": 2.223930835723877,
+      "learning_rate": 3.3042242079610075e-06,
+      "loss": 1.7816,
+      "step": 1124
+    },
+    {
+      "epoch": 1.2328767123287672,
+      "grad_norm": 2.3209826946258545,
+      "learning_rate": 3.3021933387489847e-06,
+      "loss": 1.8773,
+      "step": 1125
+    },
+    {
+      "epoch": 1.233972602739726,
+      "grad_norm": 2.2288055419921875,
+      "learning_rate": 3.300162469536962e-06,
+      "loss": 1.8931,
+      "step": 1126
+    },
+    {
+      "epoch": 1.235068493150685,
+      "grad_norm": 2.201449155807495,
+      "learning_rate": 3.2981316003249393e-06,
+      "loss": 1.8149,
+      "step": 1127
+    },
+    {
+      "epoch": 1.2361643835616438,
+      "grad_norm": 2.215824842453003,
+      "learning_rate": 3.2961007311129166e-06,
+      "loss": 1.8949,
+      "step": 1128
+    },
+    {
+      "epoch": 1.2372602739726029,
+      "grad_norm": 2.133671522140503,
+      "learning_rate": 3.294069861900894e-06,
+      "loss": 1.8669,
+      "step": 1129
+    },
+    {
+      "epoch": 1.2383561643835617,
+      "grad_norm": 2.2195792198181152,
+      "learning_rate": 3.292038992688871e-06,
+      "loss": 1.9245,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2383561643835617,
+      "eval_loss": 2.0442771911621094,
+      "eval_runtime": 36.8077,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2394520547945205,
+      "grad_norm": 2.310866117477417,
+      "learning_rate": 3.2900081234768484e-06,
+      "loss": 1.8523,
+      "step": 1131
+    },
+    {
+      "epoch": 1.2405479452054795,
+      "grad_norm": 2.318300247192383,
+      "learning_rate": 3.2879772542648257e-06,
+      "loss": 1.8665,
+      "step": 1132
+    },
+    {
+      "epoch": 1.2416438356164383,
+      "grad_norm": 2.082564115524292,
+      "learning_rate": 3.285946385052803e-06,
+      "loss": 1.8962,
+      "step": 1133
+    },
+    {
+      "epoch": 1.2427397260273974,
+      "grad_norm": 2.1982104778289795,
+      "learning_rate": 3.2839155158407802e-06,
+      "loss": 1.7722,
+      "step": 1134
+    },
+    {
+      "epoch": 1.2438356164383562,
+      "grad_norm": 2.330782890319824,
+      "learning_rate": 3.2818846466287575e-06,
+      "loss": 1.8787,
+      "step": 1135
+    },
+    {
+      "epoch": 1.244931506849315,
+      "grad_norm": 2.136747121810913,
+      "learning_rate": 3.2798537774167344e-06,
+      "loss": 1.7739,
+      "step": 1136
+    },
+    {
+      "epoch": 1.246027397260274,
+      "grad_norm": 2.205862283706665,
+      "learning_rate": 3.277822908204712e-06,
+      "loss": 1.9127,
+      "step": 1137
+    },
+    {
+      "epoch": 1.2471232876712328,
+      "grad_norm": 2.2060093879699707,
+      "learning_rate": 3.275792038992689e-06,
+      "loss": 1.8379,
+      "step": 1138
+    },
+    {
+      "epoch": 1.2482191780821918,
+      "grad_norm": 2.255772829055786,
+      "learning_rate": 3.2737611697806666e-06,
+      "loss": 1.7931,
+      "step": 1139
+    },
+    {
+      "epoch": 1.2493150684931507,
+      "grad_norm": 2.240095615386963,
+      "learning_rate": 3.2717303005686435e-06,
+      "loss": 1.8466,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2493150684931507,
+      "eval_loss": 2.0429375171661377,
+      "eval_runtime": 36.7983,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2504109589041095,
+      "grad_norm": 2.1491949558258057,
+      "learning_rate": 3.269699431356621e-06,
+      "loss": 1.8137,
+      "step": 1141
+    },
+    {
+      "epoch": 1.2515068493150685,
+      "grad_norm": 2.3854610919952393,
+      "learning_rate": 3.267668562144598e-06,
+      "loss": 1.8486,
+      "step": 1142
+    },
+    {
+      "epoch": 1.2526027397260273,
+      "grad_norm": 2.3498709201812744,
+      "learning_rate": 3.2656376929325757e-06,
+      "loss": 1.837,
+      "step": 1143
+    },
+    {
+      "epoch": 1.2536986301369863,
+      "grad_norm": 2.1651649475097656,
+      "learning_rate": 3.2636068237205526e-06,
+      "loss": 1.8,
+      "step": 1144
+    },
+    {
+      "epoch": 1.2547945205479452,
+      "grad_norm": 2.2500200271606445,
+      "learning_rate": 3.2615759545085303e-06,
+      "loss": 1.8691,
+      "step": 1145
+    },
+    {
+      "epoch": 1.2558904109589042,
+      "grad_norm": 2.2178192138671875,
+      "learning_rate": 3.259545085296507e-06,
+      "loss": 1.8345,
+      "step": 1146
+    },
+    {
+      "epoch": 1.256986301369863,
+      "grad_norm": 2.238114595413208,
+      "learning_rate": 3.257514216084484e-06,
+      "loss": 1.8935,
+      "step": 1147
+    },
+    {
+      "epoch": 1.2580821917808218,
+      "grad_norm": 2.37042498588562,
+      "learning_rate": 3.2554833468724617e-06,
+      "loss": 1.7872,
+      "step": 1148
+    },
+    {
+      "epoch": 1.2591780821917808,
+      "grad_norm": 2.1831908226013184,
+      "learning_rate": 3.2534524776604385e-06,
+      "loss": 1.8913,
+      "step": 1149
+    },
+    {
+      "epoch": 1.2602739726027397,
+      "grad_norm": 2.251382827758789,
+      "learning_rate": 3.2514216084484162e-06,
+      "loss": 1.8611,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2602739726027397,
+      "eval_loss": 2.041306257247925,
+      "eval_runtime": 36.8199,
+      "eval_samples_per_second": 27.159,
+      "eval_steps_per_second": 1.141,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2613698630136987,
+      "grad_norm": 2.2690420150756836,
+      "learning_rate": 3.249390739236393e-06,
+      "loss": 1.8283,
+      "step": 1151
+    },
+    {
+      "epoch": 1.2624657534246575,
+      "grad_norm": 2.255833387374878,
+      "learning_rate": 3.2473598700243708e-06,
+      "loss": 1.8563,
+      "step": 1152
+    },
+    {
+      "epoch": 1.2635616438356165,
+      "grad_norm": 2.1362855434417725,
+      "learning_rate": 3.2453290008123476e-06,
+      "loss": 1.809,
+      "step": 1153
+    },
+    {
+      "epoch": 1.2646575342465753,
+      "grad_norm": 2.1227219104766846,
+      "learning_rate": 3.2432981316003253e-06,
+      "loss": 1.848,
+      "step": 1154
+    },
+    {
+      "epoch": 1.2657534246575342,
+      "grad_norm": 2.3639719486236572,
+      "learning_rate": 3.241267262388302e-06,
+      "loss": 1.8472,
+      "step": 1155
+    },
+    {
+      "epoch": 1.2668493150684932,
+      "grad_norm": 2.334404468536377,
+      "learning_rate": 3.23923639317628e-06,
+      "loss": 1.8073,
+      "step": 1156
+    },
+    {
+      "epoch": 1.267945205479452,
+      "grad_norm": 2.231656312942505,
+      "learning_rate": 3.2372055239642567e-06,
+      "loss": 1.8446,
+      "step": 1157
+    },
+    {
+      "epoch": 1.269041095890411,
+      "grad_norm": 2.371375560760498,
+      "learning_rate": 3.2351746547522344e-06,
+      "loss": 1.8002,
+      "step": 1158
+    },
+    {
+      "epoch": 1.2701369863013698,
+      "grad_norm": 2.5212879180908203,
+      "learning_rate": 3.2331437855402113e-06,
+      "loss": 1.863,
+      "step": 1159
+    },
+    {
+      "epoch": 1.2712328767123289,
+      "grad_norm": 2.3285911083221436,
+      "learning_rate": 3.2311129163281885e-06,
+      "loss": 1.9101,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2712328767123289,
+      "eval_loss": 2.038487195968628,
+      "eval_runtime": 36.8002,
+      "eval_samples_per_second": 27.174,
+      "eval_steps_per_second": 1.141,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2723287671232877,
+      "grad_norm": 2.1316046714782715,
+      "learning_rate": 3.229082047116166e-06,
+      "loss": 1.7952,
+      "step": 1161
+    },
+    {
+      "epoch": 1.2734246575342465,
+      "grad_norm": 2.4078125953674316,
+      "learning_rate": 3.227051177904143e-06,
+      "loss": 1.869,
+      "step": 1162
+    },
+    {
+      "epoch": 1.2745205479452055,
+      "grad_norm": 2.363672971725464,
+      "learning_rate": 3.2250203086921204e-06,
+      "loss": 1.8478,
+      "step": 1163
+    },
+    {
+      "epoch": 1.2756164383561643,
+      "grad_norm": 2.240558385848999,
+      "learning_rate": 3.2229894394800977e-06,
+      "loss": 1.8483,
+      "step": 1164
+    },
+    {
+      "epoch": 1.2767123287671232,
+      "grad_norm": 2.9142000675201416,
+      "learning_rate": 3.220958570268075e-06,
+      "loss": 1.8789,
+      "step": 1165
+    },
+    {
+      "epoch": 1.2778082191780822,
+      "grad_norm": 2.1930558681488037,
+      "learning_rate": 3.218927701056052e-06,
+      "loss": 1.7659,
+      "step": 1166
+    },
+    {
+      "epoch": 1.2789041095890412,
+      "grad_norm": 2.312471866607666,
+      "learning_rate": 3.2168968318440295e-06,
+      "loss": 1.8374,
+      "step": 1167
+    },
+    {
+      "epoch": 1.28,
+      "grad_norm": 2.6222236156463623,
+      "learning_rate": 3.2148659626320068e-06,
+      "loss": 1.7539,
+      "step": 1168
+    },
+    {
+      "epoch": 1.2810958904109588,
+      "grad_norm": 2.3013322353363037,
+      "learning_rate": 3.212835093419984e-06,
+      "loss": 1.9011,
+      "step": 1169
+    },
+    {
+      "epoch": 1.2821917808219179,
+      "grad_norm": 2.215240716934204,
+      "learning_rate": 3.2108042242079613e-06,
+      "loss": 1.8032,
+      "step": 1170
+    },
+    {
+      "epoch": 1.2821917808219179,
+      "eval_loss": 2.0365872383117676,
+      "eval_runtime": 36.8265,
+      "eval_samples_per_second": 27.154,
+      "eval_steps_per_second": 1.14,
+      "step": 1170
+    },
+    {
+      "epoch": 1.2832876712328767,
+      "grad_norm": 2.3620481491088867,
+      "learning_rate": 3.2087733549959386e-06,
+      "loss": 1.8258,
+      "step": 1171
+    },
+    {
+      "epoch": 1.2843835616438355,
+      "grad_norm": 2.2650208473205566,
+      "learning_rate": 3.206742485783916e-06,
+      "loss": 1.8888,
+      "step": 1172
+    },
+    {
+      "epoch": 1.2854794520547945,
+      "grad_norm": 2.2564077377319336,
+      "learning_rate": 3.2047116165718927e-06,
+      "loss": 1.7874,
+      "step": 1173
+    },
+    {
+      "epoch": 1.2865753424657536,
+      "grad_norm": 2.2855281829833984,
+      "learning_rate": 3.2026807473598704e-06,
+      "loss": 1.8003,
+      "step": 1174
+    },
+    {
+      "epoch": 1.2876712328767124,
+      "grad_norm": 2.353893756866455,
+      "learning_rate": 3.2006498781478473e-06,
+      "loss": 1.8466,
+      "step": 1175
+    },
+    {
+      "epoch": 1.2887671232876712,
+      "grad_norm": 2.1893045902252197,
+      "learning_rate": 3.198619008935825e-06,
+      "loss": 1.8652,
+      "step": 1176
+    },
+    {
+      "epoch": 1.2898630136986302,
+      "grad_norm": 2.6629793643951416,
+      "learning_rate": 3.196588139723802e-06,
+      "loss": 1.797,
+      "step": 1177
+    },
+    {
+      "epoch": 1.290958904109589,
+      "grad_norm": 2.2427361011505127,
+      "learning_rate": 3.1945572705117795e-06,
+      "loss": 1.8294,
+      "step": 1178
+    },
+    {
+      "epoch": 1.2920547945205478,
+      "grad_norm": 2.321878671646118,
+      "learning_rate": 3.1925264012997564e-06,
+      "loss": 1.8494,
+      "step": 1179
+    },
+    {
+      "epoch": 1.2931506849315069,
+      "grad_norm": 2.6928281784057617,
+      "learning_rate": 3.190495532087734e-06,
+      "loss": 1.8574,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2931506849315069,
+      "eval_loss": 2.0426652431488037,
+      "eval_runtime": 36.8482,
+      "eval_samples_per_second": 27.138,
+      "eval_steps_per_second": 1.14,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2942465753424657,
+      "grad_norm": 2.382164239883423,
+      "learning_rate": 3.188464662875711e-06,
+      "loss": 1.8518,
+      "step": 1181
+    },
+    {
+      "epoch": 1.2953424657534247,
+      "grad_norm": 2.4109256267547607,
+      "learning_rate": 3.1864337936636886e-06,
+      "loss": 1.9083,
+      "step": 1182
+    },
+    {
+      "epoch": 1.2964383561643835,
+      "grad_norm": 2.468062400817871,
+      "learning_rate": 3.1844029244516655e-06,
+      "loss": 1.8247,
+      "step": 1183
+    },
+    {
+      "epoch": 1.2975342465753426,
+      "grad_norm": 2.459378242492676,
+      "learning_rate": 3.182372055239643e-06,
+      "loss": 1.791,
+      "step": 1184
+    },
+    {
+      "epoch": 1.2986301369863014,
+      "grad_norm": 2.3331072330474854,
+      "learning_rate": 3.18034118602762e-06,
+      "loss": 1.7869,
+      "step": 1185
+    },
+    {
+      "epoch": 1.2997260273972602,
+      "grad_norm": 2.339428424835205,
+      "learning_rate": 3.1783103168155973e-06,
+      "loss": 1.8729,
+      "step": 1186
+    },
+    {
+      "epoch": 1.3008219178082192,
+      "grad_norm": 2.3033406734466553,
+      "learning_rate": 3.1762794476035746e-06,
+      "loss": 1.8455,
+      "step": 1187
+    },
+    {
+      "epoch": 1.301917808219178,
+      "grad_norm": 2.2480714321136475,
+      "learning_rate": 3.174248578391552e-06,
+      "loss": 1.8233,
+      "step": 1188
+    },
+    {
+      "epoch": 1.303013698630137,
+      "grad_norm": 2.2953624725341797,
+      "learning_rate": 3.172217709179529e-06,
+      "loss": 1.8261,
+      "step": 1189
+    },
+    {
+      "epoch": 1.3041095890410959,
+      "grad_norm": 2.494307518005371,
+      "learning_rate": 3.1701868399675064e-06,
+      "loss": 1.8153,
+      "step": 1190
+    },
+    {
+      "epoch": 1.3041095890410959,
+      "eval_loss": 2.037719964981079,
+      "eval_runtime": 36.9788,
+      "eval_samples_per_second": 27.043,
+      "eval_steps_per_second": 1.136,
+      "step": 1190
+    },
+    {
+      "epoch": 1.305205479452055,
+      "grad_norm": 2.314159631729126,
+      "learning_rate": 3.1681559707554837e-06,
+      "loss": 1.8406,
+      "step": 1191
+    },
+    {
+      "epoch": 1.3063013698630137,
+      "grad_norm": 2.4262588024139404,
+      "learning_rate": 3.166125101543461e-06,
+      "loss": 1.7803,
+      "step": 1192
+    },
+    {
+      "epoch": 1.3073972602739725,
+      "grad_norm": 2.3800835609436035,
+      "learning_rate": 3.1640942323314382e-06,
+      "loss": 1.8521,
+      "step": 1193
+    },
+    {
+      "epoch": 1.3084931506849315,
+      "grad_norm": 2.234551429748535,
+      "learning_rate": 3.1620633631194155e-06,
+      "loss": 1.845,
+      "step": 1194
+    },
+    {
+      "epoch": 1.3095890410958904,
+      "grad_norm": 2.167635440826416,
+      "learning_rate": 3.1600324939073928e-06,
+      "loss": 1.789,
+      "step": 1195
+    },
+    {
+      "epoch": 1.3106849315068494,
+      "grad_norm": 2.320247173309326,
+      "learning_rate": 3.15800162469537e-06,
+      "loss": 1.8311,
+      "step": 1196
+    },
+    {
+      "epoch": 1.3117808219178082,
+      "grad_norm": 2.2715561389923096,
+      "learning_rate": 3.155970755483347e-06,
+      "loss": 1.8308,
+      "step": 1197
+    },
+    {
+      "epoch": 1.3128767123287672,
+      "grad_norm": 2.2535455226898193,
+      "learning_rate": 3.1539398862713246e-06,
+      "loss": 1.8358,
+      "step": 1198
+    },
+    {
+      "epoch": 1.313972602739726,
+      "grad_norm": 2.2399163246154785,
+      "learning_rate": 3.1519090170593015e-06,
+      "loss": 1.8513,
+      "step": 1199
+    },
+    {
+      "epoch": 1.3150684931506849,
+      "grad_norm": 2.2984132766723633,
+      "learning_rate": 3.149878147847279e-06,
+      "loss": 1.8377,
+      "step": 1200
+    },
+    {
+      "epoch": 1.3150684931506849,
+      "eval_loss": 2.0334064960479736,
+      "eval_runtime": 37.0346,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1200
+    },
+    {
+      "epoch": 1.316164383561644,
+      "grad_norm": 2.1296892166137695,
+      "learning_rate": 3.147847278635256e-06,
+      "loss": 1.9023,
+      "step": 1201
+    },
+    {
+      "epoch": 1.3172602739726027,
+      "grad_norm": 2.1034350395202637,
+      "learning_rate": 3.1458164094232337e-06,
+      "loss": 1.917,
+      "step": 1202
+    },
+    {
+      "epoch": 1.3183561643835615,
+      "grad_norm": 2.258863687515259,
+      "learning_rate": 3.1437855402112106e-06,
+      "loss": 1.8093,
+      "step": 1203
+    },
+    {
+      "epoch": 1.3194520547945205,
+      "grad_norm": 2.20119309425354,
+      "learning_rate": 3.1417546709991883e-06,
+      "loss": 1.7796,
+      "step": 1204
+    },
+    {
+      "epoch": 1.3205479452054796,
+      "grad_norm": 2.3005869388580322,
+      "learning_rate": 3.139723801787165e-06,
+      "loss": 1.8496,
+      "step": 1205
+    },
+    {
+      "epoch": 1.3216438356164384,
+      "grad_norm": 2.2009031772613525,
+      "learning_rate": 3.137692932575143e-06,
+      "loss": 1.8244,
+      "step": 1206
+    },
+    {
+      "epoch": 1.3227397260273972,
+      "grad_norm": 2.3467633724212646,
+      "learning_rate": 3.1356620633631197e-06,
+      "loss": 1.8803,
+      "step": 1207
+    },
+    {
+      "epoch": 1.3238356164383562,
+      "grad_norm": 2.318711519241333,
+      "learning_rate": 3.1336311941510974e-06,
+      "loss": 1.7999,
+      "step": 1208
+    },
+    {
+      "epoch": 1.324931506849315,
+      "grad_norm": 2.182602882385254,
+      "learning_rate": 3.1316003249390742e-06,
+      "loss": 1.836,
+      "step": 1209
+    },
+    {
+      "epoch": 1.3260273972602739,
+      "grad_norm": 2.217298984527588,
+      "learning_rate": 3.129569455727051e-06,
+      "loss": 1.8776,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3260273972602739,
+      "eval_loss": 2.033573627471924,
+      "eval_runtime": 36.8029,
+      "eval_samples_per_second": 27.172,
+      "eval_steps_per_second": 1.141,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3271232876712329,
+      "grad_norm": 2.2814879417419434,
+      "learning_rate": 3.1275385865150288e-06,
+      "loss": 1.8739,
+      "step": 1211
+    },
+    {
+      "epoch": 1.3282191780821917,
+      "grad_norm": 2.1692471504211426,
+      "learning_rate": 3.1255077173030056e-06,
+      "loss": 1.8539,
+      "step": 1212
+    },
+    {
+      "epoch": 1.3293150684931507,
+      "grad_norm": 2.2472000122070312,
+      "learning_rate": 3.1234768480909833e-06,
+      "loss": 1.8527,
+      "step": 1213
+    },
+    {
+      "epoch": 1.3304109589041095,
+      "grad_norm": 2.1969852447509766,
+      "learning_rate": 3.12144597887896e-06,
+      "loss": 1.8429,
+      "step": 1214
+    },
+    {
+      "epoch": 1.3315068493150686,
+      "grad_norm": 2.141629219055176,
+      "learning_rate": 3.119415109666938e-06,
+      "loss": 1.868,
+      "step": 1215
+    },
+    {
+      "epoch": 1.3326027397260274,
+      "grad_norm": 2.282562494277954,
+      "learning_rate": 3.1173842404549147e-06,
+      "loss": 1.8913,
+      "step": 1216
+    },
+    {
+      "epoch": 1.3336986301369862,
+      "grad_norm": 2.2375850677490234,
+      "learning_rate": 3.1153533712428924e-06,
+      "loss": 1.8462,
+      "step": 1217
+    },
+    {
+      "epoch": 1.3347945205479452,
+      "grad_norm": 2.3529837131500244,
+      "learning_rate": 3.1133225020308693e-06,
+      "loss": 1.8445,
+      "step": 1218
+    },
+    {
+      "epoch": 1.335890410958904,
+      "grad_norm": 2.385857343673706,
+      "learning_rate": 3.111291632818847e-06,
+      "loss": 1.89,
+      "step": 1219
+    },
+    {
+      "epoch": 1.336986301369863,
+      "grad_norm": 2.301253080368042,
+      "learning_rate": 3.109260763606824e-06,
+      "loss": 1.8882,
+      "step": 1220
+    },
+    {
+      "epoch": 1.336986301369863,
+      "eval_loss": 2.040025234222412,
+      "eval_runtime": 36.839,
+      "eval_samples_per_second": 27.145,
+      "eval_steps_per_second": 1.14,
+      "step": 1220
+    },
+    {
+      "epoch": 1.3380821917808219,
+      "grad_norm": 2.3772335052490234,
+      "learning_rate": 3.1072298943948015e-06,
+      "loss": 1.8545,
+      "step": 1221
+    },
+    {
+      "epoch": 1.339178082191781,
+      "grad_norm": 2.387830972671509,
+      "learning_rate": 3.1051990251827784e-06,
+      "loss": 1.8653,
+      "step": 1222
+    },
+    {
+      "epoch": 1.3402739726027397,
+      "grad_norm": 2.4285190105438232,
+      "learning_rate": 3.1031681559707557e-06,
+      "loss": 1.8268,
+      "step": 1223
+    },
+    {
+      "epoch": 1.3413698630136985,
+      "grad_norm": 2.1892075538635254,
+      "learning_rate": 3.101137286758733e-06,
+      "loss": 1.8818,
+      "step": 1224
+    },
+    {
+      "epoch": 1.3424657534246576,
+      "grad_norm": 2.2510695457458496,
+      "learning_rate": 3.09910641754671e-06,
+      "loss": 1.879,
+      "step": 1225
+    },
+    {
+      "epoch": 1.3435616438356164,
+      "grad_norm": 2.505671977996826,
+      "learning_rate": 3.0970755483346875e-06,
+      "loss": 1.872,
+      "step": 1226
+    },
+    {
+      "epoch": 1.3446575342465754,
+      "grad_norm": 2.5315067768096924,
+      "learning_rate": 3.0950446791226648e-06,
+      "loss": 1.8651,
+      "step": 1227
+    },
+    {
+      "epoch": 1.3457534246575342,
+      "grad_norm": 2.223600387573242,
+      "learning_rate": 3.093013809910642e-06,
+      "loss": 1.9167,
+      "step": 1228
+    },
+    {
+      "epoch": 1.3468493150684933,
+      "grad_norm": 2.357802391052246,
+      "learning_rate": 3.0909829406986193e-06,
+      "loss": 1.8593,
+      "step": 1229
+    },
+    {
+      "epoch": 1.347945205479452,
+      "grad_norm": 2.3102364540100098,
+      "learning_rate": 3.0889520714865966e-06,
+      "loss": 1.8994,
+      "step": 1230
+    },
+    {
+      "epoch": 1.347945205479452,
+      "eval_loss": 2.0332930088043213,
+      "eval_runtime": 36.989,
+      "eval_samples_per_second": 27.035,
+      "eval_steps_per_second": 1.135,
+      "step": 1230
+    },
+    {
+      "epoch": 1.3490410958904109,
+      "grad_norm": 2.085080623626709,
+      "learning_rate": 3.086921202274574e-06,
+      "loss": 1.8741,
+      "step": 1231
+    },
+    {
+      "epoch": 1.35013698630137,
+      "grad_norm": 2.267305612564087,
+      "learning_rate": 3.084890333062551e-06,
+      "loss": 1.8426,
+      "step": 1232
+    },
+    {
+      "epoch": 1.3512328767123287,
+      "grad_norm": 2.181033134460449,
+      "learning_rate": 3.0828594638505284e-06,
+      "loss": 1.8199,
+      "step": 1233
+    },
+    {
+      "epoch": 1.3523287671232875,
+      "grad_norm": 2.288386821746826,
+      "learning_rate": 3.0808285946385057e-06,
+      "loss": 1.8469,
+      "step": 1234
+    },
+    {
+      "epoch": 1.3534246575342466,
+      "grad_norm": 2.1516127586364746,
+      "learning_rate": 3.078797725426483e-06,
+      "loss": 1.8056,
+      "step": 1235
+    },
+    {
+      "epoch": 1.3545205479452056,
+      "grad_norm": 2.2699198722839355,
+      "learning_rate": 3.07676685621446e-06,
+      "loss": 1.7535,
+      "step": 1236
+    },
+    {
+      "epoch": 1.3556164383561644,
+      "grad_norm": 2.193251132965088,
+      "learning_rate": 3.0747359870024375e-06,
+      "loss": 1.8407,
+      "step": 1237
+    },
+    {
+      "epoch": 1.3567123287671232,
+      "grad_norm": 2.2217774391174316,
+      "learning_rate": 3.0727051177904144e-06,
+      "loss": 1.8062,
+      "step": 1238
+    },
+    {
+      "epoch": 1.3578082191780823,
+      "grad_norm": 2.2376203536987305,
+      "learning_rate": 3.070674248578392e-06,
+      "loss": 1.829,
+      "step": 1239
+    },
+    {
+      "epoch": 1.358904109589041,
+      "grad_norm": 2.387559413909912,
+      "learning_rate": 3.068643379366369e-06,
+      "loss": 1.8459,
+      "step": 1240
+    },
+    {
+      "epoch": 1.358904109589041,
+      "eval_loss": 2.029298782348633,
+      "eval_runtime": 37.0491,
+      "eval_samples_per_second": 26.991,
+      "eval_steps_per_second": 1.134,
+      "step": 1240
+    },
+    {
+      "epoch": 1.3599999999999999,
+      "grad_norm": 2.1940178871154785,
+      "learning_rate": 3.0666125101543466e-06,
+      "loss": 1.831,
+      "step": 1241
+    },
+    {
+      "epoch": 1.361095890410959,
+      "grad_norm": 2.145514488220215,
+      "learning_rate": 3.0645816409423235e-06,
+      "loss": 1.7919,
+      "step": 1242
+    },
+    {
+      "epoch": 1.362191780821918,
+      "grad_norm": 2.319164276123047,
+      "learning_rate": 3.062550771730301e-06,
+      "loss": 1.8472,
+      "step": 1243
+    },
+    {
+      "epoch": 1.3632876712328768,
+      "grad_norm": 2.218986749649048,
+      "learning_rate": 3.060519902518278e-06,
+      "loss": 1.8184,
+      "step": 1244
+    },
+    {
+      "epoch": 1.3643835616438356,
+      "grad_norm": 2.3755338191986084,
+      "learning_rate": 3.0584890333062557e-06,
+      "loss": 1.8166,
+      "step": 1245
+    },
+    {
+      "epoch": 1.3654794520547946,
+      "grad_norm": 2.3186213970184326,
+      "learning_rate": 3.0564581640942326e-06,
+      "loss": 1.7599,
+      "step": 1246
+    },
+    {
+      "epoch": 1.3665753424657534,
+      "grad_norm": 2.193958282470703,
+      "learning_rate": 3.0544272948822094e-06,
+      "loss": 1.8404,
+      "step": 1247
+    },
+    {
+      "epoch": 1.3676712328767122,
+      "grad_norm": 2.40419602394104,
+      "learning_rate": 3.052396425670187e-06,
+      "loss": 1.821,
+      "step": 1248
+    },
+    {
+      "epoch": 1.3687671232876713,
+      "grad_norm": 2.445542573928833,
+      "learning_rate": 3.050365556458164e-06,
+      "loss": 1.8518,
+      "step": 1249
+    },
+    {
+      "epoch": 1.36986301369863,
+      "grad_norm": 2.2776567935943604,
+      "learning_rate": 3.0483346872461417e-06,
+      "loss": 1.8121,
+      "step": 1250
+    },
+    {
+      "epoch": 1.36986301369863,
+      "eval_loss": 2.030043125152588,
+      "eval_runtime": 37.0721,
+      "eval_samples_per_second": 26.974,
+      "eval_steps_per_second": 1.133,
+      "step": 1250
+    },
+    {
+      "epoch": 1.370958904109589,
+      "grad_norm": 2.420448064804077,
+      "learning_rate": 3.0463038180341185e-06,
+      "loss": 1.8168,
+      "step": 1251
+    },
+    {
+      "epoch": 1.372054794520548,
+      "grad_norm": 2.252065658569336,
+      "learning_rate": 3.0442729488220962e-06,
+      "loss": 1.8757,
+      "step": 1252
+    },
+    {
+      "epoch": 1.373150684931507,
+      "grad_norm": 2.4772560596466064,
+      "learning_rate": 3.042242079610073e-06,
+      "loss": 1.8711,
+      "step": 1253
+    },
+    {
+      "epoch": 1.3742465753424657,
+      "grad_norm": 2.3986129760742188,
+      "learning_rate": 3.0402112103980508e-06,
+      "loss": 1.8636,
+      "step": 1254
+    },
+    {
+      "epoch": 1.3753424657534246,
+      "grad_norm": 2.257274866104126,
+      "learning_rate": 3.0381803411860276e-06,
+      "loss": 1.7858,
+      "step": 1255
+    },
+    {
+      "epoch": 1.3764383561643836,
+      "grad_norm": 2.2373499870300293,
+      "learning_rate": 3.0361494719740053e-06,
+      "loss": 1.817,
+      "step": 1256
+    },
+    {
+      "epoch": 1.3775342465753424,
+      "grad_norm": 2.292388916015625,
+      "learning_rate": 3.034118602761982e-06,
+      "loss": 1.827,
+      "step": 1257
+    },
+    {
+      "epoch": 1.3786301369863014,
+      "grad_norm": 2.4618642330169678,
+      "learning_rate": 3.03208773354996e-06,
+      "loss": 1.8218,
+      "step": 1258
+    },
+    {
+      "epoch": 1.3797260273972602,
+      "grad_norm": 2.271958112716675,
+      "learning_rate": 3.0300568643379367e-06,
+      "loss": 1.8092,
+      "step": 1259
+    },
+    {
+      "epoch": 1.3808219178082193,
+      "grad_norm": 2.3990538120269775,
+      "learning_rate": 3.028025995125914e-06,
+      "loss": 1.8633,
+      "step": 1260
+    },
+    {
+      "epoch": 1.3808219178082193,
+      "eval_loss": 2.028458595275879,
+      "eval_runtime": 37.0659,
+      "eval_samples_per_second": 26.979,
+      "eval_steps_per_second": 1.133,
+      "step": 1260
+    },
+    {
+      "epoch": 1.381917808219178,
+      "grad_norm": 2.208803653717041,
+      "learning_rate": 3.0259951259138913e-06,
+      "loss": 1.8463,
+      "step": 1261
+    },
+    {
+      "epoch": 1.383013698630137,
+      "grad_norm": 2.3269028663635254,
+      "learning_rate": 3.0239642567018686e-06,
+      "loss": 1.7784,
+      "step": 1262
+    },
+    {
+      "epoch": 1.384109589041096,
+      "grad_norm": 2.367910623550415,
+      "learning_rate": 3.021933387489846e-06,
+      "loss": 1.8281,
+      "step": 1263
+    },
+    {
+      "epoch": 1.3852054794520547,
+      "grad_norm": 2.273686647415161,
+      "learning_rate": 3.019902518277823e-06,
+      "loss": 1.7832,
+      "step": 1264
+    },
+    {
+      "epoch": 1.3863013698630138,
+      "grad_norm": 2.421964168548584,
+      "learning_rate": 3.0178716490658004e-06,
+      "loss": 1.8407,
+      "step": 1265
+    },
+    {
+      "epoch": 1.3873972602739726,
+      "grad_norm": 2.2627696990966797,
+      "learning_rate": 3.0158407798537777e-06,
+      "loss": 1.8718,
+      "step": 1266
+    },
+    {
+      "epoch": 1.3884931506849316,
+      "grad_norm": 2.1981618404388428,
+      "learning_rate": 3.013809910641755e-06,
+      "loss": 1.7763,
+      "step": 1267
+    },
+    {
+      "epoch": 1.3895890410958904,
+      "grad_norm": 2.245253562927246,
+      "learning_rate": 3.0117790414297322e-06,
+      "loss": 1.7915,
+      "step": 1268
+    },
+    {
+      "epoch": 1.3906849315068492,
+      "grad_norm": 2.2292444705963135,
+      "learning_rate": 3.0097481722177095e-06,
+      "loss": 1.7909,
+      "step": 1269
+    },
+    {
+      "epoch": 1.3917808219178083,
+      "grad_norm": 2.2290430068969727,
+      "learning_rate": 3.0077173030056868e-06,
+      "loss": 1.8364,
+      "step": 1270
+    },
+    {
+      "epoch": 1.3917808219178083,
+      "eval_loss": 2.0290327072143555,
+      "eval_runtime": 37.0666,
+      "eval_samples_per_second": 26.978,
+      "eval_steps_per_second": 1.133,
+      "step": 1270
+    },
+    {
+      "epoch": 1.392876712328767,
+      "grad_norm": 2.220881700515747,
+      "learning_rate": 3.005686433793664e-06,
+      "loss": 1.7968,
+      "step": 1271
+    },
+    {
+      "epoch": 1.393972602739726,
+      "grad_norm": 2.375332832336426,
+      "learning_rate": 3.0036555645816413e-06,
+      "loss": 1.8502,
+      "step": 1272
+    },
+    {
+      "epoch": 1.395068493150685,
+      "grad_norm": 2.3065083026885986,
+      "learning_rate": 3.001624695369618e-06,
+      "loss": 1.8116,
+      "step": 1273
+    },
+    {
+      "epoch": 1.396164383561644,
+      "grad_norm": 2.7262511253356934,
+      "learning_rate": 2.999593826157596e-06,
+      "loss": 1.8641,
+      "step": 1274
+    },
+    {
+      "epoch": 1.3972602739726028,
+      "grad_norm": 2.3381216526031494,
+      "learning_rate": 2.9975629569455727e-06,
+      "loss": 1.8128,
+      "step": 1275
+    },
+    {
+      "epoch": 1.3983561643835616,
+      "grad_norm": 2.1994192600250244,
+      "learning_rate": 2.9955320877335504e-06,
+      "loss": 1.8563,
+      "step": 1276
+    },
+    {
+      "epoch": 1.3994520547945206,
+      "grad_norm": 2.3810958862304688,
+      "learning_rate": 2.9935012185215273e-06,
+      "loss": 1.8196,
+      "step": 1277
+    },
+    {
+      "epoch": 1.4005479452054794,
+      "grad_norm": 2.4343955516815186,
+      "learning_rate": 2.991470349309505e-06,
+      "loss": 1.8564,
+      "step": 1278
+    },
+    {
+      "epoch": 1.4016438356164382,
+      "grad_norm": 2.1929988861083984,
+      "learning_rate": 2.989439480097482e-06,
+      "loss": 1.8119,
+      "step": 1279
+    },
+    {
+      "epoch": 1.4027397260273973,
+      "grad_norm": 2.2305309772491455,
+      "learning_rate": 2.9874086108854595e-06,
+      "loss": 1.8289,
+      "step": 1280
+    },
+    {
+      "epoch": 1.4027397260273973,
+      "eval_loss": 2.032022476196289,
+      "eval_runtime": 37.084,
+      "eval_samples_per_second": 26.966,
+      "eval_steps_per_second": 1.133,
+      "step": 1280
+    },
+    {
+      "epoch": 1.403835616438356,
+      "grad_norm": 2.6255314350128174,
+      "learning_rate": 2.9853777416734364e-06,
+      "loss": 1.8176,
+      "step": 1281
+    },
+    {
+      "epoch": 1.4049315068493151,
+      "grad_norm": 2.238891363143921,
+      "learning_rate": 2.983346872461414e-06,
+      "loss": 1.8679,
+      "step": 1282
+    },
+    {
+      "epoch": 1.406027397260274,
+      "grad_norm": 2.9972007274627686,
+      "learning_rate": 2.981316003249391e-06,
+      "loss": 1.8839,
+      "step": 1283
+    },
+    {
+      "epoch": 1.407123287671233,
+      "grad_norm": 2.3255932331085205,
+      "learning_rate": 2.9792851340373678e-06,
+      "loss": 1.8451,
+      "step": 1284
+    },
+    {
+      "epoch": 1.4082191780821918,
+      "grad_norm": 2.290395498275757,
+      "learning_rate": 2.9772542648253455e-06,
+      "loss": 1.8244,
+      "step": 1285
+    },
+    {
+      "epoch": 1.4093150684931506,
+      "grad_norm": 2.288728713989258,
+      "learning_rate": 2.9752233956133223e-06,
+      "loss": 1.8628,
+      "step": 1286
+    },
+    {
+      "epoch": 1.4104109589041096,
+      "grad_norm": 2.177381992340088,
+      "learning_rate": 2.9731925264013e-06,
+      "loss": 1.8111,
+      "step": 1287
+    },
+    {
+      "epoch": 1.4115068493150684,
+      "grad_norm": 2.214956760406494,
+      "learning_rate": 2.971161657189277e-06,
+      "loss": 1.8377,
+      "step": 1288
+    },
+    {
+      "epoch": 1.4126027397260275,
+      "grad_norm": 2.3859663009643555,
+      "learning_rate": 2.9691307879772546e-06,
+      "loss": 1.8154,
+      "step": 1289
+    },
+    {
+      "epoch": 1.4136986301369863,
+      "grad_norm": 2.0873706340789795,
+      "learning_rate": 2.9670999187652314e-06,
+      "loss": 1.8098,
+      "step": 1290
+    },
+    {
+      "epoch": 1.4136986301369863,
+      "eval_loss": 2.029996633529663,
+      "eval_runtime": 37.0979,
+      "eval_samples_per_second": 26.956,
+      "eval_steps_per_second": 1.132,
+      "step": 1290
+    },
+    {
+      "epoch": 1.4147945205479453,
+      "grad_norm": 2.2632675170898438,
+      "learning_rate": 2.965069049553209e-06,
+      "loss": 1.7974,
+      "step": 1291
+    },
+    {
+      "epoch": 1.415890410958904,
+      "grad_norm": 2.208515167236328,
+      "learning_rate": 2.963038180341186e-06,
+      "loss": 1.8856,
+      "step": 1292
+    },
+    {
+      "epoch": 1.416986301369863,
+      "grad_norm": 2.176525115966797,
+      "learning_rate": 2.9610073111291637e-06,
+      "loss": 1.7783,
+      "step": 1293
+    },
+    {
+      "epoch": 1.418082191780822,
+      "grad_norm": 2.226682186126709,
+      "learning_rate": 2.9589764419171405e-06,
+      "loss": 1.865,
+      "step": 1294
+    },
+    {
+      "epoch": 1.4191780821917808,
+      "grad_norm": 2.3544883728027344,
+      "learning_rate": 2.9569455727051182e-06,
+      "loss": 1.8808,
+      "step": 1295
+    },
+    {
+      "epoch": 1.4202739726027398,
+      "grad_norm": 2.1729962825775146,
+      "learning_rate": 2.954914703493095e-06,
+      "loss": 1.7927,
+      "step": 1296
+    },
+    {
+      "epoch": 1.4213698630136986,
+      "grad_norm": 2.168168067932129,
+      "learning_rate": 2.9528838342810724e-06,
+      "loss": 1.8253,
+      "step": 1297
+    },
+    {
+      "epoch": 1.4224657534246576,
+      "grad_norm": 2.2479910850524902,
+      "learning_rate": 2.9508529650690496e-06,
+      "loss": 1.7976,
+      "step": 1298
+    },
+    {
+      "epoch": 1.4235616438356165,
+      "grad_norm": 2.1733837127685547,
+      "learning_rate": 2.948822095857027e-06,
+      "loss": 1.8015,
+      "step": 1299
+    },
+    {
+      "epoch": 1.4246575342465753,
+      "grad_norm": 2.2553391456604004,
+      "learning_rate": 2.946791226645004e-06,
+      "loss": 1.8545,
+      "step": 1300
+    },
+    {
+      "epoch": 1.4246575342465753,
+      "eval_loss": 2.0245048999786377,
+      "eval_runtime": 37.0943,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 1300
+    },
+    {
+      "epoch": 1.4257534246575343,
+      "grad_norm": 2.2817769050598145,
+      "learning_rate": 2.9447603574329815e-06,
+      "loss": 1.8017,
+      "step": 1301
+    },
+    {
+      "epoch": 1.426849315068493,
+      "grad_norm": 2.170865297317505,
+      "learning_rate": 2.9427294882209587e-06,
+      "loss": 1.7738,
+      "step": 1302
+    },
+    {
+      "epoch": 1.427945205479452,
+      "grad_norm": 2.296919107437134,
+      "learning_rate": 2.940698619008936e-06,
+      "loss": 1.8098,
+      "step": 1303
+    },
+    {
+      "epoch": 1.429041095890411,
+      "grad_norm": 2.476266622543335,
+      "learning_rate": 2.9386677497969133e-06,
+      "loss": 1.8229,
+      "step": 1304
+    },
+    {
+      "epoch": 1.43013698630137,
+      "grad_norm": 2.400918483734131,
+      "learning_rate": 2.9366368805848906e-06,
+      "loss": 1.8057,
+      "step": 1305
+    },
+    {
+      "epoch": 1.4312328767123288,
+      "grad_norm": 2.393892765045166,
+      "learning_rate": 2.934606011372868e-06,
+      "loss": 1.8568,
+      "step": 1306
+    },
+    {
+      "epoch": 1.4323287671232876,
+      "grad_norm": 2.2551307678222656,
+      "learning_rate": 2.932575142160845e-06,
+      "loss": 1.8493,
+      "step": 1307
+    },
+    {
+      "epoch": 1.4334246575342466,
+      "grad_norm": 2.3091611862182617,
+      "learning_rate": 2.9305442729488224e-06,
+      "loss": 1.8297,
+      "step": 1308
+    },
+    {
+      "epoch": 1.4345205479452054,
+      "grad_norm": 2.231614828109741,
+      "learning_rate": 2.9285134037367997e-06,
+      "loss": 1.8345,
+      "step": 1309
+    },
+    {
+      "epoch": 1.4356164383561643,
+      "grad_norm": 2.24942946434021,
+      "learning_rate": 2.9264825345247765e-06,
+      "loss": 1.8284,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4356164383561643,
+      "eval_loss": 2.024843215942383,
+      "eval_runtime": 36.7795,
+      "eval_samples_per_second": 27.189,
+      "eval_steps_per_second": 1.142,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4367123287671233,
+      "grad_norm": 2.2798643112182617,
+      "learning_rate": 2.9244516653127542e-06,
+      "loss": 1.8554,
+      "step": 1311
+    },
+    {
+      "epoch": 1.4378082191780823,
+      "grad_norm": 2.342576265335083,
+      "learning_rate": 2.922420796100731e-06,
+      "loss": 1.8296,
+      "step": 1312
+    },
+    {
+      "epoch": 1.4389041095890411,
+      "grad_norm": 2.268016815185547,
+      "learning_rate": 2.9203899268887088e-06,
+      "loss": 1.7467,
+      "step": 1313
+    },
+    {
+      "epoch": 1.44,
+      "grad_norm": 2.2520217895507812,
+      "learning_rate": 2.9183590576766856e-06,
+      "loss": 1.8748,
+      "step": 1314
+    },
+    {
+      "epoch": 1.441095890410959,
+      "grad_norm": 2.2091598510742188,
+      "learning_rate": 2.9163281884646633e-06,
+      "loss": 1.7715,
+      "step": 1315
+    },
+    {
+      "epoch": 1.4421917808219178,
+      "grad_norm": 2.197575092315674,
+      "learning_rate": 2.91429731925264e-06,
+      "loss": 1.8265,
+      "step": 1316
+    },
+    {
+      "epoch": 1.4432876712328766,
+      "grad_norm": 2.261122465133667,
+      "learning_rate": 2.912266450040618e-06,
+      "loss": 1.8758,
+      "step": 1317
+    },
+    {
+      "epoch": 1.4443835616438356,
+      "grad_norm": 2.3291778564453125,
+      "learning_rate": 2.9102355808285947e-06,
+      "loss": 1.8213,
+      "step": 1318
+    },
+    {
+      "epoch": 1.4454794520547944,
+      "grad_norm": 2.2182250022888184,
+      "learning_rate": 2.9082047116165724e-06,
+      "loss": 1.822,
+      "step": 1319
+    },
+    {
+      "epoch": 1.4465753424657535,
+      "grad_norm": 2.295978546142578,
+      "learning_rate": 2.9061738424045493e-06,
+      "loss": 1.8008,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4465753424657535,
+      "eval_loss": 2.0230674743652344,
+      "eval_runtime": 37.0013,
+      "eval_samples_per_second": 27.026,
+      "eval_steps_per_second": 1.135,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4476712328767123,
+      "grad_norm": 2.3187007904052734,
+      "learning_rate": 2.904142973192527e-06,
+      "loss": 1.7923,
+      "step": 1321
+    },
+    {
+      "epoch": 1.4487671232876713,
+      "grad_norm": 2.242851972579956,
+      "learning_rate": 2.902112103980504e-06,
+      "loss": 1.8874,
+      "step": 1322
+    },
+    {
+      "epoch": 1.4498630136986301,
+      "grad_norm": 2.2560081481933594,
+      "learning_rate": 2.900081234768481e-06,
+      "loss": 1.8284,
+      "step": 1323
+    },
+    {
+      "epoch": 1.450958904109589,
+      "grad_norm": 2.3017799854278564,
+      "learning_rate": 2.8980503655564584e-06,
+      "loss": 1.8263,
+      "step": 1324
+    },
+    {
+      "epoch": 1.452054794520548,
+      "grad_norm": 2.3016278743743896,
+      "learning_rate": 2.8960194963444357e-06,
+      "loss": 1.8174,
+      "step": 1325
+    },
+    {
+      "epoch": 1.4531506849315068,
+      "grad_norm": 2.1965372562408447,
+      "learning_rate": 2.893988627132413e-06,
+      "loss": 1.7812,
+      "step": 1326
+    },
+    {
+      "epoch": 1.4542465753424658,
+      "grad_norm": 2.302969455718994,
+      "learning_rate": 2.8919577579203902e-06,
+      "loss": 1.8115,
+      "step": 1327
+    },
+    {
+      "epoch": 1.4553424657534246,
+      "grad_norm": 2.338066339492798,
+      "learning_rate": 2.8899268887083675e-06,
+      "loss": 1.7535,
+      "step": 1328
+    },
+    {
+      "epoch": 1.4564383561643837,
+      "grad_norm": 2.446503162384033,
+      "learning_rate": 2.8878960194963448e-06,
+      "loss": 1.8052,
+      "step": 1329
+    },
+    {
+      "epoch": 1.4575342465753425,
+      "grad_norm": 2.2319273948669434,
+      "learning_rate": 2.885865150284322e-06,
+      "loss": 1.8226,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4575342465753425,
+      "eval_loss": 2.022364854812622,
+      "eval_runtime": 37.0901,
+      "eval_samples_per_second": 26.961,
+      "eval_steps_per_second": 1.132,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4586301369863013,
+      "grad_norm": 2.391080856323242,
+      "learning_rate": 2.8838342810722993e-06,
+      "loss": 1.8529,
+      "step": 1331
+    },
+    {
+      "epoch": 1.4597260273972603,
+      "grad_norm": 2.274463415145874,
+      "learning_rate": 2.8818034118602766e-06,
+      "loss": 1.8043,
+      "step": 1332
+    },
+    {
+      "epoch": 1.4608219178082191,
+      "grad_norm": 2.4040963649749756,
+      "learning_rate": 2.879772542648254e-06,
+      "loss": 1.7842,
+      "step": 1333
+    },
+    {
+      "epoch": 1.4619178082191782,
+      "grad_norm": 2.268629312515259,
+      "learning_rate": 2.8777416734362307e-06,
+      "loss": 1.859,
+      "step": 1334
+    },
+    {
+      "epoch": 1.463013698630137,
+      "grad_norm": 2.4727842807769775,
+      "learning_rate": 2.8757108042242084e-06,
+      "loss": 1.7615,
+      "step": 1335
+    },
+    {
+      "epoch": 1.464109589041096,
+      "grad_norm": 2.179358959197998,
+      "learning_rate": 2.8736799350121853e-06,
+      "loss": 1.9085,
+      "step": 1336
+    },
+    {
+      "epoch": 1.4652054794520548,
+      "grad_norm": 2.220917224884033,
+      "learning_rate": 2.871649065800163e-06,
+      "loss": 1.8618,
+      "step": 1337
+    },
+    {
+      "epoch": 1.4663013698630136,
+      "grad_norm": 2.3807291984558105,
+      "learning_rate": 2.86961819658814e-06,
+      "loss": 1.88,
+      "step": 1338
+    },
+    {
+      "epoch": 1.4673972602739727,
+      "grad_norm": 2.272507667541504,
+      "learning_rate": 2.8675873273761175e-06,
+      "loss": 1.8045,
+      "step": 1339
+    },
+    {
+      "epoch": 1.4684931506849315,
+      "grad_norm": 2.1832377910614014,
+      "learning_rate": 2.8655564581640944e-06,
+      "loss": 1.8273,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4684931506849315,
+      "eval_loss": 2.0218358039855957,
+      "eval_runtime": 37.1223,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4695890410958903,
+      "grad_norm": 2.2655251026153564,
+      "learning_rate": 2.863525588952072e-06,
+      "loss": 1.8162,
+      "step": 1341
+    },
+    {
+      "epoch": 1.4706849315068493,
+      "grad_norm": 2.191852331161499,
+      "learning_rate": 2.861494719740049e-06,
+      "loss": 1.8778,
+      "step": 1342
+    },
+    {
+      "epoch": 1.4717808219178083,
+      "grad_norm": 2.1869187355041504,
+      "learning_rate": 2.8594638505280266e-06,
+      "loss": 1.7695,
+      "step": 1343
+    },
+    {
+      "epoch": 1.4728767123287672,
+      "grad_norm": 2.248108148574829,
+      "learning_rate": 2.8574329813160035e-06,
+      "loss": 1.7759,
+      "step": 1344
+    },
+    {
+      "epoch": 1.473972602739726,
+      "grad_norm": 2.2599704265594482,
+      "learning_rate": 2.855402112103981e-06,
+      "loss": 1.8214,
+      "step": 1345
+    },
+    {
+      "epoch": 1.475068493150685,
+      "grad_norm": 2.2920355796813965,
+      "learning_rate": 2.853371242891958e-06,
+      "loss": 1.8061,
+      "step": 1346
+    },
+    {
+      "epoch": 1.4761643835616438,
+      "grad_norm": 2.26715087890625,
+      "learning_rate": 2.851340373679935e-06,
+      "loss": 1.8244,
+      "step": 1347
+    },
+    {
+      "epoch": 1.4772602739726026,
+      "grad_norm": 2.349169969558716,
+      "learning_rate": 2.8493095044679126e-06,
+      "loss": 1.8401,
+      "step": 1348
+    },
+    {
+      "epoch": 1.4783561643835617,
+      "grad_norm": 2.3571698665618896,
+      "learning_rate": 2.8472786352558894e-06,
+      "loss": 1.8913,
+      "step": 1349
+    },
+    {
+      "epoch": 1.4794520547945205,
+      "grad_norm": 2.334057331085205,
+      "learning_rate": 2.845247766043867e-06,
+      "loss": 1.8434,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4794520547945205,
+      "eval_loss": 2.0251200199127197,
+      "eval_runtime": 37.1298,
+      "eval_samples_per_second": 26.933,
+      "eval_steps_per_second": 1.131,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4805479452054795,
+      "grad_norm": 2.363516330718994,
+      "learning_rate": 2.843216896831844e-06,
+      "loss": 1.7846,
+      "step": 1351
+    },
+    {
+      "epoch": 1.4816438356164383,
+      "grad_norm": 2.372403860092163,
+      "learning_rate": 2.8411860276198217e-06,
+      "loss": 1.7791,
+      "step": 1352
+    },
+    {
+      "epoch": 1.4827397260273973,
+      "grad_norm": 2.2666590213775635,
+      "learning_rate": 2.8391551584077985e-06,
+      "loss": 1.7969,
+      "step": 1353
+    },
+    {
+      "epoch": 1.4838356164383562,
+      "grad_norm": 2.3651468753814697,
+      "learning_rate": 2.8371242891957762e-06,
+      "loss": 1.8565,
+      "step": 1354
+    },
+    {
+      "epoch": 1.484931506849315,
+      "grad_norm": 2.3559513092041016,
+      "learning_rate": 2.835093419983753e-06,
+      "loss": 1.8765,
+      "step": 1355
+    },
+    {
+      "epoch": 1.486027397260274,
+      "grad_norm": 2.232041358947754,
+      "learning_rate": 2.833062550771731e-06,
+      "loss": 1.805,
+      "step": 1356
+    },
+    {
+      "epoch": 1.4871232876712328,
+      "grad_norm": 2.2423202991485596,
+      "learning_rate": 2.8310316815597076e-06,
+      "loss": 1.8137,
+      "step": 1357
+    },
+    {
+      "epoch": 1.4882191780821918,
+      "grad_norm": 2.3454222679138184,
+      "learning_rate": 2.8290008123476853e-06,
+      "loss": 1.7839,
+      "step": 1358
+    },
+    {
+      "epoch": 1.4893150684931507,
+      "grad_norm": 2.351266384124756,
+      "learning_rate": 2.826969943135662e-06,
+      "loss": 1.8208,
+      "step": 1359
+    },
+    {
+      "epoch": 1.4904109589041097,
+      "grad_norm": 2.3025388717651367,
+      "learning_rate": 2.8249390739236395e-06,
+      "loss": 1.8059,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4904109589041097,
+      "eval_loss": 2.0235962867736816,
+      "eval_runtime": 37.1359,
+      "eval_samples_per_second": 26.928,
+      "eval_steps_per_second": 1.131,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4915068493150685,
+      "grad_norm": 2.2277750968933105,
+      "learning_rate": 2.8229082047116167e-06,
+      "loss": 1.8238,
+      "step": 1361
+    },
+    {
+      "epoch": 1.4926027397260273,
+      "grad_norm": 2.318286180496216,
+      "learning_rate": 2.820877335499594e-06,
+      "loss": 1.7867,
+      "step": 1362
+    },
+    {
+      "epoch": 1.4936986301369863,
+      "grad_norm": 2.296259880065918,
+      "learning_rate": 2.8188464662875713e-06,
+      "loss": 1.8068,
+      "step": 1363
+    },
+    {
+      "epoch": 1.4947945205479451,
+      "grad_norm": 2.2617714405059814,
+      "learning_rate": 2.8168155970755486e-06,
+      "loss": 1.7607,
+      "step": 1364
+    },
+    {
+      "epoch": 1.4958904109589042,
+      "grad_norm": 2.318294048309326,
+      "learning_rate": 2.814784727863526e-06,
+      "loss": 1.7872,
+      "step": 1365
+    },
+    {
+      "epoch": 1.496986301369863,
+      "grad_norm": 2.260286808013916,
+      "learning_rate": 2.812753858651503e-06,
+      "loss": 1.8013,
+      "step": 1366
+    },
+    {
+      "epoch": 1.498082191780822,
+      "grad_norm": 2.222074508666992,
+      "learning_rate": 2.8107229894394804e-06,
+      "loss": 1.8338,
+      "step": 1367
+    },
+    {
+      "epoch": 1.4991780821917808,
+      "grad_norm": 2.6403002738952637,
+      "learning_rate": 2.8086921202274577e-06,
+      "loss": 1.8126,
+      "step": 1368
+    },
+    {
+      "epoch": 1.5002739726027396,
+      "grad_norm": 2.2439680099487305,
+      "learning_rate": 2.806661251015435e-06,
+      "loss": 1.7669,
+      "step": 1369
+    },
+    {
+      "epoch": 1.5013698630136987,
+      "grad_norm": 2.3914687633514404,
+      "learning_rate": 2.8046303818034122e-06,
+      "loss": 1.8421,
+      "step": 1370
+    },
+    {
+      "epoch": 1.5013698630136987,
+      "eval_loss": 2.0222702026367188,
+      "eval_runtime": 37.1598,
+      "eval_samples_per_second": 26.911,
+      "eval_steps_per_second": 1.13,
+      "step": 1370
+    },
+    {
+      "epoch": 1.5024657534246575,
+      "grad_norm": 2.3510093688964844,
+      "learning_rate": 2.802599512591389e-06,
+      "loss": 1.8902,
+      "step": 1371
+    },
+    {
+      "epoch": 1.5035616438356163,
+      "grad_norm": 2.4427363872528076,
+      "learning_rate": 2.8005686433793668e-06,
+      "loss": 1.9376,
+      "step": 1372
+    },
+    {
+      "epoch": 1.5046575342465753,
+      "grad_norm": 2.299356460571289,
+      "learning_rate": 2.7985377741673436e-06,
+      "loss": 1.812,
+      "step": 1373
+    },
+    {
+      "epoch": 1.5057534246575344,
+      "grad_norm": 2.3738906383514404,
+      "learning_rate": 2.7965069049553213e-06,
+      "loss": 1.7666,
+      "step": 1374
+    },
+    {
+      "epoch": 1.5068493150684932,
+      "grad_norm": 2.3620493412017822,
+      "learning_rate": 2.794476035743298e-06,
+      "loss": 1.824,
+      "step": 1375
+    },
+    {
+      "epoch": 1.507945205479452,
+      "grad_norm": 2.539048671722412,
+      "learning_rate": 2.792445166531276e-06,
+      "loss": 1.8712,
+      "step": 1376
+    },
+    {
+      "epoch": 1.509041095890411,
+      "grad_norm": 2.2931456565856934,
+      "learning_rate": 2.7904142973192527e-06,
+      "loss": 1.8414,
+      "step": 1377
+    },
+    {
+      "epoch": 1.5101369863013698,
+      "grad_norm": 2.4368224143981934,
+      "learning_rate": 2.7883834281072304e-06,
+      "loss": 1.8419,
+      "step": 1378
+    },
+    {
+      "epoch": 1.5112328767123286,
+      "grad_norm": 2.298302173614502,
+      "learning_rate": 2.7863525588952073e-06,
+      "loss": 1.7846,
+      "step": 1379
+    },
+    {
+      "epoch": 1.5123287671232877,
+      "grad_norm": 2.3726806640625,
+      "learning_rate": 2.784321689683185e-06,
+      "loss": 1.7733,
+      "step": 1380
+    },
+    {
+      "epoch": 1.5123287671232877,
+      "eval_loss": 2.0201807022094727,
+      "eval_runtime": 37.1012,
+      "eval_samples_per_second": 26.953,
+      "eval_steps_per_second": 1.132,
+      "step": 1380
+    },
+    {
+      "epoch": 1.5134246575342467,
+      "grad_norm": 2.6515002250671387,
+      "learning_rate": 2.782290820471162e-06,
+      "loss": 1.7764,
+      "step": 1381
+    },
+    {
+      "epoch": 1.5145205479452055,
+      "grad_norm": 2.273272752761841,
+      "learning_rate": 2.7802599512591395e-06,
+      "loss": 1.7747,
+      "step": 1382
+    },
+    {
+      "epoch": 1.5156164383561643,
+      "grad_norm": 2.570009708404541,
+      "learning_rate": 2.7782290820471164e-06,
+      "loss": 1.8612,
+      "step": 1383
+    },
+    {
+      "epoch": 1.5167123287671234,
+      "grad_norm": 2.567878007888794,
+      "learning_rate": 2.7761982128350932e-06,
+      "loss": 1.8402,
+      "step": 1384
+    },
+    {
+      "epoch": 1.5178082191780822,
+      "grad_norm": 2.3768470287323,
+      "learning_rate": 2.774167343623071e-06,
+      "loss": 1.8473,
+      "step": 1385
+    },
+    {
+      "epoch": 1.518904109589041,
+      "grad_norm": 2.1383397579193115,
+      "learning_rate": 2.772136474411048e-06,
+      "loss": 1.8593,
+      "step": 1386
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 2.4437034130096436,
+      "learning_rate": 2.7701056051990255e-06,
+      "loss": 1.792,
+      "step": 1387
+    },
+    {
+      "epoch": 1.521095890410959,
+      "grad_norm": 2.433933734893799,
+      "learning_rate": 2.7680747359870023e-06,
+      "loss": 1.7312,
+      "step": 1388
+    },
+    {
+      "epoch": 1.5221917808219179,
+      "grad_norm": 2.340994119644165,
+      "learning_rate": 2.76604386677498e-06,
+      "loss": 1.8326,
+      "step": 1389
+    },
+    {
+      "epoch": 1.5232876712328767,
+      "grad_norm": 2.4730165004730225,
+      "learning_rate": 2.764012997562957e-06,
+      "loss": 1.8129,
+      "step": 1390
+    },
+    {
+      "epoch": 1.5232876712328767,
+      "eval_loss": 2.0158324241638184,
+      "eval_runtime": 37.0997,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 1390
+    },
+    {
+      "epoch": 1.5243835616438357,
+      "grad_norm": 2.2898151874542236,
+      "learning_rate": 2.7619821283509346e-06,
+      "loss": 1.8726,
+      "step": 1391
+    },
+    {
+      "epoch": 1.5254794520547945,
+      "grad_norm": 2.255439519882202,
+      "learning_rate": 2.7599512591389114e-06,
+      "loss": 1.7587,
+      "step": 1392
+    },
+    {
+      "epoch": 1.5265753424657533,
+      "grad_norm": 2.4269862174987793,
+      "learning_rate": 2.757920389926889e-06,
+      "loss": 1.8498,
+      "step": 1393
+    },
+    {
+      "epoch": 1.5276712328767124,
+      "grad_norm": 2.3411412239074707,
+      "learning_rate": 2.755889520714866e-06,
+      "loss": 1.8103,
+      "step": 1394
+    },
+    {
+      "epoch": 1.5287671232876714,
+      "grad_norm": 2.2859277725219727,
+      "learning_rate": 2.7538586515028437e-06,
+      "loss": 1.7909,
+      "step": 1395
+    },
+    {
+      "epoch": 1.52986301369863,
+      "grad_norm": 2.2675201892852783,
+      "learning_rate": 2.7518277822908205e-06,
+      "loss": 1.7672,
+      "step": 1396
+    },
+    {
+      "epoch": 1.530958904109589,
+      "grad_norm": 2.240610361099243,
+      "learning_rate": 2.749796913078798e-06,
+      "loss": 1.8162,
+      "step": 1397
+    },
+    {
+      "epoch": 1.532054794520548,
+      "grad_norm": 2.272045373916626,
+      "learning_rate": 2.747766043866775e-06,
+      "loss": 1.8593,
+      "step": 1398
+    },
+    {
+      "epoch": 1.5331506849315069,
+      "grad_norm": 2.1736183166503906,
+      "learning_rate": 2.7457351746547524e-06,
+      "loss": 1.8154,
+      "step": 1399
+    },
+    {
+      "epoch": 1.5342465753424657,
+      "grad_norm": 2.197085380554199,
+      "learning_rate": 2.7437043054427297e-06,
+      "loss": 1.8758,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5342465753424657,
+      "eval_loss": 2.017298698425293,
+      "eval_runtime": 37.1666,
+      "eval_samples_per_second": 26.906,
+      "eval_steps_per_second": 1.13,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5353424657534247,
+      "grad_norm": 2.2274084091186523,
+      "learning_rate": 2.741673436230707e-06,
+      "loss": 1.9241,
+      "step": 1401
+    },
+    {
+      "epoch": 1.5364383561643835,
+      "grad_norm": 2.3320305347442627,
+      "learning_rate": 2.739642567018684e-06,
+      "loss": 1.8209,
+      "step": 1402
+    },
+    {
+      "epoch": 1.5375342465753423,
+      "grad_norm": 2.370651960372925,
+      "learning_rate": 2.7376116978066615e-06,
+      "loss": 1.7897,
+      "step": 1403
+    },
+    {
+      "epoch": 1.5386301369863014,
+      "grad_norm": 2.387906312942505,
+      "learning_rate": 2.7355808285946388e-06,
+      "loss": 1.8385,
+      "step": 1404
+    },
+    {
+      "epoch": 1.5397260273972604,
+      "grad_norm": 2.340228796005249,
+      "learning_rate": 2.733549959382616e-06,
+      "loss": 1.8279,
+      "step": 1405
+    },
+    {
+      "epoch": 1.5408219178082192,
+      "grad_norm": 2.338785409927368,
+      "learning_rate": 2.7315190901705933e-06,
+      "loss": 1.7887,
+      "step": 1406
+    },
+    {
+      "epoch": 1.541917808219178,
+      "grad_norm": 2.350057601928711,
+      "learning_rate": 2.7294882209585706e-06,
+      "loss": 1.8772,
+      "step": 1407
+    },
+    {
+      "epoch": 1.543013698630137,
+      "grad_norm": 2.1129322052001953,
+      "learning_rate": 2.727457351746548e-06,
+      "loss": 1.8393,
+      "step": 1408
+    },
+    {
+      "epoch": 1.5441095890410959,
+      "grad_norm": 2.282553195953369,
+      "learning_rate": 2.725426482534525e-06,
+      "loss": 1.7893,
+      "step": 1409
+    },
+    {
+      "epoch": 1.5452054794520547,
+      "grad_norm": 2.3291871547698975,
+      "learning_rate": 2.723395613322502e-06,
+      "loss": 1.7921,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5452054794520547,
+      "eval_loss": 2.0166046619415283,
+      "eval_runtime": 36.7906,
+      "eval_samples_per_second": 27.181,
+      "eval_steps_per_second": 1.142,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5463013698630137,
+      "grad_norm": 2.3050696849823,
+      "learning_rate": 2.7213647441104797e-06,
+      "loss": 1.8292,
+      "step": 1411
+    },
+    {
+      "epoch": 1.5473972602739727,
+      "grad_norm": 2.2762458324432373,
+      "learning_rate": 2.7193338748984565e-06,
+      "loss": 1.811,
+      "step": 1412
+    },
+    {
+      "epoch": 1.5484931506849315,
+      "grad_norm": 2.5544228553771973,
+      "learning_rate": 2.7173030056864342e-06,
+      "loss": 1.8203,
+      "step": 1413
+    },
+    {
+      "epoch": 1.5495890410958904,
+      "grad_norm": 2.6392276287078857,
+      "learning_rate": 2.715272136474411e-06,
+      "loss": 1.9724,
+      "step": 1414
+    },
+    {
+      "epoch": 1.5506849315068494,
+      "grad_norm": 2.445366382598877,
+      "learning_rate": 2.713241267262389e-06,
+      "loss": 1.7589,
+      "step": 1415
+    },
+    {
+      "epoch": 1.5517808219178082,
+      "grad_norm": 2.5305662155151367,
+      "learning_rate": 2.7112103980503656e-06,
+      "loss": 1.8311,
+      "step": 1416
+    },
+    {
+      "epoch": 1.552876712328767,
+      "grad_norm": 2.405482769012451,
+      "learning_rate": 2.7091795288383433e-06,
+      "loss": 1.7893,
+      "step": 1417
+    },
+    {
+      "epoch": 1.553972602739726,
+      "grad_norm": 2.258361577987671,
+      "learning_rate": 2.70714865962632e-06,
+      "loss": 1.8489,
+      "step": 1418
+    },
+    {
+      "epoch": 1.555068493150685,
+      "grad_norm": 2.1798577308654785,
+      "learning_rate": 2.705117790414298e-06,
+      "loss": 1.7983,
+      "step": 1419
+    },
+    {
+      "epoch": 1.5561643835616439,
+      "grad_norm": 2.1985363960266113,
+      "learning_rate": 2.7030869212022747e-06,
+      "loss": 1.7934,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5561643835616439,
+      "eval_loss": 2.0219907760620117,
+      "eval_runtime": 36.9416,
+      "eval_samples_per_second": 27.07,
+      "eval_steps_per_second": 1.137,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5572602739726027,
+      "grad_norm": 2.273252248764038,
+      "learning_rate": 2.7010560519902516e-06,
+      "loss": 1.8571,
+      "step": 1421
+    },
+    {
+      "epoch": 1.5583561643835617,
+      "grad_norm": 2.22920298576355,
+      "learning_rate": 2.6990251827782293e-06,
+      "loss": 1.7785,
+      "step": 1422
+    },
+    {
+      "epoch": 1.5594520547945205,
+      "grad_norm": 2.1784098148345947,
+      "learning_rate": 2.696994313566206e-06,
+      "loss": 1.8053,
+      "step": 1423
+    },
+    {
+      "epoch": 1.5605479452054793,
+      "grad_norm": 2.223404884338379,
+      "learning_rate": 2.694963444354184e-06,
+      "loss": 1.7965,
+      "step": 1424
+    },
+    {
+      "epoch": 1.5616438356164384,
+      "grad_norm": 2.2004787921905518,
+      "learning_rate": 2.6929325751421607e-06,
+      "loss": 1.8851,
+      "step": 1425
+    },
+    {
+      "epoch": 1.5627397260273974,
+      "grad_norm": 2.202467441558838,
+      "learning_rate": 2.6909017059301384e-06,
+      "loss": 1.7884,
+      "step": 1426
+    },
+    {
+      "epoch": 1.563835616438356,
+      "grad_norm": 2.337430238723755,
+      "learning_rate": 2.6888708367181153e-06,
+      "loss": 1.8543,
+      "step": 1427
+    },
+    {
+      "epoch": 1.564931506849315,
+      "grad_norm": 2.322848081588745,
+      "learning_rate": 2.686839967506093e-06,
+      "loss": 1.7607,
+      "step": 1428
+    },
+    {
+      "epoch": 1.566027397260274,
+      "grad_norm": 2.277765989303589,
+      "learning_rate": 2.68480909829407e-06,
+      "loss": 1.8354,
+      "step": 1429
+    },
+    {
+      "epoch": 1.5671232876712329,
+      "grad_norm": 2.3229448795318604,
+      "learning_rate": 2.6827782290820475e-06,
+      "loss": 1.7277,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5671232876712329,
+      "eval_loss": 2.0220985412597656,
+      "eval_runtime": 37.0054,
+      "eval_samples_per_second": 27.023,
+      "eval_steps_per_second": 1.135,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5682191780821917,
+      "grad_norm": 2.49545955657959,
+      "learning_rate": 2.6807473598700244e-06,
+      "loss": 1.7351,
+      "step": 1431
+    },
+    {
+      "epoch": 1.5693150684931507,
+      "grad_norm": 2.530553102493286,
+      "learning_rate": 2.678716490658002e-06,
+      "loss": 1.8614,
+      "step": 1432
+    },
+    {
+      "epoch": 1.5704109589041098,
+      "grad_norm": 2.3352842330932617,
+      "learning_rate": 2.676685621445979e-06,
+      "loss": 1.8218,
+      "step": 1433
+    },
+    {
+      "epoch": 1.5715068493150683,
+      "grad_norm": 2.3533058166503906,
+      "learning_rate": 2.674654752233956e-06,
+      "loss": 1.8282,
+      "step": 1434
+    },
+    {
+      "epoch": 1.5726027397260274,
+      "grad_norm": 2.5509746074676514,
+      "learning_rate": 2.6726238830219335e-06,
+      "loss": 1.8333,
+      "step": 1435
+    },
+    {
+      "epoch": 1.5736986301369864,
+      "grad_norm": 2.185175895690918,
+      "learning_rate": 2.6705930138099107e-06,
+      "loss": 1.8299,
+      "step": 1436
+    },
+    {
+      "epoch": 1.5747945205479452,
+      "grad_norm": 2.2338390350341797,
+      "learning_rate": 2.668562144597888e-06,
+      "loss": 1.857,
+      "step": 1437
+    },
+    {
+      "epoch": 1.575890410958904,
+      "grad_norm": 2.4094536304473877,
+      "learning_rate": 2.6665312753858653e-06,
+      "loss": 1.7333,
+      "step": 1438
+    },
+    {
+      "epoch": 1.576986301369863,
+      "grad_norm": 2.4650957584381104,
+      "learning_rate": 2.6645004061738426e-06,
+      "loss": 1.8237,
+      "step": 1439
+    },
+    {
+      "epoch": 1.5780821917808219,
+      "grad_norm": 2.427415132522583,
+      "learning_rate": 2.66246953696182e-06,
+      "loss": 1.8206,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5780821917808219,
+      "eval_loss": 2.0222699642181396,
+      "eval_runtime": 37.0252,
+      "eval_samples_per_second": 27.009,
+      "eval_steps_per_second": 1.134,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5791780821917807,
+      "grad_norm": 2.3878679275512695,
+      "learning_rate": 2.660438667749797e-06,
+      "loss": 1.775,
+      "step": 1441
+    },
+    {
+      "epoch": 1.5802739726027397,
+      "grad_norm": 2.317819595336914,
+      "learning_rate": 2.6584077985377744e-06,
+      "loss": 1.7479,
+      "step": 1442
+    },
+    {
+      "epoch": 1.5813698630136988,
+      "grad_norm": 2.34951114654541,
+      "learning_rate": 2.6563769293257517e-06,
+      "loss": 1.8025,
+      "step": 1443
+    },
+    {
+      "epoch": 1.5824657534246576,
+      "grad_norm": 2.2849795818328857,
+      "learning_rate": 2.654346060113729e-06,
+      "loss": 1.8152,
+      "step": 1444
+    },
+    {
+      "epoch": 1.5835616438356164,
+      "grad_norm": 2.211489677429199,
+      "learning_rate": 2.6523151909017062e-06,
+      "loss": 1.7666,
+      "step": 1445
+    },
+    {
+      "epoch": 1.5846575342465754,
+      "grad_norm": 2.1876230239868164,
+      "learning_rate": 2.6502843216896835e-06,
+      "loss": 1.7933,
+      "step": 1446
+    },
+    {
+      "epoch": 1.5857534246575342,
+      "grad_norm": 2.367628335952759,
+      "learning_rate": 2.6482534524776603e-06,
+      "loss": 1.7756,
+      "step": 1447
+    },
+    {
+      "epoch": 1.586849315068493,
+      "grad_norm": 2.1198348999023438,
+      "learning_rate": 2.646222583265638e-06,
+      "loss": 1.8908,
+      "step": 1448
+    },
+    {
+      "epoch": 1.587945205479452,
+      "grad_norm": 2.192788600921631,
+      "learning_rate": 2.644191714053615e-06,
+      "loss": 1.8432,
+      "step": 1449
+    },
+    {
+      "epoch": 1.589041095890411,
+      "grad_norm": 2.4480655193328857,
+      "learning_rate": 2.6421608448415926e-06,
+      "loss": 1.8223,
+      "step": 1450
+    },
+    {
+      "epoch": 1.589041095890411,
+      "eval_loss": 2.019976854324341,
+      "eval_runtime": 37.0117,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1450
+    },
+    {
+      "epoch": 1.59013698630137,
+      "grad_norm": 2.573078155517578,
+      "learning_rate": 2.6401299756295694e-06,
+      "loss": 1.7881,
+      "step": 1451
+    },
+    {
+      "epoch": 1.5912328767123287,
+      "grad_norm": 2.2820355892181396,
+      "learning_rate": 2.638099106417547e-06,
+      "loss": 1.7703,
+      "step": 1452
+    },
+    {
+      "epoch": 1.5923287671232877,
+      "grad_norm": 2.526477575302124,
+      "learning_rate": 2.636068237205524e-06,
+      "loss": 1.8281,
+      "step": 1453
+    },
+    {
+      "epoch": 1.5934246575342466,
+      "grad_norm": 2.873647689819336,
+      "learning_rate": 2.6340373679935017e-06,
+      "loss": 1.7771,
+      "step": 1454
+    },
+    {
+      "epoch": 1.5945205479452054,
+      "grad_norm": 2.3782413005828857,
+      "learning_rate": 2.6320064987814785e-06,
+      "loss": 1.8864,
+      "step": 1455
+    },
+    {
+      "epoch": 1.5956164383561644,
+      "grad_norm": 2.3153343200683594,
+      "learning_rate": 2.6299756295694562e-06,
+      "loss": 1.734,
+      "step": 1456
+    },
+    {
+      "epoch": 1.5967123287671234,
+      "grad_norm": 2.631880283355713,
+      "learning_rate": 2.627944760357433e-06,
+      "loss": 1.8295,
+      "step": 1457
+    },
+    {
+      "epoch": 1.5978082191780822,
+      "grad_norm": 2.5827910900115967,
+      "learning_rate": 2.6259138911454104e-06,
+      "loss": 1.8738,
+      "step": 1458
+    },
+    {
+      "epoch": 1.598904109589041,
+      "grad_norm": 2.3519935607910156,
+      "learning_rate": 2.6238830219333877e-06,
+      "loss": 1.8832,
+      "step": 1459
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 2.5236215591430664,
+      "learning_rate": 2.621852152721365e-06,
+      "loss": 1.8595,
+      "step": 1460
+    },
+    {
+      "epoch": 1.6,
+      "eval_loss": 2.0170512199401855,
+      "eval_runtime": 37.0302,
+      "eval_samples_per_second": 27.005,
+      "eval_steps_per_second": 1.134,
+      "step": 1460
+    },
+    {
+      "epoch": 1.601095890410959,
+      "grad_norm": 2.725083112716675,
+      "learning_rate": 2.619821283509342e-06,
+      "loss": 1.7476,
+      "step": 1461
+    },
+    {
+      "epoch": 1.6021917808219177,
+      "grad_norm": 2.212644100189209,
+      "learning_rate": 2.6177904142973195e-06,
+      "loss": 1.8475,
+      "step": 1462
+    },
+    {
+      "epoch": 1.6032876712328767,
+      "grad_norm": 2.346808910369873,
+      "learning_rate": 2.6157595450852968e-06,
+      "loss": 1.7762,
+      "step": 1463
+    },
+    {
+      "epoch": 1.6043835616438358,
+      "grad_norm": 2.702331304550171,
+      "learning_rate": 2.613728675873274e-06,
+      "loss": 1.8296,
+      "step": 1464
+    },
+    {
+      "epoch": 1.6054794520547944,
+      "grad_norm": 2.4175899028778076,
+      "learning_rate": 2.6116978066612513e-06,
+      "loss": 1.8476,
+      "step": 1465
+    },
+    {
+      "epoch": 1.6065753424657534,
+      "grad_norm": 2.325457811355591,
+      "learning_rate": 2.6096669374492286e-06,
+      "loss": 1.8706,
+      "step": 1466
+    },
+    {
+      "epoch": 1.6076712328767124,
+      "grad_norm": 2.383842945098877,
+      "learning_rate": 2.607636068237206e-06,
+      "loss": 1.8246,
+      "step": 1467
+    },
+    {
+      "epoch": 1.6087671232876712,
+      "grad_norm": 2.289515256881714,
+      "learning_rate": 2.605605199025183e-06,
+      "loss": 1.7728,
+      "step": 1468
+    },
+    {
+      "epoch": 1.60986301369863,
+      "grad_norm": 2.5116689205169678,
+      "learning_rate": 2.6035743298131604e-06,
+      "loss": 1.8359,
+      "step": 1469
+    },
+    {
+      "epoch": 1.610958904109589,
+      "grad_norm": 2.279695749282837,
+      "learning_rate": 2.6015434606011377e-06,
+      "loss": 1.8155,
+      "step": 1470
+    },
+    {
+      "epoch": 1.610958904109589,
+      "eval_loss": 2.0142691135406494,
+      "eval_runtime": 37.0119,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1470
+    },
+    {
+      "epoch": 1.612054794520548,
+      "grad_norm": 2.4397690296173096,
+      "learning_rate": 2.5995125913891145e-06,
+      "loss": 1.8424,
+      "step": 1471
+    },
+    {
+      "epoch": 1.6131506849315067,
+      "grad_norm": 2.433577299118042,
+      "learning_rate": 2.5974817221770922e-06,
+      "loss": 1.8196,
+      "step": 1472
+    },
+    {
+      "epoch": 1.6142465753424657,
+      "grad_norm": 2.365459442138672,
+      "learning_rate": 2.595450852965069e-06,
+      "loss": 1.8124,
+      "step": 1473
+    },
+    {
+      "epoch": 1.6153424657534248,
+      "grad_norm": 2.253133535385132,
+      "learning_rate": 2.593419983753047e-06,
+      "loss": 1.8443,
+      "step": 1474
+    },
+    {
+      "epoch": 1.6164383561643836,
+      "grad_norm": 2.227625608444214,
+      "learning_rate": 2.5913891145410236e-06,
+      "loss": 1.8524,
+      "step": 1475
+    },
+    {
+      "epoch": 1.6175342465753424,
+      "grad_norm": 2.403702974319458,
+      "learning_rate": 2.5893582453290013e-06,
+      "loss": 1.7983,
+      "step": 1476
+    },
+    {
+      "epoch": 1.6186301369863014,
+      "grad_norm": 2.24507999420166,
+      "learning_rate": 2.587327376116978e-06,
+      "loss": 1.8563,
+      "step": 1477
+    },
+    {
+      "epoch": 1.6197260273972602,
+      "grad_norm": 2.284700393676758,
+      "learning_rate": 2.585296506904956e-06,
+      "loss": 1.7958,
+      "step": 1478
+    },
+    {
+      "epoch": 1.620821917808219,
+      "grad_norm": 2.288215398788452,
+      "learning_rate": 2.5832656376929327e-06,
+      "loss": 1.8021,
+      "step": 1479
+    },
+    {
+      "epoch": 1.621917808219178,
+      "grad_norm": 2.2312939167022705,
+      "learning_rate": 2.5812347684809104e-06,
+      "loss": 1.7856,
+      "step": 1480
+    },
+    {
+      "epoch": 1.621917808219178,
+      "eval_loss": 2.01674222946167,
+      "eval_runtime": 36.9943,
+      "eval_samples_per_second": 27.031,
+      "eval_steps_per_second": 1.135,
+      "step": 1480
+    },
+    {
+      "epoch": 1.6230136986301371,
+      "grad_norm": 2.319363832473755,
+      "learning_rate": 2.5792038992688873e-06,
+      "loss": 1.7263,
+      "step": 1481
+    },
+    {
+      "epoch": 1.624109589041096,
+      "grad_norm": 2.4342033863067627,
+      "learning_rate": 2.577173030056865e-06,
+      "loss": 1.897,
+      "step": 1482
+    },
+    {
+      "epoch": 1.6252054794520547,
+      "grad_norm": 2.3052282333374023,
+      "learning_rate": 2.575142160844842e-06,
+      "loss": 1.7165,
+      "step": 1483
+    },
+    {
+      "epoch": 1.6263013698630138,
+      "grad_norm": 2.2476744651794434,
+      "learning_rate": 2.5731112916328187e-06,
+      "loss": 1.8347,
+      "step": 1484
+    },
+    {
+      "epoch": 1.6273972602739726,
+      "grad_norm": 2.3128387928009033,
+      "learning_rate": 2.5710804224207964e-06,
+      "loss": 1.8403,
+      "step": 1485
+    },
+    {
+      "epoch": 1.6284931506849314,
+      "grad_norm": 2.3483974933624268,
+      "learning_rate": 2.5690495532087733e-06,
+      "loss": 1.781,
+      "step": 1486
+    },
+    {
+      "epoch": 1.6295890410958904,
+      "grad_norm": 2.328298568725586,
+      "learning_rate": 2.567018683996751e-06,
+      "loss": 1.7356,
+      "step": 1487
+    },
+    {
+      "epoch": 1.6306849315068495,
+      "grad_norm": 2.2756922245025635,
+      "learning_rate": 2.564987814784728e-06,
+      "loss": 1.8449,
+      "step": 1488
+    },
+    {
+      "epoch": 1.6317808219178083,
+      "grad_norm": 2.363874912261963,
+      "learning_rate": 2.5629569455727055e-06,
+      "loss": 1.7767,
+      "step": 1489
+    },
+    {
+      "epoch": 1.632876712328767,
+      "grad_norm": 2.2872767448425293,
+      "learning_rate": 2.5609260763606824e-06,
+      "loss": 1.8052,
+      "step": 1490
+    },
+    {
+      "epoch": 1.632876712328767,
+      "eval_loss": 2.0134243965148926,
+      "eval_runtime": 37.029,
+      "eval_samples_per_second": 27.006,
+      "eval_steps_per_second": 1.134,
+      "step": 1490
+    },
+    {
+      "epoch": 1.633972602739726,
+      "grad_norm": 2.354851007461548,
+      "learning_rate": 2.55889520714866e-06,
+      "loss": 1.7953,
+      "step": 1491
+    },
+    {
+      "epoch": 1.635068493150685,
+      "grad_norm": 2.4440531730651855,
+      "learning_rate": 2.556864337936637e-06,
+      "loss": 1.9063,
+      "step": 1492
+    },
+    {
+      "epoch": 1.6361643835616437,
+      "grad_norm": 2.296835422515869,
+      "learning_rate": 2.5548334687246146e-06,
+      "loss": 1.7961,
+      "step": 1493
+    },
+    {
+      "epoch": 1.6372602739726028,
+      "grad_norm": 2.7421159744262695,
+      "learning_rate": 2.5528025995125915e-06,
+      "loss": 1.809,
+      "step": 1494
+    },
+    {
+      "epoch": 1.6383561643835618,
+      "grad_norm": 2.819125175476074,
+      "learning_rate": 2.550771730300569e-06,
+      "loss": 1.853,
+      "step": 1495
+    },
+    {
+      "epoch": 1.6394520547945204,
+      "grad_norm": 2.2536680698394775,
+      "learning_rate": 2.548740861088546e-06,
+      "loss": 1.8297,
+      "step": 1496
+    },
+    {
+      "epoch": 1.6405479452054794,
+      "grad_norm": 2.7130825519561768,
+      "learning_rate": 2.5467099918765233e-06,
+      "loss": 1.843,
+      "step": 1497
+    },
+    {
+      "epoch": 1.6416438356164385,
+      "grad_norm": 3.0625038146972656,
+      "learning_rate": 2.5446791226645006e-06,
+      "loss": 1.7462,
+      "step": 1498
+    },
+    {
+      "epoch": 1.6427397260273973,
+      "grad_norm": 2.3095271587371826,
+      "learning_rate": 2.542648253452478e-06,
+      "loss": 1.7363,
+      "step": 1499
+    },
+    {
+      "epoch": 1.643835616438356,
+      "grad_norm": 2.536547899246216,
+      "learning_rate": 2.540617384240455e-06,
+      "loss": 1.8422,
+      "step": 1500
+    },
+    {
+      "epoch": 1.643835616438356,
+      "eval_loss": 2.0126097202301025,
+      "eval_runtime": 37.0469,
+      "eval_samples_per_second": 26.993,
+      "eval_steps_per_second": 1.134,
+      "step": 1500
+    },
+    {
+      "epoch": 1.644931506849315,
+      "grad_norm": 2.7454309463500977,
+      "learning_rate": 2.5385865150284324e-06,
+      "loss": 1.8183,
+      "step": 1501
+    },
+    {
+      "epoch": 1.6460273972602741,
+      "grad_norm": 2.3443901538848877,
+      "learning_rate": 2.5365556458164097e-06,
+      "loss": 1.7827,
+      "step": 1502
+    },
+    {
+      "epoch": 1.6471232876712327,
+      "grad_norm": 2.2412657737731934,
+      "learning_rate": 2.534524776604387e-06,
+      "loss": 1.797,
+      "step": 1503
+    },
+    {
+      "epoch": 1.6482191780821918,
+      "grad_norm": 2.6193699836730957,
+      "learning_rate": 2.5324939073923642e-06,
+      "loss": 1.8103,
+      "step": 1504
+    },
+    {
+      "epoch": 1.6493150684931508,
+      "grad_norm": 2.363919258117676,
+      "learning_rate": 2.5304630381803415e-06,
+      "loss": 1.6987,
+      "step": 1505
+    },
+    {
+      "epoch": 1.6504109589041096,
+      "grad_norm": 2.303164482116699,
+      "learning_rate": 2.5284321689683188e-06,
+      "loss": 1.8005,
+      "step": 1506
+    },
+    {
+      "epoch": 1.6515068493150684,
+      "grad_norm": 2.444174289703369,
+      "learning_rate": 2.526401299756296e-06,
+      "loss": 1.8009,
+      "step": 1507
+    },
+    {
+      "epoch": 1.6526027397260274,
+      "grad_norm": 2.418454647064209,
+      "learning_rate": 2.524370430544273e-06,
+      "loss": 1.782,
+      "step": 1508
+    },
+    {
+      "epoch": 1.6536986301369863,
+      "grad_norm": 2.48840069770813,
+      "learning_rate": 2.5223395613322506e-06,
+      "loss": 1.8143,
+      "step": 1509
+    },
+    {
+      "epoch": 1.654794520547945,
+      "grad_norm": 2.2814347743988037,
+      "learning_rate": 2.5203086921202274e-06,
+      "loss": 1.7687,
+      "step": 1510
+    },
+    {
+      "epoch": 1.654794520547945,
+      "eval_loss": 2.0148253440856934,
+      "eval_runtime": 36.7833,
+      "eval_samples_per_second": 27.186,
+      "eval_steps_per_second": 1.142,
+      "step": 1510
+    },
+    {
+      "epoch": 1.655890410958904,
+      "grad_norm": 2.228933572769165,
+      "learning_rate": 2.518277822908205e-06,
+      "loss": 1.8416,
+      "step": 1511
+    },
+    {
+      "epoch": 1.6569863013698631,
+      "grad_norm": 2.3696742057800293,
+      "learning_rate": 2.516246953696182e-06,
+      "loss": 1.7868,
+      "step": 1512
+    },
+    {
+      "epoch": 1.658082191780822,
+      "grad_norm": 2.216360092163086,
+      "learning_rate": 2.5142160844841597e-06,
+      "loss": 1.8312,
+      "step": 1513
+    },
+    {
+      "epoch": 1.6591780821917808,
+      "grad_norm": 2.454345226287842,
+      "learning_rate": 2.5121852152721365e-06,
+      "loss": 1.8309,
+      "step": 1514
+    },
+    {
+      "epoch": 1.6602739726027398,
+      "grad_norm": 2.2106776237487793,
+      "learning_rate": 2.5101543460601142e-06,
+      "loss": 1.8155,
+      "step": 1515
+    },
+    {
+      "epoch": 1.6613698630136986,
+      "grad_norm": 2.289006233215332,
+      "learning_rate": 2.508123476848091e-06,
+      "loss": 1.8162,
+      "step": 1516
+    },
+    {
+      "epoch": 1.6624657534246574,
+      "grad_norm": 2.297246217727661,
+      "learning_rate": 2.506092607636069e-06,
+      "loss": 1.7559,
+      "step": 1517
+    },
+    {
+      "epoch": 1.6635616438356164,
+      "grad_norm": 2.3909482955932617,
+      "learning_rate": 2.5040617384240457e-06,
+      "loss": 1.8389,
+      "step": 1518
+    },
+    {
+      "epoch": 1.6646575342465755,
+      "grad_norm": 2.398171901702881,
+      "learning_rate": 2.5020308692120234e-06,
+      "loss": 1.7785,
+      "step": 1519
+    },
+    {
+      "epoch": 1.6657534246575343,
+      "grad_norm": 2.377707004547119,
+      "learning_rate": 2.5e-06,
+      "loss": 1.768,
+      "step": 1520
+    },
+    {
+      "epoch": 1.6657534246575343,
+      "eval_loss": 2.012815475463867,
+      "eval_runtime": 36.9603,
+      "eval_samples_per_second": 27.056,
+      "eval_steps_per_second": 1.136,
+      "step": 1520
+    },
+    {
+      "epoch": 1.666849315068493,
+      "grad_norm": 2.336160898208618,
+      "learning_rate": 2.4979691307879775e-06,
+      "loss": 1.8653,
+      "step": 1521
+    },
+    {
+      "epoch": 1.6679452054794521,
+      "grad_norm": 2.288424253463745,
+      "learning_rate": 2.4959382615759548e-06,
+      "loss": 1.7798,
+      "step": 1522
+    },
+    {
+      "epoch": 1.669041095890411,
+      "grad_norm": 2.2599472999572754,
+      "learning_rate": 2.493907392363932e-06,
+      "loss": 1.7736,
+      "step": 1523
+    },
+    {
+      "epoch": 1.6701369863013698,
+      "grad_norm": 2.242276430130005,
+      "learning_rate": 2.4918765231519093e-06,
+      "loss": 1.8753,
+      "step": 1524
+    },
+    {
+      "epoch": 1.6712328767123288,
+      "grad_norm": 2.1897733211517334,
+      "learning_rate": 2.4898456539398866e-06,
+      "loss": 1.8353,
+      "step": 1525
+    },
+    {
+      "epoch": 1.6723287671232878,
+      "grad_norm": 2.254009246826172,
+      "learning_rate": 2.487814784727864e-06,
+      "loss": 1.7825,
+      "step": 1526
+    },
+    {
+      "epoch": 1.6734246575342466,
+      "grad_norm": 2.3704216480255127,
+      "learning_rate": 2.4857839155158407e-06,
+      "loss": 1.8742,
+      "step": 1527
+    },
+    {
+      "epoch": 1.6745205479452054,
+      "grad_norm": 2.1590118408203125,
+      "learning_rate": 2.483753046303818e-06,
+      "loss": 1.7974,
+      "step": 1528
+    },
+    {
+      "epoch": 1.6756164383561645,
+      "grad_norm": 2.2389416694641113,
+      "learning_rate": 2.4817221770917953e-06,
+      "loss": 1.8007,
+      "step": 1529
+    },
+    {
+      "epoch": 1.6767123287671233,
+      "grad_norm": 2.324131488800049,
+      "learning_rate": 2.4796913078797725e-06,
+      "loss": 1.8079,
+      "step": 1530
+    },
+    {
+      "epoch": 1.6767123287671233,
+      "eval_loss": 2.013923168182373,
+      "eval_runtime": 37.0828,
+      "eval_samples_per_second": 26.967,
+      "eval_steps_per_second": 1.133,
+      "step": 1530
+    },
+    {
+      "epoch": 1.677808219178082,
+      "grad_norm": 2.406456708908081,
+      "learning_rate": 2.47766043866775e-06,
+      "loss": 1.8356,
+      "step": 1531
+    },
+    {
+      "epoch": 1.6789041095890411,
+      "grad_norm": 2.289919137954712,
+      "learning_rate": 2.475629569455727e-06,
+      "loss": 1.8239,
+      "step": 1532
+    },
+    {
+      "epoch": 1.6800000000000002,
+      "grad_norm": 2.2617733478546143,
+      "learning_rate": 2.4735987002437044e-06,
+      "loss": 1.8025,
+      "step": 1533
+    },
+    {
+      "epoch": 1.6810958904109587,
+      "grad_norm": 2.133652925491333,
+      "learning_rate": 2.4715678310316816e-06,
+      "loss": 1.7743,
+      "step": 1534
+    },
+    {
+      "epoch": 1.6821917808219178,
+      "grad_norm": 2.215322256088257,
+      "learning_rate": 2.469536961819659e-06,
+      "loss": 1.7979,
+      "step": 1535
+    },
+    {
+      "epoch": 1.6832876712328768,
+      "grad_norm": 2.336021900177002,
+      "learning_rate": 2.467506092607636e-06,
+      "loss": 1.8027,
+      "step": 1536
+    },
+    {
+      "epoch": 1.6843835616438356,
+      "grad_norm": 2.3567142486572266,
+      "learning_rate": 2.4654752233956135e-06,
+      "loss": 1.7877,
+      "step": 1537
+    },
+    {
+      "epoch": 1.6854794520547944,
+      "grad_norm": 2.2642927169799805,
+      "learning_rate": 2.4634443541835907e-06,
+      "loss": 1.7695,
+      "step": 1538
+    },
+    {
+      "epoch": 1.6865753424657535,
+      "grad_norm": 2.353303909301758,
+      "learning_rate": 2.461413484971568e-06,
+      "loss": 1.867,
+      "step": 1539
+    },
+    {
+      "epoch": 1.6876712328767123,
+      "grad_norm": 2.252074956893921,
+      "learning_rate": 2.4593826157595453e-06,
+      "loss": 1.8009,
+      "step": 1540
+    },
+    {
+      "epoch": 1.6876712328767123,
+      "eval_loss": 2.0113608837127686,
+      "eval_runtime": 37.0875,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 1540
+    },
+    {
+      "epoch": 1.688767123287671,
+      "grad_norm": 2.2303407192230225,
+      "learning_rate": 2.4573517465475226e-06,
+      "loss": 1.8881,
+      "step": 1541
+    },
+    {
+      "epoch": 1.6898630136986301,
+      "grad_norm": 2.3359081745147705,
+      "learning_rate": 2.4553208773355e-06,
+      "loss": 1.7388,
+      "step": 1542
+    },
+    {
+      "epoch": 1.6909589041095892,
+      "grad_norm": 2.3783583641052246,
+      "learning_rate": 2.453290008123477e-06,
+      "loss": 1.7993,
+      "step": 1543
+    },
+    {
+      "epoch": 1.692054794520548,
+      "grad_norm": 2.159264326095581,
+      "learning_rate": 2.4512591389114544e-06,
+      "loss": 1.81,
+      "step": 1544
+    },
+    {
+      "epoch": 1.6931506849315068,
+      "grad_norm": 2.2232203483581543,
+      "learning_rate": 2.4492282696994317e-06,
+      "loss": 1.8095,
+      "step": 1545
+    },
+    {
+      "epoch": 1.6942465753424658,
+      "grad_norm": 2.1755056381225586,
+      "learning_rate": 2.447197400487409e-06,
+      "loss": 1.8149,
+      "step": 1546
+    },
+    {
+      "epoch": 1.6953424657534246,
+      "grad_norm": 2.325871467590332,
+      "learning_rate": 2.4451665312753862e-06,
+      "loss": 1.8154,
+      "step": 1547
+    },
+    {
+      "epoch": 1.6964383561643834,
+      "grad_norm": 2.286848545074463,
+      "learning_rate": 2.4431356620633635e-06,
+      "loss": 1.8409,
+      "step": 1548
+    },
+    {
+      "epoch": 1.6975342465753425,
+      "grad_norm": 2.359510898590088,
+      "learning_rate": 2.4411047928513408e-06,
+      "loss": 1.843,
+      "step": 1549
+    },
+    {
+      "epoch": 1.6986301369863015,
+      "grad_norm": 2.474428176879883,
+      "learning_rate": 2.439073923639318e-06,
+      "loss": 1.8089,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6986301369863015,
+      "eval_loss": 2.0077731609344482,
+      "eval_runtime": 37.1092,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6997260273972603,
+      "grad_norm": 2.3662912845611572,
+      "learning_rate": 2.4370430544272953e-06,
+      "loss": 1.8176,
+      "step": 1551
+    },
+    {
+      "epoch": 1.7008219178082191,
+      "grad_norm": 2.448948860168457,
+      "learning_rate": 2.435012185215272e-06,
+      "loss": 1.7776,
+      "step": 1552
+    },
+    {
+      "epoch": 1.7019178082191782,
+      "grad_norm": 2.333739757537842,
+      "learning_rate": 2.4329813160032495e-06,
+      "loss": 1.7552,
+      "step": 1553
+    },
+    {
+      "epoch": 1.703013698630137,
+      "grad_norm": 2.170365571975708,
+      "learning_rate": 2.4309504467912267e-06,
+      "loss": 1.7759,
+      "step": 1554
+    },
+    {
+      "epoch": 1.7041095890410958,
+      "grad_norm": 2.2343010902404785,
+      "learning_rate": 2.428919577579204e-06,
+      "loss": 1.8073,
+      "step": 1555
+    },
+    {
+      "epoch": 1.7052054794520548,
+      "grad_norm": 2.2245287895202637,
+      "learning_rate": 2.4268887083671813e-06,
+      "loss": 1.7826,
+      "step": 1556
+    },
+    {
+      "epoch": 1.7063013698630138,
+      "grad_norm": 2.339942216873169,
+      "learning_rate": 2.4248578391551586e-06,
+      "loss": 1.8766,
+      "step": 1557
+    },
+    {
+      "epoch": 1.7073972602739727,
+      "grad_norm": 2.1585116386413574,
+      "learning_rate": 2.422826969943136e-06,
+      "loss": 1.7605,
+      "step": 1558
+    },
+    {
+      "epoch": 1.7084931506849315,
+      "grad_norm": 2.2108142375946045,
+      "learning_rate": 2.420796100731113e-06,
+      "loss": 1.7855,
+      "step": 1559
+    },
+    {
+      "epoch": 1.7095890410958905,
+      "grad_norm": 2.192457675933838,
+      "learning_rate": 2.4187652315190904e-06,
+      "loss": 1.7489,
+      "step": 1560
+    },
+    {
+      "epoch": 1.7095890410958905,
+      "eval_loss": 2.0085604190826416,
+      "eval_runtime": 37.1105,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1560
+    },
+    {
+      "epoch": 1.7106849315068493,
+      "grad_norm": 2.291106939315796,
+      "learning_rate": 2.4167343623070677e-06,
+      "loss": 1.7981,
+      "step": 1561
+    },
+    {
+      "epoch": 1.7117808219178081,
+      "grad_norm": 2.509237766265869,
+      "learning_rate": 2.414703493095045e-06,
+      "loss": 1.8146,
+      "step": 1562
+    },
+    {
+      "epoch": 1.7128767123287671,
+      "grad_norm": 2.497162342071533,
+      "learning_rate": 2.4126726238830222e-06,
+      "loss": 1.807,
+      "step": 1563
+    },
+    {
+      "epoch": 1.7139726027397262,
+      "grad_norm": 2.256833076477051,
+      "learning_rate": 2.410641754670999e-06,
+      "loss": 1.8304,
+      "step": 1564
+    },
+    {
+      "epoch": 1.7150684931506848,
+      "grad_norm": 2.2523772716522217,
+      "learning_rate": 2.4086108854589763e-06,
+      "loss": 1.7794,
+      "step": 1565
+    },
+    {
+      "epoch": 1.7161643835616438,
+      "grad_norm": 2.289182424545288,
+      "learning_rate": 2.4065800162469536e-06,
+      "loss": 1.8425,
+      "step": 1566
+    },
+    {
+      "epoch": 1.7172602739726028,
+      "grad_norm": 2.734487771987915,
+      "learning_rate": 2.404549147034931e-06,
+      "loss": 1.7791,
+      "step": 1567
+    },
+    {
+      "epoch": 1.7183561643835616,
+      "grad_norm": 2.3307888507843018,
+      "learning_rate": 2.402518277822908e-06,
+      "loss": 1.8064,
+      "step": 1568
+    },
+    {
+      "epoch": 1.7194520547945205,
+      "grad_norm": 2.5016298294067383,
+      "learning_rate": 2.4004874086108854e-06,
+      "loss": 1.8544,
+      "step": 1569
+    },
+    {
+      "epoch": 1.7205479452054795,
+      "grad_norm": 2.142246961593628,
+      "learning_rate": 2.3984565393988627e-06,
+      "loss": 1.7653,
+      "step": 1570
+    },
+    {
+      "epoch": 1.7205479452054795,
+      "eval_loss": 2.0068864822387695,
+      "eval_runtime": 37.1078,
+      "eval_samples_per_second": 26.949,
+      "eval_steps_per_second": 1.132,
+      "step": 1570
+    },
+    {
+      "epoch": 1.7216438356164385,
+      "grad_norm": 2.5692505836486816,
+      "learning_rate": 2.39642567018684e-06,
+      "loss": 1.8444,
+      "step": 1571
+    },
+    {
+      "epoch": 1.7227397260273971,
+      "grad_norm": 2.2596213817596436,
+      "learning_rate": 2.3943948009748173e-06,
+      "loss": 1.7694,
+      "step": 1572
+    },
+    {
+      "epoch": 1.7238356164383561,
+      "grad_norm": 2.230984926223755,
+      "learning_rate": 2.3923639317627945e-06,
+      "loss": 1.786,
+      "step": 1573
+    },
+    {
+      "epoch": 1.7249315068493152,
+      "grad_norm": 2.366178274154663,
+      "learning_rate": 2.390333062550772e-06,
+      "loss": 1.8138,
+      "step": 1574
+    },
+    {
+      "epoch": 1.726027397260274,
+      "grad_norm": 2.23783016204834,
+      "learning_rate": 2.388302193338749e-06,
+      "loss": 1.7615,
+      "step": 1575
+    },
+    {
+      "epoch": 1.7271232876712328,
+      "grad_norm": 2.3766677379608154,
+      "learning_rate": 2.3862713241267264e-06,
+      "loss": 1.6811,
+      "step": 1576
+    },
+    {
+      "epoch": 1.7282191780821918,
+      "grad_norm": 2.3434040546417236,
+      "learning_rate": 2.3842404549147037e-06,
+      "loss": 1.8806,
+      "step": 1577
+    },
+    {
+      "epoch": 1.7293150684931506,
+      "grad_norm": 2.2444276809692383,
+      "learning_rate": 2.382209585702681e-06,
+      "loss": 1.7968,
+      "step": 1578
+    },
+    {
+      "epoch": 1.7304109589041095,
+      "grad_norm": 2.2524919509887695,
+      "learning_rate": 2.380178716490658e-06,
+      "loss": 1.7602,
+      "step": 1579
+    },
+    {
+      "epoch": 1.7315068493150685,
+      "grad_norm": 2.3310587406158447,
+      "learning_rate": 2.3781478472786355e-06,
+      "loss": 1.8414,
+      "step": 1580
+    },
+    {
+      "epoch": 1.7315068493150685,
+      "eval_loss": 1.9981482028961182,
+      "eval_runtime": 37.0827,
+      "eval_samples_per_second": 26.967,
+      "eval_steps_per_second": 1.133,
+      "step": 1580
+    },
+    {
+      "epoch": 1.7326027397260275,
+      "grad_norm": 2.4081802368164062,
+      "learning_rate": 2.3761169780666128e-06,
+      "loss": 1.8002,
+      "step": 1581
+    },
+    {
+      "epoch": 1.7336986301369863,
+      "grad_norm": 2.293506622314453,
+      "learning_rate": 2.37408610885459e-06,
+      "loss": 1.8365,
+      "step": 1582
+    },
+    {
+      "epoch": 1.7347945205479451,
+      "grad_norm": 2.2015790939331055,
+      "learning_rate": 2.3720552396425673e-06,
+      "loss": 1.8571,
+      "step": 1583
+    },
+    {
+      "epoch": 1.7358904109589042,
+      "grad_norm": 2.308809518814087,
+      "learning_rate": 2.3700243704305446e-06,
+      "loss": 1.7821,
+      "step": 1584
+    },
+    {
+      "epoch": 1.736986301369863,
+      "grad_norm": 2.376842975616455,
+      "learning_rate": 2.367993501218522e-06,
+      "loss": 1.8701,
+      "step": 1585
+    },
+    {
+      "epoch": 1.7380821917808218,
+      "grad_norm": 2.4102022647857666,
+      "learning_rate": 2.365962632006499e-06,
+      "loss": 1.8055,
+      "step": 1586
+    },
+    {
+      "epoch": 1.7391780821917808,
+      "grad_norm": 2.2734079360961914,
+      "learning_rate": 2.3639317627944764e-06,
+      "loss": 1.8117,
+      "step": 1587
+    },
+    {
+      "epoch": 1.7402739726027399,
+      "grad_norm": 2.3426599502563477,
+      "learning_rate": 2.3619008935824537e-06,
+      "loss": 1.8337,
+      "step": 1588
+    },
+    {
+      "epoch": 1.7413698630136987,
+      "grad_norm": 2.340383768081665,
+      "learning_rate": 2.3598700243704305e-06,
+      "loss": 1.8247,
+      "step": 1589
+    },
+    {
+      "epoch": 1.7424657534246575,
+      "grad_norm": 2.2561380863189697,
+      "learning_rate": 2.357839155158408e-06,
+      "loss": 1.8249,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7424657534246575,
+      "eval_loss": 1.9979513883590698,
+      "eval_runtime": 37.0948,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7435616438356165,
+      "grad_norm": 2.2459654808044434,
+      "learning_rate": 2.355808285946385e-06,
+      "loss": 1.8273,
+      "step": 1591
+    },
+    {
+      "epoch": 1.7446575342465753,
+      "grad_norm": 2.2017974853515625,
+      "learning_rate": 2.3537774167343624e-06,
+      "loss": 1.828,
+      "step": 1592
+    },
+    {
+      "epoch": 1.7457534246575341,
+      "grad_norm": 2.286304235458374,
+      "learning_rate": 2.3517465475223396e-06,
+      "loss": 1.8497,
+      "step": 1593
+    },
+    {
+      "epoch": 1.7468493150684932,
+      "grad_norm": 2.296494722366333,
+      "learning_rate": 2.349715678310317e-06,
+      "loss": 1.8155,
+      "step": 1594
+    },
+    {
+      "epoch": 1.7479452054794522,
+      "grad_norm": 2.2989203929901123,
+      "learning_rate": 2.347684809098294e-06,
+      "loss": 1.8214,
+      "step": 1595
+    },
+    {
+      "epoch": 1.749041095890411,
+      "grad_norm": 2.3034965991973877,
+      "learning_rate": 2.3456539398862715e-06,
+      "loss": 1.7792,
+      "step": 1596
+    },
+    {
+      "epoch": 1.7501369863013698,
+      "grad_norm": 2.3823719024658203,
+      "learning_rate": 2.3436230706742487e-06,
+      "loss": 1.8326,
+      "step": 1597
+    },
+    {
+      "epoch": 1.7512328767123289,
+      "grad_norm": 2.231095552444458,
+      "learning_rate": 2.341592201462226e-06,
+      "loss": 1.7713,
+      "step": 1598
+    },
+    {
+      "epoch": 1.7523287671232877,
+      "grad_norm": 2.287524938583374,
+      "learning_rate": 2.3395613322502033e-06,
+      "loss": 1.8143,
+      "step": 1599
+    },
+    {
+      "epoch": 1.7534246575342465,
+      "grad_norm": 2.267552614212036,
+      "learning_rate": 2.3375304630381806e-06,
+      "loss": 1.7883,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7534246575342465,
+      "eval_loss": 1.998821496963501,
+      "eval_runtime": 37.1222,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7545205479452055,
+      "grad_norm": 2.2986714839935303,
+      "learning_rate": 2.335499593826158e-06,
+      "loss": 1.7615,
+      "step": 1601
+    },
+    {
+      "epoch": 1.7556164383561645,
+      "grad_norm": 2.2913246154785156,
+      "learning_rate": 2.333468724614135e-06,
+      "loss": 1.7584,
+      "step": 1602
+    },
+    {
+      "epoch": 1.7567123287671231,
+      "grad_norm": 2.2486486434936523,
+      "learning_rate": 2.3314378554021124e-06,
+      "loss": 1.8089,
+      "step": 1603
+    },
+    {
+      "epoch": 1.7578082191780822,
+      "grad_norm": 2.319581985473633,
+      "learning_rate": 2.3294069861900897e-06,
+      "loss": 1.8364,
+      "step": 1604
+    },
+    {
+      "epoch": 1.7589041095890412,
+      "grad_norm": 2.2775609493255615,
+      "learning_rate": 2.327376116978067e-06,
+      "loss": 1.7661,
+      "step": 1605
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 2.3445472717285156,
+      "learning_rate": 2.3253452477660442e-06,
+      "loss": 1.7583,
+      "step": 1606
+    },
+    {
+      "epoch": 1.7610958904109588,
+      "grad_norm": 2.420865058898926,
+      "learning_rate": 2.3233143785540215e-06,
+      "loss": 1.8411,
+      "step": 1607
+    },
+    {
+      "epoch": 1.7621917808219179,
+      "grad_norm": 2.294821262359619,
+      "learning_rate": 2.3212835093419988e-06,
+      "loss": 1.7812,
+      "step": 1608
+    },
+    {
+      "epoch": 1.7632876712328767,
+      "grad_norm": 2.332784414291382,
+      "learning_rate": 2.319252640129976e-06,
+      "loss": 1.8252,
+      "step": 1609
+    },
+    {
+      "epoch": 1.7643835616438355,
+      "grad_norm": 2.433619499206543,
+      "learning_rate": 2.3172217709179533e-06,
+      "loss": 1.7709,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7643835616438355,
+      "eval_loss": 2.0010287761688232,
+      "eval_runtime": 36.7858,
+      "eval_samples_per_second": 27.184,
+      "eval_steps_per_second": 1.142,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7654794520547945,
+      "grad_norm": 2.532618522644043,
+      "learning_rate": 2.3151909017059306e-06,
+      "loss": 1.8677,
+      "step": 1611
+    },
+    {
+      "epoch": 1.7665753424657535,
+      "grad_norm": 2.354578733444214,
+      "learning_rate": 2.313160032493908e-06,
+      "loss": 1.8412,
+      "step": 1612
+    },
+    {
+      "epoch": 1.7676712328767124,
+      "grad_norm": 2.382774591445923,
+      "learning_rate": 2.311129163281885e-06,
+      "loss": 1.7543,
+      "step": 1613
+    },
+    {
+      "epoch": 1.7687671232876712,
+      "grad_norm": 2.3443398475646973,
+      "learning_rate": 2.309098294069862e-06,
+      "loss": 1.8267,
+      "step": 1614
+    },
+    {
+      "epoch": 1.7698630136986302,
+      "grad_norm": 2.3804502487182617,
+      "learning_rate": 2.3070674248578393e-06,
+      "loss": 1.7582,
+      "step": 1615
+    },
+    {
+      "epoch": 1.770958904109589,
+      "grad_norm": 2.3163580894470215,
+      "learning_rate": 2.3050365556458166e-06,
+      "loss": 1.7539,
+      "step": 1616
+    },
+    {
+      "epoch": 1.7720547945205478,
+      "grad_norm": 2.605738639831543,
+      "learning_rate": 2.303005686433794e-06,
+      "loss": 1.8124,
+      "step": 1617
+    },
+    {
+      "epoch": 1.7731506849315068,
+      "grad_norm": 2.3859915733337402,
+      "learning_rate": 2.300974817221771e-06,
+      "loss": 1.8271,
+      "step": 1618
+    },
+    {
+      "epoch": 1.7742465753424659,
+      "grad_norm": 2.3715479373931885,
+      "learning_rate": 2.2989439480097484e-06,
+      "loss": 1.7823,
+      "step": 1619
+    },
+    {
+      "epoch": 1.7753424657534247,
+      "grad_norm": 2.441100597381592,
+      "learning_rate": 2.2969130787977257e-06,
+      "loss": 1.7461,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7753424657534247,
+      "eval_loss": 2.002152442932129,
+      "eval_runtime": 37.0246,
+      "eval_samples_per_second": 27.009,
+      "eval_steps_per_second": 1.134,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7764383561643835,
+      "grad_norm": 2.4154021739959717,
+      "learning_rate": 2.294882209585703e-06,
+      "loss": 1.8012,
+      "step": 1621
+    },
+    {
+      "epoch": 1.7775342465753425,
+      "grad_norm": 2.333400249481201,
+      "learning_rate": 2.2928513403736802e-06,
+      "loss": 1.8222,
+      "step": 1622
+    },
+    {
+      "epoch": 1.7786301369863013,
+      "grad_norm": 2.3183937072753906,
+      "learning_rate": 2.2908204711616575e-06,
+      "loss": 1.7561,
+      "step": 1623
+    },
+    {
+      "epoch": 1.7797260273972602,
+      "grad_norm": 2.302823066711426,
+      "learning_rate": 2.2887896019496348e-06,
+      "loss": 1.7975,
+      "step": 1624
+    },
+    {
+      "epoch": 1.7808219178082192,
+      "grad_norm": 2.3702917098999023,
+      "learning_rate": 2.286758732737612e-06,
+      "loss": 1.8478,
+      "step": 1625
+    },
+    {
+      "epoch": 1.7819178082191782,
+      "grad_norm": 2.4997549057006836,
+      "learning_rate": 2.284727863525589e-06,
+      "loss": 1.7855,
+      "step": 1626
+    },
+    {
+      "epoch": 1.783013698630137,
+      "grad_norm": 2.296191930770874,
+      "learning_rate": 2.282696994313566e-06,
+      "loss": 1.7922,
+      "step": 1627
+    },
+    {
+      "epoch": 1.7841095890410958,
+      "grad_norm": 2.2567873001098633,
+      "learning_rate": 2.2806661251015434e-06,
+      "loss": 1.8455,
+      "step": 1628
+    },
+    {
+      "epoch": 1.7852054794520549,
+      "grad_norm": 2.393336534500122,
+      "learning_rate": 2.2786352558895207e-06,
+      "loss": 1.8141,
+      "step": 1629
+    },
+    {
+      "epoch": 1.7863013698630137,
+      "grad_norm": 2.610466480255127,
+      "learning_rate": 2.276604386677498e-06,
+      "loss": 1.7875,
+      "step": 1630
+    },
+    {
+      "epoch": 1.7863013698630137,
+      "eval_loss": 1.9985637664794922,
+      "eval_runtime": 37.1001,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 1630
+    },
+    {
+      "epoch": 1.7873972602739725,
+      "grad_norm": 2.2163283824920654,
+      "learning_rate": 2.2745735174654753e-06,
+      "loss": 1.7479,
+      "step": 1631
+    },
+    {
+      "epoch": 1.7884931506849315,
+      "grad_norm": 2.2641165256500244,
+      "learning_rate": 2.2725426482534526e-06,
+      "loss": 1.7749,
+      "step": 1632
+    },
+    {
+      "epoch": 1.7895890410958906,
+      "grad_norm": 2.2786946296691895,
+      "learning_rate": 2.27051177904143e-06,
+      "loss": 1.8403,
+      "step": 1633
+    },
+    {
+      "epoch": 1.7906849315068492,
+      "grad_norm": 2.660356044769287,
+      "learning_rate": 2.268480909829407e-06,
+      "loss": 1.7994,
+      "step": 1634
+    },
+    {
+      "epoch": 1.7917808219178082,
+      "grad_norm": 2.322181224822998,
+      "learning_rate": 2.2664500406173844e-06,
+      "loss": 1.7917,
+      "step": 1635
+    },
+    {
+      "epoch": 1.7928767123287672,
+      "grad_norm": 2.3020682334899902,
+      "learning_rate": 2.2644191714053617e-06,
+      "loss": 1.7983,
+      "step": 1636
+    },
+    {
+      "epoch": 1.793972602739726,
+      "grad_norm": 2.5280368328094482,
+      "learning_rate": 2.262388302193339e-06,
+      "loss": 1.8378,
+      "step": 1637
+    },
+    {
+      "epoch": 1.7950684931506848,
+      "grad_norm": 2.3975653648376465,
+      "learning_rate": 2.260357432981316e-06,
+      "loss": 1.8081,
+      "step": 1638
+    },
+    {
+      "epoch": 1.7961643835616439,
+      "grad_norm": 2.419806957244873,
+      "learning_rate": 2.2583265637692935e-06,
+      "loss": 1.793,
+      "step": 1639
+    },
+    {
+      "epoch": 1.7972602739726027,
+      "grad_norm": 2.4601471424102783,
+      "learning_rate": 2.2562956945572708e-06,
+      "loss": 1.8116,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7972602739726027,
+      "eval_loss": 1.9951634407043457,
+      "eval_runtime": 37.1095,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7983561643835615,
+      "grad_norm": 2.17830753326416,
+      "learning_rate": 2.254264825345248e-06,
+      "loss": 1.7845,
+      "step": 1641
+    },
+    {
+      "epoch": 1.7994520547945205,
+      "grad_norm": 2.3332247734069824,
+      "learning_rate": 2.2522339561332253e-06,
+      "loss": 1.7945,
+      "step": 1642
+    },
+    {
+      "epoch": 1.8005479452054796,
+      "grad_norm": 2.3841960430145264,
+      "learning_rate": 2.2502030869212026e-06,
+      "loss": 1.7994,
+      "step": 1643
+    },
+    {
+      "epoch": 1.8016438356164384,
+      "grad_norm": 2.3669073581695557,
+      "learning_rate": 2.24817221770918e-06,
+      "loss": 1.8053,
+      "step": 1644
+    },
+    {
+      "epoch": 1.8027397260273972,
+      "grad_norm": 2.406858205795288,
+      "learning_rate": 2.246141348497157e-06,
+      "loss": 1.8647,
+      "step": 1645
+    },
+    {
+      "epoch": 1.8038356164383562,
+      "grad_norm": 2.2854161262512207,
+      "learning_rate": 2.2441104792851344e-06,
+      "loss": 1.8312,
+      "step": 1646
+    },
+    {
+      "epoch": 1.804931506849315,
+      "grad_norm": 2.252976655960083,
+      "learning_rate": 2.2420796100731117e-06,
+      "loss": 1.7834,
+      "step": 1647
+    },
+    {
+      "epoch": 1.8060273972602738,
+      "grad_norm": 2.4131624698638916,
+      "learning_rate": 2.240048740861089e-06,
+      "loss": 1.8204,
+      "step": 1648
+    },
+    {
+      "epoch": 1.8071232876712329,
+      "grad_norm": 2.4727940559387207,
+      "learning_rate": 2.2380178716490662e-06,
+      "loss": 1.8326,
+      "step": 1649
+    },
+    {
+      "epoch": 1.808219178082192,
+      "grad_norm": 2.4432709217071533,
+      "learning_rate": 2.2359870024370435e-06,
+      "loss": 1.6911,
+      "step": 1650
+    },
+    {
+      "epoch": 1.808219178082192,
+      "eval_loss": 1.996634840965271,
+      "eval_runtime": 37.072,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 1650
+    },
+    {
+      "epoch": 1.8093150684931507,
+      "grad_norm": 2.3810362815856934,
+      "learning_rate": 2.2339561332250204e-06,
+      "loss": 1.8256,
+      "step": 1651
+    },
+    {
+      "epoch": 1.8104109589041095,
+      "grad_norm": 2.4489142894744873,
+      "learning_rate": 2.2319252640129976e-06,
+      "loss": 1.8368,
+      "step": 1652
+    },
+    {
+      "epoch": 1.8115068493150686,
+      "grad_norm": 2.458686113357544,
+      "learning_rate": 2.229894394800975e-06,
+      "loss": 1.8176,
+      "step": 1653
+    },
+    {
+      "epoch": 1.8126027397260274,
+      "grad_norm": 2.3977842330932617,
+      "learning_rate": 2.227863525588952e-06,
+      "loss": 1.7516,
+      "step": 1654
+    },
+    {
+      "epoch": 1.8136986301369862,
+      "grad_norm": 2.3777663707733154,
+      "learning_rate": 2.2258326563769295e-06,
+      "loss": 1.732,
+      "step": 1655
+    },
+    {
+      "epoch": 1.8147945205479452,
+      "grad_norm": 2.2874817848205566,
+      "learning_rate": 2.2238017871649067e-06,
+      "loss": 1.749,
+      "step": 1656
+    },
+    {
+      "epoch": 1.8158904109589042,
+      "grad_norm": 2.288114547729492,
+      "learning_rate": 2.221770917952884e-06,
+      "loss": 1.7357,
+      "step": 1657
+    },
+    {
+      "epoch": 1.816986301369863,
+      "grad_norm": 2.3532187938690186,
+      "learning_rate": 2.2197400487408613e-06,
+      "loss": 1.7953,
+      "step": 1658
+    },
+    {
+      "epoch": 1.8180821917808219,
+      "grad_norm": 2.2829396724700928,
+      "learning_rate": 2.2177091795288386e-06,
+      "loss": 1.8358,
+      "step": 1659
+    },
+    {
+      "epoch": 1.819178082191781,
+      "grad_norm": 2.3282570838928223,
+      "learning_rate": 2.215678310316816e-06,
+      "loss": 1.7448,
+      "step": 1660
+    },
+    {
+      "epoch": 1.819178082191781,
+      "eval_loss": 1.999374270439148,
+      "eval_runtime": 37.0739,
+      "eval_samples_per_second": 26.973,
+      "eval_steps_per_second": 1.133,
+      "step": 1660
+    },
+    {
+      "epoch": 1.8202739726027397,
+      "grad_norm": 2.474106788635254,
+      "learning_rate": 2.213647441104793e-06,
+      "loss": 1.7979,
+      "step": 1661
+    },
+    {
+      "epoch": 1.8213698630136985,
+      "grad_norm": 2.159498453140259,
+      "learning_rate": 2.2116165718927704e-06,
+      "loss": 1.8075,
+      "step": 1662
+    },
+    {
+      "epoch": 1.8224657534246576,
+      "grad_norm": 2.2503573894500732,
+      "learning_rate": 2.2095857026807473e-06,
+      "loss": 1.7742,
+      "step": 1663
+    },
+    {
+      "epoch": 1.8235616438356166,
+      "grad_norm": 2.4465060234069824,
+      "learning_rate": 2.2075548334687245e-06,
+      "loss": 1.8067,
+      "step": 1664
+    },
+    {
+      "epoch": 1.8246575342465754,
+      "grad_norm": 2.462158679962158,
+      "learning_rate": 2.205523964256702e-06,
+      "loss": 1.7456,
+      "step": 1665
+    },
+    {
+      "epoch": 1.8257534246575342,
+      "grad_norm": 2.354051113128662,
+      "learning_rate": 2.203493095044679e-06,
+      "loss": 1.7688,
+      "step": 1666
+    },
+    {
+      "epoch": 1.8268493150684932,
+      "grad_norm": 2.3165128231048584,
+      "learning_rate": 2.2014622258326564e-06,
+      "loss": 1.7986,
+      "step": 1667
+    },
+    {
+      "epoch": 1.827945205479452,
+      "grad_norm": 2.3131113052368164,
+      "learning_rate": 2.1994313566206336e-06,
+      "loss": 1.8348,
+      "step": 1668
+    },
+    {
+      "epoch": 1.8290410958904109,
+      "grad_norm": 2.303086996078491,
+      "learning_rate": 2.197400487408611e-06,
+      "loss": 1.8413,
+      "step": 1669
+    },
+    {
+      "epoch": 1.83013698630137,
+      "grad_norm": 2.392995834350586,
+      "learning_rate": 2.195369618196588e-06,
+      "loss": 1.7663,
+      "step": 1670
+    },
+    {
+      "epoch": 1.83013698630137,
+      "eval_loss": 1.9936721324920654,
+      "eval_runtime": 37.0717,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 1670
+    },
+    {
+      "epoch": 1.831232876712329,
+      "grad_norm": 2.323596477508545,
+      "learning_rate": 2.1933387489845655e-06,
+      "loss": 1.7463,
+      "step": 1671
+    },
+    {
+      "epoch": 1.8323287671232875,
+      "grad_norm": 2.3553171157836914,
+      "learning_rate": 2.1913078797725427e-06,
+      "loss": 1.7844,
+      "step": 1672
+    },
+    {
+      "epoch": 1.8334246575342465,
+      "grad_norm": 2.3724591732025146,
+      "learning_rate": 2.18927701056052e-06,
+      "loss": 1.7955,
+      "step": 1673
+    },
+    {
+      "epoch": 1.8345205479452056,
+      "grad_norm": 2.4300284385681152,
+      "learning_rate": 2.1872461413484973e-06,
+      "loss": 1.8697,
+      "step": 1674
+    },
+    {
+      "epoch": 1.8356164383561644,
+      "grad_norm": 2.4300284385681152,
+      "learning_rate": 2.1872461413484973e-06,
+      "loss": 1.8065,
+      "step": 1675
+    },
+    {
+      "epoch": 1.8367123287671232,
+      "grad_norm": 2.523627519607544,
+      "learning_rate": 2.1852152721364746e-06,
+      "loss": 1.801,
+      "step": 1676
+    },
+    {
+      "epoch": 1.8378082191780822,
+      "grad_norm": 2.2884891033172607,
+      "learning_rate": 2.183184402924452e-06,
+      "loss": 1.7512,
+      "step": 1677
+    },
+    {
+      "epoch": 1.838904109589041,
+      "grad_norm": 2.3114986419677734,
+      "learning_rate": 2.181153533712429e-06,
+      "loss": 1.8245,
+      "step": 1678
+    },
+    {
+      "epoch": 1.8399999999999999,
+      "grad_norm": 2.326855421066284,
+      "learning_rate": 2.1791226645004064e-06,
+      "loss": 1.7921,
+      "step": 1679
+    },
+    {
+      "epoch": 1.841095890410959,
+      "grad_norm": 2.3448753356933594,
+      "learning_rate": 2.1770917952883837e-06,
+      "loss": 1.7709,
+      "step": 1680
+    },
+    {
+      "epoch": 1.841095890410959,
+      "eval_loss": 1.996195673942566,
+      "eval_runtime": 37.0481,
+      "eval_samples_per_second": 26.992,
+      "eval_steps_per_second": 1.134,
+      "step": 1680
+    },
+    {
+      "epoch": 1.842191780821918,
+      "grad_norm": 2.223470687866211,
+      "learning_rate": 2.175060926076361e-06,
+      "loss": 1.8468,
+      "step": 1681
+    },
+    {
+      "epoch": 1.8432876712328767,
+      "grad_norm": 2.29726505279541,
+      "learning_rate": 2.1730300568643382e-06,
+      "loss": 1.8753,
+      "step": 1682
+    },
+    {
+      "epoch": 1.8443835616438355,
+      "grad_norm": 2.3869571685791016,
+      "learning_rate": 2.1709991876523155e-06,
+      "loss": 1.7835,
+      "step": 1683
+    },
+    {
+      "epoch": 1.8454794520547946,
+      "grad_norm": 2.192608594894409,
+      "learning_rate": 2.1689683184402928e-06,
+      "loss": 1.7556,
+      "step": 1684
+    },
+    {
+      "epoch": 1.8465753424657534,
+      "grad_norm": 2.355975866317749,
+      "learning_rate": 2.16693744922827e-06,
+      "loss": 1.8323,
+      "step": 1685
+    },
+    {
+      "epoch": 1.8476712328767122,
+      "grad_norm": 2.243938446044922,
+      "learning_rate": 2.1649065800162473e-06,
+      "loss": 1.7942,
+      "step": 1686
+    },
+    {
+      "epoch": 1.8487671232876712,
+      "grad_norm": 2.337069511413574,
+      "learning_rate": 2.1628757108042246e-06,
+      "loss": 1.7709,
+      "step": 1687
+    },
+    {
+      "epoch": 1.8498630136986303,
+      "grad_norm": 2.257504463195801,
+      "learning_rate": 2.160844841592202e-06,
+      "loss": 1.7496,
+      "step": 1688
+    },
+    {
+      "epoch": 1.850958904109589,
+      "grad_norm": 2.287336826324463,
+      "learning_rate": 2.1588139723801787e-06,
+      "loss": 1.7454,
+      "step": 1689
+    },
+    {
+      "epoch": 1.8520547945205479,
+      "grad_norm": 2.2864668369293213,
+      "learning_rate": 2.156783103168156e-06,
+      "loss": 1.7807,
+      "step": 1690
+    },
+    {
+      "epoch": 1.8520547945205479,
+      "eval_loss": 1.9940879344940186,
+      "eval_runtime": 37.0261,
+      "eval_samples_per_second": 27.008,
+      "eval_steps_per_second": 1.134,
+      "step": 1690
+    },
+    {
+      "epoch": 1.853150684931507,
+      "grad_norm": 2.1782145500183105,
+      "learning_rate": 2.1547522339561333e-06,
+      "loss": 1.7567,
+      "step": 1691
+    },
+    {
+      "epoch": 1.8542465753424657,
+      "grad_norm": 2.252023458480835,
+      "learning_rate": 2.1527213647441106e-06,
+      "loss": 1.811,
+      "step": 1692
+    },
+    {
+      "epoch": 1.8553424657534245,
+      "grad_norm": 2.2532849311828613,
+      "learning_rate": 2.150690495532088e-06,
+      "loss": 1.8491,
+      "step": 1693
+    },
+    {
+      "epoch": 1.8564383561643836,
+      "grad_norm": 2.35648250579834,
+      "learning_rate": 2.148659626320065e-06,
+      "loss": 1.8125,
+      "step": 1694
+    },
+    {
+      "epoch": 1.8575342465753426,
+      "grad_norm": 2.3490521907806396,
+      "learning_rate": 2.1466287571080424e-06,
+      "loss": 1.8035,
+      "step": 1695
+    },
+    {
+      "epoch": 1.8586301369863014,
+      "grad_norm": 2.2851879596710205,
+      "learning_rate": 2.1445978878960197e-06,
+      "loss": 1.7755,
+      "step": 1696
+    },
+    {
+      "epoch": 1.8597260273972602,
+      "grad_norm": 2.166754961013794,
+      "learning_rate": 2.142567018683997e-06,
+      "loss": 1.7861,
+      "step": 1697
+    },
+    {
+      "epoch": 1.8608219178082193,
+      "grad_norm": 2.2356410026550293,
+      "learning_rate": 2.140536149471974e-06,
+      "loss": 1.8412,
+      "step": 1698
+    },
+    {
+      "epoch": 1.861917808219178,
+      "grad_norm": 2.215618133544922,
+      "learning_rate": 2.1385052802599515e-06,
+      "loss": 1.7623,
+      "step": 1699
+    },
+    {
+      "epoch": 1.8630136986301369,
+      "grad_norm": 2.283298969268799,
+      "learning_rate": 2.1364744110479288e-06,
+      "loss": 1.7692,
+      "step": 1700
+    },
+    {
+      "epoch": 1.8630136986301369,
+      "eval_loss": 1.9961501359939575,
+      "eval_runtime": 37.0338,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1700
+    },
+    {
+      "epoch": 1.864109589041096,
+      "grad_norm": 2.292818546295166,
+      "learning_rate": 2.134443541835906e-06,
+      "loss": 1.8251,
+      "step": 1701
+    },
+    {
+      "epoch": 1.865205479452055,
+      "grad_norm": 2.235895872116089,
+      "learning_rate": 2.132412672623883e-06,
+      "loss": 1.7726,
+      "step": 1702
+    },
+    {
+      "epoch": 1.8663013698630135,
+      "grad_norm": 2.239413022994995,
+      "learning_rate": 2.13038180341186e-06,
+      "loss": 1.7509,
+      "step": 1703
+    },
+    {
+      "epoch": 1.8673972602739726,
+      "grad_norm": 2.4125585556030273,
+      "learning_rate": 2.1283509341998374e-06,
+      "loss": 1.7973,
+      "step": 1704
+    },
+    {
+      "epoch": 1.8684931506849316,
+      "grad_norm": 2.3774733543395996,
+      "learning_rate": 2.1263200649878147e-06,
+      "loss": 1.8526,
+      "step": 1705
+    },
+    {
+      "epoch": 1.8695890410958904,
+      "grad_norm": 2.344918966293335,
+      "learning_rate": 2.124289195775792e-06,
+      "loss": 1.7708,
+      "step": 1706
+    },
+    {
+      "epoch": 1.8706849315068492,
+      "grad_norm": 2.1933865547180176,
+      "learning_rate": 2.1222583265637693e-06,
+      "loss": 1.7935,
+      "step": 1707
+    },
+    {
+      "epoch": 1.8717808219178083,
+      "grad_norm": 2.4939582347869873,
+      "learning_rate": 2.1202274573517465e-06,
+      "loss": 1.706,
+      "step": 1708
+    },
+    {
+      "epoch": 1.872876712328767,
+      "grad_norm": 2.407532215118408,
+      "learning_rate": 2.118196588139724e-06,
+      "loss": 1.7522,
+      "step": 1709
+    },
+    {
+      "epoch": 1.8739726027397259,
+      "grad_norm": 2.297900915145874,
+      "learning_rate": 2.116165718927701e-06,
+      "loss": 1.7558,
+      "step": 1710
+    },
+    {
+      "epoch": 1.8739726027397259,
+      "eval_loss": 1.9946404695510864,
+      "eval_runtime": 36.7569,
+      "eval_samples_per_second": 27.206,
+      "eval_steps_per_second": 1.143,
+      "step": 1710
+    },
+    {
+      "epoch": 1.875068493150685,
+      "grad_norm": 2.2298660278320312,
+      "learning_rate": 2.1141348497156784e-06,
+      "loss": 1.7836,
+      "step": 1711
+    },
+    {
+      "epoch": 1.876164383561644,
+      "grad_norm": 2.308060884475708,
+      "learning_rate": 2.1121039805036556e-06,
+      "loss": 1.7616,
+      "step": 1712
+    },
+    {
+      "epoch": 1.8772602739726028,
+      "grad_norm": 2.288381338119507,
+      "learning_rate": 2.110073111291633e-06,
+      "loss": 1.7502,
+      "step": 1713
+    },
+    {
+      "epoch": 1.8783561643835616,
+      "grad_norm": 2.2191097736358643,
+      "learning_rate": 2.10804224207961e-06,
+      "loss": 1.7946,
+      "step": 1714
+    },
+    {
+      "epoch": 1.8794520547945206,
+      "grad_norm": 2.211777925491333,
+      "learning_rate": 2.1060113728675875e-06,
+      "loss": 1.8238,
+      "step": 1715
+    },
+    {
+      "epoch": 1.8805479452054794,
+      "grad_norm": 2.3946406841278076,
+      "learning_rate": 2.1039805036555647e-06,
+      "loss": 1.8146,
+      "step": 1716
+    },
+    {
+      "epoch": 1.8816438356164382,
+      "grad_norm": 2.2949914932250977,
+      "learning_rate": 2.101949634443542e-06,
+      "loss": 1.6877,
+      "step": 1717
+    },
+    {
+      "epoch": 1.8827397260273973,
+      "grad_norm": 2.2905256748199463,
+      "learning_rate": 2.0999187652315193e-06,
+      "loss": 1.7868,
+      "step": 1718
+    },
+    {
+      "epoch": 1.8838356164383563,
+      "grad_norm": 2.289280891418457,
+      "learning_rate": 2.0978878960194966e-06,
+      "loss": 1.7387,
+      "step": 1719
+    },
+    {
+      "epoch": 1.884931506849315,
+      "grad_norm": 2.35497784614563,
+      "learning_rate": 2.095857026807474e-06,
+      "loss": 1.8002,
+      "step": 1720
+    },
+    {
+      "epoch": 1.884931506849315,
+      "eval_loss": 1.9928913116455078,
+      "eval_runtime": 36.9289,
+      "eval_samples_per_second": 27.079,
+      "eval_steps_per_second": 1.137,
+      "step": 1720
+    },
+    {
+      "epoch": 1.886027397260274,
+      "grad_norm": 2.4404468536376953,
+      "learning_rate": 2.093826157595451e-06,
+      "loss": 1.7981,
+      "step": 1721
+    },
+    {
+      "epoch": 1.887123287671233,
+      "grad_norm": 2.315718173980713,
+      "learning_rate": 2.0917952883834284e-06,
+      "loss": 1.7581,
+      "step": 1722
+    },
+    {
+      "epoch": 1.8882191780821918,
+      "grad_norm": 2.2984843254089355,
+      "learning_rate": 2.0897644191714057e-06,
+      "loss": 1.8153,
+      "step": 1723
+    },
+    {
+      "epoch": 1.8893150684931506,
+      "grad_norm": 2.3332557678222656,
+      "learning_rate": 2.087733549959383e-06,
+      "loss": 1.831,
+      "step": 1724
+    },
+    {
+      "epoch": 1.8904109589041096,
+      "grad_norm": 2.3394649028778076,
+      "learning_rate": 2.0857026807473602e-06,
+      "loss": 1.7486,
+      "step": 1725
+    },
+    {
+      "epoch": 1.8915068493150686,
+      "grad_norm": 2.2985246181488037,
+      "learning_rate": 2.0836718115353375e-06,
+      "loss": 1.721,
+      "step": 1726
+    },
+    {
+      "epoch": 1.8926027397260274,
+      "grad_norm": 2.318504810333252,
+      "learning_rate": 2.0816409423233144e-06,
+      "loss": 1.8469,
+      "step": 1727
+    },
+    {
+      "epoch": 1.8936986301369862,
+      "grad_norm": 2.4938831329345703,
+      "learning_rate": 2.0796100731112916e-06,
+      "loss": 1.7525,
+      "step": 1728
+    },
+    {
+      "epoch": 1.8947945205479453,
+      "grad_norm": 2.3391780853271484,
+      "learning_rate": 2.077579203899269e-06,
+      "loss": 1.7733,
+      "step": 1729
+    },
+    {
+      "epoch": 1.895890410958904,
+      "grad_norm": 2.437540054321289,
+      "learning_rate": 2.075548334687246e-06,
+      "loss": 1.7344,
+      "step": 1730
+    },
+    {
+      "epoch": 1.895890410958904,
+      "eval_loss": 1.990493655204773,
+      "eval_runtime": 37.0347,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1730
+    },
+    {
+      "epoch": 1.896986301369863,
+      "grad_norm": 2.413628578186035,
+      "learning_rate": 2.0735174654752235e-06,
+      "loss": 1.7978,
+      "step": 1731
+    },
+    {
+      "epoch": 1.898082191780822,
+      "grad_norm": 2.3673973083496094,
+      "learning_rate": 2.0714865962632007e-06,
+      "loss": 1.7414,
+      "step": 1732
+    },
+    {
+      "epoch": 1.899178082191781,
+      "grad_norm": 2.3961968421936035,
+      "learning_rate": 2.069455727051178e-06,
+      "loss": 1.7854,
+      "step": 1733
+    },
+    {
+      "epoch": 1.9002739726027398,
+      "grad_norm": 2.5583736896514893,
+      "learning_rate": 2.0674248578391553e-06,
+      "loss": 1.7629,
+      "step": 1734
+    },
+    {
+      "epoch": 1.9013698630136986,
+      "grad_norm": 2.256418228149414,
+      "learning_rate": 2.0653939886271326e-06,
+      "loss": 1.7845,
+      "step": 1735
+    },
+    {
+      "epoch": 1.9024657534246576,
+      "grad_norm": 2.4235012531280518,
+      "learning_rate": 2.06336311941511e-06,
+      "loss": 1.8151,
+      "step": 1736
+    },
+    {
+      "epoch": 1.9035616438356164,
+      "grad_norm": 2.2945358753204346,
+      "learning_rate": 2.061332250203087e-06,
+      "loss": 1.7665,
+      "step": 1737
+    },
+    {
+      "epoch": 1.9046575342465752,
+      "grad_norm": 2.371809959411621,
+      "learning_rate": 2.0593013809910644e-06,
+      "loss": 1.7355,
+      "step": 1738
+    },
+    {
+      "epoch": 1.9057534246575343,
+      "grad_norm": 2.26947021484375,
+      "learning_rate": 2.0572705117790417e-06,
+      "loss": 1.7702,
+      "step": 1739
+    },
+    {
+      "epoch": 1.9068493150684933,
+      "grad_norm": 2.314237356185913,
+      "learning_rate": 2.055239642567019e-06,
+      "loss": 1.8084,
+      "step": 1740
+    },
+    {
+      "epoch": 1.9068493150684933,
+      "eval_loss": 1.9961000680923462,
+      "eval_runtime": 37.0775,
+      "eval_samples_per_second": 26.971,
+      "eval_steps_per_second": 1.133,
+      "step": 1740
+    },
+    {
+      "epoch": 1.907945205479452,
+      "grad_norm": 2.248680591583252,
+      "learning_rate": 2.0532087733549962e-06,
+      "loss": 1.7898,
+      "step": 1741
+    },
+    {
+      "epoch": 1.909041095890411,
+      "grad_norm": 2.336378812789917,
+      "learning_rate": 2.0511779041429735e-06,
+      "loss": 1.795,
+      "step": 1742
+    },
+    {
+      "epoch": 1.91013698630137,
+      "grad_norm": 2.2420554161071777,
+      "learning_rate": 2.0491470349309508e-06,
+      "loss": 1.76,
+      "step": 1743
+    },
+    {
+      "epoch": 1.9112328767123288,
+      "grad_norm": 2.3305511474609375,
+      "learning_rate": 2.047116165718928e-06,
+      "loss": 1.7959,
+      "step": 1744
+    },
+    {
+      "epoch": 1.9123287671232876,
+      "grad_norm": 2.2677786350250244,
+      "learning_rate": 2.0450852965069053e-06,
+      "loss": 1.7922,
+      "step": 1745
+    },
+    {
+      "epoch": 1.9134246575342466,
+      "grad_norm": 2.224148988723755,
+      "learning_rate": 2.0430544272948826e-06,
+      "loss": 1.841,
+      "step": 1746
+    },
+    {
+      "epoch": 1.9145205479452054,
+      "grad_norm": 2.172091245651245,
+      "learning_rate": 2.04102355808286e-06,
+      "loss": 1.784,
+      "step": 1747
+    },
+    {
+      "epoch": 1.9156164383561642,
+      "grad_norm": 2.319962501525879,
+      "learning_rate": 2.038992688870837e-06,
+      "loss": 1.7238,
+      "step": 1748
+    },
+    {
+      "epoch": 1.9167123287671233,
+      "grad_norm": 2.299760103225708,
+      "learning_rate": 2.0369618196588144e-06,
+      "loss": 1.7501,
+      "step": 1749
+    },
+    {
+      "epoch": 1.9178082191780823,
+      "grad_norm": 2.218407154083252,
+      "learning_rate": 2.0349309504467917e-06,
+      "loss": 1.7628,
+      "step": 1750
+    },
+    {
+      "epoch": 1.9178082191780823,
+      "eval_loss": 1.9934331178665161,
+      "eval_runtime": 37.0986,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 1750
+    },
+    {
+      "epoch": 1.9189041095890411,
+      "grad_norm": 2.4121744632720947,
+      "learning_rate": 2.032900081234769e-06,
+      "loss": 1.8128,
+      "step": 1751
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 2.253225326538086,
+      "learning_rate": 2.030869212022746e-06,
+      "loss": 1.8161,
+      "step": 1752
+    },
+    {
+      "epoch": 1.921095890410959,
+      "grad_norm": 2.28702449798584,
+      "learning_rate": 2.028838342810723e-06,
+      "loss": 1.7769,
+      "step": 1753
+    },
+    {
+      "epoch": 1.9221917808219178,
+      "grad_norm": 2.3004465103149414,
+      "learning_rate": 2.0268074735987004e-06,
+      "loss": 1.7631,
+      "step": 1754
+    },
+    {
+      "epoch": 1.9232876712328766,
+      "grad_norm": 2.2733960151672363,
+      "learning_rate": 2.0247766043866777e-06,
+      "loss": 1.75,
+      "step": 1755
+    },
+    {
+      "epoch": 1.9243835616438356,
+      "grad_norm": 2.274019241333008,
+      "learning_rate": 2.022745735174655e-06,
+      "loss": 1.8317,
+      "step": 1756
+    },
+    {
+      "epoch": 1.9254794520547946,
+      "grad_norm": 2.3094100952148438,
+      "learning_rate": 2.020714865962632e-06,
+      "loss": 1.7841,
+      "step": 1757
+    },
+    {
+      "epoch": 1.9265753424657535,
+      "grad_norm": 2.269321918487549,
+      "learning_rate": 2.0186839967506095e-06,
+      "loss": 1.8052,
+      "step": 1758
+    },
+    {
+      "epoch": 1.9276712328767123,
+      "grad_norm": 2.5324459075927734,
+      "learning_rate": 2.0166531275385868e-06,
+      "loss": 1.7916,
+      "step": 1759
+    },
+    {
+      "epoch": 1.9287671232876713,
+      "grad_norm": 2.302569627761841,
+      "learning_rate": 2.014622258326564e-06,
+      "loss": 1.7499,
+      "step": 1760
+    },
+    {
+      "epoch": 1.9287671232876713,
+      "eval_loss": 1.9879337549209595,
+      "eval_runtime": 37.1186,
+      "eval_samples_per_second": 26.941,
+      "eval_steps_per_second": 1.132,
+      "step": 1760
+    },
+    {
+      "epoch": 1.9298630136986301,
+      "grad_norm": 2.29470157623291,
+      "learning_rate": 2.0125913891145413e-06,
+      "loss": 1.7922,
+      "step": 1761
+    },
+    {
+      "epoch": 1.930958904109589,
+      "grad_norm": 2.4307754039764404,
+      "learning_rate": 2.0105605199025186e-06,
+      "loss": 1.7802,
+      "step": 1762
+    },
+    {
+      "epoch": 1.932054794520548,
+      "grad_norm": 2.179588794708252,
+      "learning_rate": 2.008529650690496e-06,
+      "loss": 1.7391,
+      "step": 1763
+    },
+    {
+      "epoch": 1.933150684931507,
+      "grad_norm": 2.5824451446533203,
+      "learning_rate": 2.0064987814784727e-06,
+      "loss": 1.8106,
+      "step": 1764
+    },
+    {
+      "epoch": 1.9342465753424658,
+      "grad_norm": 2.32308292388916,
+      "learning_rate": 2.00446791226645e-06,
+      "loss": 1.7689,
+      "step": 1765
+    },
+    {
+      "epoch": 1.9353424657534246,
+      "grad_norm": 2.262533664703369,
+      "learning_rate": 2.0024370430544273e-06,
+      "loss": 1.7604,
+      "step": 1766
+    },
+    {
+      "epoch": 1.9364383561643836,
+      "grad_norm": 2.325253963470459,
+      "learning_rate": 2.0004061738424045e-06,
+      "loss": 1.7695,
+      "step": 1767
+    },
+    {
+      "epoch": 1.9375342465753425,
+      "grad_norm": 2.2855470180511475,
+      "learning_rate": 1.998375304630382e-06,
+      "loss": 1.8233,
+      "step": 1768
+    },
+    {
+      "epoch": 1.9386301369863013,
+      "grad_norm": 2.3688673973083496,
+      "learning_rate": 1.996344435418359e-06,
+      "loss": 1.7799,
+      "step": 1769
+    },
+    {
+      "epoch": 1.9397260273972603,
+      "grad_norm": 2.4682323932647705,
+      "learning_rate": 1.9943135662063364e-06,
+      "loss": 1.8036,
+      "step": 1770
+    },
+    {
+      "epoch": 1.9397260273972603,
+      "eval_loss": 1.9906389713287354,
+      "eval_runtime": 37.0959,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 1770
+    },
+    {
+      "epoch": 1.9408219178082193,
+      "grad_norm": 2.369166374206543,
+      "learning_rate": 1.9922826969943136e-06,
+      "loss": 1.7878,
+      "step": 1771
+    },
+    {
+      "epoch": 1.941917808219178,
+      "grad_norm": 2.4669666290283203,
+      "learning_rate": 1.990251827782291e-06,
+      "loss": 1.8228,
+      "step": 1772
+    },
+    {
+      "epoch": 1.943013698630137,
+      "grad_norm": 2.3083250522613525,
+      "learning_rate": 1.988220958570268e-06,
+      "loss": 1.7997,
+      "step": 1773
+    },
+    {
+      "epoch": 1.944109589041096,
+      "grad_norm": 2.4884281158447266,
+      "learning_rate": 1.9861900893582455e-06,
+      "loss": 1.8105,
+      "step": 1774
+    },
+    {
+      "epoch": 1.9452054794520548,
+      "grad_norm": 2.3504858016967773,
+      "learning_rate": 1.9841592201462227e-06,
+      "loss": 1.8159,
+      "step": 1775
+    },
+    {
+      "epoch": 1.9463013698630136,
+      "grad_norm": 2.618431329727173,
+      "learning_rate": 1.9821283509342e-06,
+      "loss": 1.7352,
+      "step": 1776
+    },
+    {
+      "epoch": 1.9473972602739726,
+      "grad_norm": 2.3452653884887695,
+      "learning_rate": 1.9800974817221773e-06,
+      "loss": 1.7451,
+      "step": 1777
+    },
+    {
+      "epoch": 1.9484931506849315,
+      "grad_norm": 2.5966849327087402,
+      "learning_rate": 1.9780666125101546e-06,
+      "loss": 1.7965,
+      "step": 1778
+    },
+    {
+      "epoch": 1.9495890410958903,
+      "grad_norm": 2.3017661571502686,
+      "learning_rate": 1.976035743298132e-06,
+      "loss": 1.7629,
+      "step": 1779
+    },
+    {
+      "epoch": 1.9506849315068493,
+      "grad_norm": 2.3766098022460938,
+      "learning_rate": 1.974004874086109e-06,
+      "loss": 1.8403,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9506849315068493,
+      "eval_loss": 1.9936374425888062,
+      "eval_runtime": 37.1158,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9517808219178083,
+      "grad_norm": 2.2327704429626465,
+      "learning_rate": 1.9719740048740864e-06,
+      "loss": 1.7096,
+      "step": 1781
+    },
+    {
+      "epoch": 1.9528767123287671,
+      "grad_norm": 2.458707809448242,
+      "learning_rate": 1.9699431356620637e-06,
+      "loss": 1.7636,
+      "step": 1782
+    },
+    {
+      "epoch": 1.953972602739726,
+      "grad_norm": 2.4143803119659424,
+      "learning_rate": 1.967912266450041e-06,
+      "loss": 1.757,
+      "step": 1783
+    },
+    {
+      "epoch": 1.955068493150685,
+      "grad_norm": 2.465148448944092,
+      "learning_rate": 1.9658813972380182e-06,
+      "loss": 1.7526,
+      "step": 1784
+    },
+    {
+      "epoch": 1.9561643835616438,
+      "grad_norm": 2.356210231781006,
+      "learning_rate": 1.9638505280259955e-06,
+      "loss": 1.8705,
+      "step": 1785
+    },
+    {
+      "epoch": 1.9572602739726026,
+      "grad_norm": 2.7456002235412598,
+      "learning_rate": 1.9618196588139728e-06,
+      "loss": 1.7574,
+      "step": 1786
+    },
+    {
+      "epoch": 1.9583561643835616,
+      "grad_norm": 2.3724262714385986,
+      "learning_rate": 1.95978878960195e-06,
+      "loss": 1.7854,
+      "step": 1787
+    },
+    {
+      "epoch": 1.9594520547945207,
+      "grad_norm": 2.275827646255493,
+      "learning_rate": 1.9577579203899273e-06,
+      "loss": 1.8447,
+      "step": 1788
+    },
+    {
+      "epoch": 1.9605479452054795,
+      "grad_norm": 2.369992256164551,
+      "learning_rate": 1.955727051177904e-06,
+      "loss": 1.8299,
+      "step": 1789
+    },
+    {
+      "epoch": 1.9616438356164383,
+      "grad_norm": 2.462372064590454,
+      "learning_rate": 1.9536961819658815e-06,
+      "loss": 1.7596,
+      "step": 1790
+    },
+    {
+      "epoch": 1.9616438356164383,
+      "eval_loss": 1.9947243928909302,
+      "eval_runtime": 37.111,
+      "eval_samples_per_second": 26.946,
+      "eval_steps_per_second": 1.132,
+      "step": 1790
+    },
+    {
+      "epoch": 1.9627397260273973,
+      "grad_norm": 2.3040034770965576,
+      "learning_rate": 1.9516653127538587e-06,
+      "loss": 1.7391,
+      "step": 1791
+    },
+    {
+      "epoch": 1.9638356164383561,
+      "grad_norm": 2.449328899383545,
+      "learning_rate": 1.949634443541836e-06,
+      "loss": 1.7428,
+      "step": 1792
+    },
+    {
+      "epoch": 1.964931506849315,
+      "grad_norm": 2.2679781913757324,
+      "learning_rate": 1.9476035743298133e-06,
+      "loss": 1.738,
+      "step": 1793
+    },
+    {
+      "epoch": 1.966027397260274,
+      "grad_norm": 2.3444032669067383,
+      "learning_rate": 1.9455727051177906e-06,
+      "loss": 1.8089,
+      "step": 1794
+    },
+    {
+      "epoch": 1.967123287671233,
+      "grad_norm": 2.3005306720733643,
+      "learning_rate": 1.943541835905768e-06,
+      "loss": 1.7718,
+      "step": 1795
+    },
+    {
+      "epoch": 1.9682191780821918,
+      "grad_norm": 2.6022629737854004,
+      "learning_rate": 1.941510966693745e-06,
+      "loss": 1.8325,
+      "step": 1796
+    },
+    {
+      "epoch": 1.9693150684931506,
+      "grad_norm": 2.407585382461548,
+      "learning_rate": 1.9394800974817224e-06,
+      "loss": 1.737,
+      "step": 1797
+    },
+    {
+      "epoch": 1.9704109589041097,
+      "grad_norm": 2.349339723587036,
+      "learning_rate": 1.9374492282696997e-06,
+      "loss": 1.81,
+      "step": 1798
+    },
+    {
+      "epoch": 1.9715068493150685,
+      "grad_norm": 2.296192169189453,
+      "learning_rate": 1.935418359057677e-06,
+      "loss": 1.8424,
+      "step": 1799
+    },
+    {
+      "epoch": 1.9726027397260273,
+      "grad_norm": 2.175875663757324,
+      "learning_rate": 1.9333874898456542e-06,
+      "loss": 1.7977,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9726027397260273,
+      "eval_loss": 1.9929841756820679,
+      "eval_runtime": 37.1159,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9736986301369863,
+      "grad_norm": 2.37428617477417,
+      "learning_rate": 1.931356620633631e-06,
+      "loss": 1.8248,
+      "step": 1801
+    },
+    {
+      "epoch": 1.9747945205479454,
+      "grad_norm": 2.297621965408325,
+      "learning_rate": 1.9293257514216083e-06,
+      "loss": 1.7215,
+      "step": 1802
+    },
+    {
+      "epoch": 1.9758904109589042,
+      "grad_norm": 2.217275857925415,
+      "learning_rate": 1.9272948822095856e-06,
+      "loss": 1.7828,
+      "step": 1803
+    },
+    {
+      "epoch": 1.976986301369863,
+      "grad_norm": 2.250854015350342,
+      "learning_rate": 1.925264012997563e-06,
+      "loss": 1.7245,
+      "step": 1804
+    },
+    {
+      "epoch": 1.978082191780822,
+      "grad_norm": 2.2556822299957275,
+      "learning_rate": 1.92323314378554e-06,
+      "loss": 1.7715,
+      "step": 1805
+    },
+    {
+      "epoch": 1.9791780821917808,
+      "grad_norm": 2.316389322280884,
+      "learning_rate": 1.9212022745735174e-06,
+      "loss": 1.8063,
+      "step": 1806
+    },
+    {
+      "epoch": 1.9802739726027396,
+      "grad_norm": 2.4358153343200684,
+      "learning_rate": 1.9191714053614947e-06,
+      "loss": 1.747,
+      "step": 1807
+    },
+    {
+      "epoch": 1.9813698630136987,
+      "grad_norm": 2.437631368637085,
+      "learning_rate": 1.917140536149472e-06,
+      "loss": 1.7485,
+      "step": 1808
+    },
+    {
+      "epoch": 1.9824657534246577,
+      "grad_norm": 2.3554601669311523,
+      "learning_rate": 1.9151096669374493e-06,
+      "loss": 1.7195,
+      "step": 1809
+    },
+    {
+      "epoch": 1.9835616438356163,
+      "grad_norm": 2.575559616088867,
+      "learning_rate": 1.9130787977254266e-06,
+      "loss": 1.8368,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9835616438356163,
+      "eval_loss": 1.9903514385223389,
+      "eval_runtime": 36.8082,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9846575342465753,
+      "grad_norm": 2.3074066638946533,
+      "learning_rate": 1.911047928513404e-06,
+      "loss": 1.7952,
+      "step": 1811
+    },
+    {
+      "epoch": 1.9857534246575343,
+      "grad_norm": 2.3162307739257812,
+      "learning_rate": 1.909017059301381e-06,
+      "loss": 1.7455,
+      "step": 1812
+    },
+    {
+      "epoch": 1.9868493150684932,
+      "grad_norm": 2.3085129261016846,
+      "learning_rate": 1.9069861900893586e-06,
+      "loss": 1.7513,
+      "step": 1813
+    },
+    {
+      "epoch": 1.987945205479452,
+      "grad_norm": 2.3454785346984863,
+      "learning_rate": 1.9049553208773354e-06,
+      "loss": 1.803,
+      "step": 1814
+    },
+    {
+      "epoch": 1.989041095890411,
+      "grad_norm": 2.4522876739501953,
+      "learning_rate": 1.9029244516653127e-06,
+      "loss": 1.7567,
+      "step": 1815
+    },
+    {
+      "epoch": 1.9901369863013698,
+      "grad_norm": 2.3876044750213623,
+      "learning_rate": 1.90089358245329e-06,
+      "loss": 1.7126,
+      "step": 1816
+    },
+    {
+      "epoch": 1.9912328767123286,
+      "grad_norm": 2.4061570167541504,
+      "learning_rate": 1.8988627132412673e-06,
+      "loss": 1.7618,
+      "step": 1817
+    },
+    {
+      "epoch": 1.9923287671232877,
+      "grad_norm": 2.307445764541626,
+      "learning_rate": 1.8968318440292445e-06,
+      "loss": 1.8014,
+      "step": 1818
+    },
+    {
+      "epoch": 1.9934246575342467,
+      "grad_norm": 2.4414806365966797,
+      "learning_rate": 1.8948009748172218e-06,
+      "loss": 1.7202,
+      "step": 1819
+    },
+    {
+      "epoch": 1.9945205479452055,
+      "grad_norm": 2.3933329582214355,
+      "learning_rate": 1.892770105605199e-06,
+      "loss": 1.8133,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9945205479452055,
+      "eval_loss": 1.991702675819397,
+      "eval_runtime": 37.0127,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9956164383561643,
+      "grad_norm": 2.237699508666992,
+      "learning_rate": 1.8907392363931764e-06,
+      "loss": 1.8026,
+      "step": 1821
+    },
+    {
+      "epoch": 1.9967123287671233,
+      "grad_norm": 2.290004014968872,
+      "learning_rate": 1.8887083671811536e-06,
+      "loss": 1.7689,
+      "step": 1822
+    },
+    {
+      "epoch": 1.9978082191780822,
+      "grad_norm": 2.2264676094055176,
+      "learning_rate": 1.886677497969131e-06,
+      "loss": 1.7862,
+      "step": 1823
+    },
+    {
+      "epoch": 1.998904109589041,
+      "grad_norm": 2.2764949798583984,
+      "learning_rate": 1.8846466287571082e-06,
+      "loss": 1.6763,
+      "step": 1824
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 2.3024826049804688,
+      "learning_rate": 1.8826157595450855e-06,
+      "loss": 1.8076,
+      "step": 1825
+    },
+    {
+      "epoch": 2.001095890410959,
+      "grad_norm": 2.362882614135742,
+      "learning_rate": 1.8805848903330625e-06,
+      "loss": 1.7009,
+      "step": 1826
+    },
+    {
+      "epoch": 2.0021917808219176,
+      "grad_norm": 2.432370185852051,
+      "learning_rate": 1.8785540211210398e-06,
+      "loss": 1.8712,
+      "step": 1827
+    },
+    {
+      "epoch": 2.0032876712328767,
+      "grad_norm": 2.3584744930267334,
+      "learning_rate": 1.876523151909017e-06,
+      "loss": 1.7869,
+      "step": 1828
+    },
+    {
+      "epoch": 2.0043835616438357,
+      "grad_norm": 2.534205198287964,
+      "learning_rate": 1.8744922826969944e-06,
+      "loss": 1.7376,
+      "step": 1829
+    },
+    {
+      "epoch": 2.0054794520547947,
+      "grad_norm": 2.58241605758667,
+      "learning_rate": 1.8724614134849716e-06,
+      "loss": 1.7204,
+      "step": 1830
+    },
+    {
+      "epoch": 2.0054794520547947,
+      "eval_loss": 1.9912636280059814,
+      "eval_runtime": 37.0969,
+      "eval_samples_per_second": 26.956,
+      "eval_steps_per_second": 1.132,
+      "step": 1830
+    },
+    {
+      "epoch": 2.0065753424657533,
+      "grad_norm": 2.3742504119873047,
+      "learning_rate": 1.870430544272949e-06,
+      "loss": 1.7537,
+      "step": 1831
+    },
+    {
+      "epoch": 2.0076712328767123,
+      "grad_norm": 2.6560916900634766,
+      "learning_rate": 1.8683996750609262e-06,
+      "loss": 1.8524,
+      "step": 1832
+    },
+    {
+      "epoch": 2.0087671232876714,
+      "grad_norm": 2.523041248321533,
+      "learning_rate": 1.8663688058489035e-06,
+      "loss": 1.7405,
+      "step": 1833
+    },
+    {
+      "epoch": 2.00986301369863,
+      "grad_norm": 2.5444304943084717,
+      "learning_rate": 1.8643379366368807e-06,
+      "loss": 1.7588,
+      "step": 1834
+    },
+    {
+      "epoch": 2.010958904109589,
+      "grad_norm": 2.708984613418579,
+      "learning_rate": 1.862307067424858e-06,
+      "loss": 1.7686,
+      "step": 1835
+    },
+    {
+      "epoch": 2.012054794520548,
+      "grad_norm": 2.726691961288452,
+      "learning_rate": 1.8602761982128353e-06,
+      "loss": 1.7733,
+      "step": 1836
+    },
+    {
+      "epoch": 2.013150684931507,
+      "grad_norm": 2.5085904598236084,
+      "learning_rate": 1.8582453290008126e-06,
+      "loss": 1.81,
+      "step": 1837
+    },
+    {
+      "epoch": 2.0142465753424657,
+      "grad_norm": 2.4699466228485107,
+      "learning_rate": 1.8562144597887898e-06,
+      "loss": 1.7369,
+      "step": 1838
+    },
+    {
+      "epoch": 2.0153424657534247,
+      "grad_norm": 2.531027317047119,
+      "learning_rate": 1.854183590576767e-06,
+      "loss": 1.7611,
+      "step": 1839
+    },
+    {
+      "epoch": 2.0164383561643837,
+      "grad_norm": 2.5061161518096924,
+      "learning_rate": 1.8521527213647442e-06,
+      "loss": 1.811,
+      "step": 1840
+    },
+    {
+      "epoch": 2.0164383561643837,
+      "eval_loss": 1.9893144369125366,
+      "eval_runtime": 37.1262,
+      "eval_samples_per_second": 26.935,
+      "eval_steps_per_second": 1.131,
+      "step": 1840
+    },
+    {
+      "epoch": 2.0175342465753423,
+      "grad_norm": 2.5739994049072266,
+      "learning_rate": 1.8501218521527215e-06,
+      "loss": 1.7153,
+      "step": 1841
+    },
+    {
+      "epoch": 2.0186301369863013,
+      "grad_norm": 2.5673015117645264,
+      "learning_rate": 1.8480909829406987e-06,
+      "loss": 1.6879,
+      "step": 1842
+    },
+    {
+      "epoch": 2.0197260273972604,
+      "grad_norm": 2.717430353164673,
+      "learning_rate": 1.846060113728676e-06,
+      "loss": 1.723,
+      "step": 1843
+    },
+    {
+      "epoch": 2.020821917808219,
+      "grad_norm": 2.4905285835266113,
+      "learning_rate": 1.8440292445166533e-06,
+      "loss": 1.7651,
+      "step": 1844
+    },
+    {
+      "epoch": 2.021917808219178,
+      "grad_norm": 2.6798791885375977,
+      "learning_rate": 1.8419983753046306e-06,
+      "loss": 1.7308,
+      "step": 1845
+    },
+    {
+      "epoch": 2.023013698630137,
+      "grad_norm": 2.591614007949829,
+      "learning_rate": 1.8399675060926078e-06,
+      "loss": 1.786,
+      "step": 1846
+    },
+    {
+      "epoch": 2.024109589041096,
+      "grad_norm": 2.5405824184417725,
+      "learning_rate": 1.8379366368805851e-06,
+      "loss": 1.7699,
+      "step": 1847
+    },
+    {
+      "epoch": 2.0252054794520546,
+      "grad_norm": 2.4189934730529785,
+      "learning_rate": 1.8359057676685624e-06,
+      "loss": 1.732,
+      "step": 1848
+    },
+    {
+      "epoch": 2.0263013698630137,
+      "grad_norm": 2.5857982635498047,
+      "learning_rate": 1.8338748984565397e-06,
+      "loss": 1.7123,
+      "step": 1849
+    },
+    {
+      "epoch": 2.0273972602739727,
+      "grad_norm": 2.4710893630981445,
+      "learning_rate": 1.831844029244517e-06,
+      "loss": 1.7419,
+      "step": 1850
+    },
+    {
+      "epoch": 2.0273972602739727,
+      "eval_loss": 1.9874999523162842,
+      "eval_runtime": 37.1588,
+      "eval_samples_per_second": 26.912,
+      "eval_steps_per_second": 1.13,
+      "step": 1850
+    },
+    {
+      "epoch": 2.0284931506849313,
+      "grad_norm": 2.3414645195007324,
+      "learning_rate": 1.829813160032494e-06,
+      "loss": 1.7102,
+      "step": 1851
+    },
+    {
+      "epoch": 2.0295890410958903,
+      "grad_norm": 2.4148900508880615,
+      "learning_rate": 1.8277822908204713e-06,
+      "loss": 1.7287,
+      "step": 1852
+    },
+    {
+      "epoch": 2.0306849315068494,
+      "grad_norm": 2.5501346588134766,
+      "learning_rate": 1.8257514216084486e-06,
+      "loss": 1.7633,
+      "step": 1853
+    },
+    {
+      "epoch": 2.0317808219178084,
+      "grad_norm": 2.52215576171875,
+      "learning_rate": 1.8237205523964258e-06,
+      "loss": 1.7385,
+      "step": 1854
+    },
+    {
+      "epoch": 2.032876712328767,
+      "grad_norm": 2.433335065841675,
+      "learning_rate": 1.8216896831844031e-06,
+      "loss": 1.7566,
+      "step": 1855
+    },
+    {
+      "epoch": 2.033972602739726,
+      "grad_norm": 2.441232442855835,
+      "learning_rate": 1.8196588139723804e-06,
+      "loss": 1.8027,
+      "step": 1856
+    },
+    {
+      "epoch": 2.035068493150685,
+      "grad_norm": 2.4601082801818848,
+      "learning_rate": 1.8176279447603577e-06,
+      "loss": 1.6859,
+      "step": 1857
+    },
+    {
+      "epoch": 2.0361643835616436,
+      "grad_norm": 2.5008108615875244,
+      "learning_rate": 1.815597075548335e-06,
+      "loss": 1.7294,
+      "step": 1858
+    },
+    {
+      "epoch": 2.0372602739726027,
+      "grad_norm": 2.433915615081787,
+      "learning_rate": 1.8135662063363122e-06,
+      "loss": 1.755,
+      "step": 1859
+    },
+    {
+      "epoch": 2.0383561643835617,
+      "grad_norm": 2.4938697814941406,
+      "learning_rate": 1.8115353371242895e-06,
+      "loss": 1.7904,
+      "step": 1860
+    },
+    {
+      "epoch": 2.0383561643835617,
+      "eval_loss": 1.9939266443252563,
+      "eval_runtime": 37.1539,
+      "eval_samples_per_second": 26.915,
+      "eval_steps_per_second": 1.13,
+      "step": 1860
+    },
+    {
+      "epoch": 2.0394520547945207,
+      "grad_norm": 2.382324695587158,
+      "learning_rate": 1.8095044679122668e-06,
+      "loss": 1.6856,
+      "step": 1861
+    },
+    {
+      "epoch": 2.0405479452054793,
+      "grad_norm": 2.5250437259674072,
+      "learning_rate": 1.807473598700244e-06,
+      "loss": 1.7434,
+      "step": 1862
+    },
+    {
+      "epoch": 2.0416438356164384,
+      "grad_norm": 2.63969087600708,
+      "learning_rate": 1.805442729488221e-06,
+      "loss": 1.7343,
+      "step": 1863
+    },
+    {
+      "epoch": 2.0427397260273974,
+      "grad_norm": 2.4852068424224854,
+      "learning_rate": 1.8034118602761982e-06,
+      "loss": 1.7494,
+      "step": 1864
+    },
+    {
+      "epoch": 2.043835616438356,
+      "grad_norm": 2.393967390060425,
+      "learning_rate": 1.8013809910641754e-06,
+      "loss": 1.7503,
+      "step": 1865
+    },
+    {
+      "epoch": 2.044931506849315,
+      "grad_norm": 2.4447648525238037,
+      "learning_rate": 1.7993501218521527e-06,
+      "loss": 1.7539,
+      "step": 1866
+    },
+    {
+      "epoch": 2.046027397260274,
+      "grad_norm": 2.483530044555664,
+      "learning_rate": 1.79731925264013e-06,
+      "loss": 1.7001,
+      "step": 1867
+    },
+    {
+      "epoch": 2.047123287671233,
+      "grad_norm": 2.56669282913208,
+      "learning_rate": 1.7952883834281073e-06,
+      "loss": 1.7828,
+      "step": 1868
+    },
+    {
+      "epoch": 2.0482191780821917,
+      "grad_norm": 2.4894964694976807,
+      "learning_rate": 1.7932575142160846e-06,
+      "loss": 1.7208,
+      "step": 1869
+    },
+    {
+      "epoch": 2.0493150684931507,
+      "grad_norm": 2.4713854789733887,
+      "learning_rate": 1.7912266450040618e-06,
+      "loss": 1.7687,
+      "step": 1870
+    },
+    {
+      "epoch": 2.0493150684931507,
+      "eval_loss": 1.9938212633132935,
+      "eval_runtime": 37.1298,
+      "eval_samples_per_second": 26.933,
+      "eval_steps_per_second": 1.131,
+      "step": 1870
+    },
+    {
+      "epoch": 2.0504109589041097,
+      "grad_norm": 2.6204898357391357,
+      "learning_rate": 1.789195775792039e-06,
+      "loss": 1.7718,
+      "step": 1871
+    },
+    {
+      "epoch": 2.0515068493150683,
+      "grad_norm": 2.5194199085235596,
+      "learning_rate": 1.7871649065800164e-06,
+      "loss": 1.6653,
+      "step": 1872
+    },
+    {
+      "epoch": 2.0526027397260274,
+      "grad_norm": 2.5077965259552,
+      "learning_rate": 1.7851340373679937e-06,
+      "loss": 1.8098,
+      "step": 1873
+    },
+    {
+      "epoch": 2.0536986301369864,
+      "grad_norm": 2.5841965675354004,
+      "learning_rate": 1.783103168155971e-06,
+      "loss": 1.7126,
+      "step": 1874
+    },
+    {
+      "epoch": 2.0547945205479454,
+      "grad_norm": 2.617910623550415,
+      "learning_rate": 1.7810722989439482e-06,
+      "loss": 1.7294,
+      "step": 1875
+    },
+    {
+      "epoch": 2.055890410958904,
+      "grad_norm": 2.519253969192505,
+      "learning_rate": 1.7790414297319253e-06,
+      "loss": 1.7497,
+      "step": 1876
+    },
+    {
+      "epoch": 2.056986301369863,
+      "grad_norm": 2.5518412590026855,
+      "learning_rate": 1.7770105605199025e-06,
+      "loss": 1.7188,
+      "step": 1877
+    },
+    {
+      "epoch": 2.058082191780822,
+      "grad_norm": 2.5372824668884277,
+      "learning_rate": 1.7749796913078798e-06,
+      "loss": 1.6998,
+      "step": 1878
+    },
+    {
+      "epoch": 2.0591780821917807,
+      "grad_norm": 2.6649036407470703,
+      "learning_rate": 1.772948822095857e-06,
+      "loss": 1.7621,
+      "step": 1879
+    },
+    {
+      "epoch": 2.0602739726027397,
+      "grad_norm": 2.5216212272644043,
+      "learning_rate": 1.7709179528838344e-06,
+      "loss": 1.7963,
+      "step": 1880
+    },
+    {
+      "epoch": 2.0602739726027397,
+      "eval_loss": 1.993401050567627,
+      "eval_runtime": 37.0954,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 1880
+    },
+    {
+      "epoch": 2.0613698630136987,
+      "grad_norm": 2.5226478576660156,
+      "learning_rate": 1.7688870836718116e-06,
+      "loss": 1.7841,
+      "step": 1881
+    },
+    {
+      "epoch": 2.0624657534246573,
+      "grad_norm": 2.584319829940796,
+      "learning_rate": 1.766856214459789e-06,
+      "loss": 1.7449,
+      "step": 1882
+    },
+    {
+      "epoch": 2.0635616438356164,
+      "grad_norm": 2.5305304527282715,
+      "learning_rate": 1.7648253452477662e-06,
+      "loss": 1.7373,
+      "step": 1883
+    },
+    {
+      "epoch": 2.0646575342465754,
+      "grad_norm": 2.48797607421875,
+      "learning_rate": 1.7627944760357435e-06,
+      "loss": 1.722,
+      "step": 1884
+    },
+    {
+      "epoch": 2.0657534246575344,
+      "grad_norm": 2.4461028575897217,
+      "learning_rate": 1.7607636068237208e-06,
+      "loss": 1.7983,
+      "step": 1885
+    },
+    {
+      "epoch": 2.066849315068493,
+      "grad_norm": 2.624058961868286,
+      "learning_rate": 1.758732737611698e-06,
+      "loss": 1.7709,
+      "step": 1886
+    },
+    {
+      "epoch": 2.067945205479452,
+      "grad_norm": 2.848062515258789,
+      "learning_rate": 1.7567018683996753e-06,
+      "loss": 1.7513,
+      "step": 1887
+    },
+    {
+      "epoch": 2.069041095890411,
+      "grad_norm": 2.5219593048095703,
+      "learning_rate": 1.7546709991876524e-06,
+      "loss": 1.7387,
+      "step": 1888
+    },
+    {
+      "epoch": 2.0701369863013697,
+      "grad_norm": 2.5296876430511475,
+      "learning_rate": 1.7526401299756296e-06,
+      "loss": 1.7378,
+      "step": 1889
+    },
+    {
+      "epoch": 2.0712328767123287,
+      "grad_norm": 2.5452613830566406,
+      "learning_rate": 1.750609260763607e-06,
+      "loss": 1.7402,
+      "step": 1890
+    },
+    {
+      "epoch": 2.0712328767123287,
+      "eval_loss": 1.9914054870605469,
+      "eval_runtime": 37.0548,
+      "eval_samples_per_second": 26.987,
+      "eval_steps_per_second": 1.133,
+      "step": 1890
+    },
+    {
+      "epoch": 2.0723287671232877,
+      "grad_norm": 2.653596878051758,
+      "learning_rate": 1.7485783915515842e-06,
+      "loss": 1.7369,
+      "step": 1891
+    },
+    {
+      "epoch": 2.0734246575342468,
+      "grad_norm": 2.615581750869751,
+      "learning_rate": 1.7465475223395615e-06,
+      "loss": 1.7317,
+      "step": 1892
+    },
+    {
+      "epoch": 2.0745205479452054,
+      "grad_norm": 2.610304832458496,
+      "learning_rate": 1.7445166531275387e-06,
+      "loss": 1.7121,
+      "step": 1893
+    },
+    {
+      "epoch": 2.0756164383561644,
+      "grad_norm": 2.6463494300842285,
+      "learning_rate": 1.742485783915516e-06,
+      "loss": 1.8523,
+      "step": 1894
+    },
+    {
+      "epoch": 2.0767123287671234,
+      "grad_norm": 2.4814131259918213,
+      "learning_rate": 1.7404549147034933e-06,
+      "loss": 1.8239,
+      "step": 1895
+    },
+    {
+      "epoch": 2.077808219178082,
+      "grad_norm": 2.7416090965270996,
+      "learning_rate": 1.7384240454914706e-06,
+      "loss": 1.669,
+      "step": 1896
+    },
+    {
+      "epoch": 2.078904109589041,
+      "grad_norm": 2.9458305835723877,
+      "learning_rate": 1.7363931762794478e-06,
+      "loss": 1.7504,
+      "step": 1897
+    },
+    {
+      "epoch": 2.08,
+      "grad_norm": 2.5118932723999023,
+      "learning_rate": 1.7343623070674251e-06,
+      "loss": 1.7854,
+      "step": 1898
+    },
+    {
+      "epoch": 2.081095890410959,
+      "grad_norm": 2.479421615600586,
+      "learning_rate": 1.7323314378554024e-06,
+      "loss": 1.7325,
+      "step": 1899
+    },
+    {
+      "epoch": 2.0821917808219177,
+      "grad_norm": 2.7400424480438232,
+      "learning_rate": 1.7303005686433797e-06,
+      "loss": 1.7434,
+      "step": 1900
+    },
+    {
+      "epoch": 2.0821917808219177,
+      "eval_loss": 1.9887213706970215,
+      "eval_runtime": 37.0484,
+      "eval_samples_per_second": 26.992,
+      "eval_steps_per_second": 1.134,
+      "step": 1900
+    },
+    {
+      "epoch": 2.0832876712328767,
+      "grad_norm": 2.587714433670044,
+      "learning_rate": 1.7282696994313567e-06,
+      "loss": 1.7335,
+      "step": 1901
+    },
+    {
+      "epoch": 2.0843835616438358,
+      "grad_norm": 2.526818037033081,
+      "learning_rate": 1.726238830219334e-06,
+      "loss": 1.7538,
+      "step": 1902
+    },
+    {
+      "epoch": 2.0854794520547943,
+      "grad_norm": 2.595221519470215,
+      "learning_rate": 1.7242079610073113e-06,
+      "loss": 1.7331,
+      "step": 1903
+    },
+    {
+      "epoch": 2.0865753424657534,
+      "grad_norm": 2.5018110275268555,
+      "learning_rate": 1.7221770917952886e-06,
+      "loss": 1.8069,
+      "step": 1904
+    },
+    {
+      "epoch": 2.0876712328767124,
+      "grad_norm": 2.4378256797790527,
+      "learning_rate": 1.7201462225832658e-06,
+      "loss": 1.757,
+      "step": 1905
+    },
+    {
+      "epoch": 2.0887671232876714,
+      "grad_norm": 2.651167392730713,
+      "learning_rate": 1.7181153533712431e-06,
+      "loss": 1.7696,
+      "step": 1906
+    },
+    {
+      "epoch": 2.08986301369863,
+      "grad_norm": 2.550239086151123,
+      "learning_rate": 1.7160844841592204e-06,
+      "loss": 1.6551,
+      "step": 1907
+    },
+    {
+      "epoch": 2.090958904109589,
+      "grad_norm": 2.4192144870758057,
+      "learning_rate": 1.7140536149471977e-06,
+      "loss": 1.7031,
+      "step": 1908
+    },
+    {
+      "epoch": 2.092054794520548,
+      "grad_norm": 2.5327398777008057,
+      "learning_rate": 1.712022745735175e-06,
+      "loss": 1.734,
+      "step": 1909
+    },
+    {
+      "epoch": 2.0931506849315067,
+      "grad_norm": 2.5962865352630615,
+      "learning_rate": 1.7099918765231522e-06,
+      "loss": 1.732,
+      "step": 1910
+    },
+    {
+      "epoch": 2.0931506849315067,
+      "eval_loss": 1.9915482997894287,
+      "eval_runtime": 36.7856,
+      "eval_samples_per_second": 27.185,
+      "eval_steps_per_second": 1.142,
+      "step": 1910
+    },
+    {
+      "epoch": 2.0942465753424657,
+      "grad_norm": 2.561943531036377,
+      "learning_rate": 1.7079610073111295e-06,
+      "loss": 1.7556,
+      "step": 1911
+    },
+    {
+      "epoch": 2.0953424657534248,
+      "grad_norm": 2.529829263687134,
+      "learning_rate": 1.7059301380991068e-06,
+      "loss": 1.6839,
+      "step": 1912
+    },
+    {
+      "epoch": 2.0964383561643833,
+      "grad_norm": 2.636559009552002,
+      "learning_rate": 1.7038992688870836e-06,
+      "loss": 1.7393,
+      "step": 1913
+    },
+    {
+      "epoch": 2.0975342465753424,
+      "grad_norm": 2.558166742324829,
+      "learning_rate": 1.701868399675061e-06,
+      "loss": 1.7346,
+      "step": 1914
+    },
+    {
+      "epoch": 2.0986301369863014,
+      "grad_norm": 2.5193750858306885,
+      "learning_rate": 1.6998375304630382e-06,
+      "loss": 1.7666,
+      "step": 1915
+    },
+    {
+      "epoch": 2.0997260273972604,
+      "grad_norm": 2.498244047164917,
+      "learning_rate": 1.6978066612510155e-06,
+      "loss": 1.827,
+      "step": 1916
+    },
+    {
+      "epoch": 2.100821917808219,
+      "grad_norm": 2.799603223800659,
+      "learning_rate": 1.6957757920389927e-06,
+      "loss": 1.788,
+      "step": 1917
+    },
+    {
+      "epoch": 2.101917808219178,
+      "grad_norm": 2.733391761779785,
+      "learning_rate": 1.69374492282697e-06,
+      "loss": 1.7487,
+      "step": 1918
+    },
+    {
+      "epoch": 2.103013698630137,
+      "grad_norm": 2.5135183334350586,
+      "learning_rate": 1.6917140536149473e-06,
+      "loss": 1.7114,
+      "step": 1919
+    },
+    {
+      "epoch": 2.1041095890410957,
+      "grad_norm": 2.5275800228118896,
+      "learning_rate": 1.6896831844029246e-06,
+      "loss": 1.797,
+      "step": 1920
+    },
+    {
+      "epoch": 2.1041095890410957,
+      "eval_loss": 1.9926362037658691,
+      "eval_runtime": 36.9326,
+      "eval_samples_per_second": 27.076,
+      "eval_steps_per_second": 1.137,
+      "step": 1920
+    },
+    {
+      "epoch": 2.1052054794520547,
+      "grad_norm": 2.682774782180786,
+      "learning_rate": 1.6876523151909018e-06,
+      "loss": 1.7581,
+      "step": 1921
+    },
+    {
+      "epoch": 2.1063013698630138,
+      "grad_norm": 2.625620126724243,
+      "learning_rate": 1.6856214459788791e-06,
+      "loss": 1.7606,
+      "step": 1922
+    },
+    {
+      "epoch": 2.107397260273973,
+      "grad_norm": 2.5005173683166504,
+      "learning_rate": 1.6835905767668564e-06,
+      "loss": 1.7139,
+      "step": 1923
+    },
+    {
+      "epoch": 2.1084931506849314,
+      "grad_norm": 2.6731526851654053,
+      "learning_rate": 1.6815597075548337e-06,
+      "loss": 1.7174,
+      "step": 1924
+    },
+    {
+      "epoch": 2.1095890410958904,
+      "grad_norm": 2.5407724380493164,
+      "learning_rate": 1.679528838342811e-06,
+      "loss": 1.8049,
+      "step": 1925
+    },
+    {
+      "epoch": 2.1106849315068494,
+      "grad_norm": 2.502427101135254,
+      "learning_rate": 1.677497969130788e-06,
+      "loss": 1.7068,
+      "step": 1926
+    },
+    {
+      "epoch": 2.111780821917808,
+      "grad_norm": 2.5063414573669434,
+      "learning_rate": 1.6754670999187653e-06,
+      "loss": 1.7174,
+      "step": 1927
+    },
+    {
+      "epoch": 2.112876712328767,
+      "grad_norm": 2.524588108062744,
+      "learning_rate": 1.6734362307067426e-06,
+      "loss": 1.7442,
+      "step": 1928
+    },
+    {
+      "epoch": 2.113972602739726,
+      "grad_norm": 2.531909465789795,
+      "learning_rate": 1.6714053614947198e-06,
+      "loss": 1.692,
+      "step": 1929
+    },
+    {
+      "epoch": 2.115068493150685,
+      "grad_norm": 2.5104269981384277,
+      "learning_rate": 1.669374492282697e-06,
+      "loss": 1.7039,
+      "step": 1930
+    },
+    {
+      "epoch": 2.115068493150685,
+      "eval_loss": 1.9927221536636353,
+      "eval_runtime": 37.004,
+      "eval_samples_per_second": 27.024,
+      "eval_steps_per_second": 1.135,
+      "step": 1930
+    },
+    {
+      "epoch": 2.1161643835616437,
+      "grad_norm": 2.5443670749664307,
+      "learning_rate": 1.6673436230706744e-06,
+      "loss": 1.7896,
+      "step": 1931
+    },
+    {
+      "epoch": 2.1172602739726027,
+      "grad_norm": 2.6495745182037354,
+      "learning_rate": 1.6653127538586517e-06,
+      "loss": 1.7735,
+      "step": 1932
+    },
+    {
+      "epoch": 2.118356164383562,
+      "grad_norm": 2.7233121395111084,
+      "learning_rate": 1.663281884646629e-06,
+      "loss": 1.7186,
+      "step": 1933
+    },
+    {
+      "epoch": 2.1194520547945204,
+      "grad_norm": 2.4411816596984863,
+      "learning_rate": 1.6612510154346062e-06,
+      "loss": 1.7378,
+      "step": 1934
+    },
+    {
+      "epoch": 2.1205479452054794,
+      "grad_norm": 2.6478421688079834,
+      "learning_rate": 1.6592201462225835e-06,
+      "loss": 1.7989,
+      "step": 1935
+    },
+    {
+      "epoch": 2.1216438356164384,
+      "grad_norm": 2.56093168258667,
+      "learning_rate": 1.6571892770105608e-06,
+      "loss": 1.7663,
+      "step": 1936
+    },
+    {
+      "epoch": 2.1227397260273975,
+      "grad_norm": 2.5598859786987305,
+      "learning_rate": 1.655158407798538e-06,
+      "loss": 1.7956,
+      "step": 1937
+    },
+    {
+      "epoch": 2.123835616438356,
+      "grad_norm": 2.4958112239837646,
+      "learning_rate": 1.653127538586515e-06,
+      "loss": 1.7482,
+      "step": 1938
+    },
+    {
+      "epoch": 2.124931506849315,
+      "grad_norm": 2.5133280754089355,
+      "learning_rate": 1.6510966693744924e-06,
+      "loss": 1.7811,
+      "step": 1939
+    },
+    {
+      "epoch": 2.126027397260274,
+      "grad_norm": 2.492892026901245,
+      "learning_rate": 1.6490658001624696e-06,
+      "loss": 1.7883,
+      "step": 1940
+    },
+    {
+      "epoch": 2.126027397260274,
+      "eval_loss": 1.9917834997177124,
+      "eval_runtime": 37.0222,
+      "eval_samples_per_second": 27.011,
+      "eval_steps_per_second": 1.134,
+      "step": 1940
+    },
+    {
+      "epoch": 2.1271232876712327,
+      "grad_norm": 2.560483455657959,
+      "learning_rate": 1.647034930950447e-06,
+      "loss": 1.7871,
+      "step": 1941
+    },
+    {
+      "epoch": 2.1282191780821917,
+      "grad_norm": 2.549100160598755,
+      "learning_rate": 1.6450040617384242e-06,
+      "loss": 1.7097,
+      "step": 1942
+    },
+    {
+      "epoch": 2.1293150684931508,
+      "grad_norm": 2.4832775592803955,
+      "learning_rate": 1.6429731925264015e-06,
+      "loss": 1.736,
+      "step": 1943
+    },
+    {
+      "epoch": 2.1304109589041094,
+      "grad_norm": 2.60006046295166,
+      "learning_rate": 1.6409423233143788e-06,
+      "loss": 1.765,
+      "step": 1944
+    },
+    {
+      "epoch": 2.1315068493150684,
+      "grad_norm": 2.4875144958496094,
+      "learning_rate": 1.638911454102356e-06,
+      "loss": 1.7317,
+      "step": 1945
+    },
+    {
+      "epoch": 2.1326027397260274,
+      "grad_norm": 2.5562970638275146,
+      "learning_rate": 1.6368805848903333e-06,
+      "loss": 1.6982,
+      "step": 1946
+    },
+    {
+      "epoch": 2.1336986301369865,
+      "grad_norm": 2.4928972721099854,
+      "learning_rate": 1.6348497156783106e-06,
+      "loss": 1.7821,
+      "step": 1947
+    },
+    {
+      "epoch": 2.134794520547945,
+      "grad_norm": 2.6159043312072754,
+      "learning_rate": 1.6328188464662879e-06,
+      "loss": 1.7223,
+      "step": 1948
+    },
+    {
+      "epoch": 2.135890410958904,
+      "grad_norm": 2.541520833969116,
+      "learning_rate": 1.6307879772542651e-06,
+      "loss": 1.7321,
+      "step": 1949
+    },
+    {
+      "epoch": 2.136986301369863,
+      "grad_norm": 2.5165696144104004,
+      "learning_rate": 1.628757108042242e-06,
+      "loss": 1.7571,
+      "step": 1950
+    },
+    {
+      "epoch": 2.136986301369863,
+      "eval_loss": 1.9868476390838623,
+      "eval_runtime": 37.0367,
+      "eval_samples_per_second": 27.0,
+      "eval_steps_per_second": 1.134,
+      "step": 1950
+    },
+    {
+      "epoch": 2.138082191780822,
+      "grad_norm": 2.5660274028778076,
+      "learning_rate": 1.6267262388302193e-06,
+      "loss": 1.7855,
+      "step": 1951
+    },
+    {
+      "epoch": 2.1391780821917807,
+      "grad_norm": 2.4606971740722656,
+      "learning_rate": 1.6246953696181965e-06,
+      "loss": 1.7324,
+      "step": 1952
+    },
+    {
+      "epoch": 2.1402739726027398,
+      "grad_norm": 2.6528608798980713,
+      "learning_rate": 1.6226645004061738e-06,
+      "loss": 1.7715,
+      "step": 1953
+    },
+    {
+      "epoch": 2.141369863013699,
+      "grad_norm": 2.7232985496520996,
+      "learning_rate": 1.620633631194151e-06,
+      "loss": 1.837,
+      "step": 1954
+    },
+    {
+      "epoch": 2.1424657534246574,
+      "grad_norm": 2.6264185905456543,
+      "learning_rate": 1.6186027619821284e-06,
+      "loss": 1.7678,
+      "step": 1955
+    },
+    {
+      "epoch": 2.1435616438356164,
+      "grad_norm": 2.6248252391815186,
+      "learning_rate": 1.6165718927701056e-06,
+      "loss": 1.7218,
+      "step": 1956
+    },
+    {
+      "epoch": 2.1446575342465755,
+      "grad_norm": 3.00539493560791,
+      "learning_rate": 1.614541023558083e-06,
+      "loss": 1.7678,
+      "step": 1957
+    },
+    {
+      "epoch": 2.145753424657534,
+      "grad_norm": 2.5368528366088867,
+      "learning_rate": 1.6125101543460602e-06,
+      "loss": 1.7737,
+      "step": 1958
+    },
+    {
+      "epoch": 2.146849315068493,
+      "grad_norm": 2.5805916786193848,
+      "learning_rate": 1.6104792851340375e-06,
+      "loss": 1.7394,
+      "step": 1959
+    },
+    {
+      "epoch": 2.147945205479452,
+      "grad_norm": 2.615856647491455,
+      "learning_rate": 1.6084484159220147e-06,
+      "loss": 1.766,
+      "step": 1960
+    },
+    {
+      "epoch": 2.147945205479452,
+      "eval_loss": 1.987140417098999,
+      "eval_runtime": 37.0593,
+      "eval_samples_per_second": 26.984,
+      "eval_steps_per_second": 1.133,
+      "step": 1960
+    },
+    {
+      "epoch": 2.149041095890411,
+      "grad_norm": 2.6550803184509277,
+      "learning_rate": 1.606417546709992e-06,
+      "loss": 1.7342,
+      "step": 1961
+    },
+    {
+      "epoch": 2.1501369863013697,
+      "grad_norm": 2.5999467372894287,
+      "learning_rate": 1.6043866774979693e-06,
+      "loss": 1.7175,
+      "step": 1962
+    },
+    {
+      "epoch": 2.1512328767123288,
+      "grad_norm": 2.561642646789551,
+      "learning_rate": 1.6023558082859464e-06,
+      "loss": 1.7409,
+      "step": 1963
+    },
+    {
+      "epoch": 2.152328767123288,
+      "grad_norm": 2.6062211990356445,
+      "learning_rate": 1.6003249390739236e-06,
+      "loss": 1.7052,
+      "step": 1964
+    },
+    {
+      "epoch": 2.1534246575342464,
+      "grad_norm": 2.5122230052948,
+      "learning_rate": 1.598294069861901e-06,
+      "loss": 1.7444,
+      "step": 1965
+    },
+    {
+      "epoch": 2.1545205479452054,
+      "grad_norm": 2.5838866233825684,
+      "learning_rate": 1.5962632006498782e-06,
+      "loss": 1.716,
+      "step": 1966
+    },
+    {
+      "epoch": 2.1556164383561645,
+      "grad_norm": 2.5454747676849365,
+      "learning_rate": 1.5942323314378555e-06,
+      "loss": 1.7902,
+      "step": 1967
+    },
+    {
+      "epoch": 2.1567123287671235,
+      "grad_norm": 2.57637619972229,
+      "learning_rate": 1.5922014622258327e-06,
+      "loss": 1.6963,
+      "step": 1968
+    },
+    {
+      "epoch": 2.157808219178082,
+      "grad_norm": 2.5542941093444824,
+      "learning_rate": 1.59017059301381e-06,
+      "loss": 1.7151,
+      "step": 1969
+    },
+    {
+      "epoch": 2.158904109589041,
+      "grad_norm": 2.4912102222442627,
+      "learning_rate": 1.5881397238017873e-06,
+      "loss": 1.7624,
+      "step": 1970
+    },
+    {
+      "epoch": 2.158904109589041,
+      "eval_loss": 1.9855093955993652,
+      "eval_runtime": 37.0796,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 1970
+    },
+    {
+      "epoch": 2.16,
+      "grad_norm": 2.4793057441711426,
+      "learning_rate": 1.5861088545897646e-06,
+      "loss": 1.7617,
+      "step": 1971
+    },
+    {
+      "epoch": 2.1610958904109587,
+      "grad_norm": 2.5231704711914062,
+      "learning_rate": 1.5840779853777418e-06,
+      "loss": 1.6845,
+      "step": 1972
+    },
+    {
+      "epoch": 2.1621917808219178,
+      "grad_norm": 2.5126073360443115,
+      "learning_rate": 1.5820471161657191e-06,
+      "loss": 1.7468,
+      "step": 1973
+    },
+    {
+      "epoch": 2.163287671232877,
+      "grad_norm": 2.5558865070343018,
+      "learning_rate": 1.5800162469536964e-06,
+      "loss": 1.7253,
+      "step": 1974
+    },
+    {
+      "epoch": 2.1643835616438354,
+      "grad_norm": 2.808987855911255,
+      "learning_rate": 1.5779853777416735e-06,
+      "loss": 1.7844,
+      "step": 1975
+    },
+    {
+      "epoch": 2.1654794520547944,
+      "grad_norm": 2.765275001525879,
+      "learning_rate": 1.5759545085296507e-06,
+      "loss": 1.7658,
+      "step": 1976
+    },
+    {
+      "epoch": 2.1665753424657535,
+      "grad_norm": 2.5721709728240967,
+      "learning_rate": 1.573923639317628e-06,
+      "loss": 1.74,
+      "step": 1977
+    },
+    {
+      "epoch": 2.1676712328767125,
+      "grad_norm": 2.4276933670043945,
+      "learning_rate": 1.5718927701056053e-06,
+      "loss": 1.7613,
+      "step": 1978
+    },
+    {
+      "epoch": 2.168767123287671,
+      "grad_norm": 2.5665223598480225,
+      "learning_rate": 1.5698619008935826e-06,
+      "loss": 1.7514,
+      "step": 1979
+    },
+    {
+      "epoch": 2.16986301369863,
+      "grad_norm": 2.5588295459747314,
+      "learning_rate": 1.5678310316815598e-06,
+      "loss": 1.7749,
+      "step": 1980
+    },
+    {
+      "epoch": 2.16986301369863,
+      "eval_loss": 1.9858319759368896,
+      "eval_runtime": 37.0705,
+      "eval_samples_per_second": 26.976,
+      "eval_steps_per_second": 1.133,
+      "step": 1980
+    },
+    {
+      "epoch": 2.170958904109589,
+      "grad_norm": 2.5413172245025635,
+      "learning_rate": 1.5658001624695371e-06,
+      "loss": 1.706,
+      "step": 1981
+    },
+    {
+      "epoch": 2.172054794520548,
+      "grad_norm": 2.5246310234069824,
+      "learning_rate": 1.5637692932575144e-06,
+      "loss": 1.7825,
+      "step": 1982
+    },
+    {
+      "epoch": 2.1731506849315068,
+      "grad_norm": 2.385380506515503,
+      "learning_rate": 1.5617384240454917e-06,
+      "loss": 1.7268,
+      "step": 1983
+    },
+    {
+      "epoch": 2.174246575342466,
+      "grad_norm": 2.4967970848083496,
+      "learning_rate": 1.559707554833469e-06,
+      "loss": 1.7424,
+      "step": 1984
+    },
+    {
+      "epoch": 2.175342465753425,
+      "grad_norm": 2.563683271408081,
+      "learning_rate": 1.5576766856214462e-06,
+      "loss": 1.7904,
+      "step": 1985
+    },
+    {
+      "epoch": 2.1764383561643834,
+      "grad_norm": 2.518751859664917,
+      "learning_rate": 1.5556458164094235e-06,
+      "loss": 1.6616,
+      "step": 1986
+    },
+    {
+      "epoch": 2.1775342465753424,
+      "grad_norm": 2.5789380073547363,
+      "learning_rate": 1.5536149471974008e-06,
+      "loss": 1.7363,
+      "step": 1987
+    },
+    {
+      "epoch": 2.1786301369863015,
+      "grad_norm": 2.567406415939331,
+      "learning_rate": 1.5515840779853778e-06,
+      "loss": 1.6835,
+      "step": 1988
+    },
+    {
+      "epoch": 2.17972602739726,
+      "grad_norm": 2.509427309036255,
+      "learning_rate": 1.549553208773355e-06,
+      "loss": 1.7493,
+      "step": 1989
+    },
+    {
+      "epoch": 2.180821917808219,
+      "grad_norm": 2.6355340480804443,
+      "learning_rate": 1.5475223395613324e-06,
+      "loss": 1.7292,
+      "step": 1990
+    },
+    {
+      "epoch": 2.180821917808219,
+      "eval_loss": 1.9893935918807983,
+      "eval_runtime": 37.0928,
+      "eval_samples_per_second": 26.959,
+      "eval_steps_per_second": 1.132,
+      "step": 1990
+    },
+    {
+      "epoch": 2.181917808219178,
+      "grad_norm": 2.564587116241455,
+      "learning_rate": 1.5454914703493097e-06,
+      "loss": 1.7248,
+      "step": 1991
+    },
+    {
+      "epoch": 2.183013698630137,
+      "grad_norm": 2.6708462238311768,
+      "learning_rate": 1.543460601137287e-06,
+      "loss": 1.8123,
+      "step": 1992
+    },
+    {
+      "epoch": 2.1841095890410958,
+      "grad_norm": 2.6396312713623047,
+      "learning_rate": 1.5414297319252642e-06,
+      "loss": 1.7742,
+      "step": 1993
+    },
+    {
+      "epoch": 2.185205479452055,
+      "grad_norm": 2.553544044494629,
+      "learning_rate": 1.5393988627132415e-06,
+      "loss": 1.7705,
+      "step": 1994
+    },
+    {
+      "epoch": 2.186301369863014,
+      "grad_norm": 2.554997682571411,
+      "learning_rate": 1.5373679935012188e-06,
+      "loss": 1.737,
+      "step": 1995
+    },
+    {
+      "epoch": 2.1873972602739724,
+      "grad_norm": 2.5650475025177,
+      "learning_rate": 1.535337124289196e-06,
+      "loss": 1.6865,
+      "step": 1996
+    },
+    {
+      "epoch": 2.1884931506849314,
+      "grad_norm": 2.8019111156463623,
+      "learning_rate": 1.5333062550771733e-06,
+      "loss": 1.7428,
+      "step": 1997
+    },
+    {
+      "epoch": 2.1895890410958905,
+      "grad_norm": 2.5704092979431152,
+      "learning_rate": 1.5312753858651506e-06,
+      "loss": 1.8291,
+      "step": 1998
+    },
+    {
+      "epoch": 2.1906849315068495,
+      "grad_norm": 2.5332226753234863,
+      "learning_rate": 1.5292445166531279e-06,
+      "loss": 1.7274,
+      "step": 1999
+    },
+    {
+      "epoch": 2.191780821917808,
+      "grad_norm": 2.555288553237915,
+      "learning_rate": 1.5272136474411047e-06,
+      "loss": 1.7497,
+      "step": 2000
+    },
+    {
+      "epoch": 2.191780821917808,
+      "eval_loss": 1.9881768226623535,
+      "eval_runtime": 37.0996,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 2000
+    },
+    {
+      "epoch": 2.192876712328767,
+      "grad_norm": 2.6068379878997803,
+      "learning_rate": 1.525182778229082e-06,
+      "loss": 1.7955,
+      "step": 2001
+    },
+    {
+      "epoch": 2.193972602739726,
+      "grad_norm": 2.479121208190918,
+      "learning_rate": 1.5231519090170593e-06,
+      "loss": 1.7342,
+      "step": 2002
+    },
+    {
+      "epoch": 2.1950684931506848,
+      "grad_norm": 2.495670795440674,
+      "learning_rate": 1.5211210398050365e-06,
+      "loss": 1.7125,
+      "step": 2003
+    },
+    {
+      "epoch": 2.196164383561644,
+      "grad_norm": 2.4988651275634766,
+      "learning_rate": 1.5190901705930138e-06,
+      "loss": 1.7137,
+      "step": 2004
+    },
+    {
+      "epoch": 2.197260273972603,
+      "grad_norm": 2.5191574096679688,
+      "learning_rate": 1.517059301380991e-06,
+      "loss": 1.7672,
+      "step": 2005
+    },
+    {
+      "epoch": 2.198356164383562,
+      "grad_norm": 2.5469892024993896,
+      "learning_rate": 1.5150284321689684e-06,
+      "loss": 1.7127,
+      "step": 2006
+    },
+    {
+      "epoch": 2.1994520547945204,
+      "grad_norm": 2.490896701812744,
+      "learning_rate": 1.5129975629569456e-06,
+      "loss": 1.7397,
+      "step": 2007
+    },
+    {
+      "epoch": 2.2005479452054795,
+      "grad_norm": 2.5142805576324463,
+      "learning_rate": 1.510966693744923e-06,
+      "loss": 1.7431,
+      "step": 2008
+    },
+    {
+      "epoch": 2.2016438356164385,
+      "grad_norm": 2.5397887229919434,
+      "learning_rate": 1.5089358245329002e-06,
+      "loss": 1.7439,
+      "step": 2009
+    },
+    {
+      "epoch": 2.202739726027397,
+      "grad_norm": 2.5063931941986084,
+      "learning_rate": 1.5069049553208775e-06,
+      "loss": 1.7259,
+      "step": 2010
+    },
+    {
+      "epoch": 2.202739726027397,
+      "eval_loss": 1.9853808879852295,
+      "eval_runtime": 36.8066,
+      "eval_samples_per_second": 27.169,
+      "eval_steps_per_second": 1.141,
+      "step": 2010
+    },
+    {
+      "epoch": 2.203835616438356,
+      "grad_norm": 2.6017589569091797,
+      "learning_rate": 1.5048740861088547e-06,
+      "loss": 1.7714,
+      "step": 2011
+    },
+    {
+      "epoch": 2.204931506849315,
+      "grad_norm": 2.5295040607452393,
+      "learning_rate": 1.502843216896832e-06,
+      "loss": 1.7561,
+      "step": 2012
+    },
+    {
+      "epoch": 2.206027397260274,
+      "grad_norm": 2.5187482833862305,
+      "learning_rate": 1.500812347684809e-06,
+      "loss": 1.6761,
+      "step": 2013
+    },
+    {
+      "epoch": 2.207123287671233,
+      "grad_norm": 2.5163650512695312,
+      "learning_rate": 1.4987814784727864e-06,
+      "loss": 1.7469,
+      "step": 2014
+    },
+    {
+      "epoch": 2.208219178082192,
+      "grad_norm": 2.731238603591919,
+      "learning_rate": 1.4967506092607636e-06,
+      "loss": 1.7527,
+      "step": 2015
+    },
+    {
+      "epoch": 2.209315068493151,
+      "grad_norm": 2.5298311710357666,
+      "learning_rate": 1.494719740048741e-06,
+      "loss": 1.6912,
+      "step": 2016
+    },
+    {
+      "epoch": 2.2104109589041094,
+      "grad_norm": 2.6385974884033203,
+      "learning_rate": 1.4926888708367182e-06,
+      "loss": 1.7664,
+      "step": 2017
+    },
+    {
+      "epoch": 2.2115068493150685,
+      "grad_norm": 2.6150033473968506,
+      "learning_rate": 1.4906580016246955e-06,
+      "loss": 1.7708,
+      "step": 2018
+    },
+    {
+      "epoch": 2.2126027397260275,
+      "grad_norm": 2.589329957962036,
+      "learning_rate": 1.4886271324126727e-06,
+      "loss": 1.7875,
+      "step": 2019
+    },
+    {
+      "epoch": 2.213698630136986,
+      "grad_norm": 2.6769933700561523,
+      "learning_rate": 1.48659626320065e-06,
+      "loss": 1.8136,
+      "step": 2020
+    },
+    {
+      "epoch": 2.213698630136986,
+      "eval_loss": 1.9827086925506592,
+      "eval_runtime": 37.0026,
+      "eval_samples_per_second": 27.025,
+      "eval_steps_per_second": 1.135,
+      "step": 2020
+    },
+    {
+      "epoch": 2.214794520547945,
+      "grad_norm": 2.7220497131347656,
+      "learning_rate": 1.4845653939886273e-06,
+      "loss": 1.6627,
+      "step": 2021
+    },
+    {
+      "epoch": 2.215890410958904,
+      "grad_norm": 2.6115968227386475,
+      "learning_rate": 1.4825345247766046e-06,
+      "loss": 1.6676,
+      "step": 2022
+    },
+    {
+      "epoch": 2.216986301369863,
+      "grad_norm": 2.4612581729888916,
+      "learning_rate": 1.4805036555645818e-06,
+      "loss": 1.6644,
+      "step": 2023
+    },
+    {
+      "epoch": 2.2180821917808218,
+      "grad_norm": 2.5708696842193604,
+      "learning_rate": 1.4784727863525591e-06,
+      "loss": 1.7357,
+      "step": 2024
+    },
+    {
+      "epoch": 2.219178082191781,
+      "grad_norm": 2.578158140182495,
+      "learning_rate": 1.4764419171405362e-06,
+      "loss": 1.7011,
+      "step": 2025
+    },
+    {
+      "epoch": 2.22027397260274,
+      "grad_norm": 2.547666072845459,
+      "learning_rate": 1.4744110479285135e-06,
+      "loss": 1.7787,
+      "step": 2026
+    },
+    {
+      "epoch": 2.2213698630136984,
+      "grad_norm": 2.617551326751709,
+      "learning_rate": 1.4723801787164907e-06,
+      "loss": 1.784,
+      "step": 2027
+    },
+    {
+      "epoch": 2.2224657534246575,
+      "grad_norm": 2.4385478496551514,
+      "learning_rate": 1.470349309504468e-06,
+      "loss": 1.7764,
+      "step": 2028
+    },
+    {
+      "epoch": 2.2235616438356165,
+      "grad_norm": 2.6088693141937256,
+      "learning_rate": 1.4683184402924453e-06,
+      "loss": 1.7477,
+      "step": 2029
+    },
+    {
+      "epoch": 2.2246575342465755,
+      "grad_norm": 2.6165895462036133,
+      "learning_rate": 1.4662875710804226e-06,
+      "loss": 1.7531,
+      "step": 2030
+    },
+    {
+      "epoch": 2.2246575342465755,
+      "eval_loss": 1.9826008081436157,
+      "eval_runtime": 37.085,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 2030
+    },
+    {
+      "epoch": 2.225753424657534,
+      "grad_norm": 2.6285006999969482,
+      "learning_rate": 1.4642567018683998e-06,
+      "loss": 1.8403,
+      "step": 2031
+    },
+    {
+      "epoch": 2.226849315068493,
+      "grad_norm": 2.549326181411743,
+      "learning_rate": 1.4622258326563771e-06,
+      "loss": 1.7304,
+      "step": 2032
+    },
+    {
+      "epoch": 2.227945205479452,
+      "grad_norm": 2.6198782920837402,
+      "learning_rate": 1.4601949634443544e-06,
+      "loss": 1.7147,
+      "step": 2033
+    },
+    {
+      "epoch": 2.2290410958904108,
+      "grad_norm": 2.536947727203369,
+      "learning_rate": 1.4581640942323317e-06,
+      "loss": 1.7639,
+      "step": 2034
+    },
+    {
+      "epoch": 2.23013698630137,
+      "grad_norm": 2.6757264137268066,
+      "learning_rate": 1.456133225020309e-06,
+      "loss": 1.8434,
+      "step": 2035
+    },
+    {
+      "epoch": 2.231232876712329,
+      "grad_norm": 2.598095655441284,
+      "learning_rate": 1.4541023558082862e-06,
+      "loss": 1.7593,
+      "step": 2036
+    },
+    {
+      "epoch": 2.232328767123288,
+      "grad_norm": 2.6104440689086914,
+      "learning_rate": 1.4520714865962635e-06,
+      "loss": 1.7473,
+      "step": 2037
+    },
+    {
+      "epoch": 2.2334246575342465,
+      "grad_norm": 2.6350162029266357,
+      "learning_rate": 1.4500406173842406e-06,
+      "loss": 1.777,
+      "step": 2038
+    },
+    {
+      "epoch": 2.2345205479452055,
+      "grad_norm": 2.593874216079712,
+      "learning_rate": 1.4480097481722178e-06,
+      "loss": 1.7506,
+      "step": 2039
+    },
+    {
+      "epoch": 2.2356164383561645,
+      "grad_norm": 2.6991093158721924,
+      "learning_rate": 1.4459788789601951e-06,
+      "loss": 1.7375,
+      "step": 2040
+    },
+    {
+      "epoch": 2.2356164383561645,
+      "eval_loss": 1.9815349578857422,
+      "eval_runtime": 37.1126,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 2040
+    },
+    {
+      "epoch": 2.236712328767123,
+      "grad_norm": 2.6339058876037598,
+      "learning_rate": 1.4439480097481724e-06,
+      "loss": 1.7045,
+      "step": 2041
+    },
+    {
+      "epoch": 2.237808219178082,
+      "grad_norm": 2.691631555557251,
+      "learning_rate": 1.4419171405361497e-06,
+      "loss": 1.7063,
+      "step": 2042
+    },
+    {
+      "epoch": 2.238904109589041,
+      "grad_norm": 3.021667003631592,
+      "learning_rate": 1.439886271324127e-06,
+      "loss": 1.7607,
+      "step": 2043
+    },
+    {
+      "epoch": 2.24,
+      "grad_norm": 2.690802812576294,
+      "learning_rate": 1.4378554021121042e-06,
+      "loss": 1.7594,
+      "step": 2044
+    },
+    {
+      "epoch": 2.241095890410959,
+      "grad_norm": 2.5371177196502686,
+      "learning_rate": 1.4358245329000815e-06,
+      "loss": 1.7177,
+      "step": 2045
+    },
+    {
+      "epoch": 2.242191780821918,
+      "grad_norm": 2.5586729049682617,
+      "learning_rate": 1.4337936636880588e-06,
+      "loss": 1.7864,
+      "step": 2046
+    },
+    {
+      "epoch": 2.243287671232877,
+      "grad_norm": 2.6621484756469727,
+      "learning_rate": 1.431762794476036e-06,
+      "loss": 1.7616,
+      "step": 2047
+    },
+    {
+      "epoch": 2.2443835616438355,
+      "grad_norm": 2.787508010864258,
+      "learning_rate": 1.4297319252640133e-06,
+      "loss": 1.7056,
+      "step": 2048
+    },
+    {
+      "epoch": 2.2454794520547945,
+      "grad_norm": 2.6896064281463623,
+      "learning_rate": 1.4277010560519906e-06,
+      "loss": 1.7413,
+      "step": 2049
+    },
+    {
+      "epoch": 2.2465753424657535,
+      "grad_norm": 2.540524959564209,
+      "learning_rate": 1.4256701868399674e-06,
+      "loss": 1.7096,
+      "step": 2050
+    },
+    {
+      "epoch": 2.2465753424657535,
+      "eval_loss": 1.9832353591918945,
+      "eval_runtime": 37.1202,
+      "eval_samples_per_second": 26.94,
+      "eval_steps_per_second": 1.131,
+      "step": 2050
+    },
+    {
+      "epoch": 2.247671232876712,
+      "grad_norm": 2.582171678543091,
+      "learning_rate": 1.4236393176279447e-06,
+      "loss": 1.8262,
+      "step": 2051
+    },
+    {
+      "epoch": 2.248767123287671,
+      "grad_norm": 2.659044027328491,
+      "learning_rate": 1.421608448415922e-06,
+      "loss": 1.7458,
+      "step": 2052
+    },
+    {
+      "epoch": 2.24986301369863,
+      "grad_norm": 2.7265775203704834,
+      "learning_rate": 1.4195775792038993e-06,
+      "loss": 1.8341,
+      "step": 2053
+    },
+    {
+      "epoch": 2.250958904109589,
+      "grad_norm": 2.922330379486084,
+      "learning_rate": 1.4175467099918765e-06,
+      "loss": 1.7103,
+      "step": 2054
+    },
+    {
+      "epoch": 2.252054794520548,
+      "grad_norm": 2.4665236473083496,
+      "learning_rate": 1.4155158407798538e-06,
+      "loss": 1.7461,
+      "step": 2055
+    },
+    {
+      "epoch": 2.253150684931507,
+      "grad_norm": 2.639538049697876,
+      "learning_rate": 1.413484971567831e-06,
+      "loss": 1.7683,
+      "step": 2056
+    },
+    {
+      "epoch": 2.254246575342466,
+      "grad_norm": 2.558363676071167,
+      "learning_rate": 1.4114541023558084e-06,
+      "loss": 1.6821,
+      "step": 2057
+    },
+    {
+      "epoch": 2.255342465753425,
+      "grad_norm": 2.6308321952819824,
+      "learning_rate": 1.4094232331437856e-06,
+      "loss": 1.7226,
+      "step": 2058
+    },
+    {
+      "epoch": 2.2564383561643835,
+      "grad_norm": 2.6325299739837646,
+      "learning_rate": 1.407392363931763e-06,
+      "loss": 1.783,
+      "step": 2059
+    },
+    {
+      "epoch": 2.2575342465753425,
+      "grad_norm": 2.6024723052978516,
+      "learning_rate": 1.4053614947197402e-06,
+      "loss": 1.8402,
+      "step": 2060
+    },
+    {
+      "epoch": 2.2575342465753425,
+      "eval_loss": 1.983332633972168,
+      "eval_runtime": 37.1133,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 2060
+    },
+    {
+      "epoch": 2.2586301369863016,
+      "grad_norm": 2.548522710800171,
+      "learning_rate": 1.4033306255077175e-06,
+      "loss": 1.7625,
+      "step": 2061
+    },
+    {
+      "epoch": 2.25972602739726,
+      "grad_norm": 2.5953831672668457,
+      "learning_rate": 1.4012997562956945e-06,
+      "loss": 1.6632,
+      "step": 2062
+    },
+    {
+      "epoch": 2.260821917808219,
+      "grad_norm": 2.5880215167999268,
+      "learning_rate": 1.3992688870836718e-06,
+      "loss": 1.7957,
+      "step": 2063
+    },
+    {
+      "epoch": 2.261917808219178,
+      "grad_norm": 2.439326047897339,
+      "learning_rate": 1.397238017871649e-06,
+      "loss": 1.76,
+      "step": 2064
+    },
+    {
+      "epoch": 2.263013698630137,
+      "grad_norm": 2.5185437202453613,
+      "learning_rate": 1.3952071486596264e-06,
+      "loss": 1.7385,
+      "step": 2065
+    },
+    {
+      "epoch": 2.264109589041096,
+      "grad_norm": 2.5943732261657715,
+      "learning_rate": 1.3931762794476036e-06,
+      "loss": 1.7066,
+      "step": 2066
+    },
+    {
+      "epoch": 2.265205479452055,
+      "grad_norm": 2.6375551223754883,
+      "learning_rate": 1.391145410235581e-06,
+      "loss": 1.7884,
+      "step": 2067
+    },
+    {
+      "epoch": 2.266301369863014,
+      "grad_norm": 2.4962515830993652,
+      "learning_rate": 1.3891145410235582e-06,
+      "loss": 1.7295,
+      "step": 2068
+    },
+    {
+      "epoch": 2.2673972602739725,
+      "grad_norm": 2.5873053073883057,
+      "learning_rate": 1.3870836718115355e-06,
+      "loss": 1.6697,
+      "step": 2069
+    },
+    {
+      "epoch": 2.2684931506849315,
+      "grad_norm": 2.573211908340454,
+      "learning_rate": 1.3850528025995127e-06,
+      "loss": 1.7198,
+      "step": 2070
+    },
+    {
+      "epoch": 2.2684931506849315,
+      "eval_loss": 1.9826557636260986,
+      "eval_runtime": 37.1271,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 2070
+    },
+    {
+      "epoch": 2.2695890410958905,
+      "grad_norm": 2.68015718460083,
+      "learning_rate": 1.38302193338749e-06,
+      "loss": 1.7294,
+      "step": 2071
+    },
+    {
+      "epoch": 2.270684931506849,
+      "grad_norm": 2.4882888793945312,
+      "learning_rate": 1.3809910641754673e-06,
+      "loss": 1.6592,
+      "step": 2072
+    },
+    {
+      "epoch": 2.271780821917808,
+      "grad_norm": 2.587048053741455,
+      "learning_rate": 1.3789601949634446e-06,
+      "loss": 1.7605,
+      "step": 2073
+    },
+    {
+      "epoch": 2.272876712328767,
+      "grad_norm": 2.6452062129974365,
+      "learning_rate": 1.3769293257514218e-06,
+      "loss": 1.6664,
+      "step": 2074
+    },
+    {
+      "epoch": 2.2739726027397262,
+      "grad_norm": 2.5663232803344727,
+      "learning_rate": 1.374898456539399e-06,
+      "loss": 1.771,
+      "step": 2075
+    },
+    {
+      "epoch": 2.275068493150685,
+      "grad_norm": 2.652989149093628,
+      "learning_rate": 1.3728675873273762e-06,
+      "loss": 1.7096,
+      "step": 2076
+    },
+    {
+      "epoch": 2.276164383561644,
+      "grad_norm": 2.538435220718384,
+      "learning_rate": 1.3708367181153535e-06,
+      "loss": 1.798,
+      "step": 2077
+    },
+    {
+      "epoch": 2.277260273972603,
+      "grad_norm": 2.4685096740722656,
+      "learning_rate": 1.3688058489033307e-06,
+      "loss": 1.7904,
+      "step": 2078
+    },
+    {
+      "epoch": 2.2783561643835615,
+      "grad_norm": 2.566723108291626,
+      "learning_rate": 1.366774979691308e-06,
+      "loss": 1.7569,
+      "step": 2079
+    },
+    {
+      "epoch": 2.2794520547945205,
+      "grad_norm": 2.560960531234741,
+      "learning_rate": 1.3647441104792853e-06,
+      "loss": 1.7734,
+      "step": 2080
+    },
+    {
+      "epoch": 2.2794520547945205,
+      "eval_loss": 1.9809880256652832,
+      "eval_runtime": 37.1137,
+      "eval_samples_per_second": 26.944,
+      "eval_steps_per_second": 1.132,
+      "step": 2080
+    },
+    {
+      "epoch": 2.2805479452054795,
+      "grad_norm": 2.576608419418335,
+      "learning_rate": 1.3627132412672626e-06,
+      "loss": 1.7431,
+      "step": 2081
+    },
+    {
+      "epoch": 2.281643835616438,
+      "grad_norm": 2.648301601409912,
+      "learning_rate": 1.3606823720552398e-06,
+      "loss": 1.6959,
+      "step": 2082
+    },
+    {
+      "epoch": 2.282739726027397,
+      "grad_norm": 2.516493082046509,
+      "learning_rate": 1.3586515028432171e-06,
+      "loss": 1.7519,
+      "step": 2083
+    },
+    {
+      "epoch": 2.283835616438356,
+      "grad_norm": 2.474616050720215,
+      "learning_rate": 1.3566206336311944e-06,
+      "loss": 1.6796,
+      "step": 2084
+    },
+    {
+      "epoch": 2.2849315068493152,
+      "grad_norm": 2.482044219970703,
+      "learning_rate": 1.3545897644191717e-06,
+      "loss": 1.7443,
+      "step": 2085
+    },
+    {
+      "epoch": 2.286027397260274,
+      "grad_norm": 2.6574456691741943,
+      "learning_rate": 1.352558895207149e-06,
+      "loss": 1.7145,
+      "step": 2086
+    },
+    {
+      "epoch": 2.287123287671233,
+      "grad_norm": 2.5350115299224854,
+      "learning_rate": 1.3505280259951258e-06,
+      "loss": 1.765,
+      "step": 2087
+    },
+    {
+      "epoch": 2.288219178082192,
+      "grad_norm": 2.6488730907440186,
+      "learning_rate": 1.348497156783103e-06,
+      "loss": 1.7407,
+      "step": 2088
+    },
+    {
+      "epoch": 2.289315068493151,
+      "grad_norm": 2.6020500659942627,
+      "learning_rate": 1.3464662875710804e-06,
+      "loss": 1.779,
+      "step": 2089
+    },
+    {
+      "epoch": 2.2904109589041095,
+      "grad_norm": 2.633013963699341,
+      "learning_rate": 1.3444354183590576e-06,
+      "loss": 1.6824,
+      "step": 2090
+    },
+    {
+      "epoch": 2.2904109589041095,
+      "eval_loss": 1.982380986213684,
+      "eval_runtime": 37.1166,
+      "eval_samples_per_second": 26.942,
+      "eval_steps_per_second": 1.132,
+      "step": 2090
+    },
+    {
+      "epoch": 2.2915068493150685,
+      "grad_norm": 2.6399149894714355,
+      "learning_rate": 1.342404549147035e-06,
+      "loss": 1.7634,
+      "step": 2091
+    },
+    {
+      "epoch": 2.2926027397260276,
+      "grad_norm": 2.580634832382202,
+      "learning_rate": 1.3403736799350122e-06,
+      "loss": 1.7099,
+      "step": 2092
+    },
+    {
+      "epoch": 2.293698630136986,
+      "grad_norm": 2.523400068283081,
+      "learning_rate": 1.3383428107229895e-06,
+      "loss": 1.7705,
+      "step": 2093
+    },
+    {
+      "epoch": 2.294794520547945,
+      "grad_norm": 2.5114550590515137,
+      "learning_rate": 1.3363119415109667e-06,
+      "loss": 1.7961,
+      "step": 2094
+    },
+    {
+      "epoch": 2.2958904109589042,
+      "grad_norm": 2.457369804382324,
+      "learning_rate": 1.334281072298944e-06,
+      "loss": 1.7507,
+      "step": 2095
+    },
+    {
+      "epoch": 2.296986301369863,
+      "grad_norm": 2.5671467781066895,
+      "learning_rate": 1.3322502030869213e-06,
+      "loss": 1.7607,
+      "step": 2096
+    },
+    {
+      "epoch": 2.298082191780822,
+      "grad_norm": 2.459540367126465,
+      "learning_rate": 1.3302193338748986e-06,
+      "loss": 1.7193,
+      "step": 2097
+    },
+    {
+      "epoch": 2.299178082191781,
+      "grad_norm": 2.4934308528900146,
+      "learning_rate": 1.3281884646628758e-06,
+      "loss": 1.7775,
+      "step": 2098
+    },
+    {
+      "epoch": 2.30027397260274,
+      "grad_norm": 2.520641565322876,
+      "learning_rate": 1.3261575954508531e-06,
+      "loss": 1.7847,
+      "step": 2099
+    },
+    {
+      "epoch": 2.3013698630136985,
+      "grad_norm": 2.4607160091400146,
+      "learning_rate": 1.3241267262388302e-06,
+      "loss": 1.7812,
+      "step": 2100
+    },
+    {
+      "epoch": 2.3013698630136985,
+      "eval_loss": 1.9836152791976929,
+      "eval_runtime": 37.1239,
+      "eval_samples_per_second": 26.937,
+      "eval_steps_per_second": 1.131,
+      "step": 2100
+    },
+    {
+      "epoch": 2.3024657534246575,
+      "grad_norm": 2.577761650085449,
+      "learning_rate": 1.3220958570268074e-06,
+      "loss": 1.7452,
+      "step": 2101
+    },
+    {
+      "epoch": 2.3035616438356166,
+      "grad_norm": 2.5331356525421143,
+      "learning_rate": 1.3200649878147847e-06,
+      "loss": 1.6989,
+      "step": 2102
+    },
+    {
+      "epoch": 2.304657534246575,
+      "grad_norm": 2.632581949234009,
+      "learning_rate": 1.318034118602762e-06,
+      "loss": 1.7368,
+      "step": 2103
+    },
+    {
+      "epoch": 2.305753424657534,
+      "grad_norm": 2.55940318107605,
+      "learning_rate": 1.3160032493907393e-06,
+      "loss": 1.7806,
+      "step": 2104
+    },
+    {
+      "epoch": 2.3068493150684932,
+      "grad_norm": 2.5680153369903564,
+      "learning_rate": 1.3139723801787166e-06,
+      "loss": 1.7482,
+      "step": 2105
+    },
+    {
+      "epoch": 2.3079452054794523,
+      "grad_norm": 2.7915756702423096,
+      "learning_rate": 1.3119415109666938e-06,
+      "loss": 1.832,
+      "step": 2106
+    },
+    {
+      "epoch": 2.309041095890411,
+      "grad_norm": 2.518031597137451,
+      "learning_rate": 1.309910641754671e-06,
+      "loss": 1.7401,
+      "step": 2107
+    },
+    {
+      "epoch": 2.31013698630137,
+      "grad_norm": 2.7410221099853516,
+      "learning_rate": 1.3078797725426484e-06,
+      "loss": 1.6992,
+      "step": 2108
+    },
+    {
+      "epoch": 2.311232876712329,
+      "grad_norm": 2.611825466156006,
+      "learning_rate": 1.3058489033306257e-06,
+      "loss": 1.7085,
+      "step": 2109
+    },
+    {
+      "epoch": 2.3123287671232875,
+      "grad_norm": 2.5833330154418945,
+      "learning_rate": 1.303818034118603e-06,
+      "loss": 1.7129,
+      "step": 2110
+    },
+    {
+      "epoch": 2.3123287671232875,
+      "eval_loss": 1.9824191331863403,
+      "eval_runtime": 36.773,
+      "eval_samples_per_second": 27.194,
+      "eval_steps_per_second": 1.142,
+      "step": 2110
+    },
+    {
+      "epoch": 2.3134246575342465,
+      "grad_norm": 2.587327003479004,
+      "learning_rate": 1.3017871649065802e-06,
+      "loss": 1.7439,
+      "step": 2111
+    },
+    {
+      "epoch": 2.3145205479452056,
+      "grad_norm": 2.574925422668457,
+      "learning_rate": 1.2997562956945573e-06,
+      "loss": 1.7173,
+      "step": 2112
+    },
+    {
+      "epoch": 2.315616438356164,
+      "grad_norm": 2.6404287815093994,
+      "learning_rate": 1.2977254264825345e-06,
+      "loss": 1.7052,
+      "step": 2113
+    },
+    {
+      "epoch": 2.316712328767123,
+      "grad_norm": 2.634584903717041,
+      "learning_rate": 1.2956945572705118e-06,
+      "loss": 1.741,
+      "step": 2114
+    },
+    {
+      "epoch": 2.317808219178082,
+      "grad_norm": 2.66159987449646,
+      "learning_rate": 1.293663688058489e-06,
+      "loss": 1.76,
+      "step": 2115
+    },
+    {
+      "epoch": 2.3189041095890413,
+      "grad_norm": 2.6337332725524902,
+      "learning_rate": 1.2916328188464664e-06,
+      "loss": 1.7035,
+      "step": 2116
+    },
+    {
+      "epoch": 2.32,
+      "grad_norm": 2.528524398803711,
+      "learning_rate": 1.2896019496344436e-06,
+      "loss": 1.6439,
+      "step": 2117
+    },
+    {
+      "epoch": 2.321095890410959,
+      "grad_norm": 2.5938384532928467,
+      "learning_rate": 1.287571080422421e-06,
+      "loss": 1.7134,
+      "step": 2118
+    },
+    {
+      "epoch": 2.322191780821918,
+      "grad_norm": 2.606271505355835,
+      "learning_rate": 1.2855402112103982e-06,
+      "loss": 1.7758,
+      "step": 2119
+    },
+    {
+      "epoch": 2.323287671232877,
+      "grad_norm": 2.530071258544922,
+      "learning_rate": 1.2835093419983755e-06,
+      "loss": 1.6927,
+      "step": 2120
+    },
+    {
+      "epoch": 2.323287671232877,
+      "eval_loss": 1.9856306314468384,
+      "eval_runtime": 36.9674,
+      "eval_samples_per_second": 27.051,
+      "eval_steps_per_second": 1.136,
+      "step": 2120
+    },
+    {
+      "epoch": 2.3243835616438355,
+      "grad_norm": 2.8125321865081787,
+      "learning_rate": 1.2814784727863528e-06,
+      "loss": 1.7225,
+      "step": 2121
+    },
+    {
+      "epoch": 2.3254794520547946,
+      "grad_norm": 2.57012939453125,
+      "learning_rate": 1.27944760357433e-06,
+      "loss": 1.7177,
+      "step": 2122
+    },
+    {
+      "epoch": 2.3265753424657536,
+      "grad_norm": 2.6595780849456787,
+      "learning_rate": 1.2774167343623073e-06,
+      "loss": 1.7807,
+      "step": 2123
+    },
+    {
+      "epoch": 2.327671232876712,
+      "grad_norm": 2.557115077972412,
+      "learning_rate": 1.2753858651502846e-06,
+      "loss": 1.7762,
+      "step": 2124
+    },
+    {
+      "epoch": 2.328767123287671,
+      "grad_norm": 2.763218879699707,
+      "learning_rate": 1.2733549959382616e-06,
+      "loss": 1.7638,
+      "step": 2125
+    },
+    {
+      "epoch": 2.3298630136986302,
+      "grad_norm": 2.7284505367279053,
+      "learning_rate": 1.271324126726239e-06,
+      "loss": 1.7387,
+      "step": 2126
+    },
+    {
+      "epoch": 2.330958904109589,
+      "grad_norm": 2.5893704891204834,
+      "learning_rate": 1.2692932575142162e-06,
+      "loss": 1.7322,
+      "step": 2127
+    },
+    {
+      "epoch": 2.332054794520548,
+      "grad_norm": 2.625582456588745,
+      "learning_rate": 1.2672623883021935e-06,
+      "loss": 1.6865,
+      "step": 2128
+    },
+    {
+      "epoch": 2.333150684931507,
+      "grad_norm": 2.5458641052246094,
+      "learning_rate": 1.2652315190901707e-06,
+      "loss": 1.6817,
+      "step": 2129
+    },
+    {
+      "epoch": 2.334246575342466,
+      "grad_norm": 2.5718271732330322,
+      "learning_rate": 1.263200649878148e-06,
+      "loss": 1.8076,
+      "step": 2130
+    },
+    {
+      "epoch": 2.334246575342466,
+      "eval_loss": 1.9806523323059082,
+      "eval_runtime": 36.9843,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 2130
+    },
+    {
+      "epoch": 2.3353424657534245,
+      "grad_norm": 2.4864919185638428,
+      "learning_rate": 1.2611697806661253e-06,
+      "loss": 1.7308,
+      "step": 2131
+    },
+    {
+      "epoch": 2.3364383561643836,
+      "grad_norm": 2.6007416248321533,
+      "learning_rate": 1.2591389114541026e-06,
+      "loss": 1.7523,
+      "step": 2132
+    },
+    {
+      "epoch": 2.3375342465753426,
+      "grad_norm": 2.600177049636841,
+      "learning_rate": 1.2571080422420798e-06,
+      "loss": 1.7422,
+      "step": 2133
+    },
+    {
+      "epoch": 2.338630136986301,
+      "grad_norm": 2.489025831222534,
+      "learning_rate": 1.2550771730300571e-06,
+      "loss": 1.7184,
+      "step": 2134
+    },
+    {
+      "epoch": 2.33972602739726,
+      "grad_norm": 2.5677363872528076,
+      "learning_rate": 1.2530463038180344e-06,
+      "loss": 1.7922,
+      "step": 2135
+    },
+    {
+      "epoch": 2.3408219178082192,
+      "grad_norm": 2.616976737976074,
+      "learning_rate": 1.2510154346060117e-06,
+      "loss": 1.7354,
+      "step": 2136
+    },
+    {
+      "epoch": 2.3419178082191783,
+      "grad_norm": 2.646456241607666,
+      "learning_rate": 1.2489845653939887e-06,
+      "loss": 1.7377,
+      "step": 2137
+    },
+    {
+      "epoch": 2.343013698630137,
+      "grad_norm": 2.550020933151245,
+      "learning_rate": 1.246953696181966e-06,
+      "loss": 1.7751,
+      "step": 2138
+    },
+    {
+      "epoch": 2.344109589041096,
+      "grad_norm": 2.634279251098633,
+      "learning_rate": 1.2449228269699433e-06,
+      "loss": 1.8157,
+      "step": 2139
+    },
+    {
+      "epoch": 2.345205479452055,
+      "grad_norm": 2.5621821880340576,
+      "learning_rate": 1.2428919577579204e-06,
+      "loss": 1.7383,
+      "step": 2140
+    },
+    {
+      "epoch": 2.345205479452055,
+      "eval_loss": 1.9809924364089966,
+      "eval_runtime": 37.004,
+      "eval_samples_per_second": 27.024,
+      "eval_steps_per_second": 1.135,
+      "step": 2140
+    },
+    {
+      "epoch": 2.3463013698630135,
+      "grad_norm": 2.638831615447998,
+      "learning_rate": 1.2408610885458976e-06,
+      "loss": 1.6988,
+      "step": 2141
+    },
+    {
+      "epoch": 2.3473972602739726,
+      "grad_norm": 2.634612798690796,
+      "learning_rate": 1.238830219333875e-06,
+      "loss": 1.7331,
+      "step": 2142
+    },
+    {
+      "epoch": 2.3484931506849316,
+      "grad_norm": 2.552955389022827,
+      "learning_rate": 1.2367993501218522e-06,
+      "loss": 1.7148,
+      "step": 2143
+    },
+    {
+      "epoch": 2.34958904109589,
+      "grad_norm": 2.6475753784179688,
+      "learning_rate": 1.2347684809098295e-06,
+      "loss": 1.7815,
+      "step": 2144
+    },
+    {
+      "epoch": 2.350684931506849,
+      "grad_norm": 2.7666079998016357,
+      "learning_rate": 1.2327376116978067e-06,
+      "loss": 1.7634,
+      "step": 2145
+    },
+    {
+      "epoch": 2.3517808219178082,
+      "grad_norm": 2.6810429096221924,
+      "learning_rate": 1.230706742485784e-06,
+      "loss": 1.6769,
+      "step": 2146
+    },
+    {
+      "epoch": 2.3528767123287673,
+      "grad_norm": 2.671382188796997,
+      "learning_rate": 1.2286758732737613e-06,
+      "loss": 1.6884,
+      "step": 2147
+    },
+    {
+      "epoch": 2.353972602739726,
+      "grad_norm": 2.614684820175171,
+      "learning_rate": 1.2266450040617386e-06,
+      "loss": 1.8226,
+      "step": 2148
+    },
+    {
+      "epoch": 2.355068493150685,
+      "grad_norm": 2.47189998626709,
+      "learning_rate": 1.2246141348497158e-06,
+      "loss": 1.819,
+      "step": 2149
+    },
+    {
+      "epoch": 2.356164383561644,
+      "grad_norm": 2.4641449451446533,
+      "learning_rate": 1.2225832656376931e-06,
+      "loss": 1.7494,
+      "step": 2150
+    },
+    {
+      "epoch": 2.356164383561644,
+      "eval_loss": 1.9806426763534546,
+      "eval_runtime": 37.0153,
+      "eval_samples_per_second": 27.016,
+      "eval_steps_per_second": 1.135,
+      "step": 2150
+    },
+    {
+      "epoch": 2.357260273972603,
+      "grad_norm": 2.7310373783111572,
+      "learning_rate": 1.2205523964256704e-06,
+      "loss": 1.7432,
+      "step": 2151
+    },
+    {
+      "epoch": 2.3583561643835615,
+      "grad_norm": 2.5554895401000977,
+      "learning_rate": 1.2185215272136477e-06,
+      "loss": 1.7213,
+      "step": 2152
+    },
+    {
+      "epoch": 2.3594520547945206,
+      "grad_norm": 2.5401952266693115,
+      "learning_rate": 1.2164906580016247e-06,
+      "loss": 1.7469,
+      "step": 2153
+    },
+    {
+      "epoch": 2.3605479452054796,
+      "grad_norm": 2.5213980674743652,
+      "learning_rate": 1.214459788789602e-06,
+      "loss": 1.7459,
+      "step": 2154
+    },
+    {
+      "epoch": 2.361643835616438,
+      "grad_norm": 2.5585155487060547,
+      "learning_rate": 1.2124289195775793e-06,
+      "loss": 1.6587,
+      "step": 2155
+    },
+    {
+      "epoch": 2.3627397260273972,
+      "grad_norm": 2.6329963207244873,
+      "learning_rate": 1.2103980503655566e-06,
+      "loss": 1.6925,
+      "step": 2156
+    },
+    {
+      "epoch": 2.3638356164383563,
+      "grad_norm": 2.678351640701294,
+      "learning_rate": 1.2083671811535338e-06,
+      "loss": 1.7322,
+      "step": 2157
+    },
+    {
+      "epoch": 2.364931506849315,
+      "grad_norm": 2.595313549041748,
+      "learning_rate": 1.2063363119415111e-06,
+      "loss": 1.7142,
+      "step": 2158
+    },
+    {
+      "epoch": 2.366027397260274,
+      "grad_norm": 2.5650861263275146,
+      "learning_rate": 1.2043054427294882e-06,
+      "loss": 1.7024,
+      "step": 2159
+    },
+    {
+      "epoch": 2.367123287671233,
+      "grad_norm": 2.495866537094116,
+      "learning_rate": 1.2022745735174654e-06,
+      "loss": 1.7205,
+      "step": 2160
+    },
+    {
+      "epoch": 2.367123287671233,
+      "eval_loss": 1.976922631263733,
+      "eval_runtime": 37.012,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 2160
+    },
+    {
+      "epoch": 2.368219178082192,
+      "grad_norm": 2.570307493209839,
+      "learning_rate": 1.2002437043054427e-06,
+      "loss": 1.7601,
+      "step": 2161
+    },
+    {
+      "epoch": 2.3693150684931505,
+      "grad_norm": 2.444373846054077,
+      "learning_rate": 1.19821283509342e-06,
+      "loss": 1.7857,
+      "step": 2162
+    },
+    {
+      "epoch": 2.3704109589041096,
+      "grad_norm": 2.525505781173706,
+      "learning_rate": 1.1961819658813973e-06,
+      "loss": 1.7875,
+      "step": 2163
+    },
+    {
+      "epoch": 2.3715068493150686,
+      "grad_norm": 2.622260093688965,
+      "learning_rate": 1.1941510966693746e-06,
+      "loss": 1.7596,
+      "step": 2164
+    },
+    {
+      "epoch": 2.3726027397260276,
+      "grad_norm": 2.542076349258423,
+      "learning_rate": 1.1921202274573518e-06,
+      "loss": 1.7248,
+      "step": 2165
+    },
+    {
+      "epoch": 2.3736986301369862,
+      "grad_norm": 2.510263681411743,
+      "learning_rate": 1.190089358245329e-06,
+      "loss": 1.7168,
+      "step": 2166
+    },
+    {
+      "epoch": 2.3747945205479453,
+      "grad_norm": 2.605023145675659,
+      "learning_rate": 1.1880584890333064e-06,
+      "loss": 1.798,
+      "step": 2167
+    },
+    {
+      "epoch": 2.3758904109589043,
+      "grad_norm": 2.593261480331421,
+      "learning_rate": 1.1860276198212837e-06,
+      "loss": 1.7247,
+      "step": 2168
+    },
+    {
+      "epoch": 2.376986301369863,
+      "grad_norm": 2.6010820865631104,
+      "learning_rate": 1.183996750609261e-06,
+      "loss": 1.7607,
+      "step": 2169
+    },
+    {
+      "epoch": 2.378082191780822,
+      "grad_norm": 2.62886118888855,
+      "learning_rate": 1.1819658813972382e-06,
+      "loss": 1.7641,
+      "step": 2170
+    },
+    {
+      "epoch": 2.378082191780822,
+      "eval_loss": 1.9790476560592651,
+      "eval_runtime": 37.0413,
+      "eval_samples_per_second": 26.997,
+      "eval_steps_per_second": 1.134,
+      "step": 2170
+    },
+    {
+      "epoch": 2.379178082191781,
+      "grad_norm": 2.5506582260131836,
+      "learning_rate": 1.1799350121852153e-06,
+      "loss": 1.7407,
+      "step": 2171
+    },
+    {
+      "epoch": 2.3802739726027395,
+      "grad_norm": 2.547445774078369,
+      "learning_rate": 1.1779041429731925e-06,
+      "loss": 1.7808,
+      "step": 2172
+    },
+    {
+      "epoch": 2.3813698630136986,
+      "grad_norm": 2.5443003177642822,
+      "learning_rate": 1.1758732737611698e-06,
+      "loss": 1.7842,
+      "step": 2173
+    },
+    {
+      "epoch": 2.3824657534246576,
+      "grad_norm": 2.5733375549316406,
+      "learning_rate": 1.173842404549147e-06,
+      "loss": 1.6902,
+      "step": 2174
+    },
+    {
+      "epoch": 2.383561643835616,
+      "grad_norm": 2.63702392578125,
+      "learning_rate": 1.1718115353371244e-06,
+      "loss": 1.8049,
+      "step": 2175
+    },
+    {
+      "epoch": 2.3846575342465752,
+      "grad_norm": 2.5577163696289062,
+      "learning_rate": 1.1697806661251016e-06,
+      "loss": 1.7162,
+      "step": 2176
+    },
+    {
+      "epoch": 2.3857534246575343,
+      "grad_norm": 2.4259274005889893,
+      "learning_rate": 1.167749796913079e-06,
+      "loss": 1.7883,
+      "step": 2177
+    },
+    {
+      "epoch": 2.3868493150684933,
+      "grad_norm": 2.537519931793213,
+      "learning_rate": 1.1657189277010562e-06,
+      "loss": 1.7882,
+      "step": 2178
+    },
+    {
+      "epoch": 2.387945205479452,
+      "grad_norm": 2.5480241775512695,
+      "learning_rate": 1.1636880584890335e-06,
+      "loss": 1.7245,
+      "step": 2179
+    },
+    {
+      "epoch": 2.389041095890411,
+      "grad_norm": 2.630337953567505,
+      "learning_rate": 1.1616571892770108e-06,
+      "loss": 1.7165,
+      "step": 2180
+    },
+    {
+      "epoch": 2.389041095890411,
+      "eval_loss": 1.9772392511367798,
+      "eval_runtime": 37.1283,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 2180
+    },
+    {
+      "epoch": 2.39013698630137,
+      "grad_norm": 2.5741171836853027,
+      "learning_rate": 1.159626320064988e-06,
+      "loss": 1.7806,
+      "step": 2181
+    },
+    {
+      "epoch": 2.391232876712329,
+      "grad_norm": 2.6166305541992188,
+      "learning_rate": 1.1575954508529653e-06,
+      "loss": 1.744,
+      "step": 2182
+    },
+    {
+      "epoch": 2.3923287671232876,
+      "grad_norm": 2.6965837478637695,
+      "learning_rate": 1.1555645816409426e-06,
+      "loss": 1.7736,
+      "step": 2183
+    },
+    {
+      "epoch": 2.3934246575342466,
+      "grad_norm": 2.586827516555786,
+      "learning_rate": 1.1535337124289196e-06,
+      "loss": 1.7455,
+      "step": 2184
+    },
+    {
+      "epoch": 2.3945205479452056,
+      "grad_norm": 2.587235689163208,
+      "learning_rate": 1.151502843216897e-06,
+      "loss": 1.6999,
+      "step": 2185
+    },
+    {
+      "epoch": 2.3956164383561642,
+      "grad_norm": 2.656740665435791,
+      "learning_rate": 1.1494719740048742e-06,
+      "loss": 1.6518,
+      "step": 2186
+    },
+    {
+      "epoch": 2.3967123287671233,
+      "grad_norm": 2.5353455543518066,
+      "learning_rate": 1.1474411047928515e-06,
+      "loss": 1.7389,
+      "step": 2187
+    },
+    {
+      "epoch": 2.3978082191780823,
+      "grad_norm": 2.622314453125,
+      "learning_rate": 1.1454102355808287e-06,
+      "loss": 1.7309,
+      "step": 2188
+    },
+    {
+      "epoch": 2.398904109589041,
+      "grad_norm": 2.6286797523498535,
+      "learning_rate": 1.143379366368806e-06,
+      "loss": 1.7094,
+      "step": 2189
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 2.5203921794891357,
+      "learning_rate": 1.141348497156783e-06,
+      "loss": 1.6905,
+      "step": 2190
+    },
+    {
+      "epoch": 2.4,
+      "eval_loss": 1.9822566509246826,
+      "eval_runtime": 37.0581,
+      "eval_samples_per_second": 26.985,
+      "eval_steps_per_second": 1.133,
+      "step": 2190
+    },
+    {
+      "epoch": 2.401095890410959,
+      "grad_norm": 2.6562373638153076,
+      "learning_rate": 1.1393176279447604e-06,
+      "loss": 1.8293,
+      "step": 2191
+    },
+    {
+      "epoch": 2.402191780821918,
+      "grad_norm": 2.5581510066986084,
+      "learning_rate": 1.1372867587327376e-06,
+      "loss": 1.7936,
+      "step": 2192
+    },
+    {
+      "epoch": 2.4032876712328766,
+      "grad_norm": 2.534675121307373,
+      "learning_rate": 1.135255889520715e-06,
+      "loss": 1.6648,
+      "step": 2193
+    },
+    {
+      "epoch": 2.4043835616438356,
+      "grad_norm": 2.7236168384552,
+      "learning_rate": 1.1332250203086922e-06,
+      "loss": 1.7654,
+      "step": 2194
+    },
+    {
+      "epoch": 2.4054794520547946,
+      "grad_norm": 2.5706138610839844,
+      "learning_rate": 1.1311941510966695e-06,
+      "loss": 1.7427,
+      "step": 2195
+    },
+    {
+      "epoch": 2.4065753424657537,
+      "grad_norm": 2.7030224800109863,
+      "learning_rate": 1.1291632818846467e-06,
+      "loss": 1.7216,
+      "step": 2196
+    },
+    {
+      "epoch": 2.4076712328767123,
+      "grad_norm": 2.4936161041259766,
+      "learning_rate": 1.127132412672624e-06,
+      "loss": 1.7033,
+      "step": 2197
+    },
+    {
+      "epoch": 2.4087671232876713,
+      "grad_norm": 2.5384793281555176,
+      "learning_rate": 1.1251015434606013e-06,
+      "loss": 1.7023,
+      "step": 2198
+    },
+    {
+      "epoch": 2.4098630136986303,
+      "grad_norm": 2.5271613597869873,
+      "learning_rate": 1.1230706742485786e-06,
+      "loss": 1.738,
+      "step": 2199
+    },
+    {
+      "epoch": 2.410958904109589,
+      "grad_norm": 2.557610034942627,
+      "learning_rate": 1.1210398050365558e-06,
+      "loss": 1.7927,
+      "step": 2200
+    },
+    {
+      "epoch": 2.410958904109589,
+      "eval_loss": 1.9795563220977783,
+      "eval_runtime": 37.0832,
+      "eval_samples_per_second": 26.966,
+      "eval_steps_per_second": 1.133,
+      "step": 2200
+    },
+    {
+      "epoch": 2.412054794520548,
+      "grad_norm": 2.5497145652770996,
+      "learning_rate": 1.1190089358245331e-06,
+      "loss": 1.7361,
+      "step": 2201
+    },
+    {
+      "epoch": 2.413150684931507,
+      "grad_norm": 2.622793436050415,
+      "learning_rate": 1.1169780666125102e-06,
+      "loss": 1.6624,
+      "step": 2202
+    },
+    {
+      "epoch": 2.4142465753424656,
+      "grad_norm": 2.694563865661621,
+      "learning_rate": 1.1149471974004875e-06,
+      "loss": 1.7755,
+      "step": 2203
+    },
+    {
+      "epoch": 2.4153424657534246,
+      "grad_norm": 2.68461012840271,
+      "learning_rate": 1.1129163281884647e-06,
+      "loss": 1.7132,
+      "step": 2204
+    },
+    {
+      "epoch": 2.4164383561643836,
+      "grad_norm": 2.65791654586792,
+      "learning_rate": 1.110885458976442e-06,
+      "loss": 1.7728,
+      "step": 2205
+    },
+    {
+      "epoch": 2.417534246575342,
+      "grad_norm": 2.46262788772583,
+      "learning_rate": 1.1088545897644193e-06,
+      "loss": 1.6973,
+      "step": 2206
+    },
+    {
+      "epoch": 2.4186301369863012,
+      "grad_norm": 2.5602169036865234,
+      "learning_rate": 1.1068237205523966e-06,
+      "loss": 1.7651,
+      "step": 2207
+    },
+    {
+      "epoch": 2.4197260273972603,
+      "grad_norm": 2.5492701530456543,
+      "learning_rate": 1.1047928513403736e-06,
+      "loss": 1.7674,
+      "step": 2208
+    },
+    {
+      "epoch": 2.4208219178082193,
+      "grad_norm": 2.798940420150757,
+      "learning_rate": 1.102761982128351e-06,
+      "loss": 1.7719,
+      "step": 2209
+    },
+    {
+      "epoch": 2.421917808219178,
+      "grad_norm": 2.6989545822143555,
+      "learning_rate": 1.1007311129163282e-06,
+      "loss": 1.6821,
+      "step": 2210
+    },
+    {
+      "epoch": 2.421917808219178,
+      "eval_loss": 1.9763859510421753,
+      "eval_runtime": 36.7821,
+      "eval_samples_per_second": 27.187,
+      "eval_steps_per_second": 1.142,
+      "step": 2210
+    },
+    {
+      "epoch": 2.423013698630137,
+      "grad_norm": 2.752720594406128,
+      "learning_rate": 1.0987002437043055e-06,
+      "loss": 1.7432,
+      "step": 2211
+    },
+    {
+      "epoch": 2.424109589041096,
+      "grad_norm": 2.6162850856781006,
+      "learning_rate": 1.0966693744922827e-06,
+      "loss": 1.718,
+      "step": 2212
+    },
+    {
+      "epoch": 2.425205479452055,
+      "grad_norm": 2.938995838165283,
+      "learning_rate": 1.09463850528026e-06,
+      "loss": 1.6761,
+      "step": 2213
+    },
+    {
+      "epoch": 2.4263013698630136,
+      "grad_norm": 2.7038822174072266,
+      "learning_rate": 1.0926076360682373e-06,
+      "loss": 1.7267,
+      "step": 2214
+    },
+    {
+      "epoch": 2.4273972602739726,
+      "grad_norm": 2.6025028228759766,
+      "learning_rate": 1.0905767668562146e-06,
+      "loss": 1.7021,
+      "step": 2215
+    },
+    {
+      "epoch": 2.4284931506849317,
+      "grad_norm": 2.5922977924346924,
+      "learning_rate": 1.0885458976441918e-06,
+      "loss": 1.7221,
+      "step": 2216
+    },
+    {
+      "epoch": 2.4295890410958902,
+      "grad_norm": 2.634716033935547,
+      "learning_rate": 1.0865150284321691e-06,
+      "loss": 1.8047,
+      "step": 2217
+    },
+    {
+      "epoch": 2.4306849315068493,
+      "grad_norm": 2.655155658721924,
+      "learning_rate": 1.0844841592201464e-06,
+      "loss": 1.7426,
+      "step": 2218
+    },
+    {
+      "epoch": 2.4317808219178083,
+      "grad_norm": 2.5151286125183105,
+      "learning_rate": 1.0824532900081237e-06,
+      "loss": 1.7627,
+      "step": 2219
+    },
+    {
+      "epoch": 2.432876712328767,
+      "grad_norm": 2.696597099304199,
+      "learning_rate": 1.080422420796101e-06,
+      "loss": 1.7325,
+      "step": 2220
+    },
+    {
+      "epoch": 2.432876712328767,
+      "eval_loss": 1.9818732738494873,
+      "eval_runtime": 36.9854,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 2220
+    },
+    {
+      "epoch": 2.433972602739726,
+      "grad_norm": 2.7444307804107666,
+      "learning_rate": 1.078391551584078e-06,
+      "loss": 1.743,
+      "step": 2221
+    },
+    {
+      "epoch": 2.435068493150685,
+      "grad_norm": 2.503706455230713,
+      "learning_rate": 1.0763606823720553e-06,
+      "loss": 1.7699,
+      "step": 2222
+    },
+    {
+      "epoch": 2.436164383561644,
+      "grad_norm": 2.546041965484619,
+      "learning_rate": 1.0743298131600326e-06,
+      "loss": 1.8077,
+      "step": 2223
+    },
+    {
+      "epoch": 2.4372602739726026,
+      "grad_norm": 2.5515291690826416,
+      "learning_rate": 1.0722989439480098e-06,
+      "loss": 1.7637,
+      "step": 2224
+    },
+    {
+      "epoch": 2.4383561643835616,
+      "grad_norm": 2.564180850982666,
+      "learning_rate": 1.070268074735987e-06,
+      "loss": 1.7299,
+      "step": 2225
+    },
+    {
+      "epoch": 2.4394520547945207,
+      "grad_norm": 2.532708168029785,
+      "learning_rate": 1.0682372055239644e-06,
+      "loss": 1.713,
+      "step": 2226
+    },
+    {
+      "epoch": 2.4405479452054797,
+      "grad_norm": 2.672358512878418,
+      "learning_rate": 1.0662063363119414e-06,
+      "loss": 1.7103,
+      "step": 2227
+    },
+    {
+      "epoch": 2.4416438356164383,
+      "grad_norm": 2.7836480140686035,
+      "learning_rate": 1.0641754670999187e-06,
+      "loss": 1.7129,
+      "step": 2228
+    },
+    {
+      "epoch": 2.4427397260273973,
+      "grad_norm": 2.6253440380096436,
+      "learning_rate": 1.062144597887896e-06,
+      "loss": 1.7389,
+      "step": 2229
+    },
+    {
+      "epoch": 2.4438356164383563,
+      "grad_norm": 2.5863685607910156,
+      "learning_rate": 1.0601137286758733e-06,
+      "loss": 1.7293,
+      "step": 2230
+    },
+    {
+      "epoch": 2.4438356164383563,
+      "eval_loss": 1.9771078824996948,
+      "eval_runtime": 37.0603,
+      "eval_samples_per_second": 26.983,
+      "eval_steps_per_second": 1.133,
+      "step": 2230
+    },
+    {
+      "epoch": 2.444931506849315,
+      "grad_norm": 2.732042074203491,
+      "learning_rate": 1.0580828594638505e-06,
+      "loss": 1.8012,
+      "step": 2231
+    },
+    {
+      "epoch": 2.446027397260274,
+      "grad_norm": 2.5275564193725586,
+      "learning_rate": 1.0560519902518278e-06,
+      "loss": 1.6482,
+      "step": 2232
+    },
+    {
+      "epoch": 2.447123287671233,
+      "grad_norm": 2.7054405212402344,
+      "learning_rate": 1.054021121039805e-06,
+      "loss": 1.7799,
+      "step": 2233
+    },
+    {
+      "epoch": 2.4482191780821916,
+      "grad_norm": 2.587719678878784,
+      "learning_rate": 1.0519902518277824e-06,
+      "loss": 1.6797,
+      "step": 2234
+    },
+    {
+      "epoch": 2.4493150684931506,
+      "grad_norm": 2.600907564163208,
+      "learning_rate": 1.0499593826157596e-06,
+      "loss": 1.6727,
+      "step": 2235
+    },
+    {
+      "epoch": 2.4504109589041096,
+      "grad_norm": 2.6509485244750977,
+      "learning_rate": 1.047928513403737e-06,
+      "loss": 1.7535,
+      "step": 2236
+    },
+    {
+      "epoch": 2.4515068493150687,
+      "grad_norm": 2.6827805042266846,
+      "learning_rate": 1.0458976441917142e-06,
+      "loss": 1.7285,
+      "step": 2237
+    },
+    {
+      "epoch": 2.4526027397260273,
+      "grad_norm": 2.7095749378204346,
+      "learning_rate": 1.0438667749796915e-06,
+      "loss": 1.6546,
+      "step": 2238
+    },
+    {
+      "epoch": 2.4536986301369863,
+      "grad_norm": 2.7757186889648438,
+      "learning_rate": 1.0418359057676688e-06,
+      "loss": 1.6659,
+      "step": 2239
+    },
+    {
+      "epoch": 2.4547945205479453,
+      "grad_norm": 2.7696826457977295,
+      "learning_rate": 1.0398050365556458e-06,
+      "loss": 1.7108,
+      "step": 2240
+    },
+    {
+      "epoch": 2.4547945205479453,
+      "eval_loss": 1.980337142944336,
+      "eval_runtime": 37.0883,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 2240
+    },
+    {
+      "epoch": 2.455890410958904,
+      "grad_norm": 2.96648907661438,
+      "learning_rate": 1.037774167343623e-06,
+      "loss": 1.77,
+      "step": 2241
+    },
+    {
+      "epoch": 2.456986301369863,
+      "grad_norm": 2.561461925506592,
+      "learning_rate": 1.0357432981316004e-06,
+      "loss": 1.6962,
+      "step": 2242
+    },
+    {
+      "epoch": 2.458082191780822,
+      "grad_norm": 2.5870988368988037,
+      "learning_rate": 1.0337124289195776e-06,
+      "loss": 1.7385,
+      "step": 2243
+    },
+    {
+      "epoch": 2.459178082191781,
+      "grad_norm": 2.607558488845825,
+      "learning_rate": 1.031681559707555e-06,
+      "loss": 1.7771,
+      "step": 2244
+    },
+    {
+      "epoch": 2.4602739726027396,
+      "grad_norm": 2.6373603343963623,
+      "learning_rate": 1.0296506904955322e-06,
+      "loss": 1.7845,
+      "step": 2245
+    },
+    {
+      "epoch": 2.4613698630136986,
+      "grad_norm": 2.5508599281311035,
+      "learning_rate": 1.0276198212835095e-06,
+      "loss": 1.7384,
+      "step": 2246
+    },
+    {
+      "epoch": 2.4624657534246577,
+      "grad_norm": 2.730696201324463,
+      "learning_rate": 1.0255889520714867e-06,
+      "loss": 1.7458,
+      "step": 2247
+    },
+    {
+      "epoch": 2.4635616438356163,
+      "grad_norm": 2.4905753135681152,
+      "learning_rate": 1.023558082859464e-06,
+      "loss": 1.7195,
+      "step": 2248
+    },
+    {
+      "epoch": 2.4646575342465753,
+      "grad_norm": 2.664184331893921,
+      "learning_rate": 1.0215272136474413e-06,
+      "loss": 1.6986,
+      "step": 2249
+    },
+    {
+      "epoch": 2.4657534246575343,
+      "grad_norm": 2.683645009994507,
+      "learning_rate": 1.0194963444354186e-06,
+      "loss": 1.7981,
+      "step": 2250
+    },
+    {
+      "epoch": 2.4657534246575343,
+      "eval_loss": 1.9768283367156982,
+      "eval_runtime": 37.1086,
+      "eval_samples_per_second": 26.948,
+      "eval_steps_per_second": 1.132,
+      "step": 2250
+    },
+    {
+      "epoch": 2.466849315068493,
+      "grad_norm": 2.633633852005005,
+      "learning_rate": 1.0174654752233958e-06,
+      "loss": 1.6807,
+      "step": 2251
+    },
+    {
+      "epoch": 2.467945205479452,
+      "grad_norm": 2.655794382095337,
+      "learning_rate": 1.015434606011373e-06,
+      "loss": 1.7705,
+      "step": 2252
+    },
+    {
+      "epoch": 2.469041095890411,
+      "grad_norm": 2.6457812786102295,
+      "learning_rate": 1.0134037367993502e-06,
+      "loss": 1.7201,
+      "step": 2253
+    },
+    {
+      "epoch": 2.47013698630137,
+      "grad_norm": 2.623656749725342,
+      "learning_rate": 1.0113728675873275e-06,
+      "loss": 1.7402,
+      "step": 2254
+    },
+    {
+      "epoch": 2.4712328767123286,
+      "grad_norm": 2.65552020072937,
+      "learning_rate": 1.0093419983753047e-06,
+      "loss": 1.7372,
+      "step": 2255
+    },
+    {
+      "epoch": 2.4723287671232876,
+      "grad_norm": 2.707531213760376,
+      "learning_rate": 1.007311129163282e-06,
+      "loss": 1.7299,
+      "step": 2256
+    },
+    {
+      "epoch": 2.4734246575342467,
+      "grad_norm": 2.504805564880371,
+      "learning_rate": 1.0052802599512593e-06,
+      "loss": 1.6858,
+      "step": 2257
+    },
+    {
+      "epoch": 2.4745205479452057,
+      "grad_norm": 2.584251642227173,
+      "learning_rate": 1.0032493907392364e-06,
+      "loss": 1.6867,
+      "step": 2258
+    },
+    {
+      "epoch": 2.4756164383561643,
+      "grad_norm": 2.5356390476226807,
+      "learning_rate": 1.0012185215272136e-06,
+      "loss": 1.7442,
+      "step": 2259
+    },
+    {
+      "epoch": 2.4767123287671233,
+      "grad_norm": 2.5929319858551025,
+      "learning_rate": 9.99187652315191e-07,
+      "loss": 1.6777,
+      "step": 2260
+    },
+    {
+      "epoch": 2.4767123287671233,
+      "eval_loss": 1.9772632122039795,
+      "eval_runtime": 37.0951,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 2260
+    },
+    {
+      "epoch": 2.4778082191780824,
+      "grad_norm": 2.6396026611328125,
+      "learning_rate": 9.971567831031682e-07,
+      "loss": 1.7638,
+      "step": 2261
+    },
+    {
+      "epoch": 2.478904109589041,
+      "grad_norm": 2.5710740089416504,
+      "learning_rate": 9.951259138911455e-07,
+      "loss": 1.7566,
+      "step": 2262
+    },
+    {
+      "epoch": 2.48,
+      "grad_norm": 2.5611767768859863,
+      "learning_rate": 9.930950446791227e-07,
+      "loss": 1.6639,
+      "step": 2263
+    },
+    {
+      "epoch": 2.481095890410959,
+      "grad_norm": 2.5390372276306152,
+      "learning_rate": 9.910641754671e-07,
+      "loss": 1.7887,
+      "step": 2264
+    },
+    {
+      "epoch": 2.4821917808219176,
+      "grad_norm": 2.6124560832977295,
+      "learning_rate": 9.890333062550773e-07,
+      "loss": 1.8221,
+      "step": 2265
+    },
+    {
+      "epoch": 2.4832876712328766,
+      "grad_norm": 2.6506896018981934,
+      "learning_rate": 9.870024370430546e-07,
+      "loss": 1.725,
+      "step": 2266
+    },
+    {
+      "epoch": 2.4843835616438357,
+      "grad_norm": 2.617889642715454,
+      "learning_rate": 9.849715678310318e-07,
+      "loss": 1.7566,
+      "step": 2267
+    },
+    {
+      "epoch": 2.4854794520547947,
+      "grad_norm": 2.571110486984253,
+      "learning_rate": 9.829406986190091e-07,
+      "loss": 1.7327,
+      "step": 2268
+    },
+    {
+      "epoch": 2.4865753424657533,
+      "grad_norm": 2.657665252685547,
+      "learning_rate": 9.809098294069864e-07,
+      "loss": 1.7471,
+      "step": 2269
+    },
+    {
+      "epoch": 2.4876712328767123,
+      "grad_norm": 2.5402324199676514,
+      "learning_rate": 9.788789601949637e-07,
+      "loss": 1.6597,
+      "step": 2270
+    },
+    {
+      "epoch": 2.4876712328767123,
+      "eval_loss": 1.975923776626587,
+      "eval_runtime": 37.0909,
+      "eval_samples_per_second": 26.961,
+      "eval_steps_per_second": 1.132,
+      "step": 2270
+    },
+    {
+      "epoch": 2.4887671232876714,
+      "grad_norm": 2.619765281677246,
+      "learning_rate": 9.768480909829407e-07,
+      "loss": 1.7434,
+      "step": 2271
+    },
+    {
+      "epoch": 2.48986301369863,
+      "grad_norm": 2.664238214492798,
+      "learning_rate": 9.74817221770918e-07,
+      "loss": 1.7937,
+      "step": 2272
+    },
+    {
+      "epoch": 2.490958904109589,
+      "grad_norm": 2.6743860244750977,
+      "learning_rate": 9.727863525588953e-07,
+      "loss": 1.7203,
+      "step": 2273
+    },
+    {
+      "epoch": 2.492054794520548,
+      "grad_norm": 2.695051670074463,
+      "learning_rate": 9.707554833468726e-07,
+      "loss": 1.7473,
+      "step": 2274
+    },
+    {
+      "epoch": 2.493150684931507,
+      "grad_norm": 2.6115036010742188,
+      "learning_rate": 9.687246141348498e-07,
+      "loss": 1.7358,
+      "step": 2275
+    },
+    {
+      "epoch": 2.4942465753424656,
+      "grad_norm": 2.7951300144195557,
+      "learning_rate": 9.666937449228271e-07,
+      "loss": 1.723,
+      "step": 2276
+    },
+    {
+      "epoch": 2.4953424657534247,
+      "grad_norm": 2.6242685317993164,
+      "learning_rate": 9.646628757108042e-07,
+      "loss": 1.7389,
+      "step": 2277
+    },
+    {
+      "epoch": 2.4964383561643837,
+      "grad_norm": 2.567150831222534,
+      "learning_rate": 9.626320064987814e-07,
+      "loss": 1.6775,
+      "step": 2278
+    },
+    {
+      "epoch": 2.4975342465753423,
+      "grad_norm": 2.6157355308532715,
+      "learning_rate": 9.606011372867587e-07,
+      "loss": 1.7624,
+      "step": 2279
+    },
+    {
+      "epoch": 2.4986301369863013,
+      "grad_norm": 2.5663843154907227,
+      "learning_rate": 9.58570268074736e-07,
+      "loss": 1.6955,
+      "step": 2280
+    },
+    {
+      "epoch": 2.4986301369863013,
+      "eval_loss": 1.9744558334350586,
+      "eval_runtime": 37.1094,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 2280
+    },
+    {
+      "epoch": 2.4997260273972604,
+      "grad_norm": 2.667611598968506,
+      "learning_rate": 9.565393988627133e-07,
+      "loss": 1.7261,
+      "step": 2281
+    },
+    {
+      "epoch": 2.500821917808219,
+      "grad_norm": 2.5702202320098877,
+      "learning_rate": 9.545085296506906e-07,
+      "loss": 1.7252,
+      "step": 2282
+    },
+    {
+      "epoch": 2.501917808219178,
+      "grad_norm": 2.5271430015563965,
+      "learning_rate": 9.524776604386677e-07,
+      "loss": 1.7407,
+      "step": 2283
+    },
+    {
+      "epoch": 2.503013698630137,
+      "grad_norm": 2.5804171562194824,
+      "learning_rate": 9.50446791226645e-07,
+      "loss": 1.739,
+      "step": 2284
+    },
+    {
+      "epoch": 2.504109589041096,
+      "grad_norm": 2.571345090866089,
+      "learning_rate": 9.484159220146223e-07,
+      "loss": 1.668,
+      "step": 2285
+    },
+    {
+      "epoch": 2.5052054794520546,
+      "grad_norm": 2.5697009563446045,
+      "learning_rate": 9.463850528025995e-07,
+      "loss": 1.7677,
+      "step": 2286
+    },
+    {
+      "epoch": 2.5063013698630137,
+      "grad_norm": 2.5846893787384033,
+      "learning_rate": 9.443541835905768e-07,
+      "loss": 1.7589,
+      "step": 2287
+    },
+    {
+      "epoch": 2.5073972602739727,
+      "grad_norm": 2.803380250930786,
+      "learning_rate": 9.423233143785541e-07,
+      "loss": 1.8104,
+      "step": 2288
+    },
+    {
+      "epoch": 2.5084931506849317,
+      "grad_norm": 2.6451611518859863,
+      "learning_rate": 9.402924451665313e-07,
+      "loss": 1.8061,
+      "step": 2289
+    },
+    {
+      "epoch": 2.5095890410958903,
+      "grad_norm": 2.6164097785949707,
+      "learning_rate": 9.382615759545085e-07,
+      "loss": 1.6899,
+      "step": 2290
+    },
+    {
+      "epoch": 2.5095890410958903,
+      "eval_loss": 1.9739093780517578,
+      "eval_runtime": 37.1222,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 2290
+    },
+    {
+      "epoch": 2.5106849315068493,
+      "grad_norm": 2.5874385833740234,
+      "learning_rate": 9.362307067424858e-07,
+      "loss": 1.7692,
+      "step": 2291
+    },
+    {
+      "epoch": 2.5117808219178084,
+      "grad_norm": 2.5705063343048096,
+      "learning_rate": 9.341998375304631e-07,
+      "loss": 1.7303,
+      "step": 2292
+    },
+    {
+      "epoch": 2.512876712328767,
+      "grad_norm": 2.6130380630493164,
+      "learning_rate": 9.321689683184404e-07,
+      "loss": 1.7892,
+      "step": 2293
+    },
+    {
+      "epoch": 2.513972602739726,
+      "grad_norm": 2.5991110801696777,
+      "learning_rate": 9.301380991064176e-07,
+      "loss": 1.6877,
+      "step": 2294
+    },
+    {
+      "epoch": 2.515068493150685,
+      "grad_norm": 2.6092689037323,
+      "learning_rate": 9.281072298943949e-07,
+      "loss": 1.6729,
+      "step": 2295
+    },
+    {
+      "epoch": 2.5161643835616436,
+      "grad_norm": 2.5563883781433105,
+      "learning_rate": 9.260763606823721e-07,
+      "loss": 1.7181,
+      "step": 2296
+    },
+    {
+      "epoch": 2.5172602739726027,
+      "grad_norm": 2.623948574066162,
+      "learning_rate": 9.240454914703494e-07,
+      "loss": 1.7528,
+      "step": 2297
+    },
+    {
+      "epoch": 2.5183561643835617,
+      "grad_norm": 2.661874294281006,
+      "learning_rate": 9.220146222583266e-07,
+      "loss": 1.7582,
+      "step": 2298
+    },
+    {
+      "epoch": 2.5194520547945203,
+      "grad_norm": 2.5936238765716553,
+      "learning_rate": 9.199837530463039e-07,
+      "loss": 1.6987,
+      "step": 2299
+    },
+    {
+      "epoch": 2.5205479452054793,
+      "grad_norm": 2.5134992599487305,
+      "learning_rate": 9.179528838342812e-07,
+      "loss": 1.7265,
+      "step": 2300
+    },
+    {
+      "epoch": 2.5205479452054793,
+      "eval_loss": 1.9737082719802856,
+      "eval_runtime": 37.1105,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 2300
+    },
+    {
+      "epoch": 2.5216438356164383,
+      "grad_norm": 2.627927541732788,
+      "learning_rate": 9.159220146222585e-07,
+      "loss": 1.7059,
+      "step": 2301
+    },
+    {
+      "epoch": 2.5227397260273974,
+      "grad_norm": 2.6419436931610107,
+      "learning_rate": 9.138911454102356e-07,
+      "loss": 1.7558,
+      "step": 2302
+    },
+    {
+      "epoch": 2.5238356164383564,
+      "grad_norm": 2.6184327602386475,
+      "learning_rate": 9.118602761982129e-07,
+      "loss": 1.6281,
+      "step": 2303
+    },
+    {
+      "epoch": 2.524931506849315,
+      "grad_norm": 2.6111767292022705,
+      "learning_rate": 9.098294069861902e-07,
+      "loss": 1.7408,
+      "step": 2304
+    },
+    {
+      "epoch": 2.526027397260274,
+      "grad_norm": 2.6416513919830322,
+      "learning_rate": 9.077985377741675e-07,
+      "loss": 1.7397,
+      "step": 2305
+    },
+    {
+      "epoch": 2.527123287671233,
+      "grad_norm": 2.5487682819366455,
+      "learning_rate": 9.057676685621447e-07,
+      "loss": 1.6941,
+      "step": 2306
+    },
+    {
+      "epoch": 2.5282191780821917,
+      "grad_norm": 2.6250264644622803,
+      "learning_rate": 9.03736799350122e-07,
+      "loss": 1.66,
+      "step": 2307
+    },
+    {
+      "epoch": 2.5293150684931507,
+      "grad_norm": 2.6054790019989014,
+      "learning_rate": 9.017059301380991e-07,
+      "loss": 1.7829,
+      "step": 2308
+    },
+    {
+      "epoch": 2.5304109589041097,
+      "grad_norm": 2.751016616821289,
+      "learning_rate": 8.996750609260764e-07,
+      "loss": 1.7693,
+      "step": 2309
+    },
+    {
+      "epoch": 2.5315068493150683,
+      "grad_norm": 2.6321675777435303,
+      "learning_rate": 8.976441917140536e-07,
+      "loss": 1.7971,
+      "step": 2310
+    },
+    {
+      "epoch": 2.5315068493150683,
+      "eval_loss": 1.9757944345474243,
+      "eval_runtime": 36.77,
+      "eval_samples_per_second": 27.196,
+      "eval_steps_per_second": 1.142,
+      "step": 2310
+    },
+    {
+      "epoch": 2.5326027397260273,
+      "grad_norm": 2.642702341079712,
+      "learning_rate": 8.956133225020309e-07,
+      "loss": 1.7703,
+      "step": 2311
+    },
+    {
+      "epoch": 2.5336986301369864,
+      "grad_norm": 2.614771604537964,
+      "learning_rate": 8.935824532900082e-07,
+      "loss": 1.8009,
+      "step": 2312
+    },
+    {
+      "epoch": 2.534794520547945,
+      "grad_norm": 2.514615058898926,
+      "learning_rate": 8.915515840779855e-07,
+      "loss": 1.6954,
+      "step": 2313
+    },
+    {
+      "epoch": 2.535890410958904,
+      "grad_norm": 2.644207715988159,
+      "learning_rate": 8.895207148659626e-07,
+      "loss": 1.6779,
+      "step": 2314
+    },
+    {
+      "epoch": 2.536986301369863,
+      "grad_norm": 2.7844860553741455,
+      "learning_rate": 8.874898456539399e-07,
+      "loss": 1.7595,
+      "step": 2315
+    },
+    {
+      "epoch": 2.538082191780822,
+      "grad_norm": 2.714629650115967,
+      "learning_rate": 8.854589764419172e-07,
+      "loss": 1.7448,
+      "step": 2316
+    },
+    {
+      "epoch": 2.539178082191781,
+      "grad_norm": 2.57271409034729,
+      "learning_rate": 8.834281072298945e-07,
+      "loss": 1.7438,
+      "step": 2317
+    },
+    {
+      "epoch": 2.5402739726027397,
+      "grad_norm": 2.6566264629364014,
+      "learning_rate": 8.813972380178717e-07,
+      "loss": 1.6924,
+      "step": 2318
+    },
+    {
+      "epoch": 2.5413698630136987,
+      "grad_norm": 2.619744300842285,
+      "learning_rate": 8.79366368805849e-07,
+      "loss": 1.7317,
+      "step": 2319
+    },
+    {
+      "epoch": 2.5424657534246577,
+      "grad_norm": 2.6249327659606934,
+      "learning_rate": 8.773354995938262e-07,
+      "loss": 1.7467,
+      "step": 2320
+    },
+    {
+      "epoch": 2.5424657534246577,
+      "eval_loss": 1.9757587909698486,
+      "eval_runtime": 36.9112,
+      "eval_samples_per_second": 27.092,
+      "eval_steps_per_second": 1.138,
+      "step": 2320
+    },
+    {
+      "epoch": 2.5435616438356163,
+      "grad_norm": 2.584317445755005,
+      "learning_rate": 8.753046303818035e-07,
+      "loss": 1.6962,
+      "step": 2321
+    },
+    {
+      "epoch": 2.5446575342465754,
+      "grad_norm": 2.7346856594085693,
+      "learning_rate": 8.732737611697807e-07,
+      "loss": 1.6863,
+      "step": 2322
+    },
+    {
+      "epoch": 2.5457534246575344,
+      "grad_norm": 2.574941635131836,
+      "learning_rate": 8.71242891957758e-07,
+      "loss": 1.7271,
+      "step": 2323
+    },
+    {
+      "epoch": 2.546849315068493,
+      "grad_norm": 2.546217441558838,
+      "learning_rate": 8.692120227457353e-07,
+      "loss": 1.751,
+      "step": 2324
+    },
+    {
+      "epoch": 2.547945205479452,
+      "grad_norm": 2.534414768218994,
+      "learning_rate": 8.671811535337126e-07,
+      "loss": 1.7692,
+      "step": 2325
+    },
+    {
+      "epoch": 2.549041095890411,
+      "grad_norm": 2.625565528869629,
+      "learning_rate": 8.651502843216898e-07,
+      "loss": 1.7135,
+      "step": 2326
+    },
+    {
+      "epoch": 2.5501369863013696,
+      "grad_norm": 2.715276002883911,
+      "learning_rate": 8.63119415109667e-07,
+      "loss": 1.8093,
+      "step": 2327
+    },
+    {
+      "epoch": 2.5512328767123287,
+      "grad_norm": 2.6494171619415283,
+      "learning_rate": 8.610885458976443e-07,
+      "loss": 1.7589,
+      "step": 2328
+    },
+    {
+      "epoch": 2.5523287671232877,
+      "grad_norm": 2.6840505599975586,
+      "learning_rate": 8.590576766856216e-07,
+      "loss": 1.7484,
+      "step": 2329
+    },
+    {
+      "epoch": 2.5534246575342463,
+      "grad_norm": 2.510047435760498,
+      "learning_rate": 8.570268074735988e-07,
+      "loss": 1.6985,
+      "step": 2330
+    },
+    {
+      "epoch": 2.5534246575342463,
+      "eval_loss": 1.9751781225204468,
+      "eval_runtime": 36.9998,
+      "eval_samples_per_second": 27.027,
+      "eval_steps_per_second": 1.135,
+      "step": 2330
+    },
+    {
+      "epoch": 2.5545205479452053,
+      "grad_norm": 2.6174721717834473,
+      "learning_rate": 8.549959382615761e-07,
+      "loss": 1.6873,
+      "step": 2331
+    },
+    {
+      "epoch": 2.5556164383561644,
+      "grad_norm": 2.570403814315796,
+      "learning_rate": 8.529650690495534e-07,
+      "loss": 1.7058,
+      "step": 2332
+    },
+    {
+      "epoch": 2.5567123287671234,
+      "grad_norm": 2.8619518280029297,
+      "learning_rate": 8.509341998375305e-07,
+      "loss": 1.7342,
+      "step": 2333
+    },
+    {
+      "epoch": 2.5578082191780824,
+      "grad_norm": 2.6569507122039795,
+      "learning_rate": 8.489033306255077e-07,
+      "loss": 1.6736,
+      "step": 2334
+    },
+    {
+      "epoch": 2.558904109589041,
+      "grad_norm": 2.5437800884246826,
+      "learning_rate": 8.46872461413485e-07,
+      "loss": 1.648,
+      "step": 2335
+    },
+    {
+      "epoch": 2.56,
+      "grad_norm": 2.652275323867798,
+      "learning_rate": 8.448415922014623e-07,
+      "loss": 1.692,
+      "step": 2336
+    },
+    {
+      "epoch": 2.561095890410959,
+      "grad_norm": 2.6103672981262207,
+      "learning_rate": 8.428107229894396e-07,
+      "loss": 1.7033,
+      "step": 2337
+    },
+    {
+      "epoch": 2.5621917808219177,
+      "grad_norm": 2.6765308380126953,
+      "learning_rate": 8.407798537774168e-07,
+      "loss": 1.7473,
+      "step": 2338
+    },
+    {
+      "epoch": 2.5632876712328767,
+      "grad_norm": 2.708935260772705,
+      "learning_rate": 8.38748984565394e-07,
+      "loss": 1.7346,
+      "step": 2339
+    },
+    {
+      "epoch": 2.5643835616438357,
+      "grad_norm": 2.597233772277832,
+      "learning_rate": 8.367181153533713e-07,
+      "loss": 1.6927,
+      "step": 2340
+    },
+    {
+      "epoch": 2.5643835616438357,
+      "eval_loss": 1.9768768548965454,
+      "eval_runtime": 37.0191,
+      "eval_samples_per_second": 27.013,
+      "eval_steps_per_second": 1.135,
+      "step": 2340
+    },
+    {
+      "epoch": 2.5654794520547943,
+      "grad_norm": 2.624343156814575,
+      "learning_rate": 8.346872461413486e-07,
+      "loss": 1.7219,
+      "step": 2341
+    },
+    {
+      "epoch": 2.5665753424657534,
+      "grad_norm": 2.675095796585083,
+      "learning_rate": 8.326563769293258e-07,
+      "loss": 1.7136,
+      "step": 2342
+    },
+    {
+      "epoch": 2.5676712328767124,
+      "grad_norm": 2.575009822845459,
+      "learning_rate": 8.306255077173031e-07,
+      "loss": 1.7187,
+      "step": 2343
+    },
+    {
+      "epoch": 2.568767123287671,
+      "grad_norm": 2.539958953857422,
+      "learning_rate": 8.285946385052804e-07,
+      "loss": 1.8232,
+      "step": 2344
+    },
+    {
+      "epoch": 2.56986301369863,
+      "grad_norm": 2.613976001739502,
+      "learning_rate": 8.265637692932575e-07,
+      "loss": 1.6746,
+      "step": 2345
+    },
+    {
+      "epoch": 2.570958904109589,
+      "grad_norm": 2.5001742839813232,
+      "learning_rate": 8.245329000812348e-07,
+      "loss": 1.7542,
+      "step": 2346
+    },
+    {
+      "epoch": 2.572054794520548,
+      "grad_norm": 2.6975460052490234,
+      "learning_rate": 8.225020308692121e-07,
+      "loss": 1.7434,
+      "step": 2347
+    },
+    {
+      "epoch": 2.573150684931507,
+      "grad_norm": 2.6900603771209717,
+      "learning_rate": 8.204711616571894e-07,
+      "loss": 1.6843,
+      "step": 2348
+    },
+    {
+      "epoch": 2.5742465753424657,
+      "grad_norm": 2.5895581245422363,
+      "learning_rate": 8.184402924451667e-07,
+      "loss": 1.7279,
+      "step": 2349
+    },
+    {
+      "epoch": 2.5753424657534247,
+      "grad_norm": 2.5448262691497803,
+      "learning_rate": 8.164094232331439e-07,
+      "loss": 1.6743,
+      "step": 2350
+    },
+    {
+      "epoch": 2.5753424657534247,
+      "eval_loss": 1.9762015342712402,
+      "eval_runtime": 37.0186,
+      "eval_samples_per_second": 27.013,
+      "eval_steps_per_second": 1.135,
+      "step": 2350
+    },
+    {
+      "epoch": 2.5764383561643838,
+      "grad_norm": 2.6080751419067383,
+      "learning_rate": 8.14378554021121e-07,
+      "loss": 1.7525,
+      "step": 2351
+    },
+    {
+      "epoch": 2.5775342465753424,
+      "grad_norm": 2.6307425498962402,
+      "learning_rate": 8.123476848090983e-07,
+      "loss": 1.7304,
+      "step": 2352
+    },
+    {
+      "epoch": 2.5786301369863014,
+      "grad_norm": 2.7442290782928467,
+      "learning_rate": 8.103168155970755e-07,
+      "loss": 1.7018,
+      "step": 2353
+    },
+    {
+      "epoch": 2.5797260273972604,
+      "grad_norm": 2.5672149658203125,
+      "learning_rate": 8.082859463850528e-07,
+      "loss": 1.7115,
+      "step": 2354
+    },
+    {
+      "epoch": 2.580821917808219,
+      "grad_norm": 2.7201836109161377,
+      "learning_rate": 8.062550771730301e-07,
+      "loss": 1.7461,
+      "step": 2355
+    },
+    {
+      "epoch": 2.581917808219178,
+      "grad_norm": 2.7145047187805176,
+      "learning_rate": 8.042242079610074e-07,
+      "loss": 1.7169,
+      "step": 2356
+    },
+    {
+      "epoch": 2.583013698630137,
+      "grad_norm": 2.7162413597106934,
+      "learning_rate": 8.021933387489846e-07,
+      "loss": 1.71,
+      "step": 2357
+    },
+    {
+      "epoch": 2.5841095890410957,
+      "grad_norm": 2.515169858932495,
+      "learning_rate": 8.001624695369618e-07,
+      "loss": 1.7016,
+      "step": 2358
+    },
+    {
+      "epoch": 2.5852054794520547,
+      "grad_norm": 2.620922803878784,
+      "learning_rate": 7.981316003249391e-07,
+      "loss": 1.7251,
+      "step": 2359
+    },
+    {
+      "epoch": 2.5863013698630137,
+      "grad_norm": 2.8066515922546387,
+      "learning_rate": 7.961007311129164e-07,
+      "loss": 1.787,
+      "step": 2360
+    },
+    {
+      "epoch": 2.5863013698630137,
+      "eval_loss": 1.9778543710708618,
+      "eval_runtime": 37.0094,
+      "eval_samples_per_second": 27.02,
+      "eval_steps_per_second": 1.135,
+      "step": 2360
+    },
+    {
+      "epoch": 2.5873972602739728,
+      "grad_norm": 2.5266482830047607,
+      "learning_rate": 7.940698619008936e-07,
+      "loss": 1.6654,
+      "step": 2361
+    },
+    {
+      "epoch": 2.5884931506849314,
+      "grad_norm": 2.6038002967834473,
+      "learning_rate": 7.920389926888709e-07,
+      "loss": 1.7532,
+      "step": 2362
+    },
+    {
+      "epoch": 2.5895890410958904,
+      "grad_norm": 2.716871976852417,
+      "learning_rate": 7.900081234768482e-07,
+      "loss": 1.8235,
+      "step": 2363
+    },
+    {
+      "epoch": 2.5906849315068494,
+      "grad_norm": 2.6962411403656006,
+      "learning_rate": 7.879772542648254e-07,
+      "loss": 1.6927,
+      "step": 2364
+    },
+    {
+      "epoch": 2.5917808219178085,
+      "grad_norm": 2.670116662979126,
+      "learning_rate": 7.859463850528026e-07,
+      "loss": 1.7448,
+      "step": 2365
+    },
+    {
+      "epoch": 2.592876712328767,
+      "grad_norm": 2.6886184215545654,
+      "learning_rate": 7.839155158407799e-07,
+      "loss": 1.7369,
+      "step": 2366
+    },
+    {
+      "epoch": 2.593972602739726,
+      "grad_norm": 2.6570565700531006,
+      "learning_rate": 7.818846466287572e-07,
+      "loss": 1.7368,
+      "step": 2367
+    },
+    {
+      "epoch": 2.595068493150685,
+      "grad_norm": 2.551548719406128,
+      "learning_rate": 7.798537774167345e-07,
+      "loss": 1.7357,
+      "step": 2368
+    },
+    {
+      "epoch": 2.5961643835616437,
+      "grad_norm": 2.646280527114868,
+      "learning_rate": 7.778229082047117e-07,
+      "loss": 1.7614,
+      "step": 2369
+    },
+    {
+      "epoch": 2.5972602739726027,
+      "grad_norm": 2.626239776611328,
+      "learning_rate": 7.757920389926889e-07,
+      "loss": 1.694,
+      "step": 2370
+    },
+    {
+      "epoch": 2.5972602739726027,
+      "eval_loss": 1.9757529497146606,
+      "eval_runtime": 36.9858,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 2370
+    },
+    {
+      "epoch": 2.5983561643835618,
+      "grad_norm": 2.6001148223876953,
+      "learning_rate": 7.737611697806662e-07,
+      "loss": 1.7342,
+      "step": 2371
+    },
+    {
+      "epoch": 2.5994520547945204,
+      "grad_norm": 2.5435168743133545,
+      "learning_rate": 7.717303005686435e-07,
+      "loss": 1.7031,
+      "step": 2372
+    },
+    {
+      "epoch": 2.6005479452054794,
+      "grad_norm": 2.652031183242798,
+      "learning_rate": 7.696994313566207e-07,
+      "loss": 1.7598,
+      "step": 2373
+    },
+    {
+      "epoch": 2.6016438356164384,
+      "grad_norm": 2.635340690612793,
+      "learning_rate": 7.67668562144598e-07,
+      "loss": 1.8008,
+      "step": 2374
+    },
+    {
+      "epoch": 2.602739726027397,
+      "grad_norm": 2.6935770511627197,
+      "learning_rate": 7.656376929325753e-07,
+      "loss": 1.6821,
+      "step": 2375
+    },
+    {
+      "epoch": 2.603835616438356,
+      "grad_norm": 2.621493339538574,
+      "learning_rate": 7.636068237205524e-07,
+      "loss": 1.6844,
+      "step": 2376
+    },
+    {
+      "epoch": 2.604931506849315,
+      "grad_norm": 2.6231918334960938,
+      "learning_rate": 7.615759545085296e-07,
+      "loss": 1.7622,
+      "step": 2377
+    },
+    {
+      "epoch": 2.606027397260274,
+      "grad_norm": 2.608091115951538,
+      "learning_rate": 7.595450852965069e-07,
+      "loss": 1.6772,
+      "step": 2378
+    },
+    {
+      "epoch": 2.607123287671233,
+      "grad_norm": 2.7701456546783447,
+      "learning_rate": 7.575142160844842e-07,
+      "loss": 1.7139,
+      "step": 2379
+    },
+    {
+      "epoch": 2.6082191780821917,
+      "grad_norm": 2.6648287773132324,
+      "learning_rate": 7.554833468724615e-07,
+      "loss": 1.7525,
+      "step": 2380
+    },
+    {
+      "epoch": 2.6082191780821917,
+      "eval_loss": 1.9792429208755493,
+      "eval_runtime": 37.0103,
+      "eval_samples_per_second": 27.02,
+      "eval_steps_per_second": 1.135,
+      "step": 2380
+    },
+    {
+      "epoch": 2.6093150684931508,
+      "grad_norm": 2.690938949584961,
+      "learning_rate": 7.534524776604387e-07,
+      "loss": 1.679,
+      "step": 2381
+    },
+    {
+      "epoch": 2.61041095890411,
+      "grad_norm": 2.658691167831421,
+      "learning_rate": 7.51421608448416e-07,
+      "loss": 1.7033,
+      "step": 2382
+    },
+    {
+      "epoch": 2.6115068493150684,
+      "grad_norm": 2.576317310333252,
+      "learning_rate": 7.493907392363932e-07,
+      "loss": 1.7527,
+      "step": 2383
+    },
+    {
+      "epoch": 2.6126027397260274,
+      "grad_norm": 2.5913753509521484,
+      "learning_rate": 7.473598700243705e-07,
+      "loss": 1.702,
+      "step": 2384
+    },
+    {
+      "epoch": 2.6136986301369864,
+      "grad_norm": 2.61553692817688,
+      "learning_rate": 7.453290008123477e-07,
+      "loss": 1.6568,
+      "step": 2385
+    },
+    {
+      "epoch": 2.614794520547945,
+      "grad_norm": 2.6631391048431396,
+      "learning_rate": 7.43298131600325e-07,
+      "loss": 1.7216,
+      "step": 2386
+    },
+    {
+      "epoch": 2.615890410958904,
+      "grad_norm": 2.6120619773864746,
+      "learning_rate": 7.412672623883023e-07,
+      "loss": 1.7278,
+      "step": 2387
+    },
+    {
+      "epoch": 2.616986301369863,
+      "grad_norm": 2.6166980266571045,
+      "learning_rate": 7.392363931762796e-07,
+      "loss": 1.7275,
+      "step": 2388
+    },
+    {
+      "epoch": 2.6180821917808217,
+      "grad_norm": 2.7203712463378906,
+      "learning_rate": 7.372055239642567e-07,
+      "loss": 1.6657,
+      "step": 2389
+    },
+    {
+      "epoch": 2.6191780821917807,
+      "grad_norm": 2.6785244941711426,
+      "learning_rate": 7.35174654752234e-07,
+      "loss": 1.7206,
+      "step": 2390
+    },
+    {
+      "epoch": 2.6191780821917807,
+      "eval_loss": 1.9763447046279907,
+      "eval_runtime": 37.0352,
+      "eval_samples_per_second": 27.001,
+      "eval_steps_per_second": 1.134,
+      "step": 2390
+    },
+    {
+      "epoch": 2.6202739726027398,
+      "grad_norm": 2.72511887550354,
+      "learning_rate": 7.331437855402113e-07,
+      "loss": 1.7474,
+      "step": 2391
+    },
+    {
+      "epoch": 2.621369863013699,
+      "grad_norm": 2.6428632736206055,
+      "learning_rate": 7.311129163281886e-07,
+      "loss": 1.7694,
+      "step": 2392
+    },
+    {
+      "epoch": 2.6224657534246574,
+      "grad_norm": 2.6593737602233887,
+      "learning_rate": 7.290820471161658e-07,
+      "loss": 1.6883,
+      "step": 2393
+    },
+    {
+      "epoch": 2.6235616438356164,
+      "grad_norm": 2.6048357486724854,
+      "learning_rate": 7.270511779041431e-07,
+      "loss": 1.7457,
+      "step": 2394
+    },
+    {
+      "epoch": 2.6246575342465754,
+      "grad_norm": 2.6435704231262207,
+      "learning_rate": 7.250203086921203e-07,
+      "loss": 1.7508,
+      "step": 2395
+    },
+    {
+      "epoch": 2.6257534246575345,
+      "grad_norm": 2.609815835952759,
+      "learning_rate": 7.229894394800976e-07,
+      "loss": 1.6603,
+      "step": 2396
+    },
+    {
+      "epoch": 2.626849315068493,
+      "grad_norm": 2.5431087017059326,
+      "learning_rate": 7.209585702680748e-07,
+      "loss": 1.725,
+      "step": 2397
+    },
+    {
+      "epoch": 2.627945205479452,
+      "grad_norm": 2.6712050437927246,
+      "learning_rate": 7.189277010560521e-07,
+      "loss": 1.7428,
+      "step": 2398
+    },
+    {
+      "epoch": 2.629041095890411,
+      "grad_norm": 2.6398820877075195,
+      "learning_rate": 7.168968318440294e-07,
+      "loss": 1.6834,
+      "step": 2399
+    },
+    {
+      "epoch": 2.6301369863013697,
+      "grad_norm": 2.5986127853393555,
+      "learning_rate": 7.148659626320067e-07,
+      "loss": 1.6985,
+      "step": 2400
+    },
+    {
+      "epoch": 2.6301369863013697,
+      "eval_loss": 1.9727513790130615,
+      "eval_runtime": 37.0552,
+      "eval_samples_per_second": 26.987,
+      "eval_steps_per_second": 1.133,
+      "step": 2400
+    },
+    {
+      "epoch": 2.6312328767123287,
+      "grad_norm": 2.6422410011291504,
+      "learning_rate": 7.128350934199837e-07,
+      "loss": 1.6732,
+      "step": 2401
+    },
+    {
+      "epoch": 2.632328767123288,
+      "grad_norm": 2.59894061088562,
+      "learning_rate": 7.10804224207961e-07,
+      "loss": 1.7855,
+      "step": 2402
+    },
+    {
+      "epoch": 2.6334246575342464,
+      "grad_norm": 2.6503429412841797,
+      "learning_rate": 7.087733549959383e-07,
+      "loss": 1.7768,
+      "step": 2403
+    },
+    {
+      "epoch": 2.6345205479452054,
+      "grad_norm": 2.5734922885894775,
+      "learning_rate": 7.067424857839155e-07,
+      "loss": 1.7221,
+      "step": 2404
+    },
+    {
+      "epoch": 2.6356164383561644,
+      "grad_norm": 2.5473709106445312,
+      "learning_rate": 7.047116165718928e-07,
+      "loss": 1.7116,
+      "step": 2405
+    },
+    {
+      "epoch": 2.636712328767123,
+      "grad_norm": 2.655651330947876,
+      "learning_rate": 7.026807473598701e-07,
+      "loss": 1.7085,
+      "step": 2406
+    },
+    {
+      "epoch": 2.637808219178082,
+      "grad_norm": 2.615900993347168,
+      "learning_rate": 7.006498781478473e-07,
+      "loss": 1.7189,
+      "step": 2407
+    },
+    {
+      "epoch": 2.638904109589041,
+      "grad_norm": 2.5976428985595703,
+      "learning_rate": 6.986190089358245e-07,
+      "loss": 1.7902,
+      "step": 2408
+    },
+    {
+      "epoch": 2.64,
+      "grad_norm": 2.6213088035583496,
+      "learning_rate": 6.965881397238018e-07,
+      "loss": 1.6842,
+      "step": 2409
+    },
+    {
+      "epoch": 2.641095890410959,
+      "grad_norm": 2.476685047149658,
+      "learning_rate": 6.945572705117791e-07,
+      "loss": 1.7297,
+      "step": 2410
+    },
+    {
+      "epoch": 2.641095890410959,
+      "eval_loss": 1.9702810049057007,
+      "eval_runtime": 36.7953,
+      "eval_samples_per_second": 27.177,
+      "eval_steps_per_second": 1.141,
+      "step": 2410
+    },
+    {
+      "epoch": 2.6421917808219177,
+      "grad_norm": 2.625504732131958,
+      "learning_rate": 6.925264012997564e-07,
+      "loss": 1.7335,
+      "step": 2411
+    },
+    {
+      "epoch": 2.643287671232877,
+      "grad_norm": 2.849334716796875,
+      "learning_rate": 6.904955320877336e-07,
+      "loss": 1.7339,
+      "step": 2412
+    },
+    {
+      "epoch": 2.644383561643836,
+      "grad_norm": 2.7533011436462402,
+      "learning_rate": 6.884646628757109e-07,
+      "loss": 1.7431,
+      "step": 2413
+    },
+    {
+      "epoch": 2.6454794520547944,
+      "grad_norm": 2.6793112754821777,
+      "learning_rate": 6.864337936636881e-07,
+      "loss": 1.7474,
+      "step": 2414
+    },
+    {
+      "epoch": 2.6465753424657534,
+      "grad_norm": 2.68094539642334,
+      "learning_rate": 6.844029244516654e-07,
+      "loss": 1.8092,
+      "step": 2415
+    },
+    {
+      "epoch": 2.6476712328767125,
+      "grad_norm": 2.7346346378326416,
+      "learning_rate": 6.823720552396426e-07,
+      "loss": 1.7838,
+      "step": 2416
+    },
+    {
+      "epoch": 2.648767123287671,
+      "grad_norm": 2.6850287914276123,
+      "learning_rate": 6.803411860276199e-07,
+      "loss": 1.6927,
+      "step": 2417
+    },
+    {
+      "epoch": 2.64986301369863,
+      "grad_norm": 2.5687966346740723,
+      "learning_rate": 6.783103168155972e-07,
+      "loss": 1.7505,
+      "step": 2418
+    },
+    {
+      "epoch": 2.650958904109589,
+      "grad_norm": 2.515760660171509,
+      "learning_rate": 6.762794476035745e-07,
+      "loss": 1.7285,
+      "step": 2419
+    },
+    {
+      "epoch": 2.6520547945205477,
+      "grad_norm": 2.604621171951294,
+      "learning_rate": 6.742485783915515e-07,
+      "loss": 1.7148,
+      "step": 2420
+    },
+    {
+      "epoch": 2.6520547945205477,
+      "eval_loss": 1.9678161144256592,
+      "eval_runtime": 36.9829,
+      "eval_samples_per_second": 27.04,
+      "eval_steps_per_second": 1.136,
+      "step": 2420
+    },
+    {
+      "epoch": 2.6531506849315067,
+      "grad_norm": 2.5340449810028076,
+      "learning_rate": 6.722177091795288e-07,
+      "loss": 1.756,
+      "step": 2421
+    },
+    {
+      "epoch": 2.6542465753424658,
+      "grad_norm": 2.5800235271453857,
+      "learning_rate": 6.701868399675061e-07,
+      "loss": 1.7439,
+      "step": 2422
+    },
+    {
+      "epoch": 2.655342465753425,
+      "grad_norm": 2.6596486568450928,
+      "learning_rate": 6.681559707554834e-07,
+      "loss": 1.7351,
+      "step": 2423
+    },
+    {
+      "epoch": 2.6564383561643834,
+      "grad_norm": 2.55759334564209,
+      "learning_rate": 6.661251015434606e-07,
+      "loss": 1.6978,
+      "step": 2424
+    },
+    {
+      "epoch": 2.6575342465753424,
+      "grad_norm": 2.6761016845703125,
+      "learning_rate": 6.640942323314379e-07,
+      "loss": 1.7011,
+      "step": 2425
+    },
+    {
+      "epoch": 2.6586301369863015,
+      "grad_norm": 2.5663347244262695,
+      "learning_rate": 6.620633631194151e-07,
+      "loss": 1.6891,
+      "step": 2426
+    },
+    {
+      "epoch": 2.6597260273972605,
+      "grad_norm": 2.64680552482605,
+      "learning_rate": 6.600324939073924e-07,
+      "loss": 1.703,
+      "step": 2427
+    },
+    {
+      "epoch": 2.660821917808219,
+      "grad_norm": 2.730487585067749,
+      "learning_rate": 6.580016246953696e-07,
+      "loss": 1.7329,
+      "step": 2428
+    },
+    {
+      "epoch": 2.661917808219178,
+      "grad_norm": 2.6785171031951904,
+      "learning_rate": 6.559707554833469e-07,
+      "loss": 1.7126,
+      "step": 2429
+    },
+    {
+      "epoch": 2.663013698630137,
+      "grad_norm": 2.657883882522583,
+      "learning_rate": 6.539398862713242e-07,
+      "loss": 1.7663,
+      "step": 2430
+    },
+    {
+      "epoch": 2.663013698630137,
+      "eval_loss": 1.968948245048523,
+      "eval_runtime": 37.0622,
+      "eval_samples_per_second": 26.982,
+      "eval_steps_per_second": 1.133,
+      "step": 2430
+    },
+    {
+      "epoch": 2.6641095890410957,
+      "grad_norm": 2.5575273036956787,
+      "learning_rate": 6.519090170593015e-07,
+      "loss": 1.7067,
+      "step": 2431
+    },
+    {
+      "epoch": 2.6652054794520548,
+      "grad_norm": 2.664853811264038,
+      "learning_rate": 6.498781478472786e-07,
+      "loss": 1.7674,
+      "step": 2432
+    },
+    {
+      "epoch": 2.666301369863014,
+      "grad_norm": 2.571324348449707,
+      "learning_rate": 6.478472786352559e-07,
+      "loss": 1.7328,
+      "step": 2433
+    },
+    {
+      "epoch": 2.6673972602739724,
+      "grad_norm": 2.7143940925598145,
+      "learning_rate": 6.458164094232332e-07,
+      "loss": 1.7736,
+      "step": 2434
+    },
+    {
+      "epoch": 2.6684931506849314,
+      "grad_norm": 2.6819112300872803,
+      "learning_rate": 6.437855402112105e-07,
+      "loss": 1.7266,
+      "step": 2435
+    },
+    {
+      "epoch": 2.6695890410958905,
+      "grad_norm": 2.701815128326416,
+      "learning_rate": 6.417546709991877e-07,
+      "loss": 1.6662,
+      "step": 2436
+    },
+    {
+      "epoch": 2.670684931506849,
+      "grad_norm": 2.6434922218322754,
+      "learning_rate": 6.39723801787165e-07,
+      "loss": 1.6896,
+      "step": 2437
+    },
+    {
+      "epoch": 2.671780821917808,
+      "grad_norm": 2.738567590713501,
+      "learning_rate": 6.376929325751423e-07,
+      "loss": 1.7794,
+      "step": 2438
+    },
+    {
+      "epoch": 2.672876712328767,
+      "grad_norm": 2.6100597381591797,
+      "learning_rate": 6.356620633631195e-07,
+      "loss": 1.7159,
+      "step": 2439
+    },
+    {
+      "epoch": 2.673972602739726,
+      "grad_norm": 2.6096298694610596,
+      "learning_rate": 6.336311941510967e-07,
+      "loss": 1.7106,
+      "step": 2440
+    },
+    {
+      "epoch": 2.673972602739726,
+      "eval_loss": 1.967889428138733,
+      "eval_runtime": 37.0763,
+      "eval_samples_per_second": 26.971,
+      "eval_steps_per_second": 1.133,
+      "step": 2440
+    },
+    {
+      "epoch": 2.675068493150685,
+      "grad_norm": 2.671086072921753,
+      "learning_rate": 6.31600324939074e-07,
+      "loss": 1.7106,
+      "step": 2441
+    },
+    {
+      "epoch": 2.6761643835616438,
+      "grad_norm": 2.7339978218078613,
+      "learning_rate": 6.295694557270513e-07,
+      "loss": 1.734,
+      "step": 2442
+    },
+    {
+      "epoch": 2.677260273972603,
+      "grad_norm": 2.626713752746582,
+      "learning_rate": 6.275385865150286e-07,
+      "loss": 1.7677,
+      "step": 2443
+    },
+    {
+      "epoch": 2.678356164383562,
+      "grad_norm": 2.542891502380371,
+      "learning_rate": 6.255077173030058e-07,
+      "loss": 1.7415,
+      "step": 2444
+    },
+    {
+      "epoch": 2.6794520547945204,
+      "grad_norm": 2.657703161239624,
+      "learning_rate": 6.23476848090983e-07,
+      "loss": 1.724,
+      "step": 2445
+    },
+    {
+      "epoch": 2.6805479452054795,
+      "grad_norm": 2.6566591262817383,
+      "learning_rate": 6.214459788789602e-07,
+      "loss": 1.7319,
+      "step": 2446
+    },
+    {
+      "epoch": 2.6816438356164385,
+      "grad_norm": 2.684117078781128,
+      "learning_rate": 6.194151096669375e-07,
+      "loss": 1.7148,
+      "step": 2447
+    },
+    {
+      "epoch": 2.682739726027397,
+      "grad_norm": 2.6989500522613525,
+      "learning_rate": 6.173842404549147e-07,
+      "loss": 1.7515,
+      "step": 2448
+    },
+    {
+      "epoch": 2.683835616438356,
+      "grad_norm": 2.731233835220337,
+      "learning_rate": 6.15353371242892e-07,
+      "loss": 1.7195,
+      "step": 2449
+    },
+    {
+      "epoch": 2.684931506849315,
+      "grad_norm": 2.6298322677612305,
+      "learning_rate": 6.133225020308693e-07,
+      "loss": 1.7311,
+      "step": 2450
+    },
+    {
+      "epoch": 2.684931506849315,
+      "eval_loss": 1.9676637649536133,
+      "eval_runtime": 37.0926,
+      "eval_samples_per_second": 26.96,
+      "eval_steps_per_second": 1.132,
+      "step": 2450
+    },
+    {
+      "epoch": 2.6860273972602737,
+      "grad_norm": 2.57230806350708,
+      "learning_rate": 6.112916328188466e-07,
+      "loss": 1.6388,
+      "step": 2451
+    },
+    {
+      "epoch": 2.6871232876712328,
+      "grad_norm": 2.5881690979003906,
+      "learning_rate": 6.092607636068238e-07,
+      "loss": 1.7108,
+      "step": 2452
+    },
+    {
+      "epoch": 2.688219178082192,
+      "grad_norm": 2.732182264328003,
+      "learning_rate": 6.07229894394801e-07,
+      "loss": 1.7814,
+      "step": 2453
+    },
+    {
+      "epoch": 2.689315068493151,
+      "grad_norm": 2.6008358001708984,
+      "learning_rate": 6.051990251827783e-07,
+      "loss": 1.6969,
+      "step": 2454
+    },
+    {
+      "epoch": 2.69041095890411,
+      "grad_norm": 2.638119697570801,
+      "learning_rate": 6.031681559707556e-07,
+      "loss": 1.6973,
+      "step": 2455
+    },
+    {
+      "epoch": 2.6915068493150685,
+      "grad_norm": 2.5765764713287354,
+      "learning_rate": 6.011372867587327e-07,
+      "loss": 1.7385,
+      "step": 2456
+    },
+    {
+      "epoch": 2.6926027397260275,
+      "grad_norm": 2.6819193363189697,
+      "learning_rate": 5.9910641754671e-07,
+      "loss": 1.7311,
+      "step": 2457
+    },
+    {
+      "epoch": 2.6936986301369865,
+      "grad_norm": 2.562638521194458,
+      "learning_rate": 5.970755483346873e-07,
+      "loss": 1.6935,
+      "step": 2458
+    },
+    {
+      "epoch": 2.694794520547945,
+      "grad_norm": 2.6333518028259277,
+      "learning_rate": 5.950446791226646e-07,
+      "loss": 1.7083,
+      "step": 2459
+    },
+    {
+      "epoch": 2.695890410958904,
+      "grad_norm": 2.6318745613098145,
+      "learning_rate": 5.930138099106418e-07,
+      "loss": 1.7082,
+      "step": 2460
+    },
+    {
+      "epoch": 2.695890410958904,
+      "eval_loss": 1.9667798280715942,
+      "eval_runtime": 37.0761,
+      "eval_samples_per_second": 26.972,
+      "eval_steps_per_second": 1.133,
+      "step": 2460
+    },
+    {
+      "epoch": 2.696986301369863,
+      "grad_norm": 2.7109830379486084,
+      "learning_rate": 5.909829406986191e-07,
+      "loss": 1.7251,
+      "step": 2461
+    },
+    {
+      "epoch": 2.6980821917808218,
+      "grad_norm": 2.5945379734039307,
+      "learning_rate": 5.889520714865963e-07,
+      "loss": 1.7057,
+      "step": 2462
+    },
+    {
+      "epoch": 2.699178082191781,
+      "grad_norm": 2.658888816833496,
+      "learning_rate": 5.869212022745735e-07,
+      "loss": 1.7358,
+      "step": 2463
+    },
+    {
+      "epoch": 2.70027397260274,
+      "grad_norm": 2.6142160892486572,
+      "learning_rate": 5.848903330625508e-07,
+      "loss": 1.6777,
+      "step": 2464
+    },
+    {
+      "epoch": 2.7013698630136984,
+      "grad_norm": 2.6961355209350586,
+      "learning_rate": 5.828594638505281e-07,
+      "loss": 1.7342,
+      "step": 2465
+    },
+    {
+      "epoch": 2.7024657534246574,
+      "grad_norm": 2.527519464492798,
+      "learning_rate": 5.808285946385054e-07,
+      "loss": 1.7343,
+      "step": 2466
+    },
+    {
+      "epoch": 2.7035616438356165,
+      "grad_norm": 2.661269187927246,
+      "learning_rate": 5.787977254264827e-07,
+      "loss": 1.7084,
+      "step": 2467
+    },
+    {
+      "epoch": 2.704657534246575,
+      "grad_norm": 2.648587703704834,
+      "learning_rate": 5.767668562144598e-07,
+      "loss": 1.7417,
+      "step": 2468
+    },
+    {
+      "epoch": 2.705753424657534,
+      "grad_norm": 2.690467596054077,
+      "learning_rate": 5.747359870024371e-07,
+      "loss": 1.7651,
+      "step": 2469
+    },
+    {
+      "epoch": 2.706849315068493,
+      "grad_norm": 2.6351239681243896,
+      "learning_rate": 5.727051177904144e-07,
+      "loss": 1.7156,
+      "step": 2470
+    },
+    {
+      "epoch": 2.706849315068493,
+      "eval_loss": 1.9673000574111938,
+      "eval_runtime": 37.0884,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 2470
+    },
+    {
+      "epoch": 2.707945205479452,
+      "grad_norm": 2.674419403076172,
+      "learning_rate": 5.706742485783915e-07,
+      "loss": 1.7276,
+      "step": 2471
+    },
+    {
+      "epoch": 2.709041095890411,
+      "grad_norm": 2.626304864883423,
+      "learning_rate": 5.686433793663688e-07,
+      "loss": 1.7578,
+      "step": 2472
+    },
+    {
+      "epoch": 2.71013698630137,
+      "grad_norm": 2.6923720836639404,
+      "learning_rate": 5.666125101543461e-07,
+      "loss": 1.6909,
+      "step": 2473
+    },
+    {
+      "epoch": 2.711232876712329,
+      "grad_norm": 2.6496634483337402,
+      "learning_rate": 5.645816409423234e-07,
+      "loss": 1.6691,
+      "step": 2474
+    },
+    {
+      "epoch": 2.712328767123288,
+      "grad_norm": 2.640418529510498,
+      "learning_rate": 5.625507717303006e-07,
+      "loss": 1.7705,
+      "step": 2475
+    },
+    {
+      "epoch": 2.7134246575342464,
+      "grad_norm": 2.5379440784454346,
+      "learning_rate": 5.605199025182779e-07,
+      "loss": 1.7356,
+      "step": 2476
+    },
+    {
+      "epoch": 2.7145205479452055,
+      "grad_norm": 2.6212170124053955,
+      "learning_rate": 5.584890333062551e-07,
+      "loss": 1.7128,
+      "step": 2477
+    },
+    {
+      "epoch": 2.7156164383561645,
+      "grad_norm": 2.676152467727661,
+      "learning_rate": 5.564581640942324e-07,
+      "loss": 1.6933,
+      "step": 2478
+    },
+    {
+      "epoch": 2.716712328767123,
+      "grad_norm": 2.684361457824707,
+      "learning_rate": 5.544272948822096e-07,
+      "loss": 1.7651,
+      "step": 2479
+    },
+    {
+      "epoch": 2.717808219178082,
+      "grad_norm": 2.630239486694336,
+      "learning_rate": 5.523964256701868e-07,
+      "loss": 1.7054,
+      "step": 2480
+    },
+    {
+      "epoch": 2.717808219178082,
+      "eval_loss": 1.967214822769165,
+      "eval_runtime": 37.0987,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 2480
+    },
+    {
+      "epoch": 2.718904109589041,
+      "grad_norm": 2.5909178256988525,
+      "learning_rate": 5.503655564581641e-07,
+      "loss": 1.7152,
+      "step": 2481
+    },
+    {
+      "epoch": 2.7199999999999998,
+      "grad_norm": 2.6395678520202637,
+      "learning_rate": 5.483346872461414e-07,
+      "loss": 1.7383,
+      "step": 2482
+    },
+    {
+      "epoch": 2.721095890410959,
+      "grad_norm": 2.5965633392333984,
+      "learning_rate": 5.463038180341186e-07,
+      "loss": 1.7111,
+      "step": 2483
+    },
+    {
+      "epoch": 2.722191780821918,
+      "grad_norm": 2.5583207607269287,
+      "learning_rate": 5.442729488220959e-07,
+      "loss": 1.7126,
+      "step": 2484
+    },
+    {
+      "epoch": 2.723287671232877,
+      "grad_norm": 2.691272258758545,
+      "learning_rate": 5.422420796100732e-07,
+      "loss": 1.698,
+      "step": 2485
+    },
+    {
+      "epoch": 2.724383561643836,
+      "grad_norm": 2.5674681663513184,
+      "learning_rate": 5.402112103980505e-07,
+      "loss": 1.7397,
+      "step": 2486
+    },
+    {
+      "epoch": 2.7254794520547945,
+      "grad_norm": 2.626338243484497,
+      "learning_rate": 5.381803411860276e-07,
+      "loss": 1.7095,
+      "step": 2487
+    },
+    {
+      "epoch": 2.7265753424657535,
+      "grad_norm": 2.636225700378418,
+      "learning_rate": 5.361494719740049e-07,
+      "loss": 1.758,
+      "step": 2488
+    },
+    {
+      "epoch": 2.7276712328767125,
+      "grad_norm": 2.7304928302764893,
+      "learning_rate": 5.341186027619822e-07,
+      "loss": 1.7969,
+      "step": 2489
+    },
+    {
+      "epoch": 2.728767123287671,
+      "grad_norm": 2.660935640335083,
+      "learning_rate": 5.320877335499594e-07,
+      "loss": 1.6876,
+      "step": 2490
+    },
+    {
+      "epoch": 2.728767123287671,
+      "eval_loss": 1.9665592908859253,
+      "eval_runtime": 37.0848,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 2490
+    },
+    {
+      "epoch": 2.72986301369863,
+      "grad_norm": 2.623068332672119,
+      "learning_rate": 5.300568643379366e-07,
+      "loss": 1.699,
+      "step": 2491
+    },
+    {
+      "epoch": 2.730958904109589,
+      "grad_norm": 2.528489112854004,
+      "learning_rate": 5.280259951259139e-07,
+      "loss": 1.7789,
+      "step": 2492
+    },
+    {
+      "epoch": 2.732054794520548,
+      "grad_norm": 2.6441330909729004,
+      "learning_rate": 5.259951259138912e-07,
+      "loss": 1.7358,
+      "step": 2493
+    },
+    {
+      "epoch": 2.733150684931507,
+      "grad_norm": 2.6011250019073486,
+      "learning_rate": 5.239642567018685e-07,
+      "loss": 1.6946,
+      "step": 2494
+    },
+    {
+      "epoch": 2.734246575342466,
+      "grad_norm": 2.669063091278076,
+      "learning_rate": 5.219333874898457e-07,
+      "loss": 1.6922,
+      "step": 2495
+    },
+    {
+      "epoch": 2.7353424657534244,
+      "grad_norm": 2.7551729679107666,
+      "learning_rate": 5.199025182778229e-07,
+      "loss": 1.7191,
+      "step": 2496
+    },
+    {
+      "epoch": 2.7364383561643835,
+      "grad_norm": 2.730877637863159,
+      "learning_rate": 5.178716490658002e-07,
+      "loss": 1.7319,
+      "step": 2497
+    },
+    {
+      "epoch": 2.7375342465753425,
+      "grad_norm": 2.6159300804138184,
+      "learning_rate": 5.158407798537775e-07,
+      "loss": 1.6891,
+      "step": 2498
+    },
+    {
+      "epoch": 2.7386301369863015,
+      "grad_norm": 2.7436540126800537,
+      "learning_rate": 5.138099106417547e-07,
+      "loss": 1.651,
+      "step": 2499
+    },
+    {
+      "epoch": 2.73972602739726,
+      "grad_norm": 2.6271872520446777,
+      "learning_rate": 5.11779041429732e-07,
+      "loss": 1.6241,
+      "step": 2500
+    },
+    {
+      "epoch": 2.73972602739726,
+      "eval_loss": 1.9660766124725342,
+      "eval_runtime": 37.0963,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 2500
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2736,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.7932735234048e+17,
+  "train_batch_size": 3,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/training_args.bin b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..b7adbe5a2dbf27f05befec6cc65fbc93e90eefd0
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aaa882279512a8c278376c0c601206933f3c5662ddcd3c13c53bf32ae3680d14
+size 6011
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2500/zero_to_fp32.py
@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/config.json
@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "meta-llama/Llama-3.2-3B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "torch_dtype": "float16",
+  "transformers_version": "4.45.1",
+  "use_cache": true,
+  "vocab_size": 128256
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/generation_config.json
@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": 128001,
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.45.1"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/latest b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/latest
new file mode 100644
index 0000000000000000000000000000000000000000..cfe584fa09082a54d48545d05e97aee6fd00673d
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/latest
@@ -0,0 +1 @@
+global_step2600
\ No newline at end of file
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/model.safetensors.index.json
new file mode 100644
index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/model.safetensors.index.json
@@ -0,0 +1,262 @@
+{
+  "metadata": {
+    "total_size": 7213504512
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00002-of-00002.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/rng_state_0.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..320a5e760ae3ee4a341a9c842fdf11ca16991278
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b29ad903afb6749549a67ba4698ef3df31a1e3e9dfb1b5e1608cdc0c7f134fed
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/rng_state_1.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..1b64f430f12bc12bab631d5df0987bab9572751a
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbfc100a7ea6b8e39a9f3b3bd74445fc7ec7c756194072251f759b6ef0f7f0d9
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/rng_state_2.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..d92f61d03abfc0e5e45cc65b64d441bb21661caa
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:65c1c60da2259cc4f2bb221937d352d1d8d838bc52aff6746e8e832ad9930d34
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/scheduler.pt b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..7d941352c8c18291b41a41e57924503d624d57fa
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b40f83a17a29486af7309757ba81a563b7660434d8b54e89ca159b547d13d88d
+size 627
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/tokenizer.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb
+size 17210374
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128256": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|end_of_text|>",
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "[PAD]",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..01a5c793376dafdcdb1cf57399a8df2d803cc52d
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/trainer_state.json
@@ -0,0 +1,20313 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.8493150684931505,
+  "eval_steps": 10,
+  "global_step": 2600,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001095890410958904,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0636,
+      "step": 1
+    },
+    {
+      "epoch": 0.002191780821917808,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0501,
+      "step": 2
+    },
+    {
+      "epoch": 0.0032876712328767125,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0238,
+      "step": 3
+    },
+    {
+      "epoch": 0.004383561643835616,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0346,
+      "step": 4
+    },
+    {
+      "epoch": 0.005479452054794521,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0108,
+      "step": 5
+    },
+    {
+      "epoch": 0.006575342465753425,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.131,
+      "step": 6
+    },
+    {
+      "epoch": 0.007671232876712329,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0488,
+      "step": 7
+    },
+    {
+      "epoch": 0.008767123287671232,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0905,
+      "step": 8
+    },
+    {
+      "epoch": 0.009863013698630137,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0493,
+      "step": 9
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.1038,
+      "step": 10
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "eval_loss": 3.100078821182251,
+      "eval_runtime": 37.5838,
+      "eval_samples_per_second": 26.607,
+      "eval_steps_per_second": 1.118,
+      "step": 10
+    },
+    {
+      "epoch": 0.012054794520547946,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0804,
+      "step": 11
+    },
+    {
+      "epoch": 0.01315068493150685,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0053,
+      "step": 12
+    },
+    {
+      "epoch": 0.014246575342465753,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0607,
+      "step": 13
+    },
+    {
+      "epoch": 0.015342465753424657,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0539,
+      "step": 14
+    },
+    {
+      "epoch": 0.01643835616438356,
+      "grad_norm": 5.420984745025635,
+      "learning_rate": 1.8248175182481753e-08,
+      "loss": 3.0941,
+      "step": 15
+    },
+    {
+      "epoch": 0.017534246575342465,
+      "grad_norm": 5.76840877532959,
+      "learning_rate": 3.6496350364963505e-08,
+      "loss": 3.036,
+      "step": 16
+    },
+    {
+      "epoch": 0.01863013698630137,
+      "grad_norm": 5.905357360839844,
+      "learning_rate": 5.4744525547445264e-08,
+      "loss": 3.0491,
+      "step": 17
+    },
+    {
+      "epoch": 0.019726027397260273,
+      "grad_norm": 5.981553554534912,
+      "learning_rate": 7.299270072992701e-08,
+      "loss": 3.0775,
+      "step": 18
+    },
+    {
+      "epoch": 0.020821917808219178,
+      "grad_norm": 6.240172863006592,
+      "learning_rate": 9.124087591240876e-08,
+      "loss": 3.0345,
+      "step": 19
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "grad_norm": 5.59375,
+      "learning_rate": 1.0948905109489053e-07,
+      "loss": 3.1063,
+      "step": 20
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "eval_loss": 3.098820686340332,
+      "eval_runtime": 37.2695,
+      "eval_samples_per_second": 26.832,
+      "eval_steps_per_second": 1.127,
+      "step": 20
+    },
+    {
+      "epoch": 0.023013698630136987,
+      "grad_norm": 5.819795608520508,
+      "learning_rate": 1.2773722627737227e-07,
+      "loss": 3.0808,
+      "step": 21
+    },
+    {
+      "epoch": 0.02410958904109589,
+      "grad_norm": 5.7714762687683105,
+      "learning_rate": 1.4598540145985402e-07,
+      "loss": 3.0055,
+      "step": 22
+    },
+    {
+      "epoch": 0.025205479452054796,
+      "grad_norm": 5.51552152633667,
+      "learning_rate": 1.642335766423358e-07,
+      "loss": 3.0396,
+      "step": 23
+    },
+    {
+      "epoch": 0.0263013698630137,
+      "grad_norm": 5.594012260437012,
+      "learning_rate": 1.8248175182481753e-07,
+      "loss": 3.0824,
+      "step": 24
+    },
+    {
+      "epoch": 0.0273972602739726,
+      "grad_norm": 5.1974029541015625,
+      "learning_rate": 2.0072992700729928e-07,
+      "loss": 3.08,
+      "step": 25
+    },
+    {
+      "epoch": 0.028493150684931506,
+      "grad_norm": 4.733968257904053,
+      "learning_rate": 2.1897810218978106e-07,
+      "loss": 3.0597,
+      "step": 26
+    },
+    {
+      "epoch": 0.02958904109589041,
+      "grad_norm": 4.717319011688232,
+      "learning_rate": 2.3722627737226278e-07,
+      "loss": 3.0696,
+      "step": 27
+    },
+    {
+      "epoch": 0.030684931506849315,
+      "grad_norm": 4.6303534507751465,
+      "learning_rate": 2.5547445255474454e-07,
+      "loss": 3.0273,
+      "step": 28
+    },
+    {
+      "epoch": 0.03178082191780822,
+      "grad_norm": 4.413999557495117,
+      "learning_rate": 2.737226277372263e-07,
+      "loss": 3.0202,
+      "step": 29
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "grad_norm": 4.516043186187744,
+      "learning_rate": 2.9197080291970804e-07,
+      "loss": 2.9926,
+      "step": 30
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "eval_loss": 3.0561797618865967,
+      "eval_runtime": 37.0351,
+      "eval_samples_per_second": 27.001,
+      "eval_steps_per_second": 1.134,
+      "step": 30
+    },
+    {
+      "epoch": 0.03397260273972603,
+      "grad_norm": 3.7179207801818848,
+      "learning_rate": 3.102189781021898e-07,
+      "loss": 3.0086,
+      "step": 31
+    },
+    {
+      "epoch": 0.03506849315068493,
+      "grad_norm": 3.7840425968170166,
+      "learning_rate": 3.284671532846716e-07,
+      "loss": 2.9986,
+      "step": 32
+    },
+    {
+      "epoch": 0.03616438356164384,
+      "grad_norm": 3.843116521835327,
+      "learning_rate": 3.467153284671533e-07,
+      "loss": 2.9677,
+      "step": 33
+    },
+    {
+      "epoch": 0.03726027397260274,
+      "grad_norm": 3.8528714179992676,
+      "learning_rate": 3.6496350364963505e-07,
+      "loss": 2.9732,
+      "step": 34
+    },
+    {
+      "epoch": 0.038356164383561646,
+      "grad_norm": 3.76963210105896,
+      "learning_rate": 3.8321167883211686e-07,
+      "loss": 2.9855,
+      "step": 35
+    },
+    {
+      "epoch": 0.03945205479452055,
+      "grad_norm": 3.6240015029907227,
+      "learning_rate": 4.0145985401459856e-07,
+      "loss": 3.0117,
+      "step": 36
+    },
+    {
+      "epoch": 0.040547945205479455,
+      "grad_norm": 3.1380624771118164,
+      "learning_rate": 4.197080291970803e-07,
+      "loss": 2.9481,
+      "step": 37
+    },
+    {
+      "epoch": 0.041643835616438356,
+      "grad_norm": 3.565329074859619,
+      "learning_rate": 4.379562043795621e-07,
+      "loss": 2.9175,
+      "step": 38
+    },
+    {
+      "epoch": 0.042739726027397264,
+      "grad_norm": 3.8062448501586914,
+      "learning_rate": 4.562043795620438e-07,
+      "loss": 3.0074,
+      "step": 39
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "grad_norm": 3.412046432495117,
+      "learning_rate": 4.7445255474452557e-07,
+      "loss": 2.9119,
+      "step": 40
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "eval_loss": 2.9874370098114014,
+      "eval_runtime": 37.0514,
+      "eval_samples_per_second": 26.99,
+      "eval_steps_per_second": 1.134,
+      "step": 40
+    },
+    {
+      "epoch": 0.044931506849315066,
+      "grad_norm": 3.227437973022461,
+      "learning_rate": 4.927007299270073e-07,
+      "loss": 2.9436,
+      "step": 41
+    },
+    {
+      "epoch": 0.046027397260273974,
+      "grad_norm": 2.9633443355560303,
+      "learning_rate": 5.109489051094891e-07,
+      "loss": 2.967,
+      "step": 42
+    },
+    {
+      "epoch": 0.047123287671232875,
+      "grad_norm": 2.942678928375244,
+      "learning_rate": 5.291970802919708e-07,
+      "loss": 2.9375,
+      "step": 43
+    },
+    {
+      "epoch": 0.04821917808219178,
+      "grad_norm": 2.9806888103485107,
+      "learning_rate": 5.474452554744526e-07,
+      "loss": 2.9474,
+      "step": 44
+    },
+    {
+      "epoch": 0.049315068493150684,
+      "grad_norm": 3.000063419342041,
+      "learning_rate": 5.656934306569343e-07,
+      "loss": 2.9289,
+      "step": 45
+    },
+    {
+      "epoch": 0.05041095890410959,
+      "grad_norm": 3.1481194496154785,
+      "learning_rate": 5.839416058394161e-07,
+      "loss": 2.8869,
+      "step": 46
+    },
+    {
+      "epoch": 0.05150684931506849,
+      "grad_norm": 2.949202060699463,
+      "learning_rate": 6.021897810218979e-07,
+      "loss": 2.8783,
+      "step": 47
+    },
+    {
+      "epoch": 0.0526027397260274,
+      "grad_norm": 2.706526279449463,
+      "learning_rate": 6.204379562043796e-07,
+      "loss": 2.9024,
+      "step": 48
+    },
+    {
+      "epoch": 0.0536986301369863,
+      "grad_norm": 2.7317655086517334,
+      "learning_rate": 6.386861313868614e-07,
+      "loss": 2.8325,
+      "step": 49
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "grad_norm": 2.25644588470459,
+      "learning_rate": 6.569343065693432e-07,
+      "loss": 2.9409,
+      "step": 50
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "eval_loss": 2.909320116043091,
+      "eval_runtime": 37.0714,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 50
+    },
+    {
+      "epoch": 0.05589041095890411,
+      "grad_norm": 2.8789613246917725,
+      "learning_rate": 6.751824817518248e-07,
+      "loss": 2.8459,
+      "step": 51
+    },
+    {
+      "epoch": 0.05698630136986301,
+      "grad_norm": 2.686873435974121,
+      "learning_rate": 6.934306569343066e-07,
+      "loss": 2.8653,
+      "step": 52
+    },
+    {
+      "epoch": 0.05808219178082192,
+      "grad_norm": 2.54382586479187,
+      "learning_rate": 7.116788321167883e-07,
+      "loss": 2.7973,
+      "step": 53
+    },
+    {
+      "epoch": 0.05917808219178082,
+      "grad_norm": 2.1871724128723145,
+      "learning_rate": 7.299270072992701e-07,
+      "loss": 2.7982,
+      "step": 54
+    },
+    {
+      "epoch": 0.06027397260273973,
+      "grad_norm": 2.145303249359131,
+      "learning_rate": 7.48175182481752e-07,
+      "loss": 2.8195,
+      "step": 55
+    },
+    {
+      "epoch": 0.06136986301369863,
+      "grad_norm": 2.1635069847106934,
+      "learning_rate": 7.664233576642337e-07,
+      "loss": 2.825,
+      "step": 56
+    },
+    {
+      "epoch": 0.06246575342465754,
+      "grad_norm": 2.3530752658843994,
+      "learning_rate": 7.846715328467154e-07,
+      "loss": 2.8407,
+      "step": 57
+    },
+    {
+      "epoch": 0.06356164383561644,
+      "grad_norm": 2.2455942630767822,
+      "learning_rate": 8.029197080291971e-07,
+      "loss": 2.8182,
+      "step": 58
+    },
+    {
+      "epoch": 0.06465753424657535,
+      "grad_norm": 1.8985673189163208,
+      "learning_rate": 8.211678832116789e-07,
+      "loss": 2.772,
+      "step": 59
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "grad_norm": 1.9699499607086182,
+      "learning_rate": 8.394160583941606e-07,
+      "loss": 2.7367,
+      "step": 60
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "eval_loss": 2.8451859951019287,
+      "eval_runtime": 37.0793,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 60
+    },
+    {
+      "epoch": 0.06684931506849315,
+      "grad_norm": 2.159547805786133,
+      "learning_rate": 8.576642335766425e-07,
+      "loss": 2.7635,
+      "step": 61
+    },
+    {
+      "epoch": 0.06794520547945206,
+      "grad_norm": 2.12705659866333,
+      "learning_rate": 8.759124087591242e-07,
+      "loss": 2.7624,
+      "step": 62
+    },
+    {
+      "epoch": 0.06904109589041096,
+      "grad_norm": 1.8393241167068481,
+      "learning_rate": 8.94160583941606e-07,
+      "loss": 2.7763,
+      "step": 63
+    },
+    {
+      "epoch": 0.07013698630136986,
+      "grad_norm": 1.683522343635559,
+      "learning_rate": 9.124087591240876e-07,
+      "loss": 2.7762,
+      "step": 64
+    },
+    {
+      "epoch": 0.07123287671232877,
+      "grad_norm": 1.9260376691818237,
+      "learning_rate": 9.306569343065694e-07,
+      "loss": 2.7418,
+      "step": 65
+    },
+    {
+      "epoch": 0.07232876712328767,
+      "grad_norm": 2.09322190284729,
+      "learning_rate": 9.489051094890511e-07,
+      "loss": 2.7625,
+      "step": 66
+    },
+    {
+      "epoch": 0.07342465753424658,
+      "grad_norm": 1.9183429479599,
+      "learning_rate": 9.67153284671533e-07,
+      "loss": 2.7453,
+      "step": 67
+    },
+    {
+      "epoch": 0.07452054794520548,
+      "grad_norm": 1.9792778491973877,
+      "learning_rate": 9.854014598540146e-07,
+      "loss": 2.7858,
+      "step": 68
+    },
+    {
+      "epoch": 0.07561643835616438,
+      "grad_norm": 2.1397311687469482,
+      "learning_rate": 1.0036496350364965e-06,
+      "loss": 2.7242,
+      "step": 69
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "grad_norm": 1.8556463718414307,
+      "learning_rate": 1.0218978102189781e-06,
+      "loss": 2.6853,
+      "step": 70
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "eval_loss": 2.7787046432495117,
+      "eval_runtime": 37.1017,
+      "eval_samples_per_second": 26.953,
+      "eval_steps_per_second": 1.132,
+      "step": 70
+    },
+    {
+      "epoch": 0.07780821917808219,
+      "grad_norm": 1.9776800870895386,
+      "learning_rate": 1.04014598540146e-06,
+      "loss": 2.749,
+      "step": 71
+    },
+    {
+      "epoch": 0.0789041095890411,
+      "grad_norm": 2.3972630500793457,
+      "learning_rate": 1.0583941605839416e-06,
+      "loss": 2.6457,
+      "step": 72
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 2.090210199356079,
+      "learning_rate": 1.0766423357664233e-06,
+      "loss": 2.6401,
+      "step": 73
+    },
+    {
+      "epoch": 0.08109589041095891,
+      "grad_norm": 2.4117701053619385,
+      "learning_rate": 1.0948905109489052e-06,
+      "loss": 2.6649,
+      "step": 74
+    },
+    {
+      "epoch": 0.0821917808219178,
+      "grad_norm": 2.4845173358917236,
+      "learning_rate": 1.113138686131387e-06,
+      "loss": 2.6855,
+      "step": 75
+    },
+    {
+      "epoch": 0.08328767123287671,
+      "grad_norm": 1.7779242992401123,
+      "learning_rate": 1.1313868613138687e-06,
+      "loss": 2.6926,
+      "step": 76
+    },
+    {
+      "epoch": 0.08438356164383562,
+      "grad_norm": 2.3163535594940186,
+      "learning_rate": 1.1496350364963505e-06,
+      "loss": 2.6302,
+      "step": 77
+    },
+    {
+      "epoch": 0.08547945205479453,
+      "grad_norm": 2.2296528816223145,
+      "learning_rate": 1.1678832116788322e-06,
+      "loss": 2.6054,
+      "step": 78
+    },
+    {
+      "epoch": 0.08657534246575342,
+      "grad_norm": 2.253431797027588,
+      "learning_rate": 1.1861313868613138e-06,
+      "loss": 2.6104,
+      "step": 79
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "grad_norm": 2.490609645843506,
+      "learning_rate": 1.2043795620437959e-06,
+      "loss": 2.6078,
+      "step": 80
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "eval_loss": 2.706103563308716,
+      "eval_runtime": 37.0892,
+      "eval_samples_per_second": 26.962,
+      "eval_steps_per_second": 1.132,
+      "step": 80
+    },
+    {
+      "epoch": 0.08876712328767124,
+      "grad_norm": 2.3048248291015625,
+      "learning_rate": 1.2226277372262775e-06,
+      "loss": 2.6096,
+      "step": 81
+    },
+    {
+      "epoch": 0.08986301369863013,
+      "grad_norm": 2.440666913986206,
+      "learning_rate": 1.2408759124087592e-06,
+      "loss": 2.6412,
+      "step": 82
+    },
+    {
+      "epoch": 0.09095890410958904,
+      "grad_norm": 1.9456374645233154,
+      "learning_rate": 1.259124087591241e-06,
+      "loss": 2.591,
+      "step": 83
+    },
+    {
+      "epoch": 0.09205479452054795,
+      "grad_norm": 2.598172664642334,
+      "learning_rate": 1.2773722627737229e-06,
+      "loss": 2.6093,
+      "step": 84
+    },
+    {
+      "epoch": 0.09315068493150686,
+      "grad_norm": 2.2587032318115234,
+      "learning_rate": 1.2956204379562045e-06,
+      "loss": 2.5949,
+      "step": 85
+    },
+    {
+      "epoch": 0.09424657534246575,
+      "grad_norm": 2.2927448749542236,
+      "learning_rate": 1.3138686131386864e-06,
+      "loss": 2.5464,
+      "step": 86
+    },
+    {
+      "epoch": 0.09534246575342466,
+      "grad_norm": 2.5940468311309814,
+      "learning_rate": 1.332116788321168e-06,
+      "loss": 2.5613,
+      "step": 87
+    },
+    {
+      "epoch": 0.09643835616438357,
+      "grad_norm": 2.073207378387451,
+      "learning_rate": 1.3503649635036497e-06,
+      "loss": 2.4955,
+      "step": 88
+    },
+    {
+      "epoch": 0.09753424657534246,
+      "grad_norm": 3.211766242980957,
+      "learning_rate": 1.3686131386861315e-06,
+      "loss": 2.5315,
+      "step": 89
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "grad_norm": 2.2590038776397705,
+      "learning_rate": 1.3868613138686132e-06,
+      "loss": 2.4641,
+      "step": 90
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "eval_loss": 2.625087261199951,
+      "eval_runtime": 37.0852,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 90
+    },
+    {
+      "epoch": 0.09972602739726028,
+      "grad_norm": 2.3959836959838867,
+      "learning_rate": 1.405109489051095e-06,
+      "loss": 2.5537,
+      "step": 91
+    },
+    {
+      "epoch": 0.10082191780821918,
+      "grad_norm": 2.596761465072632,
+      "learning_rate": 1.4233576642335767e-06,
+      "loss": 2.5135,
+      "step": 92
+    },
+    {
+      "epoch": 0.10191780821917808,
+      "grad_norm": 2.475759267807007,
+      "learning_rate": 1.4416058394160583e-06,
+      "loss": 2.5436,
+      "step": 93
+    },
+    {
+      "epoch": 0.10301369863013699,
+      "grad_norm": 3.0613298416137695,
+      "learning_rate": 1.4598540145985402e-06,
+      "loss": 2.5066,
+      "step": 94
+    },
+    {
+      "epoch": 0.10410958904109589,
+      "grad_norm": 2.406097650527954,
+      "learning_rate": 1.478102189781022e-06,
+      "loss": 2.4646,
+      "step": 95
+    },
+    {
+      "epoch": 0.1052054794520548,
+      "grad_norm": 2.5872504711151123,
+      "learning_rate": 1.496350364963504e-06,
+      "loss": 2.5351,
+      "step": 96
+    },
+    {
+      "epoch": 0.1063013698630137,
+      "grad_norm": 2.4294471740722656,
+      "learning_rate": 1.5145985401459856e-06,
+      "loss": 2.4953,
+      "step": 97
+    },
+    {
+      "epoch": 0.1073972602739726,
+      "grad_norm": 2.674884080886841,
+      "learning_rate": 1.5328467153284674e-06,
+      "loss": 2.5133,
+      "step": 98
+    },
+    {
+      "epoch": 0.10849315068493151,
+      "grad_norm": 2.5096852779388428,
+      "learning_rate": 1.551094890510949e-06,
+      "loss": 2.4693,
+      "step": 99
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "grad_norm": 2.7390334606170654,
+      "learning_rate": 1.5693430656934307e-06,
+      "loss": 2.4195,
+      "step": 100
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "eval_loss": 2.55997371673584,
+      "eval_runtime": 37.0732,
+      "eval_samples_per_second": 26.974,
+      "eval_steps_per_second": 1.133,
+      "step": 100
+    },
+    {
+      "epoch": 0.11068493150684931,
+      "grad_norm": 2.8887312412261963,
+      "learning_rate": 1.5875912408759126e-06,
+      "loss": 2.4193,
+      "step": 101
+    },
+    {
+      "epoch": 0.11178082191780822,
+      "grad_norm": 3.0721421241760254,
+      "learning_rate": 1.6058394160583942e-06,
+      "loss": 2.4709,
+      "step": 102
+    },
+    {
+      "epoch": 0.11287671232876713,
+      "grad_norm": 2.728156566619873,
+      "learning_rate": 1.624087591240876e-06,
+      "loss": 2.5105,
+      "step": 103
+    },
+    {
+      "epoch": 0.11397260273972602,
+      "grad_norm": 3.1214346885681152,
+      "learning_rate": 1.6423357664233577e-06,
+      "loss": 2.4316,
+      "step": 104
+    },
+    {
+      "epoch": 0.11506849315068493,
+      "grad_norm": 2.9249517917633057,
+      "learning_rate": 1.6605839416058394e-06,
+      "loss": 2.3737,
+      "step": 105
+    },
+    {
+      "epoch": 0.11616438356164384,
+      "grad_norm": 2.6646173000335693,
+      "learning_rate": 1.6788321167883212e-06,
+      "loss": 2.4165,
+      "step": 106
+    },
+    {
+      "epoch": 0.11726027397260275,
+      "grad_norm": 3.108067750930786,
+      "learning_rate": 1.6970802919708029e-06,
+      "loss": 2.4217,
+      "step": 107
+    },
+    {
+      "epoch": 0.11835616438356164,
+      "grad_norm": 3.142338991165161,
+      "learning_rate": 1.715328467153285e-06,
+      "loss": 2.3527,
+      "step": 108
+    },
+    {
+      "epoch": 0.11945205479452055,
+      "grad_norm": 2.9078011512756348,
+      "learning_rate": 1.7335766423357666e-06,
+      "loss": 2.3927,
+      "step": 109
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "grad_norm": 3.40766978263855,
+      "learning_rate": 1.7518248175182485e-06,
+      "loss": 2.4021,
+      "step": 110
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "eval_loss": 2.5031485557556152,
+      "eval_runtime": 36.8147,
+      "eval_samples_per_second": 27.163,
+      "eval_steps_per_second": 1.141,
+      "step": 110
+    },
+    {
+      "epoch": 0.12164383561643835,
+      "grad_norm": 4.396062850952148,
+      "learning_rate": 1.77007299270073e-06,
+      "loss": 2.4373,
+      "step": 111
+    },
+    {
+      "epoch": 0.12273972602739726,
+      "grad_norm": 4.0441365242004395,
+      "learning_rate": 1.788321167883212e-06,
+      "loss": 2.4162,
+      "step": 112
+    },
+    {
+      "epoch": 0.12383561643835617,
+      "grad_norm": 2.8869969844818115,
+      "learning_rate": 1.8065693430656936e-06,
+      "loss": 2.4515,
+      "step": 113
+    },
+    {
+      "epoch": 0.12493150684931507,
+      "grad_norm": 3.4833669662475586,
+      "learning_rate": 1.8248175182481753e-06,
+      "loss": 2.3965,
+      "step": 114
+    },
+    {
+      "epoch": 0.12602739726027398,
+      "grad_norm": 2.9996068477630615,
+      "learning_rate": 1.8430656934306571e-06,
+      "loss": 2.3374,
+      "step": 115
+    },
+    {
+      "epoch": 0.12712328767123288,
+      "grad_norm": 2.701371192932129,
+      "learning_rate": 1.8613138686131388e-06,
+      "loss": 2.3608,
+      "step": 116
+    },
+    {
+      "epoch": 0.12821917808219177,
+      "grad_norm": 2.9491288661956787,
+      "learning_rate": 1.8795620437956206e-06,
+      "loss": 2.355,
+      "step": 117
+    },
+    {
+      "epoch": 0.1293150684931507,
+      "grad_norm": 2.9266560077667236,
+      "learning_rate": 1.8978102189781023e-06,
+      "loss": 2.3475,
+      "step": 118
+    },
+    {
+      "epoch": 0.1304109589041096,
+      "grad_norm": 3.0419797897338867,
+      "learning_rate": 1.916058394160584e-06,
+      "loss": 2.3475,
+      "step": 119
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "grad_norm": 2.4531214237213135,
+      "learning_rate": 1.934306569343066e-06,
+      "loss": 2.3724,
+      "step": 120
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "eval_loss": 2.455423593521118,
+      "eval_runtime": 37.0304,
+      "eval_samples_per_second": 27.005,
+      "eval_steps_per_second": 1.134,
+      "step": 120
+    },
+    {
+      "epoch": 0.1326027397260274,
+      "grad_norm": 3.572502851486206,
+      "learning_rate": 1.952554744525548e-06,
+      "loss": 2.2671,
+      "step": 121
+    },
+    {
+      "epoch": 0.1336986301369863,
+      "grad_norm": 3.5408504009246826,
+      "learning_rate": 1.9708029197080293e-06,
+      "loss": 2.3133,
+      "step": 122
+    },
+    {
+      "epoch": 0.13479452054794522,
+      "grad_norm": 2.488725423812866,
+      "learning_rate": 1.989051094890511e-06,
+      "loss": 2.3141,
+      "step": 123
+    },
+    {
+      "epoch": 0.1358904109589041,
+      "grad_norm": 3.241276264190674,
+      "learning_rate": 2.007299270072993e-06,
+      "loss": 2.318,
+      "step": 124
+    },
+    {
+      "epoch": 0.136986301369863,
+      "grad_norm": 2.853959798812866,
+      "learning_rate": 2.025547445255475e-06,
+      "loss": 2.3712,
+      "step": 125
+    },
+    {
+      "epoch": 0.13808219178082193,
+      "grad_norm": 2.9798169136047363,
+      "learning_rate": 2.0437956204379563e-06,
+      "loss": 2.3072,
+      "step": 126
+    },
+    {
+      "epoch": 0.13917808219178082,
+      "grad_norm": 2.8529601097106934,
+      "learning_rate": 2.062043795620438e-06,
+      "loss": 2.318,
+      "step": 127
+    },
+    {
+      "epoch": 0.14027397260273972,
+      "grad_norm": 3.475337028503418,
+      "learning_rate": 2.08029197080292e-06,
+      "loss": 2.31,
+      "step": 128
+    },
+    {
+      "epoch": 0.14136986301369864,
+      "grad_norm": 4.213955879211426,
+      "learning_rate": 2.0985401459854014e-06,
+      "loss": 2.3734,
+      "step": 129
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "grad_norm": 3.0833330154418945,
+      "learning_rate": 2.1167883211678833e-06,
+      "loss": 2.2602,
+      "step": 130
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "eval_loss": 2.4205682277679443,
+      "eval_runtime": 37.0798,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 130
+    },
+    {
+      "epoch": 0.14356164383561643,
+      "grad_norm": 3.411235809326172,
+      "learning_rate": 2.135036496350365e-06,
+      "loss": 2.3255,
+      "step": 131
+    },
+    {
+      "epoch": 0.14465753424657535,
+      "grad_norm": 2.882053852081299,
+      "learning_rate": 2.1532846715328466e-06,
+      "loss": 2.3483,
+      "step": 132
+    },
+    {
+      "epoch": 0.14575342465753424,
+      "grad_norm": 3.2436420917510986,
+      "learning_rate": 2.171532846715329e-06,
+      "loss": 2.2492,
+      "step": 133
+    },
+    {
+      "epoch": 0.14684931506849316,
+      "grad_norm": 2.929466962814331,
+      "learning_rate": 2.1897810218978103e-06,
+      "loss": 2.2647,
+      "step": 134
+    },
+    {
+      "epoch": 0.14794520547945206,
+      "grad_norm": 2.9344966411590576,
+      "learning_rate": 2.208029197080292e-06,
+      "loss": 2.3061,
+      "step": 135
+    },
+    {
+      "epoch": 0.14904109589041095,
+      "grad_norm": 3.2567203044891357,
+      "learning_rate": 2.226277372262774e-06,
+      "loss": 2.2746,
+      "step": 136
+    },
+    {
+      "epoch": 0.15013698630136987,
+      "grad_norm": 3.1494054794311523,
+      "learning_rate": 2.244525547445256e-06,
+      "loss": 2.2866,
+      "step": 137
+    },
+    {
+      "epoch": 0.15123287671232877,
+      "grad_norm": 2.723390817642212,
+      "learning_rate": 2.2627737226277373e-06,
+      "loss": 2.3109,
+      "step": 138
+    },
+    {
+      "epoch": 0.15232876712328766,
+      "grad_norm": 2.8904967308044434,
+      "learning_rate": 2.281021897810219e-06,
+      "loss": 2.301,
+      "step": 139
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "grad_norm": 2.8899483680725098,
+      "learning_rate": 2.299270072992701e-06,
+      "loss": 2.2637,
+      "step": 140
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "eval_loss": 2.3952724933624268,
+      "eval_runtime": 37.1272,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 140
+    },
+    {
+      "epoch": 0.15452054794520548,
+      "grad_norm": 3.5309066772460938,
+      "learning_rate": 2.3175182481751825e-06,
+      "loss": 2.2572,
+      "step": 141
+    },
+    {
+      "epoch": 0.15561643835616437,
+      "grad_norm": 2.8369667530059814,
+      "learning_rate": 2.3357664233576643e-06,
+      "loss": 2.2712,
+      "step": 142
+    },
+    {
+      "epoch": 0.1567123287671233,
+      "grad_norm": 3.629161834716797,
+      "learning_rate": 2.354014598540146e-06,
+      "loss": 2.2301,
+      "step": 143
+    },
+    {
+      "epoch": 0.1578082191780822,
+      "grad_norm": 3.5313000679016113,
+      "learning_rate": 2.3722627737226276e-06,
+      "loss": 2.2712,
+      "step": 144
+    },
+    {
+      "epoch": 0.1589041095890411,
+      "grad_norm": 3.1328771114349365,
+      "learning_rate": 2.39051094890511e-06,
+      "loss": 2.2706,
+      "step": 145
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 4.6992950439453125,
+      "learning_rate": 2.4087591240875918e-06,
+      "loss": 2.219,
+      "step": 146
+    },
+    {
+      "epoch": 0.1610958904109589,
+      "grad_norm": 5.230221748352051,
+      "learning_rate": 2.427007299270073e-06,
+      "loss": 2.2396,
+      "step": 147
+    },
+    {
+      "epoch": 0.16219178082191782,
+      "grad_norm": 3.5680761337280273,
+      "learning_rate": 2.445255474452555e-06,
+      "loss": 2.2434,
+      "step": 148
+    },
+    {
+      "epoch": 0.1632876712328767,
+      "grad_norm": 3.1642441749572754,
+      "learning_rate": 2.463503649635037e-06,
+      "loss": 2.2856,
+      "step": 149
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "grad_norm": 2.694085121154785,
+      "learning_rate": 2.4817518248175183e-06,
+      "loss": 2.2006,
+      "step": 150
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "eval_loss": 2.37199068069458,
+      "eval_runtime": 37.1342,
+      "eval_samples_per_second": 26.929,
+      "eval_steps_per_second": 1.131,
+      "step": 150
+    },
+    {
+      "epoch": 0.16547945205479453,
+      "grad_norm": 2.9868853092193604,
+      "learning_rate": 2.5e-06,
+      "loss": 2.1694,
+      "step": 151
+    },
+    {
+      "epoch": 0.16657534246575342,
+      "grad_norm": 2.793532371520996,
+      "learning_rate": 2.518248175182482e-06,
+      "loss": 2.2965,
+      "step": 152
+    },
+    {
+      "epoch": 0.16767123287671232,
+      "grad_norm": 2.60152530670166,
+      "learning_rate": 2.5364963503649635e-06,
+      "loss": 2.1954,
+      "step": 153
+    },
+    {
+      "epoch": 0.16876712328767124,
+      "grad_norm": 2.6192219257354736,
+      "learning_rate": 2.5547445255474458e-06,
+      "loss": 2.1909,
+      "step": 154
+    },
+    {
+      "epoch": 0.16986301369863013,
+      "grad_norm": 2.467344045639038,
+      "learning_rate": 2.5729927007299272e-06,
+      "loss": 2.2438,
+      "step": 155
+    },
+    {
+      "epoch": 0.17095890410958905,
+      "grad_norm": 2.772012948989868,
+      "learning_rate": 2.591240875912409e-06,
+      "loss": 2.2213,
+      "step": 156
+    },
+    {
+      "epoch": 0.17205479452054795,
+      "grad_norm": 2.613675355911255,
+      "learning_rate": 2.6094890510948905e-06,
+      "loss": 2.2213,
+      "step": 157
+    },
+    {
+      "epoch": 0.17315068493150684,
+      "grad_norm": 2.849245071411133,
+      "learning_rate": 2.627737226277373e-06,
+      "loss": 2.255,
+      "step": 158
+    },
+    {
+      "epoch": 0.17424657534246576,
+      "grad_norm": 2.8388969898223877,
+      "learning_rate": 2.6459854014598542e-06,
+      "loss": 2.2572,
+      "step": 159
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "grad_norm": 2.5219321250915527,
+      "learning_rate": 2.664233576642336e-06,
+      "loss": 2.2253,
+      "step": 160
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "eval_loss": 2.3509061336517334,
+      "eval_runtime": 37.0953,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 160
+    },
+    {
+      "epoch": 0.17643835616438355,
+      "grad_norm": 2.8460891246795654,
+      "learning_rate": 2.6824817518248175e-06,
+      "loss": 2.2009,
+      "step": 161
+    },
+    {
+      "epoch": 0.17753424657534247,
+      "grad_norm": 2.673377275466919,
+      "learning_rate": 2.7007299270072994e-06,
+      "loss": 2.2211,
+      "step": 162
+    },
+    {
+      "epoch": 0.17863013698630137,
+      "grad_norm": 2.8514182567596436,
+      "learning_rate": 2.718978102189781e-06,
+      "loss": 2.1955,
+      "step": 163
+    },
+    {
+      "epoch": 0.17972602739726026,
+      "grad_norm": 3.530385971069336,
+      "learning_rate": 2.737226277372263e-06,
+      "loss": 2.2536,
+      "step": 164
+    },
+    {
+      "epoch": 0.18082191780821918,
+      "grad_norm": 2.733394145965576,
+      "learning_rate": 2.755474452554745e-06,
+      "loss": 2.2367,
+      "step": 165
+    },
+    {
+      "epoch": 0.18191780821917808,
+      "grad_norm": 3.871150016784668,
+      "learning_rate": 2.7737226277372264e-06,
+      "loss": 2.1949,
+      "step": 166
+    },
+    {
+      "epoch": 0.18301369863013697,
+      "grad_norm": 3.7782530784606934,
+      "learning_rate": 2.7919708029197082e-06,
+      "loss": 2.2979,
+      "step": 167
+    },
+    {
+      "epoch": 0.1841095890410959,
+      "grad_norm": 3.094599485397339,
+      "learning_rate": 2.81021897810219e-06,
+      "loss": 2.2512,
+      "step": 168
+    },
+    {
+      "epoch": 0.1852054794520548,
+      "grad_norm": 3.2166848182678223,
+      "learning_rate": 2.828467153284672e-06,
+      "loss": 2.1784,
+      "step": 169
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "grad_norm": 3.4127073287963867,
+      "learning_rate": 2.8467153284671534e-06,
+      "loss": 2.2241,
+      "step": 170
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "eval_loss": 2.333245038986206,
+      "eval_runtime": 37.0402,
+      "eval_samples_per_second": 26.998,
+      "eval_steps_per_second": 1.134,
+      "step": 170
+    },
+    {
+      "epoch": 0.1873972602739726,
+      "grad_norm": 2.677579879760742,
+      "learning_rate": 2.8649635036496353e-06,
+      "loss": 2.135,
+      "step": 171
+    },
+    {
+      "epoch": 0.1884931506849315,
+      "grad_norm": 2.9602150917053223,
+      "learning_rate": 2.8832116788321167e-06,
+      "loss": 2.1916,
+      "step": 172
+    },
+    {
+      "epoch": 0.18958904109589042,
+      "grad_norm": 2.580122232437134,
+      "learning_rate": 2.901459854014599e-06,
+      "loss": 2.1477,
+      "step": 173
+    },
+    {
+      "epoch": 0.19068493150684931,
+      "grad_norm": 2.652217388153076,
+      "learning_rate": 2.9197080291970804e-06,
+      "loss": 2.2394,
+      "step": 174
+    },
+    {
+      "epoch": 0.1917808219178082,
+      "grad_norm": 2.6887567043304443,
+      "learning_rate": 2.9379562043795623e-06,
+      "loss": 2.2245,
+      "step": 175
+    },
+    {
+      "epoch": 0.19287671232876713,
+      "grad_norm": 2.768165111541748,
+      "learning_rate": 2.956204379562044e-06,
+      "loss": 2.2017,
+      "step": 176
+    },
+    {
+      "epoch": 0.19397260273972602,
+      "grad_norm": 2.9375691413879395,
+      "learning_rate": 2.9744525547445256e-06,
+      "loss": 2.2464,
+      "step": 177
+    },
+    {
+      "epoch": 0.19506849315068492,
+      "grad_norm": 2.8364248275756836,
+      "learning_rate": 2.992700729927008e-06,
+      "loss": 2.2155,
+      "step": 178
+    },
+    {
+      "epoch": 0.19616438356164384,
+      "grad_norm": 2.843487024307251,
+      "learning_rate": 3.0109489051094893e-06,
+      "loss": 2.1577,
+      "step": 179
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "grad_norm": 3.005230665206909,
+      "learning_rate": 3.029197080291971e-06,
+      "loss": 2.1549,
+      "step": 180
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "eval_loss": 2.3209409713745117,
+      "eval_runtime": 37.022,
+      "eval_samples_per_second": 27.011,
+      "eval_steps_per_second": 1.134,
+      "step": 180
+    },
+    {
+      "epoch": 0.19835616438356166,
+      "grad_norm": 2.490260601043701,
+      "learning_rate": 3.0474452554744526e-06,
+      "loss": 2.2313,
+      "step": 181
+    },
+    {
+      "epoch": 0.19945205479452055,
+      "grad_norm": 2.973118543624878,
+      "learning_rate": 3.065693430656935e-06,
+      "loss": 2.2316,
+      "step": 182
+    },
+    {
+      "epoch": 0.20054794520547944,
+      "grad_norm": 3.3308680057525635,
+      "learning_rate": 3.0839416058394163e-06,
+      "loss": 2.2087,
+      "step": 183
+    },
+    {
+      "epoch": 0.20164383561643837,
+      "grad_norm": 3.1440727710723877,
+      "learning_rate": 3.102189781021898e-06,
+      "loss": 2.2256,
+      "step": 184
+    },
+    {
+      "epoch": 0.20273972602739726,
+      "grad_norm": 2.5696325302124023,
+      "learning_rate": 3.1204379562043796e-06,
+      "loss": 2.2093,
+      "step": 185
+    },
+    {
+      "epoch": 0.20383561643835615,
+      "grad_norm": 3.1225037574768066,
+      "learning_rate": 3.1386861313868614e-06,
+      "loss": 2.1425,
+      "step": 186
+    },
+    {
+      "epoch": 0.20493150684931508,
+      "grad_norm": 3.650743246078491,
+      "learning_rate": 3.156934306569343e-06,
+      "loss": 2.1339,
+      "step": 187
+    },
+    {
+      "epoch": 0.20602739726027397,
+      "grad_norm": 2.924072027206421,
+      "learning_rate": 3.175182481751825e-06,
+      "loss": 2.2363,
+      "step": 188
+    },
+    {
+      "epoch": 0.20712328767123286,
+      "grad_norm": 2.9561989307403564,
+      "learning_rate": 3.193430656934307e-06,
+      "loss": 2.1934,
+      "step": 189
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "grad_norm": 2.731393814086914,
+      "learning_rate": 3.2116788321167884e-06,
+      "loss": 2.1718,
+      "step": 190
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "eval_loss": 2.3087782859802246,
+      "eval_runtime": 37.0161,
+      "eval_samples_per_second": 27.015,
+      "eval_steps_per_second": 1.135,
+      "step": 190
+    },
+    {
+      "epoch": 0.20931506849315068,
+      "grad_norm": 2.774442434310913,
+      "learning_rate": 3.2299270072992707e-06,
+      "loss": 2.1841,
+      "step": 191
+    },
+    {
+      "epoch": 0.2104109589041096,
+      "grad_norm": 3.0579135417938232,
+      "learning_rate": 3.248175182481752e-06,
+      "loss": 2.1677,
+      "step": 192
+    },
+    {
+      "epoch": 0.2115068493150685,
+      "grad_norm": 2.732984781265259,
+      "learning_rate": 3.266423357664234e-06,
+      "loss": 2.0975,
+      "step": 193
+    },
+    {
+      "epoch": 0.2126027397260274,
+      "grad_norm": 3.442659854888916,
+      "learning_rate": 3.2846715328467155e-06,
+      "loss": 2.1923,
+      "step": 194
+    },
+    {
+      "epoch": 0.2136986301369863,
+      "grad_norm": 3.064223527908325,
+      "learning_rate": 3.3029197080291973e-06,
+      "loss": 2.1707,
+      "step": 195
+    },
+    {
+      "epoch": 0.2147945205479452,
+      "grad_norm": 2.5138468742370605,
+      "learning_rate": 3.3211678832116788e-06,
+      "loss": 2.1933,
+      "step": 196
+    },
+    {
+      "epoch": 0.2158904109589041,
+      "grad_norm": 3.1076464653015137,
+      "learning_rate": 3.339416058394161e-06,
+      "loss": 2.1725,
+      "step": 197
+    },
+    {
+      "epoch": 0.21698630136986302,
+      "grad_norm": 2.7291388511657715,
+      "learning_rate": 3.3576642335766425e-06,
+      "loss": 2.1722,
+      "step": 198
+    },
+    {
+      "epoch": 0.21808219178082192,
+      "grad_norm": 2.805506467819214,
+      "learning_rate": 3.3759124087591243e-06,
+      "loss": 2.1688,
+      "step": 199
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "grad_norm": 2.6190085411071777,
+      "learning_rate": 3.3941605839416058e-06,
+      "loss": 2.1708,
+      "step": 200
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "eval_loss": 2.295546054840088,
+      "eval_runtime": 37.0133,
+      "eval_samples_per_second": 27.017,
+      "eval_steps_per_second": 1.135,
+      "step": 200
+    },
+    {
+      "epoch": 0.22027397260273973,
+      "grad_norm": 3.1270673274993896,
+      "learning_rate": 3.412408759124088e-06,
+      "loss": 2.1759,
+      "step": 201
+    },
+    {
+      "epoch": 0.22136986301369863,
+      "grad_norm": 3.494178533554077,
+      "learning_rate": 3.43065693430657e-06,
+      "loss": 2.1795,
+      "step": 202
+    },
+    {
+      "epoch": 0.22246575342465755,
+      "grad_norm": 2.7398736476898193,
+      "learning_rate": 3.4489051094890513e-06,
+      "loss": 2.1201,
+      "step": 203
+    },
+    {
+      "epoch": 0.22356164383561644,
+      "grad_norm": 2.981506824493408,
+      "learning_rate": 3.467153284671533e-06,
+      "loss": 2.1693,
+      "step": 204
+    },
+    {
+      "epoch": 0.22465753424657534,
+      "grad_norm": 2.4129178524017334,
+      "learning_rate": 3.4854014598540146e-06,
+      "loss": 2.1125,
+      "step": 205
+    },
+    {
+      "epoch": 0.22575342465753426,
+      "grad_norm": 2.701197385787964,
+      "learning_rate": 3.503649635036497e-06,
+      "loss": 2.1565,
+      "step": 206
+    },
+    {
+      "epoch": 0.22684931506849315,
+      "grad_norm": 2.533306837081909,
+      "learning_rate": 3.5218978102189783e-06,
+      "loss": 2.1487,
+      "step": 207
+    },
+    {
+      "epoch": 0.22794520547945205,
+      "grad_norm": 2.8030788898468018,
+      "learning_rate": 3.54014598540146e-06,
+      "loss": 2.1801,
+      "step": 208
+    },
+    {
+      "epoch": 0.22904109589041097,
+      "grad_norm": 2.895414352416992,
+      "learning_rate": 3.5583941605839416e-06,
+      "loss": 2.189,
+      "step": 209
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "grad_norm": 2.508390426635742,
+      "learning_rate": 3.576642335766424e-06,
+      "loss": 2.2376,
+      "step": 210
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "eval_loss": 2.2900516986846924,
+      "eval_runtime": 36.7645,
+      "eval_samples_per_second": 27.2,
+      "eval_steps_per_second": 1.142,
+      "step": 210
+    },
+    {
+      "epoch": 0.23123287671232876,
+      "grad_norm": 2.6393816471099854,
+      "learning_rate": 3.5948905109489054e-06,
+      "loss": 2.2267,
+      "step": 211
+    },
+    {
+      "epoch": 0.23232876712328768,
+      "grad_norm": 3.0687549114227295,
+      "learning_rate": 3.6131386861313872e-06,
+      "loss": 2.0744,
+      "step": 212
+    },
+    {
+      "epoch": 0.23342465753424657,
+      "grad_norm": 2.6342105865478516,
+      "learning_rate": 3.6313868613138687e-06,
+      "loss": 2.0737,
+      "step": 213
+    },
+    {
+      "epoch": 0.2345205479452055,
+      "grad_norm": 3.5662875175476074,
+      "learning_rate": 3.6496350364963505e-06,
+      "loss": 2.1147,
+      "step": 214
+    },
+    {
+      "epoch": 0.2356164383561644,
+      "grad_norm": 2.5357561111450195,
+      "learning_rate": 3.667883211678833e-06,
+      "loss": 2.1268,
+      "step": 215
+    },
+    {
+      "epoch": 0.23671232876712328,
+      "grad_norm": 3.439711093902588,
+      "learning_rate": 3.6861313868613142e-06,
+      "loss": 2.1118,
+      "step": 216
+    },
+    {
+      "epoch": 0.2378082191780822,
+      "grad_norm": 2.9410204887390137,
+      "learning_rate": 3.704379562043796e-06,
+      "loss": 2.1087,
+      "step": 217
+    },
+    {
+      "epoch": 0.2389041095890411,
+      "grad_norm": 3.3182294368743896,
+      "learning_rate": 3.7226277372262775e-06,
+      "loss": 2.196,
+      "step": 218
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 3.3834011554718018,
+      "learning_rate": 3.7408759124087594e-06,
+      "loss": 2.1633,
+      "step": 219
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "grad_norm": 3.114952325820923,
+      "learning_rate": 3.7591240875912412e-06,
+      "loss": 2.1107,
+      "step": 220
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "eval_loss": 2.277580976486206,
+      "eval_runtime": 36.9846,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 220
+    },
+    {
+      "epoch": 0.2421917808219178,
+      "grad_norm": 3.1151392459869385,
+      "learning_rate": 3.777372262773723e-06,
+      "loss": 2.1471,
+      "step": 221
+    },
+    {
+      "epoch": 0.2432876712328767,
+      "grad_norm": 3.798450469970703,
+      "learning_rate": 3.7956204379562045e-06,
+      "loss": 2.1324,
+      "step": 222
+    },
+    {
+      "epoch": 0.24438356164383562,
+      "grad_norm": 2.6187098026275635,
+      "learning_rate": 3.8138686131386864e-06,
+      "loss": 2.1061,
+      "step": 223
+    },
+    {
+      "epoch": 0.24547945205479452,
+      "grad_norm": 4.014018535614014,
+      "learning_rate": 3.832116788321168e-06,
+      "loss": 2.166,
+      "step": 224
+    },
+    {
+      "epoch": 0.2465753424657534,
+      "grad_norm": 3.0043938159942627,
+      "learning_rate": 3.85036496350365e-06,
+      "loss": 2.1197,
+      "step": 225
+    },
+    {
+      "epoch": 0.24767123287671233,
+      "grad_norm": 3.507066249847412,
+      "learning_rate": 3.868613138686132e-06,
+      "loss": 2.0848,
+      "step": 226
+    },
+    {
+      "epoch": 0.24876712328767123,
+      "grad_norm": 3.5713281631469727,
+      "learning_rate": 3.886861313868613e-06,
+      "loss": 2.1239,
+      "step": 227
+    },
+    {
+      "epoch": 0.24986301369863015,
+      "grad_norm": 3.961082696914673,
+      "learning_rate": 3.905109489051096e-06,
+      "loss": 2.1406,
+      "step": 228
+    },
+    {
+      "epoch": 0.25095890410958904,
+      "grad_norm": 3.1514859199523926,
+      "learning_rate": 3.923357664233577e-06,
+      "loss": 2.1689,
+      "step": 229
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "grad_norm": 3.8486833572387695,
+      "learning_rate": 3.9416058394160585e-06,
+      "loss": 2.1321,
+      "step": 230
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "eval_loss": 2.2620601654052734,
+      "eval_runtime": 37.058,
+      "eval_samples_per_second": 26.985,
+      "eval_steps_per_second": 1.133,
+      "step": 230
+    },
+    {
+      "epoch": 0.25315068493150683,
+      "grad_norm": 2.5626907348632812,
+      "learning_rate": 3.95985401459854e-06,
+      "loss": 2.1777,
+      "step": 231
+    },
+    {
+      "epoch": 0.25424657534246575,
+      "grad_norm": 4.105278015136719,
+      "learning_rate": 3.978102189781022e-06,
+      "loss": 2.2099,
+      "step": 232
+    },
+    {
+      "epoch": 0.2553424657534247,
+      "grad_norm": 2.9818739891052246,
+      "learning_rate": 3.996350364963504e-06,
+      "loss": 2.0705,
+      "step": 233
+    },
+    {
+      "epoch": 0.25643835616438354,
+      "grad_norm": 2.87030291557312,
+      "learning_rate": 4.014598540145986e-06,
+      "loss": 2.1453,
+      "step": 234
+    },
+    {
+      "epoch": 0.25753424657534246,
+      "grad_norm": 2.4847915172576904,
+      "learning_rate": 4.032846715328467e-06,
+      "loss": 2.0461,
+      "step": 235
+    },
+    {
+      "epoch": 0.2586301369863014,
+      "grad_norm": 3.451420783996582,
+      "learning_rate": 4.05109489051095e-06,
+      "loss": 2.1133,
+      "step": 236
+    },
+    {
+      "epoch": 0.25972602739726025,
+      "grad_norm": 2.954723358154297,
+      "learning_rate": 4.069343065693431e-06,
+      "loss": 2.1316,
+      "step": 237
+    },
+    {
+      "epoch": 0.2608219178082192,
+      "grad_norm": 2.5295896530151367,
+      "learning_rate": 4.0875912408759126e-06,
+      "loss": 2.0901,
+      "step": 238
+    },
+    {
+      "epoch": 0.2619178082191781,
+      "grad_norm": 2.6359453201293945,
+      "learning_rate": 4.105839416058395e-06,
+      "loss": 2.117,
+      "step": 239
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "grad_norm": 2.5236105918884277,
+      "learning_rate": 4.124087591240876e-06,
+      "loss": 2.1958,
+      "step": 240
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "eval_loss": 2.250295400619507,
+      "eval_runtime": 37.0811,
+      "eval_samples_per_second": 26.968,
+      "eval_steps_per_second": 1.133,
+      "step": 240
+    },
+    {
+      "epoch": 0.2641095890410959,
+      "grad_norm": 2.581918954849243,
+      "learning_rate": 4.1423357664233586e-06,
+      "loss": 2.1054,
+      "step": 241
+    },
+    {
+      "epoch": 0.2652054794520548,
+      "grad_norm": 2.501332998275757,
+      "learning_rate": 4.16058394160584e-06,
+      "loss": 2.0878,
+      "step": 242
+    },
+    {
+      "epoch": 0.2663013698630137,
+      "grad_norm": 2.6477770805358887,
+      "learning_rate": 4.1788321167883214e-06,
+      "loss": 2.1656,
+      "step": 243
+    },
+    {
+      "epoch": 0.2673972602739726,
+      "grad_norm": 2.7252438068389893,
+      "learning_rate": 4.197080291970803e-06,
+      "loss": 2.0497,
+      "step": 244
+    },
+    {
+      "epoch": 0.2684931506849315,
+      "grad_norm": 2.6056301593780518,
+      "learning_rate": 4.215328467153285e-06,
+      "loss": 2.1411,
+      "step": 245
+    },
+    {
+      "epoch": 0.26958904109589044,
+      "grad_norm": 2.6581931114196777,
+      "learning_rate": 4.233576642335767e-06,
+      "loss": 2.0829,
+      "step": 246
+    },
+    {
+      "epoch": 0.2706849315068493,
+      "grad_norm": 2.6316752433776855,
+      "learning_rate": 4.251824817518249e-06,
+      "loss": 2.0882,
+      "step": 247
+    },
+    {
+      "epoch": 0.2717808219178082,
+      "grad_norm": 2.803823947906494,
+      "learning_rate": 4.27007299270073e-06,
+      "loss": 2.1203,
+      "step": 248
+    },
+    {
+      "epoch": 0.27287671232876715,
+      "grad_norm": 2.9551174640655518,
+      "learning_rate": 4.288321167883212e-06,
+      "loss": 2.0897,
+      "step": 249
+    },
+    {
+      "epoch": 0.273972602739726,
+      "grad_norm": 2.4847161769866943,
+      "learning_rate": 4.306569343065693e-06,
+      "loss": 2.0845,
+      "step": 250
+    },
+    {
+      "epoch": 0.273972602739726,
+      "eval_loss": 2.2491259574890137,
+      "eval_runtime": 37.1024,
+      "eval_samples_per_second": 26.952,
+      "eval_steps_per_second": 1.132,
+      "step": 250
+    },
+    {
+      "epoch": 0.27506849315068493,
+      "grad_norm": 2.6398351192474365,
+      "learning_rate": 4.3248175182481755e-06,
+      "loss": 2.0479,
+      "step": 251
+    },
+    {
+      "epoch": 0.27616438356164386,
+      "grad_norm": 3.0948562622070312,
+      "learning_rate": 4.343065693430658e-06,
+      "loss": 2.1184,
+      "step": 252
+    },
+    {
+      "epoch": 0.2772602739726027,
+      "grad_norm": 2.5378012657165527,
+      "learning_rate": 4.361313868613139e-06,
+      "loss": 2.1099,
+      "step": 253
+    },
+    {
+      "epoch": 0.27835616438356164,
+      "grad_norm": 3.0282225608825684,
+      "learning_rate": 4.379562043795621e-06,
+      "loss": 2.0982,
+      "step": 254
+    },
+    {
+      "epoch": 0.27945205479452057,
+      "grad_norm": 2.6205334663391113,
+      "learning_rate": 4.397810218978102e-06,
+      "loss": 2.1285,
+      "step": 255
+    },
+    {
+      "epoch": 0.28054794520547943,
+      "grad_norm": 3.2056095600128174,
+      "learning_rate": 4.416058394160584e-06,
+      "loss": 2.1221,
+      "step": 256
+    },
+    {
+      "epoch": 0.28164383561643835,
+      "grad_norm": 3.1370127201080322,
+      "learning_rate": 4.434306569343066e-06,
+      "loss": 2.1016,
+      "step": 257
+    },
+    {
+      "epoch": 0.2827397260273973,
+      "grad_norm": 2.494168519973755,
+      "learning_rate": 4.452554744525548e-06,
+      "loss": 2.0754,
+      "step": 258
+    },
+    {
+      "epoch": 0.28383561643835614,
+      "grad_norm": 2.7979183197021484,
+      "learning_rate": 4.4708029197080295e-06,
+      "loss": 2.1218,
+      "step": 259
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "grad_norm": 2.755682945251465,
+      "learning_rate": 4.489051094890512e-06,
+      "loss": 2.1136,
+      "step": 260
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "eval_loss": 2.2433817386627197,
+      "eval_runtime": 37.1121,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 260
+    },
+    {
+      "epoch": 0.286027397260274,
+      "grad_norm": 2.3641891479492188,
+      "learning_rate": 4.507299270072993e-06,
+      "loss": 2.1488,
+      "step": 261
+    },
+    {
+      "epoch": 0.28712328767123285,
+      "grad_norm": 2.741292953491211,
+      "learning_rate": 4.525547445255475e-06,
+      "loss": 2.1048,
+      "step": 262
+    },
+    {
+      "epoch": 0.2882191780821918,
+      "grad_norm": 2.811483383178711,
+      "learning_rate": 4.543795620437956e-06,
+      "loss": 2.0996,
+      "step": 263
+    },
+    {
+      "epoch": 0.2893150684931507,
+      "grad_norm": 2.6705479621887207,
+      "learning_rate": 4.562043795620438e-06,
+      "loss": 2.0652,
+      "step": 264
+    },
+    {
+      "epoch": 0.29041095890410956,
+      "grad_norm": 2.9383747577667236,
+      "learning_rate": 4.580291970802921e-06,
+      "loss": 2.1574,
+      "step": 265
+    },
+    {
+      "epoch": 0.2915068493150685,
+      "grad_norm": 2.591373920440674,
+      "learning_rate": 4.598540145985402e-06,
+      "loss": 2.1655,
+      "step": 266
+    },
+    {
+      "epoch": 0.2926027397260274,
+      "grad_norm": 3.16955828666687,
+      "learning_rate": 4.6167883211678835e-06,
+      "loss": 2.1285,
+      "step": 267
+    },
+    {
+      "epoch": 0.2936986301369863,
+      "grad_norm": 2.66941499710083,
+      "learning_rate": 4.635036496350365e-06,
+      "loss": 2.0777,
+      "step": 268
+    },
+    {
+      "epoch": 0.2947945205479452,
+      "grad_norm": 2.768139600753784,
+      "learning_rate": 4.653284671532847e-06,
+      "loss": 2.0748,
+      "step": 269
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "grad_norm": 2.5951924324035645,
+      "learning_rate": 4.671532846715329e-06,
+      "loss": 2.1289,
+      "step": 270
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "eval_loss": 2.2393550872802734,
+      "eval_runtime": 37.0993,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 270
+    },
+    {
+      "epoch": 0.29698630136986304,
+      "grad_norm": 2.5422160625457764,
+      "learning_rate": 4.689781021897811e-06,
+      "loss": 2.1342,
+      "step": 271
+    },
+    {
+      "epoch": 0.2980821917808219,
+      "grad_norm": 2.540555477142334,
+      "learning_rate": 4.708029197080292e-06,
+      "loss": 2.097,
+      "step": 272
+    },
+    {
+      "epoch": 0.2991780821917808,
+      "grad_norm": 2.995366334915161,
+      "learning_rate": 4.726277372262774e-06,
+      "loss": 2.1452,
+      "step": 273
+    },
+    {
+      "epoch": 0.30027397260273975,
+      "grad_norm": 2.2561020851135254,
+      "learning_rate": 4.744525547445255e-06,
+      "loss": 2.1069,
+      "step": 274
+    },
+    {
+      "epoch": 0.3013698630136986,
+      "grad_norm": 2.53476619720459,
+      "learning_rate": 4.7627737226277375e-06,
+      "loss": 2.0381,
+      "step": 275
+    },
+    {
+      "epoch": 0.30246575342465754,
+      "grad_norm": 2.3415749073028564,
+      "learning_rate": 4.78102189781022e-06,
+      "loss": 2.1027,
+      "step": 276
+    },
+    {
+      "epoch": 0.30356164383561646,
+      "grad_norm": 2.9756431579589844,
+      "learning_rate": 4.799270072992701e-06,
+      "loss": 2.1136,
+      "step": 277
+    },
+    {
+      "epoch": 0.3046575342465753,
+      "grad_norm": 2.606208324432373,
+      "learning_rate": 4.8175182481751835e-06,
+      "loss": 2.0797,
+      "step": 278
+    },
+    {
+      "epoch": 0.30575342465753425,
+      "grad_norm": 2.5089313983917236,
+      "learning_rate": 4.835766423357665e-06,
+      "loss": 2.1026,
+      "step": 279
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "grad_norm": 2.5003838539123535,
+      "learning_rate": 4.854014598540146e-06,
+      "loss": 2.0923,
+      "step": 280
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "eval_loss": 2.2351036071777344,
+      "eval_runtime": 37.111,
+      "eval_samples_per_second": 26.946,
+      "eval_steps_per_second": 1.132,
+      "step": 280
+    },
+    {
+      "epoch": 0.30794520547945203,
+      "grad_norm": 2.403602361679077,
+      "learning_rate": 4.872262773722628e-06,
+      "loss": 2.0388,
+      "step": 281
+    },
+    {
+      "epoch": 0.30904109589041096,
+      "grad_norm": 2.7775959968566895,
+      "learning_rate": 4.89051094890511e-06,
+      "loss": 2.0561,
+      "step": 282
+    },
+    {
+      "epoch": 0.3101369863013699,
+      "grad_norm": 2.7128653526306152,
+      "learning_rate": 4.9087591240875915e-06,
+      "loss": 2.042,
+      "step": 283
+    },
+    {
+      "epoch": 0.31123287671232874,
+      "grad_norm": 2.7029991149902344,
+      "learning_rate": 4.927007299270074e-06,
+      "loss": 2.0814,
+      "step": 284
+    },
+    {
+      "epoch": 0.31232876712328766,
+      "grad_norm": 2.3942933082580566,
+      "learning_rate": 4.945255474452555e-06,
+      "loss": 2.0906,
+      "step": 285
+    },
+    {
+      "epoch": 0.3134246575342466,
+      "grad_norm": 2.5086541175842285,
+      "learning_rate": 4.963503649635037e-06,
+      "loss": 2.0625,
+      "step": 286
+    },
+    {
+      "epoch": 0.31452054794520545,
+      "grad_norm": 2.6668760776519775,
+      "learning_rate": 4.981751824817518e-06,
+      "loss": 2.1538,
+      "step": 287
+    },
+    {
+      "epoch": 0.3156164383561644,
+      "grad_norm": 2.5066750049591064,
+      "learning_rate": 5e-06,
+      "loss": 2.1423,
+      "step": 288
+    },
+    {
+      "epoch": 0.3167123287671233,
+      "grad_norm": 2.84967041015625,
+      "learning_rate": 4.997969130787978e-06,
+      "loss": 2.0943,
+      "step": 289
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "grad_norm": 2.151946544647217,
+      "learning_rate": 4.995938261575955e-06,
+      "loss": 2.1098,
+      "step": 290
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "eval_loss": 2.223247528076172,
+      "eval_runtime": 37.1755,
+      "eval_samples_per_second": 26.899,
+      "eval_steps_per_second": 1.13,
+      "step": 290
+    },
+    {
+      "epoch": 0.3189041095890411,
+      "grad_norm": 2.887166976928711,
+      "learning_rate": 4.993907392363932e-06,
+      "loss": 2.0281,
+      "step": 291
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 2.5930211544036865,
+      "learning_rate": 4.9918765231519095e-06,
+      "loss": 2.1063,
+      "step": 292
+    },
+    {
+      "epoch": 0.32109589041095893,
+      "grad_norm": 2.5568413734436035,
+      "learning_rate": 4.989845653939887e-06,
+      "loss": 2.0313,
+      "step": 293
+    },
+    {
+      "epoch": 0.3221917808219178,
+      "grad_norm": 2.536095380783081,
+      "learning_rate": 4.987814784727864e-06,
+      "loss": 2.0749,
+      "step": 294
+    },
+    {
+      "epoch": 0.3232876712328767,
+      "grad_norm": 2.3385651111602783,
+      "learning_rate": 4.985783915515841e-06,
+      "loss": 2.0666,
+      "step": 295
+    },
+    {
+      "epoch": 0.32438356164383564,
+      "grad_norm": 2.504826068878174,
+      "learning_rate": 4.983753046303819e-06,
+      "loss": 2.0677,
+      "step": 296
+    },
+    {
+      "epoch": 0.3254794520547945,
+      "grad_norm": 2.5289409160614014,
+      "learning_rate": 4.981722177091796e-06,
+      "loss": 2.0828,
+      "step": 297
+    },
+    {
+      "epoch": 0.3265753424657534,
+      "grad_norm": 2.803795099258423,
+      "learning_rate": 4.979691307879773e-06,
+      "loss": 2.0733,
+      "step": 298
+    },
+    {
+      "epoch": 0.32767123287671235,
+      "grad_norm": 2.651273012161255,
+      "learning_rate": 4.9776604386677504e-06,
+      "loss": 2.0987,
+      "step": 299
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "grad_norm": 2.4715864658355713,
+      "learning_rate": 4.975629569455728e-06,
+      "loss": 2.0733,
+      "step": 300
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "eval_loss": 2.214858293533325,
+      "eval_runtime": 37.0559,
+      "eval_samples_per_second": 26.986,
+      "eval_steps_per_second": 1.133,
+      "step": 300
+    },
+    {
+      "epoch": 0.32986301369863014,
+      "grad_norm": 2.2573442459106445,
+      "learning_rate": 4.973598700243705e-06,
+      "loss": 2.1291,
+      "step": 301
+    },
+    {
+      "epoch": 0.33095890410958906,
+      "grad_norm": 2.645966053009033,
+      "learning_rate": 4.971567831031681e-06,
+      "loss": 2.046,
+      "step": 302
+    },
+    {
+      "epoch": 0.3320547945205479,
+      "grad_norm": 2.361392021179199,
+      "learning_rate": 4.9695369618196595e-06,
+      "loss": 2.0447,
+      "step": 303
+    },
+    {
+      "epoch": 0.33315068493150685,
+      "grad_norm": 2.272714376449585,
+      "learning_rate": 4.967506092607636e-06,
+      "loss": 1.9272,
+      "step": 304
+    },
+    {
+      "epoch": 0.33424657534246577,
+      "grad_norm": 2.5846943855285645,
+      "learning_rate": 4.965475223395614e-06,
+      "loss": 2.0199,
+      "step": 305
+    },
+    {
+      "epoch": 0.33534246575342463,
+      "grad_norm": 2.38985013961792,
+      "learning_rate": 4.9634443541835905e-06,
+      "loss": 2.0394,
+      "step": 306
+    },
+    {
+      "epoch": 0.33643835616438356,
+      "grad_norm": 2.3526251316070557,
+      "learning_rate": 4.961413484971569e-06,
+      "loss": 2.0403,
+      "step": 307
+    },
+    {
+      "epoch": 0.3375342465753425,
+      "grad_norm": 2.399531602859497,
+      "learning_rate": 4.959382615759545e-06,
+      "loss": 2.0607,
+      "step": 308
+    },
+    {
+      "epoch": 0.33863013698630134,
+      "grad_norm": 2.50803804397583,
+      "learning_rate": 4.957351746547523e-06,
+      "loss": 2.0688,
+      "step": 309
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "grad_norm": 2.278088092803955,
+      "learning_rate": 4.9553208773355e-06,
+      "loss": 2.0453,
+      "step": 310
+    },
+    {
+      "epoch": 0.33972602739726027,
+      "eval_loss": 2.2185235023498535,
+      "eval_runtime": 36.9712,
+      "eval_samples_per_second": 27.048,
+      "eval_steps_per_second": 1.136,
+      "step": 310
+    },
+    {
+      "epoch": 0.3408219178082192,
+      "grad_norm": 2.555300712585449,
+      "learning_rate": 4.953290008123478e-06,
+      "loss": 2.1726,
+      "step": 311
+    },
+    {
+      "epoch": 0.3419178082191781,
+      "grad_norm": 2.614870548248291,
+      "learning_rate": 4.951259138911454e-06,
+      "loss": 2.0659,
+      "step": 312
+    },
+    {
+      "epoch": 0.343013698630137,
+      "grad_norm": 2.1932661533355713,
+      "learning_rate": 4.9492282696994315e-06,
+      "loss": 1.9981,
+      "step": 313
+    },
+    {
+      "epoch": 0.3441095890410959,
+      "grad_norm": 2.707451581954956,
+      "learning_rate": 4.947197400487409e-06,
+      "loss": 2.1321,
+      "step": 314
+    },
+    {
+      "epoch": 0.3452054794520548,
+      "grad_norm": 2.4317946434020996,
+      "learning_rate": 4.945166531275386e-06,
+      "loss": 2.0548,
+      "step": 315
+    },
+    {
+      "epoch": 0.3463013698630137,
+      "grad_norm": 2.639883279800415,
+      "learning_rate": 4.943135662063363e-06,
+      "loss": 2.0512,
+      "step": 316
+    },
+    {
+      "epoch": 0.3473972602739726,
+      "grad_norm": 3.1968822479248047,
+      "learning_rate": 4.9411047928513406e-06,
+      "loss": 2.0611,
+      "step": 317
+    },
+    {
+      "epoch": 0.34849315068493153,
+      "grad_norm": 2.4361510276794434,
+      "learning_rate": 4.939073923639318e-06,
+      "loss": 2.0974,
+      "step": 318
+    },
+    {
+      "epoch": 0.3495890410958904,
+      "grad_norm": 2.8278262615203857,
+      "learning_rate": 4.937043054427295e-06,
+      "loss": 2.0992,
+      "step": 319
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "grad_norm": 3.01247239112854,
+      "learning_rate": 4.935012185215272e-06,
+      "loss": 2.0372,
+      "step": 320
+    },
+    {
+      "epoch": 0.3506849315068493,
+      "eval_loss": 2.206442356109619,
+      "eval_runtime": 36.8797,
+      "eval_samples_per_second": 27.115,
+      "eval_steps_per_second": 1.139,
+      "step": 320
+    },
+    {
+      "epoch": 0.35178082191780824,
+      "grad_norm": 2.3372156620025635,
+      "learning_rate": 4.93298131600325e-06,
+      "loss": 2.0211,
+      "step": 321
+    },
+    {
+      "epoch": 0.3528767123287671,
+      "grad_norm": 2.91377329826355,
+      "learning_rate": 4.930950446791227e-06,
+      "loss": 2.0879,
+      "step": 322
+    },
+    {
+      "epoch": 0.353972602739726,
+      "grad_norm": 2.3207666873931885,
+      "learning_rate": 4.928919577579204e-06,
+      "loss": 2.056,
+      "step": 323
+    },
+    {
+      "epoch": 0.35506849315068495,
+      "grad_norm": 2.3260514736175537,
+      "learning_rate": 4.9268887083671815e-06,
+      "loss": 2.0684,
+      "step": 324
+    },
+    {
+      "epoch": 0.3561643835616438,
+      "grad_norm": 2.3070554733276367,
+      "learning_rate": 4.924857839155159e-06,
+      "loss": 2.0494,
+      "step": 325
+    },
+    {
+      "epoch": 0.35726027397260274,
+      "grad_norm": 2.189871072769165,
+      "learning_rate": 4.922826969943136e-06,
+      "loss": 2.0531,
+      "step": 326
+    },
+    {
+      "epoch": 0.35835616438356166,
+      "grad_norm": 2.553520917892456,
+      "learning_rate": 4.920796100731113e-06,
+      "loss": 2.0749,
+      "step": 327
+    },
+    {
+      "epoch": 0.3594520547945205,
+      "grad_norm": 2.62296199798584,
+      "learning_rate": 4.918765231519091e-06,
+      "loss": 1.9983,
+      "step": 328
+    },
+    {
+      "epoch": 0.36054794520547945,
+      "grad_norm": 2.4612369537353516,
+      "learning_rate": 4.916734362307068e-06,
+      "loss": 2.0641,
+      "step": 329
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "grad_norm": 2.656157970428467,
+      "learning_rate": 4.914703493095045e-06,
+      "loss": 2.0148,
+      "step": 330
+    },
+    {
+      "epoch": 0.36164383561643837,
+      "eval_loss": 2.2054214477539062,
+      "eval_runtime": 36.886,
+      "eval_samples_per_second": 27.111,
+      "eval_steps_per_second": 1.139,
+      "step": 330
+    },
+    {
+      "epoch": 0.36273972602739724,
+      "grad_norm": 2.2823736667633057,
+      "learning_rate": 4.912672623883022e-06,
+      "loss": 2.0221,
+      "step": 331
+    },
+    {
+      "epoch": 0.36383561643835616,
+      "grad_norm": 2.8403289318084717,
+      "learning_rate": 4.910641754671e-06,
+      "loss": 2.0149,
+      "step": 332
+    },
+    {
+      "epoch": 0.3649315068493151,
+      "grad_norm": 2.453887462615967,
+      "learning_rate": 4.908610885458977e-06,
+      "loss": 2.0411,
+      "step": 333
+    },
+    {
+      "epoch": 0.36602739726027395,
+      "grad_norm": 2.6094300746917725,
+      "learning_rate": 4.906580016246954e-06,
+      "loss": 2.0293,
+      "step": 334
+    },
+    {
+      "epoch": 0.36712328767123287,
+      "grad_norm": 2.4412384033203125,
+      "learning_rate": 4.9045491470349315e-06,
+      "loss": 2.0778,
+      "step": 335
+    },
+    {
+      "epoch": 0.3682191780821918,
+      "grad_norm": 2.4435057640075684,
+      "learning_rate": 4.902518277822909e-06,
+      "loss": 2.0597,
+      "step": 336
+    },
+    {
+      "epoch": 0.3693150684931507,
+      "grad_norm": 2.2390244007110596,
+      "learning_rate": 4.900487408610886e-06,
+      "loss": 2.0654,
+      "step": 337
+    },
+    {
+      "epoch": 0.3704109589041096,
+      "grad_norm": 2.4243953227996826,
+      "learning_rate": 4.898456539398863e-06,
+      "loss": 2.0696,
+      "step": 338
+    },
+    {
+      "epoch": 0.3715068493150685,
+      "grad_norm": 2.375720262527466,
+      "learning_rate": 4.89642567018684e-06,
+      "loss": 2.0866,
+      "step": 339
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "grad_norm": 2.6379668712615967,
+      "learning_rate": 4.894394800974818e-06,
+      "loss": 2.1552,
+      "step": 340
+    },
+    {
+      "epoch": 0.3726027397260274,
+      "eval_loss": 2.196237802505493,
+      "eval_runtime": 36.8953,
+      "eval_samples_per_second": 27.104,
+      "eval_steps_per_second": 1.138,
+      "step": 340
+    },
+    {
+      "epoch": 0.3736986301369863,
+      "grad_norm": 2.417145252227783,
+      "learning_rate": 4.892363931762794e-06,
+      "loss": 2.0024,
+      "step": 341
+    },
+    {
+      "epoch": 0.3747945205479452,
+      "grad_norm": 2.234590768814087,
+      "learning_rate": 4.8903330625507725e-06,
+      "loss": 2.0837,
+      "step": 342
+    },
+    {
+      "epoch": 0.37589041095890413,
+      "grad_norm": 2.404266595840454,
+      "learning_rate": 4.888302193338749e-06,
+      "loss": 2.018,
+      "step": 343
+    },
+    {
+      "epoch": 0.376986301369863,
+      "grad_norm": 2.364802837371826,
+      "learning_rate": 4.886271324126727e-06,
+      "loss": 1.994,
+      "step": 344
+    },
+    {
+      "epoch": 0.3780821917808219,
+      "grad_norm": 2.2262251377105713,
+      "learning_rate": 4.8842404549147034e-06,
+      "loss": 2.0929,
+      "step": 345
+    },
+    {
+      "epoch": 0.37917808219178084,
+      "grad_norm": 2.4854862689971924,
+      "learning_rate": 4.8822095857026816e-06,
+      "loss": 2.0801,
+      "step": 346
+    },
+    {
+      "epoch": 0.3802739726027397,
+      "grad_norm": 2.3861470222473145,
+      "learning_rate": 4.880178716490658e-06,
+      "loss": 2.0496,
+      "step": 347
+    },
+    {
+      "epoch": 0.38136986301369863,
+      "grad_norm": 2.372326612472534,
+      "learning_rate": 4.878147847278636e-06,
+      "loss": 1.9939,
+      "step": 348
+    },
+    {
+      "epoch": 0.38246575342465755,
+      "grad_norm": 2.3593149185180664,
+      "learning_rate": 4.8761169780666125e-06,
+      "loss": 2.0108,
+      "step": 349
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "grad_norm": 2.1802141666412354,
+      "learning_rate": 4.874086108854591e-06,
+      "loss": 1.983,
+      "step": 350
+    },
+    {
+      "epoch": 0.3835616438356164,
+      "eval_loss": 2.1947567462921143,
+      "eval_runtime": 36.8798,
+      "eval_samples_per_second": 27.115,
+      "eval_steps_per_second": 1.139,
+      "step": 350
+    },
+    {
+      "epoch": 0.38465753424657534,
+      "grad_norm": 2.47067928314209,
+      "learning_rate": 4.872055239642567e-06,
+      "loss": 2.0002,
+      "step": 351
+    },
+    {
+      "epoch": 0.38575342465753426,
+      "grad_norm": 2.359609365463257,
+      "learning_rate": 4.870024370430544e-06,
+      "loss": 1.9799,
+      "step": 352
+    },
+    {
+      "epoch": 0.3868493150684931,
+      "grad_norm": 2.3711702823638916,
+      "learning_rate": 4.867993501218522e-06,
+      "loss": 2.0341,
+      "step": 353
+    },
+    {
+      "epoch": 0.38794520547945205,
+      "grad_norm": 2.246647834777832,
+      "learning_rate": 4.865962632006499e-06,
+      "loss": 2.0687,
+      "step": 354
+    },
+    {
+      "epoch": 0.38904109589041097,
+      "grad_norm": 2.3031938076019287,
+      "learning_rate": 4.863931762794476e-06,
+      "loss": 1.9994,
+      "step": 355
+    },
+    {
+      "epoch": 0.39013698630136984,
+      "grad_norm": 2.492159605026245,
+      "learning_rate": 4.8619008935824535e-06,
+      "loss": 2.0195,
+      "step": 356
+    },
+    {
+      "epoch": 0.39123287671232876,
+      "grad_norm": 2.2878570556640625,
+      "learning_rate": 4.859870024370431e-06,
+      "loss": 2.0589,
+      "step": 357
+    },
+    {
+      "epoch": 0.3923287671232877,
+      "grad_norm": 2.444096088409424,
+      "learning_rate": 4.857839155158408e-06,
+      "loss": 2.0796,
+      "step": 358
+    },
+    {
+      "epoch": 0.3934246575342466,
+      "grad_norm": 2.463857650756836,
+      "learning_rate": 4.855808285946385e-06,
+      "loss": 2.0824,
+      "step": 359
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "grad_norm": 2.0893168449401855,
+      "learning_rate": 4.8537774167343626e-06,
+      "loss": 2.0651,
+      "step": 360
+    },
+    {
+      "epoch": 0.39452054794520547,
+      "eval_loss": 2.188523054122925,
+      "eval_runtime": 36.8679,
+      "eval_samples_per_second": 27.124,
+      "eval_steps_per_second": 1.139,
+      "step": 360
+    },
+    {
+      "epoch": 0.3956164383561644,
+      "grad_norm": 2.602850914001465,
+      "learning_rate": 4.85174654752234e-06,
+      "loss": 2.0507,
+      "step": 361
+    },
+    {
+      "epoch": 0.3967123287671233,
+      "grad_norm": 2.4185140132904053,
+      "learning_rate": 4.849715678310317e-06,
+      "loss": 2.0386,
+      "step": 362
+    },
+    {
+      "epoch": 0.3978082191780822,
+      "grad_norm": 2.1492245197296143,
+      "learning_rate": 4.847684809098294e-06,
+      "loss": 2.0376,
+      "step": 363
+    },
+    {
+      "epoch": 0.3989041095890411,
+      "grad_norm": 2.5710153579711914,
+      "learning_rate": 4.845653939886272e-06,
+      "loss": 2.0383,
+      "step": 364
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 2.4986608028411865,
+      "learning_rate": 4.843623070674249e-06,
+      "loss": 2.0464,
+      "step": 365
+    },
+    {
+      "epoch": 0.4010958904109589,
+      "grad_norm": 2.260388135910034,
+      "learning_rate": 4.841592201462226e-06,
+      "loss": 2.0217,
+      "step": 366
+    },
+    {
+      "epoch": 0.4021917808219178,
+      "grad_norm": 2.6500394344329834,
+      "learning_rate": 4.8395613322502035e-06,
+      "loss": 2.0777,
+      "step": 367
+    },
+    {
+      "epoch": 0.40328767123287673,
+      "grad_norm": 2.159274101257324,
+      "learning_rate": 4.837530463038181e-06,
+      "loss": 1.9747,
+      "step": 368
+    },
+    {
+      "epoch": 0.4043835616438356,
+      "grad_norm": 2.4623043537139893,
+      "learning_rate": 4.835499593826158e-06,
+      "loss": 2.039,
+      "step": 369
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "grad_norm": 2.2265007495880127,
+      "learning_rate": 4.833468724614135e-06,
+      "loss": 2.0618,
+      "step": 370
+    },
+    {
+      "epoch": 0.4054794520547945,
+      "eval_loss": 2.184051275253296,
+      "eval_runtime": 36.9284,
+      "eval_samples_per_second": 27.079,
+      "eval_steps_per_second": 1.137,
+      "step": 370
+    },
+    {
+      "epoch": 0.40657534246575344,
+      "grad_norm": 2.300276517868042,
+      "learning_rate": 4.831437855402113e-06,
+      "loss": 2.0382,
+      "step": 371
+    },
+    {
+      "epoch": 0.4076712328767123,
+      "grad_norm": 2.341156482696533,
+      "learning_rate": 4.82940698619009e-06,
+      "loss": 1.9361,
+      "step": 372
+    },
+    {
+      "epoch": 0.40876712328767123,
+      "grad_norm": 2.4196999073028564,
+      "learning_rate": 4.827376116978067e-06,
+      "loss": 2.0087,
+      "step": 373
+    },
+    {
+      "epoch": 0.40986301369863015,
+      "grad_norm": 2.4650514125823975,
+      "learning_rate": 4.8253452477660444e-06,
+      "loss": 2.0687,
+      "step": 374
+    },
+    {
+      "epoch": 0.410958904109589,
+      "grad_norm": 2.393739700317383,
+      "learning_rate": 4.823314378554022e-06,
+      "loss": 2.0421,
+      "step": 375
+    },
+    {
+      "epoch": 0.41205479452054794,
+      "grad_norm": 2.2436327934265137,
+      "learning_rate": 4.821283509341998e-06,
+      "loss": 2.0358,
+      "step": 376
+    },
+    {
+      "epoch": 0.41315068493150686,
+      "grad_norm": 2.1801650524139404,
+      "learning_rate": 4.819252640129976e-06,
+      "loss": 2.0081,
+      "step": 377
+    },
+    {
+      "epoch": 0.41424657534246573,
+      "grad_norm": 2.300882339477539,
+      "learning_rate": 4.817221770917953e-06,
+      "loss": 2.0321,
+      "step": 378
+    },
+    {
+      "epoch": 0.41534246575342465,
+      "grad_norm": 2.2136058807373047,
+      "learning_rate": 4.815190901705931e-06,
+      "loss": 1.9999,
+      "step": 379
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "grad_norm": 2.168034076690674,
+      "learning_rate": 4.813160032493907e-06,
+      "loss": 2.0336,
+      "step": 380
+    },
+    {
+      "epoch": 0.41643835616438357,
+      "eval_loss": 2.1770179271698,
+      "eval_runtime": 37.0015,
+      "eval_samples_per_second": 27.026,
+      "eval_steps_per_second": 1.135,
+      "step": 380
+    },
+    {
+      "epoch": 0.41753424657534244,
+      "grad_norm": 2.1604163646698,
+      "learning_rate": 4.811129163281885e-06,
+      "loss": 2.0212,
+      "step": 381
+    },
+    {
+      "epoch": 0.41863013698630136,
+      "grad_norm": 2.23923659324646,
+      "learning_rate": 4.809098294069862e-06,
+      "loss": 2.0495,
+      "step": 382
+    },
+    {
+      "epoch": 0.4197260273972603,
+      "grad_norm": 2.149949312210083,
+      "learning_rate": 4.80706742485784e-06,
+      "loss": 2.041,
+      "step": 383
+    },
+    {
+      "epoch": 0.4208219178082192,
+      "grad_norm": 2.49120831489563,
+      "learning_rate": 4.805036555645816e-06,
+      "loss": 2.0931,
+      "step": 384
+    },
+    {
+      "epoch": 0.42191780821917807,
+      "grad_norm": 2.2123563289642334,
+      "learning_rate": 4.8030056864337945e-06,
+      "loss": 2.0087,
+      "step": 385
+    },
+    {
+      "epoch": 0.423013698630137,
+      "grad_norm": 2.4235594272613525,
+      "learning_rate": 4.800974817221771e-06,
+      "loss": 2.0656,
+      "step": 386
+    },
+    {
+      "epoch": 0.4241095890410959,
+      "grad_norm": 2.158252239227295,
+      "learning_rate": 4.798943948009749e-06,
+      "loss": 2.0181,
+      "step": 387
+    },
+    {
+      "epoch": 0.4252054794520548,
+      "grad_norm": 2.2412970066070557,
+      "learning_rate": 4.7969130787977254e-06,
+      "loss": 2.0323,
+      "step": 388
+    },
+    {
+      "epoch": 0.4263013698630137,
+      "grad_norm": 2.2520956993103027,
+      "learning_rate": 4.794882209585703e-06,
+      "loss": 2.026,
+      "step": 389
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "grad_norm": 2.168954849243164,
+      "learning_rate": 4.79285134037368e-06,
+      "loss": 2.0459,
+      "step": 390
+    },
+    {
+      "epoch": 0.4273972602739726,
+      "eval_loss": 2.1705472469329834,
+      "eval_runtime": 37.0913,
+      "eval_samples_per_second": 26.96,
+      "eval_steps_per_second": 1.132,
+      "step": 390
+    },
+    {
+      "epoch": 0.4284931506849315,
+      "grad_norm": 2.1397407054901123,
+      "learning_rate": 4.790820471161657e-06,
+      "loss": 1.9995,
+      "step": 391
+    },
+    {
+      "epoch": 0.4295890410958904,
+      "grad_norm": 2.1747751235961914,
+      "learning_rate": 4.7887896019496345e-06,
+      "loss": 2.0088,
+      "step": 392
+    },
+    {
+      "epoch": 0.43068493150684933,
+      "grad_norm": 2.2860991954803467,
+      "learning_rate": 4.786758732737612e-06,
+      "loss": 2.0134,
+      "step": 393
+    },
+    {
+      "epoch": 0.4317808219178082,
+      "grad_norm": 2.2133519649505615,
+      "learning_rate": 4.784727863525589e-06,
+      "loss": 2.0586,
+      "step": 394
+    },
+    {
+      "epoch": 0.4328767123287671,
+      "grad_norm": 2.2147040367126465,
+      "learning_rate": 4.782696994313566e-06,
+      "loss": 1.9748,
+      "step": 395
+    },
+    {
+      "epoch": 0.43397260273972604,
+      "grad_norm": 2.343418836593628,
+      "learning_rate": 4.780666125101544e-06,
+      "loss": 2.0223,
+      "step": 396
+    },
+    {
+      "epoch": 0.4350684931506849,
+      "grad_norm": 2.198935031890869,
+      "learning_rate": 4.778635255889521e-06,
+      "loss": 1.9627,
+      "step": 397
+    },
+    {
+      "epoch": 0.43616438356164383,
+      "grad_norm": 2.1314704418182373,
+      "learning_rate": 4.776604386677498e-06,
+      "loss": 1.9938,
+      "step": 398
+    },
+    {
+      "epoch": 0.43726027397260275,
+      "grad_norm": 2.3005123138427734,
+      "learning_rate": 4.7745735174654755e-06,
+      "loss": 2.0029,
+      "step": 399
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "grad_norm": 2.2826664447784424,
+      "learning_rate": 4.772542648253453e-06,
+      "loss": 2.0585,
+      "step": 400
+    },
+    {
+      "epoch": 0.4383561643835616,
+      "eval_loss": 2.1691415309906006,
+      "eval_runtime": 37.0852,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 400
+    },
+    {
+      "epoch": 0.43945205479452054,
+      "grad_norm": 2.0547878742218018,
+      "learning_rate": 4.77051177904143e-06,
+      "loss": 2.0582,
+      "step": 401
+    },
+    {
+      "epoch": 0.44054794520547946,
+      "grad_norm": 2.200042486190796,
+      "learning_rate": 4.768480909829407e-06,
+      "loss": 1.9936,
+      "step": 402
+    },
+    {
+      "epoch": 0.44164383561643833,
+      "grad_norm": 2.354637861251831,
+      "learning_rate": 4.766450040617385e-06,
+      "loss": 2.0327,
+      "step": 403
+    },
+    {
+      "epoch": 0.44273972602739725,
+      "grad_norm": 2.2332348823547363,
+      "learning_rate": 4.764419171405362e-06,
+      "loss": 2.0052,
+      "step": 404
+    },
+    {
+      "epoch": 0.4438356164383562,
+      "grad_norm": 2.288060426712036,
+      "learning_rate": 4.762388302193339e-06,
+      "loss": 2.0472,
+      "step": 405
+    },
+    {
+      "epoch": 0.4449315068493151,
+      "grad_norm": 2.1368887424468994,
+      "learning_rate": 4.760357432981316e-06,
+      "loss": 2.0305,
+      "step": 406
+    },
+    {
+      "epoch": 0.44602739726027396,
+      "grad_norm": 2.362630844116211,
+      "learning_rate": 4.758326563769294e-06,
+      "loss": 2.002,
+      "step": 407
+    },
+    {
+      "epoch": 0.4471232876712329,
+      "grad_norm": 2.283419370651245,
+      "learning_rate": 4.756295694557271e-06,
+      "loss": 2.0271,
+      "step": 408
+    },
+    {
+      "epoch": 0.4482191780821918,
+      "grad_norm": 2.0400819778442383,
+      "learning_rate": 4.754264825345248e-06,
+      "loss": 2.0256,
+      "step": 409
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "grad_norm": 2.588416814804077,
+      "learning_rate": 4.7522339561332255e-06,
+      "loss": 1.9756,
+      "step": 410
+    },
+    {
+      "epoch": 0.44931506849315067,
+      "eval_loss": 2.17041015625,
+      "eval_runtime": 36.958,
+      "eval_samples_per_second": 27.058,
+      "eval_steps_per_second": 1.136,
+      "step": 410
+    },
+    {
+      "epoch": 0.4504109589041096,
+      "grad_norm": 2.2397494316101074,
+      "learning_rate": 4.750203086921203e-06,
+      "loss": 1.9876,
+      "step": 411
+    },
+    {
+      "epoch": 0.4515068493150685,
+      "grad_norm": 2.2085509300231934,
+      "learning_rate": 4.74817221770918e-06,
+      "loss": 1.9464,
+      "step": 412
+    },
+    {
+      "epoch": 0.4526027397260274,
+      "grad_norm": 2.140974760055542,
+      "learning_rate": 4.7461413484971565e-06,
+      "loss": 2.0179,
+      "step": 413
+    },
+    {
+      "epoch": 0.4536986301369863,
+      "grad_norm": 2.124366044998169,
+      "learning_rate": 4.744110479285135e-06,
+      "loss": 2.0732,
+      "step": 414
+    },
+    {
+      "epoch": 0.4547945205479452,
+      "grad_norm": 2.2463996410369873,
+      "learning_rate": 4.742079610073111e-06,
+      "loss": 1.9616,
+      "step": 415
+    },
+    {
+      "epoch": 0.4558904109589041,
+      "grad_norm": 2.1249961853027344,
+      "learning_rate": 4.740048740861089e-06,
+      "loss": 1.9889,
+      "step": 416
+    },
+    {
+      "epoch": 0.456986301369863,
+      "grad_norm": 2.25842022895813,
+      "learning_rate": 4.738017871649066e-06,
+      "loss": 2.0033,
+      "step": 417
+    },
+    {
+      "epoch": 0.45808219178082193,
+      "grad_norm": 2.271935224533081,
+      "learning_rate": 4.735987002437044e-06,
+      "loss": 2.0148,
+      "step": 418
+    },
+    {
+      "epoch": 0.4591780821917808,
+      "grad_norm": 2.4797163009643555,
+      "learning_rate": 4.73395613322502e-06,
+      "loss": 2.0736,
+      "step": 419
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "grad_norm": 2.1901395320892334,
+      "learning_rate": 4.731925264012998e-06,
+      "loss": 2.0763,
+      "step": 420
+    },
+    {
+      "epoch": 0.4602739726027397,
+      "eval_loss": 2.161660671234131,
+      "eval_runtime": 36.9099,
+      "eval_samples_per_second": 27.093,
+      "eval_steps_per_second": 1.138,
+      "step": 420
+    },
+    {
+      "epoch": 0.46136986301369864,
+      "grad_norm": 2.083536386489868,
+      "learning_rate": 4.729894394800975e-06,
+      "loss": 2.0299,
+      "step": 421
+    },
+    {
+      "epoch": 0.4624657534246575,
+      "grad_norm": 2.1118173599243164,
+      "learning_rate": 4.727863525588953e-06,
+      "loss": 1.9742,
+      "step": 422
+    },
+    {
+      "epoch": 0.46356164383561643,
+      "grad_norm": 2.195802927017212,
+      "learning_rate": 4.725832656376929e-06,
+      "loss": 1.9318,
+      "step": 423
+    },
+    {
+      "epoch": 0.46465753424657535,
+      "grad_norm": 2.318098545074463,
+      "learning_rate": 4.723801787164907e-06,
+      "loss": 1.9875,
+      "step": 424
+    },
+    {
+      "epoch": 0.4657534246575342,
+      "grad_norm": 2.1698496341705322,
+      "learning_rate": 4.721770917952885e-06,
+      "loss": 2.031,
+      "step": 425
+    },
+    {
+      "epoch": 0.46684931506849314,
+      "grad_norm": 2.283935785293579,
+      "learning_rate": 4.719740048740861e-06,
+      "loss": 1.9766,
+      "step": 426
+    },
+    {
+      "epoch": 0.46794520547945206,
+      "grad_norm": 2.4180757999420166,
+      "learning_rate": 4.717709179528839e-06,
+      "loss": 1.9976,
+      "step": 427
+    },
+    {
+      "epoch": 0.469041095890411,
+      "grad_norm": 2.2689030170440674,
+      "learning_rate": 4.715678310316816e-06,
+      "loss": 2.03,
+      "step": 428
+    },
+    {
+      "epoch": 0.47013698630136985,
+      "grad_norm": 2.423708200454712,
+      "learning_rate": 4.713647441104794e-06,
+      "loss": 2.0202,
+      "step": 429
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "grad_norm": 2.184705972671509,
+      "learning_rate": 4.71161657189277e-06,
+      "loss": 1.9655,
+      "step": 430
+    },
+    {
+      "epoch": 0.4712328767123288,
+      "eval_loss": 2.1548638343811035,
+      "eval_runtime": 36.9426,
+      "eval_samples_per_second": 27.069,
+      "eval_steps_per_second": 1.137,
+      "step": 430
+    },
+    {
+      "epoch": 0.4723287671232877,
+      "grad_norm": 2.295851945877075,
+      "learning_rate": 4.709585702680748e-06,
+      "loss": 1.9561,
+      "step": 431
+    },
+    {
+      "epoch": 0.47342465753424656,
+      "grad_norm": 2.1762046813964844,
+      "learning_rate": 4.707554833468725e-06,
+      "loss": 2.0341,
+      "step": 432
+    },
+    {
+      "epoch": 0.4745205479452055,
+      "grad_norm": 2.4684343338012695,
+      "learning_rate": 4.705523964256703e-06,
+      "loss": 2.0898,
+      "step": 433
+    },
+    {
+      "epoch": 0.4756164383561644,
+      "grad_norm": 2.293318271636963,
+      "learning_rate": 4.703493095044679e-06,
+      "loss": 1.9794,
+      "step": 434
+    },
+    {
+      "epoch": 0.4767123287671233,
+      "grad_norm": 2.1884210109710693,
+      "learning_rate": 4.701462225832657e-06,
+      "loss": 1.9979,
+      "step": 435
+    },
+    {
+      "epoch": 0.4778082191780822,
+      "grad_norm": 2.412504196166992,
+      "learning_rate": 4.699431356620634e-06,
+      "loss": 1.9237,
+      "step": 436
+    },
+    {
+      "epoch": 0.4789041095890411,
+      "grad_norm": 2.192824602127075,
+      "learning_rate": 4.697400487408612e-06,
+      "loss": 2.0756,
+      "step": 437
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 2.368100643157959,
+      "learning_rate": 4.695369618196588e-06,
+      "loss": 2.028,
+      "step": 438
+    },
+    {
+      "epoch": 0.4810958904109589,
+      "grad_norm": 2.218454122543335,
+      "learning_rate": 4.693338748984566e-06,
+      "loss": 2.0049,
+      "step": 439
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "grad_norm": 2.356168031692505,
+      "learning_rate": 4.691307879772543e-06,
+      "loss": 1.9958,
+      "step": 440
+    },
+    {
+      "epoch": 0.4821917808219178,
+      "eval_loss": 2.1540727615356445,
+      "eval_runtime": 36.9557,
+      "eval_samples_per_second": 27.059,
+      "eval_steps_per_second": 1.136,
+      "step": 440
+    },
+    {
+      "epoch": 0.4832876712328767,
+      "grad_norm": 2.09956693649292,
+      "learning_rate": 4.68927701056052e-06,
+      "loss": 1.9388,
+      "step": 441
+    },
+    {
+      "epoch": 0.4843835616438356,
+      "grad_norm": 2.5746614933013916,
+      "learning_rate": 4.6872461413484975e-06,
+      "loss": 2.0425,
+      "step": 442
+    },
+    {
+      "epoch": 0.48547945205479454,
+      "grad_norm": 2.2064783573150635,
+      "learning_rate": 4.685215272136475e-06,
+      "loss": 2.0082,
+      "step": 443
+    },
+    {
+      "epoch": 0.4865753424657534,
+      "grad_norm": 2.159029006958008,
+      "learning_rate": 4.683184402924452e-06,
+      "loss": 2.0055,
+      "step": 444
+    },
+    {
+      "epoch": 0.4876712328767123,
+      "grad_norm": 2.5222008228302,
+      "learning_rate": 4.681153533712429e-06,
+      "loss": 2.0007,
+      "step": 445
+    },
+    {
+      "epoch": 0.48876712328767125,
+      "grad_norm": 2.163487672805786,
+      "learning_rate": 4.679122664500407e-06,
+      "loss": 2.0719,
+      "step": 446
+    },
+    {
+      "epoch": 0.4898630136986301,
+      "grad_norm": 2.3383803367614746,
+      "learning_rate": 4.677091795288384e-06,
+      "loss": 1.9623,
+      "step": 447
+    },
+    {
+      "epoch": 0.49095890410958903,
+      "grad_norm": 2.1056675910949707,
+      "learning_rate": 4.675060926076361e-06,
+      "loss": 1.9443,
+      "step": 448
+    },
+    {
+      "epoch": 0.49205479452054796,
+      "grad_norm": 2.145270824432373,
+      "learning_rate": 4.673030056864338e-06,
+      "loss": 2.0469,
+      "step": 449
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "grad_norm": 2.578706741333008,
+      "learning_rate": 4.670999187652316e-06,
+      "loss": 1.9569,
+      "step": 450
+    },
+    {
+      "epoch": 0.4931506849315068,
+      "eval_loss": 2.1527743339538574,
+      "eval_runtime": 36.8961,
+      "eval_samples_per_second": 27.103,
+      "eval_steps_per_second": 1.138,
+      "step": 450
+    },
+    {
+      "epoch": 0.49424657534246574,
+      "grad_norm": 2.1328542232513428,
+      "learning_rate": 4.668968318440293e-06,
+      "loss": 1.9559,
+      "step": 451
+    },
+    {
+      "epoch": 0.49534246575342467,
+      "grad_norm": 2.148820638656616,
+      "learning_rate": 4.66693744922827e-06,
+      "loss": 1.9545,
+      "step": 452
+    },
+    {
+      "epoch": 0.4964383561643836,
+      "grad_norm": 2.250340461730957,
+      "learning_rate": 4.6649065800162475e-06,
+      "loss": 2.0358,
+      "step": 453
+    },
+    {
+      "epoch": 0.49753424657534245,
+      "grad_norm": 2.452397584915161,
+      "learning_rate": 4.662875710804225e-06,
+      "loss": 1.9294,
+      "step": 454
+    },
+    {
+      "epoch": 0.4986301369863014,
+      "grad_norm": 2.140615701675415,
+      "learning_rate": 4.660844841592202e-06,
+      "loss": 2.0084,
+      "step": 455
+    },
+    {
+      "epoch": 0.4997260273972603,
+      "grad_norm": 2.176806926727295,
+      "learning_rate": 4.658813972380179e-06,
+      "loss": 1.9174,
+      "step": 456
+    },
+    {
+      "epoch": 0.5008219178082192,
+      "grad_norm": 2.349031925201416,
+      "learning_rate": 4.656783103168157e-06,
+      "loss": 2.0377,
+      "step": 457
+    },
+    {
+      "epoch": 0.5019178082191781,
+      "grad_norm": 2.436889171600342,
+      "learning_rate": 4.654752233956134e-06,
+      "loss": 1.9133,
+      "step": 458
+    },
+    {
+      "epoch": 0.503013698630137,
+      "grad_norm": 2.116610288619995,
+      "learning_rate": 4.652721364744111e-06,
+      "loss": 1.9993,
+      "step": 459
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "grad_norm": 2.186915636062622,
+      "learning_rate": 4.6506904955320885e-06,
+      "loss": 1.9744,
+      "step": 460
+    },
+    {
+      "epoch": 0.5041095890410959,
+      "eval_loss": 2.1580278873443604,
+      "eval_runtime": 36.9097,
+      "eval_samples_per_second": 27.093,
+      "eval_steps_per_second": 1.138,
+      "step": 460
+    },
+    {
+      "epoch": 0.5052054794520547,
+      "grad_norm": 2.2009212970733643,
+      "learning_rate": 4.648659626320066e-06,
+      "loss": 2.0173,
+      "step": 461
+    },
+    {
+      "epoch": 0.5063013698630137,
+      "grad_norm": 2.117286443710327,
+      "learning_rate": 4.646628757108043e-06,
+      "loss": 2.0238,
+      "step": 462
+    },
+    {
+      "epoch": 0.5073972602739726,
+      "grad_norm": 2.238790273666382,
+      "learning_rate": 4.6445978878960194e-06,
+      "loss": 1.9885,
+      "step": 463
+    },
+    {
+      "epoch": 0.5084931506849315,
+      "grad_norm": 2.3493096828460693,
+      "learning_rate": 4.6425670186839976e-06,
+      "loss": 1.9593,
+      "step": 464
+    },
+    {
+      "epoch": 0.5095890410958904,
+      "grad_norm": 2.3276610374450684,
+      "learning_rate": 4.640536149471974e-06,
+      "loss": 1.9993,
+      "step": 465
+    },
+    {
+      "epoch": 0.5106849315068493,
+      "grad_norm": 2.3671715259552,
+      "learning_rate": 4.638505280259952e-06,
+      "loss": 2.0256,
+      "step": 466
+    },
+    {
+      "epoch": 0.5117808219178083,
+      "grad_norm": 2.251803159713745,
+      "learning_rate": 4.6364744110479285e-06,
+      "loss": 2.0436,
+      "step": 467
+    },
+    {
+      "epoch": 0.5128767123287671,
+      "grad_norm": 2.3023312091827393,
+      "learning_rate": 4.634443541835907e-06,
+      "loss": 1.9968,
+      "step": 468
+    },
+    {
+      "epoch": 0.513972602739726,
+      "grad_norm": 2.3660333156585693,
+      "learning_rate": 4.632412672623883e-06,
+      "loss": 2.0492,
+      "step": 469
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "grad_norm": 2.262179136276245,
+      "learning_rate": 4.630381803411861e-06,
+      "loss": 1.9698,
+      "step": 470
+    },
+    {
+      "epoch": 0.5150684931506849,
+      "eval_loss": 2.1516528129577637,
+      "eval_runtime": 36.7956,
+      "eval_samples_per_second": 27.177,
+      "eval_steps_per_second": 1.141,
+      "step": 470
+    },
+    {
+      "epoch": 0.5161643835616438,
+      "grad_norm": 2.450770616531372,
+      "learning_rate": 4.628350934199838e-06,
+      "loss": 1.8799,
+      "step": 471
+    },
+    {
+      "epoch": 0.5172602739726028,
+      "grad_norm": 2.187458038330078,
+      "learning_rate": 4.626320064987816e-06,
+      "loss": 1.9829,
+      "step": 472
+    },
+    {
+      "epoch": 0.5183561643835617,
+      "grad_norm": 2.1726722717285156,
+      "learning_rate": 4.624289195775792e-06,
+      "loss": 2.012,
+      "step": 473
+    },
+    {
+      "epoch": 0.5194520547945205,
+      "grad_norm": 2.4177136421203613,
+      "learning_rate": 4.62225832656377e-06,
+      "loss": 1.9796,
+      "step": 474
+    },
+    {
+      "epoch": 0.5205479452054794,
+      "grad_norm": 2.2353951930999756,
+      "learning_rate": 4.620227457351747e-06,
+      "loss": 1.99,
+      "step": 475
+    },
+    {
+      "epoch": 0.5216438356164383,
+      "grad_norm": 2.369591236114502,
+      "learning_rate": 4.618196588139724e-06,
+      "loss": 1.9847,
+      "step": 476
+    },
+    {
+      "epoch": 0.5227397260273973,
+      "grad_norm": 2.546703577041626,
+      "learning_rate": 4.616165718927701e-06,
+      "loss": 2.0047,
+      "step": 477
+    },
+    {
+      "epoch": 0.5238356164383562,
+      "grad_norm": 2.3509185314178467,
+      "learning_rate": 4.6141348497156786e-06,
+      "loss": 1.9589,
+      "step": 478
+    },
+    {
+      "epoch": 0.5249315068493151,
+      "grad_norm": 2.3063488006591797,
+      "learning_rate": 4.612103980503656e-06,
+      "loss": 2.045,
+      "step": 479
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "grad_norm": 2.285572052001953,
+      "learning_rate": 4.610073111291633e-06,
+      "loss": 2.0218,
+      "step": 480
+    },
+    {
+      "epoch": 0.5260273972602739,
+      "eval_loss": 2.1472301483154297,
+      "eval_runtime": 36.7689,
+      "eval_samples_per_second": 27.197,
+      "eval_steps_per_second": 1.142,
+      "step": 480
+    },
+    {
+      "epoch": 0.5271232876712328,
+      "grad_norm": 2.22987961769104,
+      "learning_rate": 4.60804224207961e-06,
+      "loss": 1.9955,
+      "step": 481
+    },
+    {
+      "epoch": 0.5282191780821918,
+      "grad_norm": 2.200918197631836,
+      "learning_rate": 4.606011372867588e-06,
+      "loss": 1.9768,
+      "step": 482
+    },
+    {
+      "epoch": 0.5293150684931507,
+      "grad_norm": 2.2441930770874023,
+      "learning_rate": 4.603980503655565e-06,
+      "loss": 2.0088,
+      "step": 483
+    },
+    {
+      "epoch": 0.5304109589041096,
+      "grad_norm": 2.380383014678955,
+      "learning_rate": 4.601949634443542e-06,
+      "loss": 1.9853,
+      "step": 484
+    },
+    {
+      "epoch": 0.5315068493150685,
+      "grad_norm": 2.317995071411133,
+      "learning_rate": 4.5999187652315195e-06,
+      "loss": 1.9183,
+      "step": 485
+    },
+    {
+      "epoch": 0.5326027397260275,
+      "grad_norm": 2.2462668418884277,
+      "learning_rate": 4.597887896019497e-06,
+      "loss": 1.9201,
+      "step": 486
+    },
+    {
+      "epoch": 0.5336986301369863,
+      "grad_norm": 2.1482582092285156,
+      "learning_rate": 4.595857026807474e-06,
+      "loss": 1.9722,
+      "step": 487
+    },
+    {
+      "epoch": 0.5347945205479452,
+      "grad_norm": 2.079831123352051,
+      "learning_rate": 4.593826157595451e-06,
+      "loss": 1.977,
+      "step": 488
+    },
+    {
+      "epoch": 0.5358904109589041,
+      "grad_norm": 2.2169578075408936,
+      "learning_rate": 4.591795288383429e-06,
+      "loss": 1.9615,
+      "step": 489
+    },
+    {
+      "epoch": 0.536986301369863,
+      "grad_norm": 2.087846279144287,
+      "learning_rate": 4.589764419171406e-06,
+      "loss": 2.0402,
+      "step": 490
+    },
+    {
+      "epoch": 0.536986301369863,
+      "eval_loss": 2.1416702270507812,
+      "eval_runtime": 37.1191,
+      "eval_samples_per_second": 26.94,
+      "eval_steps_per_second": 1.131,
+      "step": 490
+    },
+    {
+      "epoch": 0.538082191780822,
+      "grad_norm": 2.2161977291107178,
+      "learning_rate": 4.587733549959383e-06,
+      "loss": 1.9959,
+      "step": 491
+    },
+    {
+      "epoch": 0.5391780821917809,
+      "grad_norm": 2.19530987739563,
+      "learning_rate": 4.5857026807473604e-06,
+      "loss": 1.961,
+      "step": 492
+    },
+    {
+      "epoch": 0.5402739726027397,
+      "grad_norm": 2.1150615215301514,
+      "learning_rate": 4.583671811535338e-06,
+      "loss": 1.946,
+      "step": 493
+    },
+    {
+      "epoch": 0.5413698630136986,
+      "grad_norm": 2.186352491378784,
+      "learning_rate": 4.581640942323315e-06,
+      "loss": 1.9137,
+      "step": 494
+    },
+    {
+      "epoch": 0.5424657534246575,
+      "grad_norm": 2.233370780944824,
+      "learning_rate": 4.579610073111292e-06,
+      "loss": 2.021,
+      "step": 495
+    },
+    {
+      "epoch": 0.5435616438356164,
+      "grad_norm": 2.1326382160186768,
+      "learning_rate": 4.5775792038992695e-06,
+      "loss": 2.0242,
+      "step": 496
+    },
+    {
+      "epoch": 0.5446575342465754,
+      "grad_norm": 2.536837577819824,
+      "learning_rate": 4.575548334687247e-06,
+      "loss": 2.0185,
+      "step": 497
+    },
+    {
+      "epoch": 0.5457534246575343,
+      "grad_norm": 2.193598747253418,
+      "learning_rate": 4.573517465475224e-06,
+      "loss": 2.0533,
+      "step": 498
+    },
+    {
+      "epoch": 0.5468493150684931,
+      "grad_norm": 2.1346311569213867,
+      "learning_rate": 4.571486596263201e-06,
+      "loss": 1.9564,
+      "step": 499
+    },
+    {
+      "epoch": 0.547945205479452,
+      "grad_norm": 2.4294722080230713,
+      "learning_rate": 4.569455727051178e-06,
+      "loss": 2.0003,
+      "step": 500
+    },
+    {
+      "epoch": 0.547945205479452,
+      "eval_loss": 2.1430764198303223,
+      "eval_runtime": 37.0591,
+      "eval_samples_per_second": 26.984,
+      "eval_steps_per_second": 1.133,
+      "step": 500
+    },
+    {
+      "epoch": 0.549041095890411,
+      "grad_norm": 2.2394182682037354,
+      "learning_rate": 4.567424857839156e-06,
+      "loss": 2.0055,
+      "step": 501
+    },
+    {
+      "epoch": 0.5501369863013699,
+      "grad_norm": 2.2543067932128906,
+      "learning_rate": 4.565393988627132e-06,
+      "loss": 1.9837,
+      "step": 502
+    },
+    {
+      "epoch": 0.5512328767123288,
+      "grad_norm": 2.2805464267730713,
+      "learning_rate": 4.5633631194151105e-06,
+      "loss": 2.0495,
+      "step": 503
+    },
+    {
+      "epoch": 0.5523287671232877,
+      "grad_norm": 2.2838146686553955,
+      "learning_rate": 4.561332250203087e-06,
+      "loss": 1.9866,
+      "step": 504
+    },
+    {
+      "epoch": 0.5534246575342465,
+      "grad_norm": 2.286147356033325,
+      "learning_rate": 4.559301380991065e-06,
+      "loss": 1.9805,
+      "step": 505
+    },
+    {
+      "epoch": 0.5545205479452054,
+      "grad_norm": 2.064378023147583,
+      "learning_rate": 4.5572705117790414e-06,
+      "loss": 1.9259,
+      "step": 506
+    },
+    {
+      "epoch": 0.5556164383561644,
+      "grad_norm": 2.4020493030548096,
+      "learning_rate": 4.5552396425670196e-06,
+      "loss": 1.9731,
+      "step": 507
+    },
+    {
+      "epoch": 0.5567123287671233,
+      "grad_norm": 2.11594557762146,
+      "learning_rate": 4.553208773354996e-06,
+      "loss": 2.0126,
+      "step": 508
+    },
+    {
+      "epoch": 0.5578082191780822,
+      "grad_norm": 2.4520673751831055,
+      "learning_rate": 4.551177904142974e-06,
+      "loss": 1.9396,
+      "step": 509
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "grad_norm": 2.302771806716919,
+      "learning_rate": 4.5491470349309505e-06,
+      "loss": 2.0361,
+      "step": 510
+    },
+    {
+      "epoch": 0.5589041095890411,
+      "eval_loss": 2.1408755779266357,
+      "eval_runtime": 36.9881,
+      "eval_samples_per_second": 27.036,
+      "eval_steps_per_second": 1.136,
+      "step": 510
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 2.250927686691284,
+      "learning_rate": 4.547116165718929e-06,
+      "loss": 1.9616,
+      "step": 511
+    },
+    {
+      "epoch": 0.5610958904109589,
+      "grad_norm": 2.4257938861846924,
+      "learning_rate": 4.545085296506905e-06,
+      "loss": 1.9836,
+      "step": 512
+    },
+    {
+      "epoch": 0.5621917808219178,
+      "grad_norm": 2.08693265914917,
+      "learning_rate": 4.543054427294882e-06,
+      "loss": 1.9541,
+      "step": 513
+    },
+    {
+      "epoch": 0.5632876712328767,
+      "grad_norm": 2.133875846862793,
+      "learning_rate": 4.54102355808286e-06,
+      "loss": 1.9377,
+      "step": 514
+    },
+    {
+      "epoch": 0.5643835616438356,
+      "grad_norm": 2.1217594146728516,
+      "learning_rate": 4.538992688870837e-06,
+      "loss": 1.9369,
+      "step": 515
+    },
+    {
+      "epoch": 0.5654794520547946,
+      "grad_norm": 2.117994546890259,
+      "learning_rate": 4.536961819658814e-06,
+      "loss": 1.9666,
+      "step": 516
+    },
+    {
+      "epoch": 0.5665753424657535,
+      "grad_norm": 2.1514523029327393,
+      "learning_rate": 4.5349309504467915e-06,
+      "loss": 1.9704,
+      "step": 517
+    },
+    {
+      "epoch": 0.5676712328767123,
+      "grad_norm": 2.2309377193450928,
+      "learning_rate": 4.532900081234769e-06,
+      "loss": 1.9858,
+      "step": 518
+    },
+    {
+      "epoch": 0.5687671232876712,
+      "grad_norm": 2.100398540496826,
+      "learning_rate": 4.530869212022746e-06,
+      "loss": 1.9573,
+      "step": 519
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "grad_norm": 2.122840166091919,
+      "learning_rate": 4.528838342810723e-06,
+      "loss": 1.9813,
+      "step": 520
+    },
+    {
+      "epoch": 0.5698630136986301,
+      "eval_loss": 2.135481357574463,
+      "eval_runtime": 36.893,
+      "eval_samples_per_second": 27.105,
+      "eval_steps_per_second": 1.138,
+      "step": 520
+    },
+    {
+      "epoch": 0.570958904109589,
+      "grad_norm": 2.1103999614715576,
+      "learning_rate": 4.526807473598701e-06,
+      "loss": 1.9691,
+      "step": 521
+    },
+    {
+      "epoch": 0.572054794520548,
+      "grad_norm": 2.059232473373413,
+      "learning_rate": 4.524776604386678e-06,
+      "loss": 1.9687,
+      "step": 522
+    },
+    {
+      "epoch": 0.5731506849315069,
+      "grad_norm": 2.207000732421875,
+      "learning_rate": 4.522745735174655e-06,
+      "loss": 1.949,
+      "step": 523
+    },
+    {
+      "epoch": 0.5742465753424657,
+      "grad_norm": 2.264000177383423,
+      "learning_rate": 4.520714865962632e-06,
+      "loss": 2.0066,
+      "step": 524
+    },
+    {
+      "epoch": 0.5753424657534246,
+      "grad_norm": 2.121701717376709,
+      "learning_rate": 4.51868399675061e-06,
+      "loss": 1.9879,
+      "step": 525
+    },
+    {
+      "epoch": 0.5764383561643835,
+      "grad_norm": 2.0745644569396973,
+      "learning_rate": 4.516653127538587e-06,
+      "loss": 1.9924,
+      "step": 526
+    },
+    {
+      "epoch": 0.5775342465753425,
+      "grad_norm": 2.2257349491119385,
+      "learning_rate": 4.514622258326564e-06,
+      "loss": 1.9525,
+      "step": 527
+    },
+    {
+      "epoch": 0.5786301369863014,
+      "grad_norm": 2.1678507328033447,
+      "learning_rate": 4.5125913891145415e-06,
+      "loss": 1.9378,
+      "step": 528
+    },
+    {
+      "epoch": 0.5797260273972603,
+      "grad_norm": 2.2800841331481934,
+      "learning_rate": 4.510560519902519e-06,
+      "loss": 1.9685,
+      "step": 529
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "grad_norm": 2.1471869945526123,
+      "learning_rate": 4.508529650690496e-06,
+      "loss": 2.0027,
+      "step": 530
+    },
+    {
+      "epoch": 0.5808219178082191,
+      "eval_loss": 2.1354892253875732,
+      "eval_runtime": 36.8434,
+      "eval_samples_per_second": 27.142,
+      "eval_steps_per_second": 1.14,
+      "step": 530
+    },
+    {
+      "epoch": 0.581917808219178,
+      "grad_norm": 2.2238621711730957,
+      "learning_rate": 4.506498781478473e-06,
+      "loss": 1.9258,
+      "step": 531
+    },
+    {
+      "epoch": 0.583013698630137,
+      "grad_norm": 2.2224624156951904,
+      "learning_rate": 4.504467912266451e-06,
+      "loss": 1.9168,
+      "step": 532
+    },
+    {
+      "epoch": 0.5841095890410959,
+      "grad_norm": 1.9850294589996338,
+      "learning_rate": 4.502437043054428e-06,
+      "loss": 1.9656,
+      "step": 533
+    },
+    {
+      "epoch": 0.5852054794520548,
+      "grad_norm": 2.1303517818450928,
+      "learning_rate": 4.500406173842405e-06,
+      "loss": 2.0349,
+      "step": 534
+    },
+    {
+      "epoch": 0.5863013698630137,
+      "grad_norm": 2.2114832401275635,
+      "learning_rate": 4.4983753046303824e-06,
+      "loss": 1.9878,
+      "step": 535
+    },
+    {
+      "epoch": 0.5873972602739727,
+      "grad_norm": 1.9375886917114258,
+      "learning_rate": 4.49634443541836e-06,
+      "loss": 1.9724,
+      "step": 536
+    },
+    {
+      "epoch": 0.5884931506849315,
+      "grad_norm": 2.185518741607666,
+      "learning_rate": 4.494313566206336e-06,
+      "loss": 1.9584,
+      "step": 537
+    },
+    {
+      "epoch": 0.5895890410958904,
+      "grad_norm": 2.125638484954834,
+      "learning_rate": 4.492282696994314e-06,
+      "loss": 2.0203,
+      "step": 538
+    },
+    {
+      "epoch": 0.5906849315068493,
+      "grad_norm": 1.9456262588500977,
+      "learning_rate": 4.490251827782291e-06,
+      "loss": 1.9523,
+      "step": 539
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "grad_norm": 2.088125705718994,
+      "learning_rate": 4.488220958570269e-06,
+      "loss": 1.9271,
+      "step": 540
+    },
+    {
+      "epoch": 0.5917808219178082,
+      "eval_loss": 2.134958505630493,
+      "eval_runtime": 36.8342,
+      "eval_samples_per_second": 27.149,
+      "eval_steps_per_second": 1.14,
+      "step": 540
+    },
+    {
+      "epoch": 0.5928767123287672,
+      "grad_norm": 2.1423559188842773,
+      "learning_rate": 4.486190089358245e-06,
+      "loss": 1.9467,
+      "step": 541
+    },
+    {
+      "epoch": 0.5939726027397261,
+      "grad_norm": 2.2169737815856934,
+      "learning_rate": 4.484159220146223e-06,
+      "loss": 1.8901,
+      "step": 542
+    },
+    {
+      "epoch": 0.5950684931506849,
+      "grad_norm": 2.419947385787964,
+      "learning_rate": 4.4821283509342e-06,
+      "loss": 2.0064,
+      "step": 543
+    },
+    {
+      "epoch": 0.5961643835616438,
+      "grad_norm": 2.0396435260772705,
+      "learning_rate": 4.480097481722178e-06,
+      "loss": 1.9819,
+      "step": 544
+    },
+    {
+      "epoch": 0.5972602739726027,
+      "grad_norm": 2.210700511932373,
+      "learning_rate": 4.478066612510154e-06,
+      "loss": 1.9578,
+      "step": 545
+    },
+    {
+      "epoch": 0.5983561643835617,
+      "grad_norm": 2.159698009490967,
+      "learning_rate": 4.4760357432981325e-06,
+      "loss": 1.9313,
+      "step": 546
+    },
+    {
+      "epoch": 0.5994520547945206,
+      "grad_norm": 2.125058650970459,
+      "learning_rate": 4.474004874086109e-06,
+      "loss": 1.9549,
+      "step": 547
+    },
+    {
+      "epoch": 0.6005479452054795,
+      "grad_norm": 2.477320432662964,
+      "learning_rate": 4.471974004874087e-06,
+      "loss": 2.0883,
+      "step": 548
+    },
+    {
+      "epoch": 0.6016438356164383,
+      "grad_norm": 2.214901924133301,
+      "learning_rate": 4.4699431356620635e-06,
+      "loss": 1.9184,
+      "step": 549
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "grad_norm": 2.1001312732696533,
+      "learning_rate": 4.467912266450041e-06,
+      "loss": 2.0231,
+      "step": 550
+    },
+    {
+      "epoch": 0.6027397260273972,
+      "eval_loss": 2.1315157413482666,
+      "eval_runtime": 36.8131,
+      "eval_samples_per_second": 27.164,
+      "eval_steps_per_second": 1.141,
+      "step": 550
+    },
+    {
+      "epoch": 0.6038356164383561,
+      "grad_norm": 2.1030538082122803,
+      "learning_rate": 4.465881397238018e-06,
+      "loss": 1.967,
+      "step": 551
+    },
+    {
+      "epoch": 0.6049315068493151,
+      "grad_norm": 2.0390465259552,
+      "learning_rate": 4.463850528025995e-06,
+      "loss": 1.9615,
+      "step": 552
+    },
+    {
+      "epoch": 0.606027397260274,
+      "grad_norm": 2.1331627368927,
+      "learning_rate": 4.4618196588139726e-06,
+      "loss": 1.9667,
+      "step": 553
+    },
+    {
+      "epoch": 0.6071232876712329,
+      "grad_norm": 2.1987671852111816,
+      "learning_rate": 4.45978878960195e-06,
+      "loss": 1.9922,
+      "step": 554
+    },
+    {
+      "epoch": 0.6082191780821918,
+      "grad_norm": 2.127434253692627,
+      "learning_rate": 4.457757920389927e-06,
+      "loss": 1.9913,
+      "step": 555
+    },
+    {
+      "epoch": 0.6093150684931506,
+      "grad_norm": 2.2486324310302734,
+      "learning_rate": 4.455727051177904e-06,
+      "loss": 1.9239,
+      "step": 556
+    },
+    {
+      "epoch": 0.6104109589041096,
+      "grad_norm": 2.127575397491455,
+      "learning_rate": 4.453696181965882e-06,
+      "loss": 1.9521,
+      "step": 557
+    },
+    {
+      "epoch": 0.6115068493150685,
+      "grad_norm": 2.0798254013061523,
+      "learning_rate": 4.451665312753859e-06,
+      "loss": 2.0085,
+      "step": 558
+    },
+    {
+      "epoch": 0.6126027397260274,
+      "grad_norm": 2.287574291229248,
+      "learning_rate": 4.449634443541836e-06,
+      "loss": 1.98,
+      "step": 559
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "grad_norm": 2.2056541442871094,
+      "learning_rate": 4.4476035743298135e-06,
+      "loss": 2.0282,
+      "step": 560
+    },
+    {
+      "epoch": 0.6136986301369863,
+      "eval_loss": 2.1287026405334473,
+      "eval_runtime": 36.7788,
+      "eval_samples_per_second": 27.19,
+      "eval_steps_per_second": 1.142,
+      "step": 560
+    },
+    {
+      "epoch": 0.6147945205479453,
+      "grad_norm": 2.169157028198242,
+      "learning_rate": 4.445572705117791e-06,
+      "loss": 1.9308,
+      "step": 561
+    },
+    {
+      "epoch": 0.6158904109589041,
+      "grad_norm": 2.1035149097442627,
+      "learning_rate": 4.443541835905768e-06,
+      "loss": 1.9507,
+      "step": 562
+    },
+    {
+      "epoch": 0.616986301369863,
+      "grad_norm": 2.1174328327178955,
+      "learning_rate": 4.441510966693745e-06,
+      "loss": 2.0162,
+      "step": 563
+    },
+    {
+      "epoch": 0.6180821917808219,
+      "grad_norm": 2.201641321182251,
+      "learning_rate": 4.439480097481723e-06,
+      "loss": 1.9573,
+      "step": 564
+    },
+    {
+      "epoch": 0.6191780821917808,
+      "grad_norm": 2.1763603687286377,
+      "learning_rate": 4.4374492282697e-06,
+      "loss": 2.0149,
+      "step": 565
+    },
+    {
+      "epoch": 0.6202739726027398,
+      "grad_norm": 2.2866439819335938,
+      "learning_rate": 4.435418359057677e-06,
+      "loss": 1.8836,
+      "step": 566
+    },
+    {
+      "epoch": 0.6213698630136987,
+      "grad_norm": 2.2060160636901855,
+      "learning_rate": 4.433387489845654e-06,
+      "loss": 1.9806,
+      "step": 567
+    },
+    {
+      "epoch": 0.6224657534246575,
+      "grad_norm": 2.0614662170410156,
+      "learning_rate": 4.431356620633632e-06,
+      "loss": 1.9821,
+      "step": 568
+    },
+    {
+      "epoch": 0.6235616438356164,
+      "grad_norm": 2.1055402755737305,
+      "learning_rate": 4.429325751421609e-06,
+      "loss": 1.9365,
+      "step": 569
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "grad_norm": 1.9907679557800293,
+      "learning_rate": 4.427294882209586e-06,
+      "loss": 1.9429,
+      "step": 570
+    },
+    {
+      "epoch": 0.6246575342465753,
+      "eval_loss": 2.1286206245422363,
+      "eval_runtime": 36.7748,
+      "eval_samples_per_second": 27.193,
+      "eval_steps_per_second": 1.142,
+      "step": 570
+    },
+    {
+      "epoch": 0.6257534246575343,
+      "grad_norm": 2.0169243812561035,
+      "learning_rate": 4.4252640129975635e-06,
+      "loss": 1.9643,
+      "step": 571
+    },
+    {
+      "epoch": 0.6268493150684932,
+      "grad_norm": 2.1082401275634766,
+      "learning_rate": 4.423233143785541e-06,
+      "loss": 1.9009,
+      "step": 572
+    },
+    {
+      "epoch": 0.6279452054794521,
+      "grad_norm": 2.202542543411255,
+      "learning_rate": 4.421202274573518e-06,
+      "loss": 1.9694,
+      "step": 573
+    },
+    {
+      "epoch": 0.6290410958904109,
+      "grad_norm": 2.028489828109741,
+      "learning_rate": 4.4191714053614945e-06,
+      "loss": 1.9697,
+      "step": 574
+    },
+    {
+      "epoch": 0.6301369863013698,
+      "grad_norm": 2.134274482727051,
+      "learning_rate": 4.417140536149473e-06,
+      "loss": 1.8672,
+      "step": 575
+    },
+    {
+      "epoch": 0.6312328767123287,
+      "grad_norm": 2.063542366027832,
+      "learning_rate": 4.415109666937449e-06,
+      "loss": 1.909,
+      "step": 576
+    },
+    {
+      "epoch": 0.6323287671232877,
+      "grad_norm": 1.9747039079666138,
+      "learning_rate": 4.413078797725427e-06,
+      "loss": 1.9795,
+      "step": 577
+    },
+    {
+      "epoch": 0.6334246575342466,
+      "grad_norm": 2.063673496246338,
+      "learning_rate": 4.411047928513404e-06,
+      "loss": 1.9283,
+      "step": 578
+    },
+    {
+      "epoch": 0.6345205479452055,
+      "grad_norm": 2.024611711502075,
+      "learning_rate": 4.409017059301382e-06,
+      "loss": 1.9682,
+      "step": 579
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "grad_norm": 2.0471599102020264,
+      "learning_rate": 4.406986190089358e-06,
+      "loss": 1.969,
+      "step": 580
+    },
+    {
+      "epoch": 0.6356164383561644,
+      "eval_loss": 2.1233692169189453,
+      "eval_runtime": 36.8185,
+      "eval_samples_per_second": 27.16,
+      "eval_steps_per_second": 1.141,
+      "step": 580
+    },
+    {
+      "epoch": 0.6367123287671232,
+      "grad_norm": 2.1038553714752197,
+      "learning_rate": 4.404955320877336e-06,
+      "loss": 1.9915,
+      "step": 581
+    },
+    {
+      "epoch": 0.6378082191780822,
+      "grad_norm": 2.1222152709960938,
+      "learning_rate": 4.402924451665313e-06,
+      "loss": 1.9892,
+      "step": 582
+    },
+    {
+      "epoch": 0.6389041095890411,
+      "grad_norm": 2.09137225151062,
+      "learning_rate": 4.400893582453291e-06,
+      "loss": 1.934,
+      "step": 583
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 2.2158281803131104,
+      "learning_rate": 4.398862713241267e-06,
+      "loss": 1.9188,
+      "step": 584
+    },
+    {
+      "epoch": 0.6410958904109589,
+      "grad_norm": 2.34619140625,
+      "learning_rate": 4.396831844029245e-06,
+      "loss": 1.9495,
+      "step": 585
+    },
+    {
+      "epoch": 0.6421917808219179,
+      "grad_norm": 2.2732315063476562,
+      "learning_rate": 4.394800974817222e-06,
+      "loss": 1.9632,
+      "step": 586
+    },
+    {
+      "epoch": 0.6432876712328767,
+      "grad_norm": 2.153338670730591,
+      "learning_rate": 4.392770105605199e-06,
+      "loss": 1.9551,
+      "step": 587
+    },
+    {
+      "epoch": 0.6443835616438356,
+      "grad_norm": 2.2201521396636963,
+      "learning_rate": 4.390739236393176e-06,
+      "loss": 1.9677,
+      "step": 588
+    },
+    {
+      "epoch": 0.6454794520547945,
+      "grad_norm": 1.9894570112228394,
+      "learning_rate": 4.388708367181154e-06,
+      "loss": 1.9826,
+      "step": 589
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "grad_norm": 2.145061492919922,
+      "learning_rate": 4.386677497969131e-06,
+      "loss": 1.9826,
+      "step": 590
+    },
+    {
+      "epoch": 0.6465753424657534,
+      "eval_loss": 2.122312545776367,
+      "eval_runtime": 37.1153,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 590
+    },
+    {
+      "epoch": 0.6476712328767124,
+      "grad_norm": 2.257237672805786,
+      "learning_rate": 4.384646628757108e-06,
+      "loss": 1.9722,
+      "step": 591
+    },
+    {
+      "epoch": 0.6487671232876713,
+      "grad_norm": 2.2406375408172607,
+      "learning_rate": 4.3826157595450855e-06,
+      "loss": 2.0141,
+      "step": 592
+    },
+    {
+      "epoch": 0.6498630136986301,
+      "grad_norm": 1.9656853675842285,
+      "learning_rate": 4.380584890333063e-06,
+      "loss": 1.9435,
+      "step": 593
+    },
+    {
+      "epoch": 0.650958904109589,
+      "grad_norm": 2.1063120365142822,
+      "learning_rate": 4.37855402112104e-06,
+      "loss": 1.9001,
+      "step": 594
+    },
+    {
+      "epoch": 0.6520547945205479,
+      "grad_norm": 2.1594269275665283,
+      "learning_rate": 4.376523151909017e-06,
+      "loss": 1.8596,
+      "step": 595
+    },
+    {
+      "epoch": 0.6531506849315069,
+      "grad_norm": 2.2261908054351807,
+      "learning_rate": 4.3744922826969946e-06,
+      "loss": 2.016,
+      "step": 596
+    },
+    {
+      "epoch": 0.6542465753424658,
+      "grad_norm": 1.9940093755722046,
+      "learning_rate": 4.372461413484972e-06,
+      "loss": 1.9293,
+      "step": 597
+    },
+    {
+      "epoch": 0.6553424657534247,
+      "grad_norm": 2.62186861038208,
+      "learning_rate": 4.370430544272949e-06,
+      "loss": 1.9656,
+      "step": 598
+    },
+    {
+      "epoch": 0.6564383561643835,
+      "grad_norm": 2.3531494140625,
+      "learning_rate": 4.368399675060926e-06,
+      "loss": 1.9984,
+      "step": 599
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "grad_norm": 2.362769365310669,
+      "learning_rate": 4.366368805848904e-06,
+      "loss": 1.9971,
+      "step": 600
+    },
+    {
+      "epoch": 0.6575342465753424,
+      "eval_loss": 2.1172940731048584,
+      "eval_runtime": 37.0498,
+      "eval_samples_per_second": 26.991,
+      "eval_steps_per_second": 1.134,
+      "step": 600
+    },
+    {
+      "epoch": 0.6586301369863014,
+      "grad_norm": 2.3806607723236084,
+      "learning_rate": 4.364337936636881e-06,
+      "loss": 1.9566,
+      "step": 601
+    },
+    {
+      "epoch": 0.6597260273972603,
+      "grad_norm": 2.5531814098358154,
+      "learning_rate": 4.362307067424858e-06,
+      "loss": 1.9002,
+      "step": 602
+    },
+    {
+      "epoch": 0.6608219178082192,
+      "grad_norm": 2.283364772796631,
+      "learning_rate": 4.3602761982128355e-06,
+      "loss": 1.9633,
+      "step": 603
+    },
+    {
+      "epoch": 0.6619178082191781,
+      "grad_norm": 2.180171489715576,
+      "learning_rate": 4.358245329000813e-06,
+      "loss": 1.9277,
+      "step": 604
+    },
+    {
+      "epoch": 0.663013698630137,
+      "grad_norm": 2.0771350860595703,
+      "learning_rate": 4.35621445978879e-06,
+      "loss": 1.9484,
+      "step": 605
+    },
+    {
+      "epoch": 0.6641095890410958,
+      "grad_norm": 2.3772811889648438,
+      "learning_rate": 4.354183590576767e-06,
+      "loss": 1.9564,
+      "step": 606
+    },
+    {
+      "epoch": 0.6652054794520548,
+      "grad_norm": 2.205486297607422,
+      "learning_rate": 4.352152721364745e-06,
+      "loss": 1.9389,
+      "step": 607
+    },
+    {
+      "epoch": 0.6663013698630137,
+      "grad_norm": 2.085190773010254,
+      "learning_rate": 4.350121852152722e-06,
+      "loss": 1.8948,
+      "step": 608
+    },
+    {
+      "epoch": 0.6673972602739726,
+      "grad_norm": 2.4398982524871826,
+      "learning_rate": 4.348090982940699e-06,
+      "loss": 1.9736,
+      "step": 609
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "grad_norm": 2.194087028503418,
+      "learning_rate": 4.3460601137286764e-06,
+      "loss": 1.9607,
+      "step": 610
+    },
+    {
+      "epoch": 0.6684931506849315,
+      "eval_loss": 2.115041971206665,
+      "eval_runtime": 36.9738,
+      "eval_samples_per_second": 27.046,
+      "eval_steps_per_second": 1.136,
+      "step": 610
+    },
+    {
+      "epoch": 0.6695890410958905,
+      "grad_norm": 2.143054485321045,
+      "learning_rate": 4.344029244516654e-06,
+      "loss": 1.9733,
+      "step": 611
+    },
+    {
+      "epoch": 0.6706849315068493,
+      "grad_norm": 2.2484617233276367,
+      "learning_rate": 4.341998375304631e-06,
+      "loss": 1.9417,
+      "step": 612
+    },
+    {
+      "epoch": 0.6717808219178082,
+      "grad_norm": 2.0430805683135986,
+      "learning_rate": 4.339967506092607e-06,
+      "loss": 2.0192,
+      "step": 613
+    },
+    {
+      "epoch": 0.6728767123287671,
+      "grad_norm": 2.061089277267456,
+      "learning_rate": 4.3379366368805855e-06,
+      "loss": 1.9543,
+      "step": 614
+    },
+    {
+      "epoch": 0.673972602739726,
+      "grad_norm": 2.298022747039795,
+      "learning_rate": 4.335905767668562e-06,
+      "loss": 1.9579,
+      "step": 615
+    },
+    {
+      "epoch": 0.675068493150685,
+      "grad_norm": 2.1418962478637695,
+      "learning_rate": 4.33387489845654e-06,
+      "loss": 1.938,
+      "step": 616
+    },
+    {
+      "epoch": 0.6761643835616439,
+      "grad_norm": 2.1005501747131348,
+      "learning_rate": 4.3318440292445165e-06,
+      "loss": 1.972,
+      "step": 617
+    },
+    {
+      "epoch": 0.6772602739726027,
+      "grad_norm": 2.1665523052215576,
+      "learning_rate": 4.329813160032495e-06,
+      "loss": 1.9535,
+      "step": 618
+    },
+    {
+      "epoch": 0.6783561643835616,
+      "grad_norm": 2.144580841064453,
+      "learning_rate": 4.327782290820471e-06,
+      "loss": 1.9188,
+      "step": 619
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "grad_norm": 1.9855908155441284,
+      "learning_rate": 4.325751421608449e-06,
+      "loss": 1.912,
+      "step": 620
+    },
+    {
+      "epoch": 0.6794520547945205,
+      "eval_loss": 2.118319272994995,
+      "eval_runtime": 36.9677,
+      "eval_samples_per_second": 27.051,
+      "eval_steps_per_second": 1.136,
+      "step": 620
+    },
+    {
+      "epoch": 0.6805479452054795,
+      "grad_norm": 2.395538091659546,
+      "learning_rate": 4.323720552396426e-06,
+      "loss": 1.9886,
+      "step": 621
+    },
+    {
+      "epoch": 0.6816438356164384,
+      "grad_norm": 2.190574884414673,
+      "learning_rate": 4.321689683184404e-06,
+      "loss": 1.9326,
+      "step": 622
+    },
+    {
+      "epoch": 0.6827397260273973,
+      "grad_norm": 2.1826560497283936,
+      "learning_rate": 4.31965881397238e-06,
+      "loss": 1.9278,
+      "step": 623
+    },
+    {
+      "epoch": 0.6838356164383562,
+      "grad_norm": 2.151097536087036,
+      "learning_rate": 4.3176279447603574e-06,
+      "loss": 1.9359,
+      "step": 624
+    },
+    {
+      "epoch": 0.684931506849315,
+      "grad_norm": 2.1830122470855713,
+      "learning_rate": 4.315597075548335e-06,
+      "loss": 1.9728,
+      "step": 625
+    },
+    {
+      "epoch": 0.686027397260274,
+      "grad_norm": 2.203325033187866,
+      "learning_rate": 4.313566206336312e-06,
+      "loss": 1.9638,
+      "step": 626
+    },
+    {
+      "epoch": 0.6871232876712329,
+      "grad_norm": 2.7240309715270996,
+      "learning_rate": 4.311535337124289e-06,
+      "loss": 1.9907,
+      "step": 627
+    },
+    {
+      "epoch": 0.6882191780821918,
+      "grad_norm": 2.133599281311035,
+      "learning_rate": 4.3095044679122665e-06,
+      "loss": 1.9352,
+      "step": 628
+    },
+    {
+      "epoch": 0.6893150684931507,
+      "grad_norm": 2.0318551063537598,
+      "learning_rate": 4.307473598700244e-06,
+      "loss": 1.9432,
+      "step": 629
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "grad_norm": 2.502101182937622,
+      "learning_rate": 4.305442729488221e-06,
+      "loss": 1.9641,
+      "step": 630
+    },
+    {
+      "epoch": 0.6904109589041096,
+      "eval_loss": 2.1158926486968994,
+      "eval_runtime": 36.8818,
+      "eval_samples_per_second": 27.114,
+      "eval_steps_per_second": 1.139,
+      "step": 630
+    },
+    {
+      "epoch": 0.6915068493150685,
+      "grad_norm": 2.3620944023132324,
+      "learning_rate": 4.303411860276198e-06,
+      "loss": 1.8841,
+      "step": 631
+    },
+    {
+      "epoch": 0.6926027397260274,
+      "grad_norm": 2.3139240741729736,
+      "learning_rate": 4.301380991064176e-06,
+      "loss": 1.9896,
+      "step": 632
+    },
+    {
+      "epoch": 0.6936986301369863,
+      "grad_norm": 2.314877510070801,
+      "learning_rate": 4.299350121852153e-06,
+      "loss": 1.9655,
+      "step": 633
+    },
+    {
+      "epoch": 0.6947945205479452,
+      "grad_norm": 2.036371946334839,
+      "learning_rate": 4.29731925264013e-06,
+      "loss": 1.9917,
+      "step": 634
+    },
+    {
+      "epoch": 0.6958904109589041,
+      "grad_norm": 2.1859662532806396,
+      "learning_rate": 4.2952883834281075e-06,
+      "loss": 1.9805,
+      "step": 635
+    },
+    {
+      "epoch": 0.6969863013698631,
+      "grad_norm": 2.2317097187042236,
+      "learning_rate": 4.293257514216085e-06,
+      "loss": 1.9402,
+      "step": 636
+    },
+    {
+      "epoch": 0.6980821917808219,
+      "grad_norm": 2.3249905109405518,
+      "learning_rate": 4.291226645004062e-06,
+      "loss": 1.9013,
+      "step": 637
+    },
+    {
+      "epoch": 0.6991780821917808,
+      "grad_norm": 2.199758768081665,
+      "learning_rate": 4.289195775792039e-06,
+      "loss": 1.9783,
+      "step": 638
+    },
+    {
+      "epoch": 0.7002739726027397,
+      "grad_norm": 2.3598549365997314,
+      "learning_rate": 4.287164906580017e-06,
+      "loss": 1.9936,
+      "step": 639
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "grad_norm": 2.1705522537231445,
+      "learning_rate": 4.285134037367994e-06,
+      "loss": 1.9304,
+      "step": 640
+    },
+    {
+      "epoch": 0.7013698630136986,
+      "eval_loss": 2.1055190563201904,
+      "eval_runtime": 36.8673,
+      "eval_samples_per_second": 27.124,
+      "eval_steps_per_second": 1.139,
+      "step": 640
+    },
+    {
+      "epoch": 0.7024657534246576,
+      "grad_norm": 2.3253612518310547,
+      "learning_rate": 4.283103168155971e-06,
+      "loss": 1.9196,
+      "step": 641
+    },
+    {
+      "epoch": 0.7035616438356165,
+      "grad_norm": 2.3054087162017822,
+      "learning_rate": 4.281072298943948e-06,
+      "loss": 1.9053,
+      "step": 642
+    },
+    {
+      "epoch": 0.7046575342465753,
+      "grad_norm": 2.262057304382324,
+      "learning_rate": 4.279041429731926e-06,
+      "loss": 1.8999,
+      "step": 643
+    },
+    {
+      "epoch": 0.7057534246575342,
+      "grad_norm": 2.421391248703003,
+      "learning_rate": 4.277010560519903e-06,
+      "loss": 1.9438,
+      "step": 644
+    },
+    {
+      "epoch": 0.7068493150684931,
+      "grad_norm": 2.2841498851776123,
+      "learning_rate": 4.27497969130788e-06,
+      "loss": 1.9187,
+      "step": 645
+    },
+    {
+      "epoch": 0.707945205479452,
+      "grad_norm": 2.3226938247680664,
+      "learning_rate": 4.2729488220958575e-06,
+      "loss": 1.9275,
+      "step": 646
+    },
+    {
+      "epoch": 0.709041095890411,
+      "grad_norm": 2.2306301593780518,
+      "learning_rate": 4.270917952883835e-06,
+      "loss": 1.9684,
+      "step": 647
+    },
+    {
+      "epoch": 0.7101369863013699,
+      "grad_norm": 2.0551652908325195,
+      "learning_rate": 4.268887083671812e-06,
+      "loss": 1.9849,
+      "step": 648
+    },
+    {
+      "epoch": 0.7112328767123288,
+      "grad_norm": 2.2952210903167725,
+      "learning_rate": 4.266856214459789e-06,
+      "loss": 1.9026,
+      "step": 649
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "grad_norm": 2.1630797386169434,
+      "learning_rate": 4.264825345247766e-06,
+      "loss": 2.0184,
+      "step": 650
+    },
+    {
+      "epoch": 0.7123287671232876,
+      "eval_loss": 2.1037871837615967,
+      "eval_runtime": 36.8221,
+      "eval_samples_per_second": 27.158,
+      "eval_steps_per_second": 1.141,
+      "step": 650
+    },
+    {
+      "epoch": 0.7134246575342466,
+      "grad_norm": 1.964150309562683,
+      "learning_rate": 4.262794476035744e-06,
+      "loss": 1.8932,
+      "step": 651
+    },
+    {
+      "epoch": 0.7145205479452055,
+      "grad_norm": 2.1474685668945312,
+      "learning_rate": 4.26076360682372e-06,
+      "loss": 1.9383,
+      "step": 652
+    },
+    {
+      "epoch": 0.7156164383561644,
+      "grad_norm": 2.199824094772339,
+      "learning_rate": 4.2587327376116984e-06,
+      "loss": 1.8599,
+      "step": 653
+    },
+    {
+      "epoch": 0.7167123287671233,
+      "grad_norm": 2.12156343460083,
+      "learning_rate": 4.256701868399675e-06,
+      "loss": 1.9162,
+      "step": 654
+    },
+    {
+      "epoch": 0.7178082191780822,
+      "grad_norm": 2.3832154273986816,
+      "learning_rate": 4.254670999187653e-06,
+      "loss": 1.9613,
+      "step": 655
+    },
+    {
+      "epoch": 0.718904109589041,
+      "grad_norm": 2.063906192779541,
+      "learning_rate": 4.252640129975629e-06,
+      "loss": 1.9421,
+      "step": 656
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 2.184842586517334,
+      "learning_rate": 4.2506092607636075e-06,
+      "loss": 1.9471,
+      "step": 657
+    },
+    {
+      "epoch": 0.7210958904109589,
+      "grad_norm": 2.1340036392211914,
+      "learning_rate": 4.248578391551584e-06,
+      "loss": 1.931,
+      "step": 658
+    },
+    {
+      "epoch": 0.7221917808219178,
+      "grad_norm": 2.117544651031494,
+      "learning_rate": 4.246547522339562e-06,
+      "loss": 1.8698,
+      "step": 659
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "grad_norm": 2.308344841003418,
+      "learning_rate": 4.2445166531275385e-06,
+      "loss": 1.9217,
+      "step": 660
+    },
+    {
+      "epoch": 0.7232876712328767,
+      "eval_loss": 2.1086323261260986,
+      "eval_runtime": 36.8082,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 660
+    },
+    {
+      "epoch": 0.7243835616438357,
+      "grad_norm": 2.2239017486572266,
+      "learning_rate": 4.242485783915517e-06,
+      "loss": 1.9286,
+      "step": 661
+    },
+    {
+      "epoch": 0.7254794520547945,
+      "grad_norm": 2.263697624206543,
+      "learning_rate": 4.240454914703493e-06,
+      "loss": 1.8821,
+      "step": 662
+    },
+    {
+      "epoch": 0.7265753424657534,
+      "grad_norm": 2.4305059909820557,
+      "learning_rate": 4.23842404549147e-06,
+      "loss": 1.8996,
+      "step": 663
+    },
+    {
+      "epoch": 0.7276712328767123,
+      "grad_norm": 2.2333695888519287,
+      "learning_rate": 4.236393176279448e-06,
+      "loss": 1.8906,
+      "step": 664
+    },
+    {
+      "epoch": 0.7287671232876712,
+      "grad_norm": 2.160982608795166,
+      "learning_rate": 4.234362307067425e-06,
+      "loss": 1.879,
+      "step": 665
+    },
+    {
+      "epoch": 0.7298630136986302,
+      "grad_norm": 2.3707094192504883,
+      "learning_rate": 4.232331437855402e-06,
+      "loss": 1.9577,
+      "step": 666
+    },
+    {
+      "epoch": 0.7309589041095891,
+      "grad_norm": 2.0502827167510986,
+      "learning_rate": 4.2303005686433795e-06,
+      "loss": 1.9801,
+      "step": 667
+    },
+    {
+      "epoch": 0.7320547945205479,
+      "grad_norm": 2.0648176670074463,
+      "learning_rate": 4.228269699431357e-06,
+      "loss": 1.978,
+      "step": 668
+    },
+    {
+      "epoch": 0.7331506849315068,
+      "grad_norm": 2.346316337585449,
+      "learning_rate": 4.226238830219334e-06,
+      "loss": 1.9387,
+      "step": 669
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "grad_norm": 1.9734236001968384,
+      "learning_rate": 4.224207961007311e-06,
+      "loss": 1.9396,
+      "step": 670
+    },
+    {
+      "epoch": 0.7342465753424657,
+      "eval_loss": 2.1104705333709717,
+      "eval_runtime": 36.8144,
+      "eval_samples_per_second": 27.163,
+      "eval_steps_per_second": 1.141,
+      "step": 670
+    },
+    {
+      "epoch": 0.7353424657534247,
+      "grad_norm": 2.3170549869537354,
+      "learning_rate": 4.2221770917952886e-06,
+      "loss": 1.9398,
+      "step": 671
+    },
+    {
+      "epoch": 0.7364383561643836,
+      "grad_norm": 2.0437264442443848,
+      "learning_rate": 4.220146222583266e-06,
+      "loss": 1.9284,
+      "step": 672
+    },
+    {
+      "epoch": 0.7375342465753425,
+      "grad_norm": 2.0808510780334473,
+      "learning_rate": 4.218115353371243e-06,
+      "loss": 1.9425,
+      "step": 673
+    },
+    {
+      "epoch": 0.7386301369863014,
+      "grad_norm": 2.0854742527008057,
+      "learning_rate": 4.21608448415922e-06,
+      "loss": 1.9358,
+      "step": 674
+    },
+    {
+      "epoch": 0.7397260273972602,
+      "grad_norm": 2.131622552871704,
+      "learning_rate": 4.214053614947198e-06,
+      "loss": 1.9048,
+      "step": 675
+    },
+    {
+      "epoch": 0.7408219178082192,
+      "grad_norm": 2.276359796524048,
+      "learning_rate": 4.212022745735175e-06,
+      "loss": 1.9871,
+      "step": 676
+    },
+    {
+      "epoch": 0.7419178082191781,
+      "grad_norm": 2.2194812297821045,
+      "learning_rate": 4.209991876523152e-06,
+      "loss": 1.882,
+      "step": 677
+    },
+    {
+      "epoch": 0.743013698630137,
+      "grad_norm": 2.181891679763794,
+      "learning_rate": 4.2079610073111295e-06,
+      "loss": 1.9546,
+      "step": 678
+    },
+    {
+      "epoch": 0.7441095890410959,
+      "grad_norm": 2.0956673622131348,
+      "learning_rate": 4.205930138099107e-06,
+      "loss": 1.8836,
+      "step": 679
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "grad_norm": 2.057894229888916,
+      "learning_rate": 4.203899268887084e-06,
+      "loss": 1.9842,
+      "step": 680
+    },
+    {
+      "epoch": 0.7452054794520548,
+      "eval_loss": 2.104182720184326,
+      "eval_runtime": 36.8189,
+      "eval_samples_per_second": 27.16,
+      "eval_steps_per_second": 1.141,
+      "step": 680
+    },
+    {
+      "epoch": 0.7463013698630137,
+      "grad_norm": 2.0965662002563477,
+      "learning_rate": 4.201868399675061e-06,
+      "loss": 1.9378,
+      "step": 681
+    },
+    {
+      "epoch": 0.7473972602739726,
+      "grad_norm": 2.1843457221984863,
+      "learning_rate": 4.199837530463039e-06,
+      "loss": 1.9784,
+      "step": 682
+    },
+    {
+      "epoch": 0.7484931506849315,
+      "grad_norm": 2.2044246196746826,
+      "learning_rate": 4.197806661251016e-06,
+      "loss": 1.8642,
+      "step": 683
+    },
+    {
+      "epoch": 0.7495890410958904,
+      "grad_norm": 2.202115058898926,
+      "learning_rate": 4.195775792038993e-06,
+      "loss": 1.929,
+      "step": 684
+    },
+    {
+      "epoch": 0.7506849315068493,
+      "grad_norm": 2.3671462535858154,
+      "learning_rate": 4.19374492282697e-06,
+      "loss": 1.9172,
+      "step": 685
+    },
+    {
+      "epoch": 0.7517808219178083,
+      "grad_norm": 2.0602803230285645,
+      "learning_rate": 4.191714053614948e-06,
+      "loss": 1.8503,
+      "step": 686
+    },
+    {
+      "epoch": 0.7528767123287671,
+      "grad_norm": 2.12145733833313,
+      "learning_rate": 4.189683184402924e-06,
+      "loss": 1.9511,
+      "step": 687
+    },
+    {
+      "epoch": 0.753972602739726,
+      "grad_norm": 2.1049914360046387,
+      "learning_rate": 4.187652315190902e-06,
+      "loss": 1.9043,
+      "step": 688
+    },
+    {
+      "epoch": 0.7550684931506849,
+      "grad_norm": 2.235497236251831,
+      "learning_rate": 4.185621445978879e-06,
+      "loss": 1.9784,
+      "step": 689
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "grad_norm": 2.173222303390503,
+      "learning_rate": 4.183590576766857e-06,
+      "loss": 1.9052,
+      "step": 690
+    },
+    {
+      "epoch": 0.7561643835616438,
+      "eval_loss": 2.0982251167297363,
+      "eval_runtime": 37.1446,
+      "eval_samples_per_second": 26.922,
+      "eval_steps_per_second": 1.131,
+      "step": 690
+    },
+    {
+      "epoch": 0.7572602739726028,
+      "grad_norm": 2.0693562030792236,
+      "learning_rate": 4.181559707554833e-06,
+      "loss": 1.9223,
+      "step": 691
+    },
+    {
+      "epoch": 0.7583561643835617,
+      "grad_norm": 2.209419012069702,
+      "learning_rate": 4.179528838342811e-06,
+      "loss": 1.9709,
+      "step": 692
+    },
+    {
+      "epoch": 0.7594520547945205,
+      "grad_norm": 2.066041946411133,
+      "learning_rate": 4.177497969130788e-06,
+      "loss": 1.889,
+      "step": 693
+    },
+    {
+      "epoch": 0.7605479452054794,
+      "grad_norm": 2.4114325046539307,
+      "learning_rate": 4.175467099918766e-06,
+      "loss": 1.9059,
+      "step": 694
+    },
+    {
+      "epoch": 0.7616438356164383,
+      "grad_norm": 2.0185043811798096,
+      "learning_rate": 4.173436230706742e-06,
+      "loss": 1.9877,
+      "step": 695
+    },
+    {
+      "epoch": 0.7627397260273973,
+      "grad_norm": 2.23392915725708,
+      "learning_rate": 4.1714053614947205e-06,
+      "loss": 1.9199,
+      "step": 696
+    },
+    {
+      "epoch": 0.7638356164383562,
+      "grad_norm": 2.2158737182617188,
+      "learning_rate": 4.169374492282697e-06,
+      "loss": 1.9001,
+      "step": 697
+    },
+    {
+      "epoch": 0.7649315068493151,
+      "grad_norm": 2.1511690616607666,
+      "learning_rate": 4.167343623070675e-06,
+      "loss": 1.9166,
+      "step": 698
+    },
+    {
+      "epoch": 0.766027397260274,
+      "grad_norm": 2.374741554260254,
+      "learning_rate": 4.165312753858652e-06,
+      "loss": 1.9684,
+      "step": 699
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "grad_norm": 2.0409998893737793,
+      "learning_rate": 4.163281884646629e-06,
+      "loss": 1.9177,
+      "step": 700
+    },
+    {
+      "epoch": 0.7671232876712328,
+      "eval_loss": 2.096895933151245,
+      "eval_runtime": 37.066,
+      "eval_samples_per_second": 26.979,
+      "eval_steps_per_second": 1.133,
+      "step": 700
+    },
+    {
+      "epoch": 0.7682191780821918,
+      "grad_norm": 2.3173422813415527,
+      "learning_rate": 4.161251015434607e-06,
+      "loss": 1.9071,
+      "step": 701
+    },
+    {
+      "epoch": 0.7693150684931507,
+      "grad_norm": 2.077842950820923,
+      "learning_rate": 4.159220146222583e-06,
+      "loss": 1.935,
+      "step": 702
+    },
+    {
+      "epoch": 0.7704109589041096,
+      "grad_norm": 2.0518617630004883,
+      "learning_rate": 4.157189277010561e-06,
+      "loss": 1.8942,
+      "step": 703
+    },
+    {
+      "epoch": 0.7715068493150685,
+      "grad_norm": 1.9950882196426392,
+      "learning_rate": 4.155158407798538e-06,
+      "loss": 1.917,
+      "step": 704
+    },
+    {
+      "epoch": 0.7726027397260274,
+      "grad_norm": 2.3222219944000244,
+      "learning_rate": 4.153127538586516e-06,
+      "loss": 1.9551,
+      "step": 705
+    },
+    {
+      "epoch": 0.7736986301369863,
+      "grad_norm": 2.129333972930908,
+      "learning_rate": 4.151096669374492e-06,
+      "loss": 1.9204,
+      "step": 706
+    },
+    {
+      "epoch": 0.7747945205479452,
+      "grad_norm": 2.0357553958892822,
+      "learning_rate": 4.1490658001624705e-06,
+      "loss": 1.9399,
+      "step": 707
+    },
+    {
+      "epoch": 0.7758904109589041,
+      "grad_norm": 2.233694553375244,
+      "learning_rate": 4.147034930950447e-06,
+      "loss": 1.9182,
+      "step": 708
+    },
+    {
+      "epoch": 0.776986301369863,
+      "grad_norm": 2.1391239166259766,
+      "learning_rate": 4.145004061738425e-06,
+      "loss": 1.9157,
+      "step": 709
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "grad_norm": 2.162320613861084,
+      "learning_rate": 4.1429731925264015e-06,
+      "loss": 1.9115,
+      "step": 710
+    },
+    {
+      "epoch": 0.7780821917808219,
+      "eval_loss": 2.0954039096832275,
+      "eval_runtime": 36.9862,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 710
+    },
+    {
+      "epoch": 0.7791780821917809,
+      "grad_norm": 2.2382194995880127,
+      "learning_rate": 4.140942323314379e-06,
+      "loss": 1.8428,
+      "step": 711
+    },
+    {
+      "epoch": 0.7802739726027397,
+      "grad_norm": 2.3286185264587402,
+      "learning_rate": 4.138911454102356e-06,
+      "loss": 1.8906,
+      "step": 712
+    },
+    {
+      "epoch": 0.7813698630136986,
+      "grad_norm": 2.182708501815796,
+      "learning_rate": 4.136880584890333e-06,
+      "loss": 1.9132,
+      "step": 713
+    },
+    {
+      "epoch": 0.7824657534246575,
+      "grad_norm": 1.9978337287902832,
+      "learning_rate": 4.1348497156783106e-06,
+      "loss": 1.9433,
+      "step": 714
+    },
+    {
+      "epoch": 0.7835616438356164,
+      "grad_norm": 2.3092565536499023,
+      "learning_rate": 4.132818846466288e-06,
+      "loss": 1.9374,
+      "step": 715
+    },
+    {
+      "epoch": 0.7846575342465754,
+      "grad_norm": 2.0351054668426514,
+      "learning_rate": 4.130787977254265e-06,
+      "loss": 1.9361,
+      "step": 716
+    },
+    {
+      "epoch": 0.7857534246575343,
+      "grad_norm": 2.14770770072937,
+      "learning_rate": 4.128757108042242e-06,
+      "loss": 1.9553,
+      "step": 717
+    },
+    {
+      "epoch": 0.7868493150684932,
+      "grad_norm": 2.0613555908203125,
+      "learning_rate": 4.12672623883022e-06,
+      "loss": 1.9506,
+      "step": 718
+    },
+    {
+      "epoch": 0.787945205479452,
+      "grad_norm": 2.1312081813812256,
+      "learning_rate": 4.124695369618197e-06,
+      "loss": 2.0079,
+      "step": 719
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "grad_norm": 2.088552474975586,
+      "learning_rate": 4.122664500406174e-06,
+      "loss": 1.9166,
+      "step": 720
+    },
+    {
+      "epoch": 0.7890410958904109,
+      "eval_loss": 2.093032121658325,
+      "eval_runtime": 36.8821,
+      "eval_samples_per_second": 27.113,
+      "eval_steps_per_second": 1.139,
+      "step": 720
+    },
+    {
+      "epoch": 0.7901369863013699,
+      "grad_norm": 2.1748929023742676,
+      "learning_rate": 4.1206336311941515e-06,
+      "loss": 1.9212,
+      "step": 721
+    },
+    {
+      "epoch": 0.7912328767123288,
+      "grad_norm": 2.1492016315460205,
+      "learning_rate": 4.118602761982129e-06,
+      "loss": 1.9452,
+      "step": 722
+    },
+    {
+      "epoch": 0.7923287671232877,
+      "grad_norm": 2.3364970684051514,
+      "learning_rate": 4.116571892770106e-06,
+      "loss": 1.8821,
+      "step": 723
+    },
+    {
+      "epoch": 0.7934246575342466,
+      "grad_norm": 2.318488597869873,
+      "learning_rate": 4.114541023558083e-06,
+      "loss": 1.8597,
+      "step": 724
+    },
+    {
+      "epoch": 0.7945205479452054,
+      "grad_norm": 2.153841257095337,
+      "learning_rate": 4.112510154346061e-06,
+      "loss": 1.9843,
+      "step": 725
+    },
+    {
+      "epoch": 0.7956164383561644,
+      "grad_norm": 2.3464651107788086,
+      "learning_rate": 4.110479285134038e-06,
+      "loss": 1.9105,
+      "step": 726
+    },
+    {
+      "epoch": 0.7967123287671233,
+      "grad_norm": 2.34413743019104,
+      "learning_rate": 4.108448415922015e-06,
+      "loss": 1.9285,
+      "step": 727
+    },
+    {
+      "epoch": 0.7978082191780822,
+      "grad_norm": 2.0822196006774902,
+      "learning_rate": 4.1064175467099924e-06,
+      "loss": 1.9536,
+      "step": 728
+    },
+    {
+      "epoch": 0.7989041095890411,
+      "grad_norm": 2.1978065967559814,
+      "learning_rate": 4.10438667749797e-06,
+      "loss": 1.9487,
+      "step": 729
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 2.4218783378601074,
+      "learning_rate": 4.102355808285947e-06,
+      "loss": 1.8113,
+      "step": 730
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 2.0963101387023926,
+      "eval_runtime": 36.8853,
+      "eval_samples_per_second": 27.111,
+      "eval_steps_per_second": 1.139,
+      "step": 730
+    },
+    {
+      "epoch": 0.8010958904109589,
+      "grad_norm": 2.3887739181518555,
+      "learning_rate": 4.100324939073924e-06,
+      "loss": 2.0024,
+      "step": 731
+    },
+    {
+      "epoch": 0.8021917808219178,
+      "grad_norm": 2.296579360961914,
+      "learning_rate": 4.0982940698619015e-06,
+      "loss": 1.9092,
+      "step": 732
+    },
+    {
+      "epoch": 0.8032876712328767,
+      "grad_norm": 2.3881773948669434,
+      "learning_rate": 4.096263200649879e-06,
+      "loss": 1.9112,
+      "step": 733
+    },
+    {
+      "epoch": 0.8043835616438356,
+      "grad_norm": 2.40120530128479,
+      "learning_rate": 4.094232331437856e-06,
+      "loss": 1.9191,
+      "step": 734
+    },
+    {
+      "epoch": 0.8054794520547945,
+      "grad_norm": 2.184706211090088,
+      "learning_rate": 4.092201462225833e-06,
+      "loss": 1.8857,
+      "step": 735
+    },
+    {
+      "epoch": 0.8065753424657535,
+      "grad_norm": 2.3338732719421387,
+      "learning_rate": 4.090170593013811e-06,
+      "loss": 1.8889,
+      "step": 736
+    },
+    {
+      "epoch": 0.8076712328767123,
+      "grad_norm": 2.4065468311309814,
+      "learning_rate": 4.088139723801787e-06,
+      "loss": 1.9309,
+      "step": 737
+    },
+    {
+      "epoch": 0.8087671232876712,
+      "grad_norm": 2.044992446899414,
+      "learning_rate": 4.086108854589765e-06,
+      "loss": 1.8727,
+      "step": 738
+    },
+    {
+      "epoch": 0.8098630136986301,
+      "grad_norm": 2.23860239982605,
+      "learning_rate": 4.084077985377742e-06,
+      "loss": 1.8936,
+      "step": 739
+    },
+    {
+      "epoch": 0.810958904109589,
+      "grad_norm": 2.1324410438537598,
+      "learning_rate": 4.08204711616572e-06,
+      "loss": 1.9035,
+      "step": 740
+    },
+    {
+      "epoch": 0.810958904109589,
+      "eval_loss": 2.092885732650757,
+      "eval_runtime": 36.7991,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 740
+    },
+    {
+      "epoch": 0.812054794520548,
+      "grad_norm": 2.337340831756592,
+      "learning_rate": 4.080016246953696e-06,
+      "loss": 1.9747,
+      "step": 741
+    },
+    {
+      "epoch": 0.8131506849315069,
+      "grad_norm": 2.262801170349121,
+      "learning_rate": 4.077985377741674e-06,
+      "loss": 1.8817,
+      "step": 742
+    },
+    {
+      "epoch": 0.8142465753424658,
+      "grad_norm": 2.0494258403778076,
+      "learning_rate": 4.075954508529651e-06,
+      "loss": 1.9493,
+      "step": 743
+    },
+    {
+      "epoch": 0.8153424657534246,
+      "grad_norm": 2.085829019546509,
+      "learning_rate": 4.073923639317629e-06,
+      "loss": 1.9611,
+      "step": 744
+    },
+    {
+      "epoch": 0.8164383561643835,
+      "grad_norm": 2.105806827545166,
+      "learning_rate": 4.071892770105605e-06,
+      "loss": 1.9779,
+      "step": 745
+    },
+    {
+      "epoch": 0.8175342465753425,
+      "grad_norm": 2.2664732933044434,
+      "learning_rate": 4.069861900893583e-06,
+      "loss": 1.9009,
+      "step": 746
+    },
+    {
+      "epoch": 0.8186301369863014,
+      "grad_norm": 1.9899768829345703,
+      "learning_rate": 4.06783103168156e-06,
+      "loss": 1.8585,
+      "step": 747
+    },
+    {
+      "epoch": 0.8197260273972603,
+      "grad_norm": 2.242236614227295,
+      "learning_rate": 4.065800162469538e-06,
+      "loss": 1.8904,
+      "step": 748
+    },
+    {
+      "epoch": 0.8208219178082192,
+      "grad_norm": 2.4481139183044434,
+      "learning_rate": 4.063769293257514e-06,
+      "loss": 1.8922,
+      "step": 749
+    },
+    {
+      "epoch": 0.821917808219178,
+      "grad_norm": 2.231722354888916,
+      "learning_rate": 4.061738424045492e-06,
+      "loss": 1.9292,
+      "step": 750
+    },
+    {
+      "epoch": 0.821917808219178,
+      "eval_loss": 2.0909852981567383,
+      "eval_runtime": 36.8255,
+      "eval_samples_per_second": 27.155,
+      "eval_steps_per_second": 1.141,
+      "step": 750
+    },
+    {
+      "epoch": 0.823013698630137,
+      "grad_norm": 2.323070764541626,
+      "learning_rate": 4.059707554833469e-06,
+      "loss": 1.9053,
+      "step": 751
+    },
+    {
+      "epoch": 0.8241095890410959,
+      "grad_norm": 2.137535572052002,
+      "learning_rate": 4.057676685621446e-06,
+      "loss": 1.9272,
+      "step": 752
+    },
+    {
+      "epoch": 0.8252054794520548,
+      "grad_norm": 2.2087643146514893,
+      "learning_rate": 4.0556458164094235e-06,
+      "loss": 1.9541,
+      "step": 753
+    },
+    {
+      "epoch": 0.8263013698630137,
+      "grad_norm": 2.04811692237854,
+      "learning_rate": 4.053614947197401e-06,
+      "loss": 1.922,
+      "step": 754
+    },
+    {
+      "epoch": 0.8273972602739726,
+      "grad_norm": 2.062622308731079,
+      "learning_rate": 4.051584077985378e-06,
+      "loss": 1.8388,
+      "step": 755
+    },
+    {
+      "epoch": 0.8284931506849315,
+      "grad_norm": 2.14631986618042,
+      "learning_rate": 4.049553208773355e-06,
+      "loss": 1.9153,
+      "step": 756
+    },
+    {
+      "epoch": 0.8295890410958904,
+      "grad_norm": 2.353529214859009,
+      "learning_rate": 4.047522339561333e-06,
+      "loss": 1.9623,
+      "step": 757
+    },
+    {
+      "epoch": 0.8306849315068493,
+      "grad_norm": 2.122342348098755,
+      "learning_rate": 4.04549147034931e-06,
+      "loss": 1.9095,
+      "step": 758
+    },
+    {
+      "epoch": 0.8317808219178082,
+      "grad_norm": 2.2540230751037598,
+      "learning_rate": 4.043460601137287e-06,
+      "loss": 1.9068,
+      "step": 759
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "grad_norm": 2.164384365081787,
+      "learning_rate": 4.041429731925264e-06,
+      "loss": 2.0239,
+      "step": 760
+    },
+    {
+      "epoch": 0.8328767123287671,
+      "eval_loss": 2.0904648303985596,
+      "eval_runtime": 36.7617,
+      "eval_samples_per_second": 27.202,
+      "eval_steps_per_second": 1.142,
+      "step": 760
+    },
+    {
+      "epoch": 0.8339726027397261,
+      "grad_norm": 2.2487905025482178,
+      "learning_rate": 4.039398862713242e-06,
+      "loss": 1.9109,
+      "step": 761
+    },
+    {
+      "epoch": 0.8350684931506849,
+      "grad_norm": 2.0947930812835693,
+      "learning_rate": 4.037367993501219e-06,
+      "loss": 1.9133,
+      "step": 762
+    },
+    {
+      "epoch": 0.8361643835616438,
+      "grad_norm": 2.2259862422943115,
+      "learning_rate": 4.035337124289196e-06,
+      "loss": 1.9821,
+      "step": 763
+    },
+    {
+      "epoch": 0.8372602739726027,
+      "grad_norm": 2.0616455078125,
+      "learning_rate": 4.0333062550771735e-06,
+      "loss": 1.9037,
+      "step": 764
+    },
+    {
+      "epoch": 0.8383561643835616,
+      "grad_norm": 2.2704474925994873,
+      "learning_rate": 4.031275385865151e-06,
+      "loss": 1.8664,
+      "step": 765
+    },
+    {
+      "epoch": 0.8394520547945206,
+      "grad_norm": 2.0059332847595215,
+      "learning_rate": 4.029244516653128e-06,
+      "loss": 1.908,
+      "step": 766
+    },
+    {
+      "epoch": 0.8405479452054795,
+      "grad_norm": 2.0462095737457275,
+      "learning_rate": 4.027213647441105e-06,
+      "loss": 1.9397,
+      "step": 767
+    },
+    {
+      "epoch": 0.8416438356164384,
+      "grad_norm": 2.150207996368408,
+      "learning_rate": 4.025182778229083e-06,
+      "loss": 1.8505,
+      "step": 768
+    },
+    {
+      "epoch": 0.8427397260273972,
+      "grad_norm": 2.174863815307617,
+      "learning_rate": 4.02315190901706e-06,
+      "loss": 1.8895,
+      "step": 769
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "grad_norm": 2.1907355785369873,
+      "learning_rate": 4.021121039805037e-06,
+      "loss": 1.9156,
+      "step": 770
+    },
+    {
+      "epoch": 0.8438356164383561,
+      "eval_loss": 2.0900955200195312,
+      "eval_runtime": 36.8004,
+      "eval_samples_per_second": 27.174,
+      "eval_steps_per_second": 1.141,
+      "step": 770
+    },
+    {
+      "epoch": 0.8449315068493151,
+      "grad_norm": 2.2831835746765137,
+      "learning_rate": 4.0190901705930144e-06,
+      "loss": 1.9115,
+      "step": 771
+    },
+    {
+      "epoch": 0.846027397260274,
+      "grad_norm": 2.039520263671875,
+      "learning_rate": 4.017059301380992e-06,
+      "loss": 2.0047,
+      "step": 772
+    },
+    {
+      "epoch": 0.8471232876712329,
+      "grad_norm": 2.147209405899048,
+      "learning_rate": 4.015028432168969e-06,
+      "loss": 1.8845,
+      "step": 773
+    },
+    {
+      "epoch": 0.8482191780821918,
+      "grad_norm": 2.168997287750244,
+      "learning_rate": 4.012997562956945e-06,
+      "loss": 1.9606,
+      "step": 774
+    },
+    {
+      "epoch": 0.8493150684931506,
+      "grad_norm": 2.038966417312622,
+      "learning_rate": 4.0109666937449235e-06,
+      "loss": 1.8734,
+      "step": 775
+    },
+    {
+      "epoch": 0.8504109589041096,
+      "grad_norm": 2.0531768798828125,
+      "learning_rate": 4.0089358245329e-06,
+      "loss": 1.889,
+      "step": 776
+    },
+    {
+      "epoch": 0.8515068493150685,
+      "grad_norm": 2.348008155822754,
+      "learning_rate": 4.006904955320878e-06,
+      "loss": 1.8972,
+      "step": 777
+    },
+    {
+      "epoch": 0.8526027397260274,
+      "grad_norm": 2.1647253036499023,
+      "learning_rate": 4.0048740861088545e-06,
+      "loss": 1.8652,
+      "step": 778
+    },
+    {
+      "epoch": 0.8536986301369863,
+      "grad_norm": 2.1971640586853027,
+      "learning_rate": 4.002843216896833e-06,
+      "loss": 1.9495,
+      "step": 779
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "grad_norm": 2.169632911682129,
+      "learning_rate": 4.000812347684809e-06,
+      "loss": 1.9307,
+      "step": 780
+    },
+    {
+      "epoch": 0.8547945205479452,
+      "eval_loss": 2.087860107421875,
+      "eval_runtime": 36.7982,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 780
+    },
+    {
+      "epoch": 0.8558904109589041,
+      "grad_norm": 2.4424400329589844,
+      "learning_rate": 3.998781478472787e-06,
+      "loss": 1.8849,
+      "step": 781
+    },
+    {
+      "epoch": 0.856986301369863,
+      "grad_norm": 2.2226672172546387,
+      "learning_rate": 3.996750609260764e-06,
+      "loss": 1.8975,
+      "step": 782
+    },
+    {
+      "epoch": 0.8580821917808219,
+      "grad_norm": 2.1898491382598877,
+      "learning_rate": 3.994719740048742e-06,
+      "loss": 1.8751,
+      "step": 783
+    },
+    {
+      "epoch": 0.8591780821917808,
+      "grad_norm": 2.466963052749634,
+      "learning_rate": 3.992688870836718e-06,
+      "loss": 2.0155,
+      "step": 784
+    },
+    {
+      "epoch": 0.8602739726027397,
+      "grad_norm": 2.184039831161499,
+      "learning_rate": 3.990658001624696e-06,
+      "loss": 1.9133,
+      "step": 785
+    },
+    {
+      "epoch": 0.8613698630136987,
+      "grad_norm": 2.089228868484497,
+      "learning_rate": 3.988627132412673e-06,
+      "loss": 1.8707,
+      "step": 786
+    },
+    {
+      "epoch": 0.8624657534246576,
+      "grad_norm": 2.0700666904449463,
+      "learning_rate": 3.98659626320065e-06,
+      "loss": 1.9487,
+      "step": 787
+    },
+    {
+      "epoch": 0.8635616438356164,
+      "grad_norm": 2.1495554447174072,
+      "learning_rate": 3.984565393988627e-06,
+      "loss": 1.9037,
+      "step": 788
+    },
+    {
+      "epoch": 0.8646575342465753,
+      "grad_norm": 2.0916988849639893,
+      "learning_rate": 3.9825345247766046e-06,
+      "loss": 1.9096,
+      "step": 789
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "grad_norm": 2.072387933731079,
+      "learning_rate": 3.980503655564582e-06,
+      "loss": 1.9139,
+      "step": 790
+    },
+    {
+      "epoch": 0.8657534246575342,
+      "eval_loss": 2.0876142978668213,
+      "eval_runtime": 37.1068,
+      "eval_samples_per_second": 26.949,
+      "eval_steps_per_second": 1.132,
+      "step": 790
+    },
+    {
+      "epoch": 0.8668493150684932,
+      "grad_norm": 2.1206343173980713,
+      "learning_rate": 3.978472786352559e-06,
+      "loss": 1.9331,
+      "step": 791
+    },
+    {
+      "epoch": 0.8679452054794521,
+      "grad_norm": 2.20263934135437,
+      "learning_rate": 3.976441917140536e-06,
+      "loss": 1.9297,
+      "step": 792
+    },
+    {
+      "epoch": 0.869041095890411,
+      "grad_norm": 2.027554750442505,
+      "learning_rate": 3.974411047928514e-06,
+      "loss": 1.919,
+      "step": 793
+    },
+    {
+      "epoch": 0.8701369863013698,
+      "grad_norm": 2.0291616916656494,
+      "learning_rate": 3.972380178716491e-06,
+      "loss": 1.9682,
+      "step": 794
+    },
+    {
+      "epoch": 0.8712328767123287,
+      "grad_norm": 2.2719762325286865,
+      "learning_rate": 3.970349309504468e-06,
+      "loss": 1.9407,
+      "step": 795
+    },
+    {
+      "epoch": 0.8723287671232877,
+      "grad_norm": 2.0452375411987305,
+      "learning_rate": 3.9683184402924455e-06,
+      "loss": 1.9487,
+      "step": 796
+    },
+    {
+      "epoch": 0.8734246575342466,
+      "grad_norm": 2.170441150665283,
+      "learning_rate": 3.966287571080423e-06,
+      "loss": 1.8651,
+      "step": 797
+    },
+    {
+      "epoch": 0.8745205479452055,
+      "grad_norm": 2.3370516300201416,
+      "learning_rate": 3.9642567018684e-06,
+      "loss": 1.8599,
+      "step": 798
+    },
+    {
+      "epoch": 0.8756164383561644,
+      "grad_norm": 2.336737632751465,
+      "learning_rate": 3.962225832656377e-06,
+      "loss": 1.9304,
+      "step": 799
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "grad_norm": 1.9708369970321655,
+      "learning_rate": 3.960194963444355e-06,
+      "loss": 1.9125,
+      "step": 800
+    },
+    {
+      "epoch": 0.8767123287671232,
+      "eval_loss": 2.0812273025512695,
+      "eval_runtime": 37.0752,
+      "eval_samples_per_second": 26.972,
+      "eval_steps_per_second": 1.133,
+      "step": 800
+    },
+    {
+      "epoch": 0.8778082191780822,
+      "grad_norm": 2.154219627380371,
+      "learning_rate": 3.958164094232332e-06,
+      "loss": 1.8304,
+      "step": 801
+    },
+    {
+      "epoch": 0.8789041095890411,
+      "grad_norm": 2.237786054611206,
+      "learning_rate": 3.956133225020309e-06,
+      "loss": 1.9147,
+      "step": 802
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 2.07143235206604,
+      "learning_rate": 3.954102355808286e-06,
+      "loss": 1.9128,
+      "step": 803
+    },
+    {
+      "epoch": 0.8810958904109589,
+      "grad_norm": 2.2780392169952393,
+      "learning_rate": 3.952071486596264e-06,
+      "loss": 1.9145,
+      "step": 804
+    },
+    {
+      "epoch": 0.8821917808219178,
+      "grad_norm": 2.2396292686462402,
+      "learning_rate": 3.950040617384241e-06,
+      "loss": 1.9049,
+      "step": 805
+    },
+    {
+      "epoch": 0.8832876712328767,
+      "grad_norm": 2.426511287689209,
+      "learning_rate": 3.948009748172218e-06,
+      "loss": 1.8984,
+      "step": 806
+    },
+    {
+      "epoch": 0.8843835616438356,
+      "grad_norm": 2.1874325275421143,
+      "learning_rate": 3.9459788789601955e-06,
+      "loss": 1.8888,
+      "step": 807
+    },
+    {
+      "epoch": 0.8854794520547945,
+      "grad_norm": 2.038098096847534,
+      "learning_rate": 3.943948009748173e-06,
+      "loss": 1.9582,
+      "step": 808
+    },
+    {
+      "epoch": 0.8865753424657534,
+      "grad_norm": 2.3748130798339844,
+      "learning_rate": 3.94191714053615e-06,
+      "loss": 1.9091,
+      "step": 809
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "grad_norm": 2.2457494735717773,
+      "learning_rate": 3.939886271324127e-06,
+      "loss": 1.9067,
+      "step": 810
+    },
+    {
+      "epoch": 0.8876712328767123,
+      "eval_loss": 2.0766079425811768,
+      "eval_runtime": 36.9896,
+      "eval_samples_per_second": 27.035,
+      "eval_steps_per_second": 1.135,
+      "step": 810
+    },
+    {
+      "epoch": 0.8887671232876713,
+      "grad_norm": 2.156644821166992,
+      "learning_rate": 3.937855402112104e-06,
+      "loss": 1.9347,
+      "step": 811
+    },
+    {
+      "epoch": 0.8898630136986302,
+      "grad_norm": 2.3991472721099854,
+      "learning_rate": 3.935824532900082e-06,
+      "loss": 1.8645,
+      "step": 812
+    },
+    {
+      "epoch": 0.890958904109589,
+      "grad_norm": 2.252729654312134,
+      "learning_rate": 3.933793663688058e-06,
+      "loss": 1.8537,
+      "step": 813
+    },
+    {
+      "epoch": 0.8920547945205479,
+      "grad_norm": 2.2917964458465576,
+      "learning_rate": 3.9317627944760365e-06,
+      "loss": 1.9271,
+      "step": 814
+    },
+    {
+      "epoch": 0.8931506849315068,
+      "grad_norm": 2.217310667037964,
+      "learning_rate": 3.929731925264013e-06,
+      "loss": 1.9406,
+      "step": 815
+    },
+    {
+      "epoch": 0.8942465753424658,
+      "grad_norm": 2.167931079864502,
+      "learning_rate": 3.927701056051991e-06,
+      "loss": 1.8554,
+      "step": 816
+    },
+    {
+      "epoch": 0.8953424657534247,
+      "grad_norm": 2.18337082862854,
+      "learning_rate": 3.9256701868399674e-06,
+      "loss": 1.8795,
+      "step": 817
+    },
+    {
+      "epoch": 0.8964383561643836,
+      "grad_norm": 2.2348973751068115,
+      "learning_rate": 3.9236393176279456e-06,
+      "loss": 1.8831,
+      "step": 818
+    },
+    {
+      "epoch": 0.8975342465753424,
+      "grad_norm": 2.1822657585144043,
+      "learning_rate": 3.921608448415922e-06,
+      "loss": 1.8728,
+      "step": 819
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "grad_norm": 2.5253243446350098,
+      "learning_rate": 3.9195775792039e-06,
+      "loss": 1.8633,
+      "step": 820
+    },
+    {
+      "epoch": 0.8986301369863013,
+      "eval_loss": 2.0769834518432617,
+      "eval_runtime": 36.838,
+      "eval_samples_per_second": 27.146,
+      "eval_steps_per_second": 1.14,
+      "step": 820
+    },
+    {
+      "epoch": 0.8997260273972603,
+      "grad_norm": 2.2109973430633545,
+      "learning_rate": 3.9175467099918765e-06,
+      "loss": 1.912,
+      "step": 821
+    },
+    {
+      "epoch": 0.9008219178082192,
+      "grad_norm": 2.053692102432251,
+      "learning_rate": 3.915515840779855e-06,
+      "loss": 1.8928,
+      "step": 822
+    },
+    {
+      "epoch": 0.9019178082191781,
+      "grad_norm": 2.1418213844299316,
+      "learning_rate": 3.913484971567831e-06,
+      "loss": 1.9399,
+      "step": 823
+    },
+    {
+      "epoch": 0.903013698630137,
+      "grad_norm": 2.3615458011627197,
+      "learning_rate": 3.911454102355808e-06,
+      "loss": 1.9293,
+      "step": 824
+    },
+    {
+      "epoch": 0.9041095890410958,
+      "grad_norm": 2.098254680633545,
+      "learning_rate": 3.909423233143786e-06,
+      "loss": 1.8301,
+      "step": 825
+    },
+    {
+      "epoch": 0.9052054794520548,
+      "grad_norm": 2.4473679065704346,
+      "learning_rate": 3.907392363931763e-06,
+      "loss": 1.9254,
+      "step": 826
+    },
+    {
+      "epoch": 0.9063013698630137,
+      "grad_norm": 2.252843141555786,
+      "learning_rate": 3.90536149471974e-06,
+      "loss": 1.9155,
+      "step": 827
+    },
+    {
+      "epoch": 0.9073972602739726,
+      "grad_norm": 2.1574337482452393,
+      "learning_rate": 3.9033306255077175e-06,
+      "loss": 1.9809,
+      "step": 828
+    },
+    {
+      "epoch": 0.9084931506849315,
+      "grad_norm": 2.2899720668792725,
+      "learning_rate": 3.901299756295695e-06,
+      "loss": 2.0045,
+      "step": 829
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "grad_norm": 2.227285146713257,
+      "learning_rate": 3.899268887083672e-06,
+      "loss": 2.026,
+      "step": 830
+    },
+    {
+      "epoch": 0.9095890410958904,
+      "eval_loss": 2.0764760971069336,
+      "eval_runtime": 36.9195,
+      "eval_samples_per_second": 27.086,
+      "eval_steps_per_second": 1.138,
+      "step": 830
+    },
+    {
+      "epoch": 0.9106849315068493,
+      "grad_norm": 2.1935312747955322,
+      "learning_rate": 3.897238017871649e-06,
+      "loss": 1.8922,
+      "step": 831
+    },
+    {
+      "epoch": 0.9117808219178082,
+      "grad_norm": 2.129692316055298,
+      "learning_rate": 3.8952071486596266e-06,
+      "loss": 1.8981,
+      "step": 832
+    },
+    {
+      "epoch": 0.9128767123287671,
+      "grad_norm": 2.4435646533966064,
+      "learning_rate": 3.893176279447604e-06,
+      "loss": 1.91,
+      "step": 833
+    },
+    {
+      "epoch": 0.913972602739726,
+      "grad_norm": 2.1556732654571533,
+      "learning_rate": 3.891145410235581e-06,
+      "loss": 1.8688,
+      "step": 834
+    },
+    {
+      "epoch": 0.915068493150685,
+      "grad_norm": 2.175891160964966,
+      "learning_rate": 3.889114541023558e-06,
+      "loss": 1.9777,
+      "step": 835
+    },
+    {
+      "epoch": 0.9161643835616439,
+      "grad_norm": 2.220466375350952,
+      "learning_rate": 3.887083671811536e-06,
+      "loss": 1.9152,
+      "step": 836
+    },
+    {
+      "epoch": 0.9172602739726028,
+      "grad_norm": 2.331397294998169,
+      "learning_rate": 3.885052802599513e-06,
+      "loss": 1.892,
+      "step": 837
+    },
+    {
+      "epoch": 0.9183561643835616,
+      "grad_norm": 2.2434773445129395,
+      "learning_rate": 3.88302193338749e-06,
+      "loss": 1.841,
+      "step": 838
+    },
+    {
+      "epoch": 0.9194520547945205,
+      "grad_norm": 2.3248002529144287,
+      "learning_rate": 3.8809910641754675e-06,
+      "loss": 1.9657,
+      "step": 839
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "grad_norm": 2.090252637863159,
+      "learning_rate": 3.878960194963445e-06,
+      "loss": 1.9121,
+      "step": 840
+    },
+    {
+      "epoch": 0.9205479452054794,
+      "eval_loss": 2.0798065662384033,
+      "eval_runtime": 36.8557,
+      "eval_samples_per_second": 27.133,
+      "eval_steps_per_second": 1.14,
+      "step": 840
+    },
+    {
+      "epoch": 0.9216438356164384,
+      "grad_norm": 2.6968672275543213,
+      "learning_rate": 3.876929325751422e-06,
+      "loss": 1.9003,
+      "step": 841
+    },
+    {
+      "epoch": 0.9227397260273973,
+      "grad_norm": 2.356321096420288,
+      "learning_rate": 3.874898456539399e-06,
+      "loss": 1.9471,
+      "step": 842
+    },
+    {
+      "epoch": 0.9238356164383562,
+      "grad_norm": 2.2678606510162354,
+      "learning_rate": 3.872867587327377e-06,
+      "loss": 1.8488,
+      "step": 843
+    },
+    {
+      "epoch": 0.924931506849315,
+      "grad_norm": 2.1122310161590576,
+      "learning_rate": 3.870836718115354e-06,
+      "loss": 1.9247,
+      "step": 844
+    },
+    {
+      "epoch": 0.9260273972602739,
+      "grad_norm": 2.1219136714935303,
+      "learning_rate": 3.868805848903331e-06,
+      "loss": 1.8829,
+      "step": 845
+    },
+    {
+      "epoch": 0.9271232876712329,
+      "grad_norm": 2.175607204437256,
+      "learning_rate": 3.8667749796913084e-06,
+      "loss": 1.887,
+      "step": 846
+    },
+    {
+      "epoch": 0.9282191780821918,
+      "grad_norm": 2.156214714050293,
+      "learning_rate": 3.864744110479286e-06,
+      "loss": 1.9119,
+      "step": 847
+    },
+    {
+      "epoch": 0.9293150684931507,
+      "grad_norm": 2.158461570739746,
+      "learning_rate": 3.862713241267262e-06,
+      "loss": 1.9092,
+      "step": 848
+    },
+    {
+      "epoch": 0.9304109589041096,
+      "grad_norm": 2.180551767349243,
+      "learning_rate": 3.86068237205524e-06,
+      "loss": 1.946,
+      "step": 849
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "grad_norm": 2.1030542850494385,
+      "learning_rate": 3.858651502843217e-06,
+      "loss": 1.9484,
+      "step": 850
+    },
+    {
+      "epoch": 0.9315068493150684,
+      "eval_loss": 2.0767791271209717,
+      "eval_runtime": 36.8047,
+      "eval_samples_per_second": 27.17,
+      "eval_steps_per_second": 1.141,
+      "step": 850
+    },
+    {
+      "epoch": 0.9326027397260274,
+      "grad_norm": 2.240769386291504,
+      "learning_rate": 3.856620633631195e-06,
+      "loss": 1.9114,
+      "step": 851
+    },
+    {
+      "epoch": 0.9336986301369863,
+      "grad_norm": 2.1964619159698486,
+      "learning_rate": 3.854589764419171e-06,
+      "loss": 1.8339,
+      "step": 852
+    },
+    {
+      "epoch": 0.9347945205479452,
+      "grad_norm": 2.2875845432281494,
+      "learning_rate": 3.852558895207149e-06,
+      "loss": 1.8942,
+      "step": 853
+    },
+    {
+      "epoch": 0.9358904109589041,
+      "grad_norm": 2.1462507247924805,
+      "learning_rate": 3.850528025995126e-06,
+      "loss": 1.9838,
+      "step": 854
+    },
+    {
+      "epoch": 0.936986301369863,
+      "grad_norm": 2.0982913970947266,
+      "learning_rate": 3.848497156783104e-06,
+      "loss": 1.9351,
+      "step": 855
+    },
+    {
+      "epoch": 0.938082191780822,
+      "grad_norm": 2.260263442993164,
+      "learning_rate": 3.84646628757108e-06,
+      "loss": 1.8745,
+      "step": 856
+    },
+    {
+      "epoch": 0.9391780821917808,
+      "grad_norm": 2.091778039932251,
+      "learning_rate": 3.8444354183590585e-06,
+      "loss": 1.8888,
+      "step": 857
+    },
+    {
+      "epoch": 0.9402739726027397,
+      "grad_norm": 2.054974317550659,
+      "learning_rate": 3.842404549147035e-06,
+      "loss": 1.8801,
+      "step": 858
+    },
+    {
+      "epoch": 0.9413698630136986,
+      "grad_norm": 2.079641580581665,
+      "learning_rate": 3.840373679935013e-06,
+      "loss": 1.8879,
+      "step": 859
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "grad_norm": 2.3441810607910156,
+      "learning_rate": 3.8383428107229894e-06,
+      "loss": 1.8571,
+      "step": 860
+    },
+    {
+      "epoch": 0.9424657534246575,
+      "eval_loss": 2.0766854286193848,
+      "eval_runtime": 36.8414,
+      "eval_samples_per_second": 27.143,
+      "eval_steps_per_second": 1.14,
+      "step": 860
+    },
+    {
+      "epoch": 0.9435616438356165,
+      "grad_norm": 2.123061180114746,
+      "learning_rate": 3.836311941510967e-06,
+      "loss": 1.8922,
+      "step": 861
+    },
+    {
+      "epoch": 0.9446575342465754,
+      "grad_norm": 2.1573386192321777,
+      "learning_rate": 3.834281072298944e-06,
+      "loss": 1.8713,
+      "step": 862
+    },
+    {
+      "epoch": 0.9457534246575342,
+      "grad_norm": 2.225186824798584,
+      "learning_rate": 3.832250203086921e-06,
+      "loss": 1.925,
+      "step": 863
+    },
+    {
+      "epoch": 0.9468493150684931,
+      "grad_norm": 2.1539998054504395,
+      "learning_rate": 3.8302193338748985e-06,
+      "loss": 1.8721,
+      "step": 864
+    },
+    {
+      "epoch": 0.947945205479452,
+      "grad_norm": 2.075580358505249,
+      "learning_rate": 3.828188464662876e-06,
+      "loss": 1.9489,
+      "step": 865
+    },
+    {
+      "epoch": 0.949041095890411,
+      "grad_norm": 2.1789908409118652,
+      "learning_rate": 3.826157595450853e-06,
+      "loss": 1.9883,
+      "step": 866
+    },
+    {
+      "epoch": 0.9501369863013699,
+      "grad_norm": 2.0838615894317627,
+      "learning_rate": 3.82412672623883e-06,
+      "loss": 1.9149,
+      "step": 867
+    },
+    {
+      "epoch": 0.9512328767123288,
+      "grad_norm": 2.1321213245391846,
+      "learning_rate": 3.822095857026808e-06,
+      "loss": 1.8505,
+      "step": 868
+    },
+    {
+      "epoch": 0.9523287671232876,
+      "grad_norm": 2.1419479846954346,
+      "learning_rate": 3.820064987814785e-06,
+      "loss": 1.9226,
+      "step": 869
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "grad_norm": 2.290163040161133,
+      "learning_rate": 3.818034118602762e-06,
+      "loss": 1.8737,
+      "step": 870
+    },
+    {
+      "epoch": 0.9534246575342465,
+      "eval_loss": 2.073598623275757,
+      "eval_runtime": 36.7896,
+      "eval_samples_per_second": 27.182,
+      "eval_steps_per_second": 1.142,
+      "step": 870
+    },
+    {
+      "epoch": 0.9545205479452055,
+      "grad_norm": 2.1951074600219727,
+      "learning_rate": 3.8160032493907395e-06,
+      "loss": 1.8699,
+      "step": 871
+    },
+    {
+      "epoch": 0.9556164383561644,
+      "grad_norm": 2.25960111618042,
+      "learning_rate": 3.813972380178717e-06,
+      "loss": 1.8813,
+      "step": 872
+    },
+    {
+      "epoch": 0.9567123287671233,
+      "grad_norm": 2.4376907348632812,
+      "learning_rate": 3.811941510966694e-06,
+      "loss": 1.8782,
+      "step": 873
+    },
+    {
+      "epoch": 0.9578082191780822,
+      "grad_norm": 1.980772614479065,
+      "learning_rate": 3.809910641754671e-06,
+      "loss": 1.9205,
+      "step": 874
+    },
+    {
+      "epoch": 0.958904109589041,
+      "grad_norm": 2.0699028968811035,
+      "learning_rate": 3.8078797725426486e-06,
+      "loss": 1.8674,
+      "step": 875
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 2.2765867710113525,
+      "learning_rate": 3.8058489033306254e-06,
+      "loss": 1.81,
+      "step": 876
+    },
+    {
+      "epoch": 0.9610958904109589,
+      "grad_norm": 2.0998799800872803,
+      "learning_rate": 3.803818034118603e-06,
+      "loss": 1.9061,
+      "step": 877
+    },
+    {
+      "epoch": 0.9621917808219178,
+      "grad_norm": 2.2362914085388184,
+      "learning_rate": 3.80178716490658e-06,
+      "loss": 1.8798,
+      "step": 878
+    },
+    {
+      "epoch": 0.9632876712328767,
+      "grad_norm": 2.0789835453033447,
+      "learning_rate": 3.7997562956945577e-06,
+      "loss": 1.9491,
+      "step": 879
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "grad_norm": 2.0458688735961914,
+      "learning_rate": 3.7977254264825345e-06,
+      "loss": 1.931,
+      "step": 880
+    },
+    {
+      "epoch": 0.9643835616438357,
+      "eval_loss": 2.0672221183776855,
+      "eval_runtime": 36.8218,
+      "eval_samples_per_second": 27.158,
+      "eval_steps_per_second": 1.141,
+      "step": 880
+    },
+    {
+      "epoch": 0.9654794520547946,
+      "grad_norm": 2.361626625061035,
+      "learning_rate": 3.7956945572705122e-06,
+      "loss": 1.8815,
+      "step": 881
+    },
+    {
+      "epoch": 0.9665753424657534,
+      "grad_norm": 2.294160842895508,
+      "learning_rate": 3.793663688058489e-06,
+      "loss": 1.8649,
+      "step": 882
+    },
+    {
+      "epoch": 0.9676712328767123,
+      "grad_norm": 2.2242612838745117,
+      "learning_rate": 3.791632818846467e-06,
+      "loss": 1.9535,
+      "step": 883
+    },
+    {
+      "epoch": 0.9687671232876712,
+      "grad_norm": 2.220269203186035,
+      "learning_rate": 3.7896019496344436e-06,
+      "loss": 1.9506,
+      "step": 884
+    },
+    {
+      "epoch": 0.9698630136986301,
+      "grad_norm": 2.112518072128296,
+      "learning_rate": 3.787571080422421e-06,
+      "loss": 1.9487,
+      "step": 885
+    },
+    {
+      "epoch": 0.9709589041095891,
+      "grad_norm": 2.17695951461792,
+      "learning_rate": 3.785540211210398e-06,
+      "loss": 1.9467,
+      "step": 886
+    },
+    {
+      "epoch": 0.972054794520548,
+      "grad_norm": 2.0990567207336426,
+      "learning_rate": 3.7835093419983755e-06,
+      "loss": 1.9216,
+      "step": 887
+    },
+    {
+      "epoch": 0.9731506849315068,
+      "grad_norm": 2.222566604614258,
+      "learning_rate": 3.7814784727863527e-06,
+      "loss": 1.9043,
+      "step": 888
+    },
+    {
+      "epoch": 0.9742465753424657,
+      "grad_norm": 2.09890079498291,
+      "learning_rate": 3.77944760357433e-06,
+      "loss": 1.884,
+      "step": 889
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "grad_norm": 2.136141061782837,
+      "learning_rate": 3.7774167343623073e-06,
+      "loss": 2.0032,
+      "step": 890
+    },
+    {
+      "epoch": 0.9753424657534246,
+      "eval_loss": 2.0665438175201416,
+      "eval_runtime": 37.1207,
+      "eval_samples_per_second": 26.939,
+      "eval_steps_per_second": 1.131,
+      "step": 890
+    },
+    {
+      "epoch": 0.9764383561643836,
+      "grad_norm": 2.267444372177124,
+      "learning_rate": 3.7753858651502846e-06,
+      "loss": 1.947,
+      "step": 891
+    },
+    {
+      "epoch": 0.9775342465753425,
+      "grad_norm": 2.1141815185546875,
+      "learning_rate": 3.773354995938262e-06,
+      "loss": 1.8705,
+      "step": 892
+    },
+    {
+      "epoch": 0.9786301369863014,
+      "grad_norm": 2.1036782264709473,
+      "learning_rate": 3.771324126726239e-06,
+      "loss": 1.9381,
+      "step": 893
+    },
+    {
+      "epoch": 0.9797260273972602,
+      "grad_norm": 2.081982135772705,
+      "learning_rate": 3.7692932575142164e-06,
+      "loss": 1.9079,
+      "step": 894
+    },
+    {
+      "epoch": 0.9808219178082191,
+      "grad_norm": 2.0815446376800537,
+      "learning_rate": 3.7672623883021937e-06,
+      "loss": 1.9417,
+      "step": 895
+    },
+    {
+      "epoch": 0.9819178082191781,
+      "grad_norm": 2.0531511306762695,
+      "learning_rate": 3.765231519090171e-06,
+      "loss": 1.876,
+      "step": 896
+    },
+    {
+      "epoch": 0.983013698630137,
+      "grad_norm": 2.212940216064453,
+      "learning_rate": 3.7632006498781482e-06,
+      "loss": 1.8008,
+      "step": 897
+    },
+    {
+      "epoch": 0.9841095890410959,
+      "grad_norm": 2.1541531085968018,
+      "learning_rate": 3.761169780666125e-06,
+      "loss": 1.9039,
+      "step": 898
+    },
+    {
+      "epoch": 0.9852054794520548,
+      "grad_norm": 2.113884925842285,
+      "learning_rate": 3.7591389114541028e-06,
+      "loss": 1.8562,
+      "step": 899
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "grad_norm": 2.1503701210021973,
+      "learning_rate": 3.7571080422420796e-06,
+      "loss": 1.9052,
+      "step": 900
+    },
+    {
+      "epoch": 0.9863013698630136,
+      "eval_loss": 2.0664408206939697,
+      "eval_runtime": 37.0876,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 900
+    },
+    {
+      "epoch": 0.9873972602739726,
+      "grad_norm": 2.1584348678588867,
+      "learning_rate": 3.7550771730300573e-06,
+      "loss": 1.9296,
+      "step": 901
+    },
+    {
+      "epoch": 0.9884931506849315,
+      "grad_norm": 2.164318084716797,
+      "learning_rate": 3.753046303818034e-06,
+      "loss": 1.8834,
+      "step": 902
+    },
+    {
+      "epoch": 0.9895890410958904,
+      "grad_norm": 2.0222678184509277,
+      "learning_rate": 3.751015434606012e-06,
+      "loss": 1.9426,
+      "step": 903
+    },
+    {
+      "epoch": 0.9906849315068493,
+      "grad_norm": 2.116835832595825,
+      "learning_rate": 3.7489845653939887e-06,
+      "loss": 1.9195,
+      "step": 904
+    },
+    {
+      "epoch": 0.9917808219178083,
+      "grad_norm": 2.2708873748779297,
+      "learning_rate": 3.7469536961819664e-06,
+      "loss": 1.9631,
+      "step": 905
+    },
+    {
+      "epoch": 0.9928767123287672,
+      "grad_norm": 2.1174001693725586,
+      "learning_rate": 3.7449228269699433e-06,
+      "loss": 1.8602,
+      "step": 906
+    },
+    {
+      "epoch": 0.993972602739726,
+      "grad_norm": 2.091830015182495,
+      "learning_rate": 3.742891957757921e-06,
+      "loss": 1.8906,
+      "step": 907
+    },
+    {
+      "epoch": 0.9950684931506849,
+      "grad_norm": 2.007119655609131,
+      "learning_rate": 3.740861088545898e-06,
+      "loss": 1.9122,
+      "step": 908
+    },
+    {
+      "epoch": 0.9961643835616438,
+      "grad_norm": 2.070396661758423,
+      "learning_rate": 3.7388302193338755e-06,
+      "loss": 1.9106,
+      "step": 909
+    },
+    {
+      "epoch": 0.9972602739726028,
+      "grad_norm": 2.080331325531006,
+      "learning_rate": 3.7367993501218524e-06,
+      "loss": 1.934,
+      "step": 910
+    },
+    {
+      "epoch": 0.9972602739726028,
+      "eval_loss": 2.0694408416748047,
+      "eval_runtime": 36.9953,
+      "eval_samples_per_second": 27.03,
+      "eval_steps_per_second": 1.135,
+      "step": 910
+    },
+    {
+      "epoch": 0.9983561643835617,
+      "grad_norm": 2.0920615196228027,
+      "learning_rate": 3.7347684809098297e-06,
+      "loss": 1.8679,
+      "step": 911
+    },
+    {
+      "epoch": 0.9994520547945206,
+      "grad_norm": 2.112732172012329,
+      "learning_rate": 3.732737611697807e-06,
+      "loss": 1.8723,
+      "step": 912
+    },
+    {
+      "epoch": 1.0005479452054795,
+      "grad_norm": 2.1449220180511475,
+      "learning_rate": 3.7307067424857842e-06,
+      "loss": 1.8509,
+      "step": 913
+    },
+    {
+      "epoch": 1.0016438356164383,
+      "grad_norm": 2.2179951667785645,
+      "learning_rate": 3.7286758732737615e-06,
+      "loss": 1.7911,
+      "step": 914
+    },
+    {
+      "epoch": 1.0027397260273974,
+      "grad_norm": 2.1392273902893066,
+      "learning_rate": 3.7266450040617388e-06,
+      "loss": 1.8475,
+      "step": 915
+    },
+    {
+      "epoch": 1.0038356164383562,
+      "grad_norm": 2.190265417098999,
+      "learning_rate": 3.724614134849716e-06,
+      "loss": 1.8394,
+      "step": 916
+    },
+    {
+      "epoch": 1.004931506849315,
+      "grad_norm": 2.1833760738372803,
+      "learning_rate": 3.7225832656376933e-06,
+      "loss": 1.9696,
+      "step": 917
+    },
+    {
+      "epoch": 1.006027397260274,
+      "grad_norm": 2.2504308223724365,
+      "learning_rate": 3.7205523964256706e-06,
+      "loss": 1.8135,
+      "step": 918
+    },
+    {
+      "epoch": 1.0071232876712328,
+      "grad_norm": 2.3909389972686768,
+      "learning_rate": 3.718521527213648e-06,
+      "loss": 1.9331,
+      "step": 919
+    },
+    {
+      "epoch": 1.0082191780821919,
+      "grad_norm": 2.3367602825164795,
+      "learning_rate": 3.716490658001625e-06,
+      "loss": 1.9026,
+      "step": 920
+    },
+    {
+      "epoch": 1.0082191780821919,
+      "eval_loss": 2.066087484359741,
+      "eval_runtime": 36.8701,
+      "eval_samples_per_second": 27.122,
+      "eval_steps_per_second": 1.139,
+      "step": 920
+    },
+    {
+      "epoch": 1.0093150684931507,
+      "grad_norm": 2.1732144355773926,
+      "learning_rate": 3.7144597887896024e-06,
+      "loss": 1.8888,
+      "step": 921
+    },
+    {
+      "epoch": 1.0104109589041095,
+      "grad_norm": 2.4632201194763184,
+      "learning_rate": 3.7124289195775797e-06,
+      "loss": 1.7725,
+      "step": 922
+    },
+    {
+      "epoch": 1.0115068493150685,
+      "grad_norm": 2.4887173175811768,
+      "learning_rate": 3.710398050365557e-06,
+      "loss": 1.9125,
+      "step": 923
+    },
+    {
+      "epoch": 1.0126027397260273,
+      "grad_norm": 2.4684619903564453,
+      "learning_rate": 3.708367181153534e-06,
+      "loss": 1.8829,
+      "step": 924
+    },
+    {
+      "epoch": 1.0136986301369864,
+      "grad_norm": 2.348788022994995,
+      "learning_rate": 3.7063363119415115e-06,
+      "loss": 1.8542,
+      "step": 925
+    },
+    {
+      "epoch": 1.0147945205479452,
+      "grad_norm": 2.4432766437530518,
+      "learning_rate": 3.7043054427294884e-06,
+      "loss": 1.903,
+      "step": 926
+    },
+    {
+      "epoch": 1.0158904109589042,
+      "grad_norm": 2.425910234451294,
+      "learning_rate": 3.702274573517466e-06,
+      "loss": 1.8532,
+      "step": 927
+    },
+    {
+      "epoch": 1.016986301369863,
+      "grad_norm": 2.447777032852173,
+      "learning_rate": 3.700243704305443e-06,
+      "loss": 1.8277,
+      "step": 928
+    },
+    {
+      "epoch": 1.0180821917808218,
+      "grad_norm": 2.1830408573150635,
+      "learning_rate": 3.6982128350934206e-06,
+      "loss": 1.8546,
+      "step": 929
+    },
+    {
+      "epoch": 1.0191780821917809,
+      "grad_norm": 2.2692818641662598,
+      "learning_rate": 3.6961819658813975e-06,
+      "loss": 1.8356,
+      "step": 930
+    },
+    {
+      "epoch": 1.0191780821917809,
+      "eval_loss": 2.0714614391326904,
+      "eval_runtime": 36.8661,
+      "eval_samples_per_second": 27.125,
+      "eval_steps_per_second": 1.139,
+      "step": 930
+    },
+    {
+      "epoch": 1.0202739726027397,
+      "grad_norm": 2.5044169425964355,
+      "learning_rate": 3.694151096669375e-06,
+      "loss": 1.8173,
+      "step": 931
+    },
+    {
+      "epoch": 1.0213698630136987,
+      "grad_norm": 2.419637680053711,
+      "learning_rate": 3.692120227457352e-06,
+      "loss": 1.9418,
+      "step": 932
+    },
+    {
+      "epoch": 1.0224657534246575,
+      "grad_norm": 2.282288074493408,
+      "learning_rate": 3.6900893582453297e-06,
+      "loss": 1.8698,
+      "step": 933
+    },
+    {
+      "epoch": 1.0235616438356165,
+      "grad_norm": 2.3215551376342773,
+      "learning_rate": 3.6880584890333066e-06,
+      "loss": 1.8935,
+      "step": 934
+    },
+    {
+      "epoch": 1.0246575342465754,
+      "grad_norm": 2.5938258171081543,
+      "learning_rate": 3.6860276198212834e-06,
+      "loss": 1.79,
+      "step": 935
+    },
+    {
+      "epoch": 1.0257534246575342,
+      "grad_norm": 2.215024471282959,
+      "learning_rate": 3.683996750609261e-06,
+      "loss": 1.8768,
+      "step": 936
+    },
+    {
+      "epoch": 1.0268493150684932,
+      "grad_norm": 2.239609956741333,
+      "learning_rate": 3.681965881397238e-06,
+      "loss": 1.8887,
+      "step": 937
+    },
+    {
+      "epoch": 1.027945205479452,
+      "grad_norm": 2.3134827613830566,
+      "learning_rate": 3.6799350121852157e-06,
+      "loss": 1.8897,
+      "step": 938
+    },
+    {
+      "epoch": 1.029041095890411,
+      "grad_norm": 2.4100260734558105,
+      "learning_rate": 3.6779041429731925e-06,
+      "loss": 1.909,
+      "step": 939
+    },
+    {
+      "epoch": 1.0301369863013699,
+      "grad_norm": 2.26204776763916,
+      "learning_rate": 3.6758732737611702e-06,
+      "loss": 1.8514,
+      "step": 940
+    },
+    {
+      "epoch": 1.0301369863013699,
+      "eval_loss": 2.0666027069091797,
+      "eval_runtime": 36.8336,
+      "eval_samples_per_second": 27.149,
+      "eval_steps_per_second": 1.14,
+      "step": 940
+    },
+    {
+      "epoch": 1.0312328767123287,
+      "grad_norm": 2.22210955619812,
+      "learning_rate": 3.673842404549147e-06,
+      "loss": 1.8743,
+      "step": 941
+    },
+    {
+      "epoch": 1.0323287671232877,
+      "grad_norm": 2.2371468544006348,
+      "learning_rate": 3.671811535337125e-06,
+      "loss": 1.8573,
+      "step": 942
+    },
+    {
+      "epoch": 1.0334246575342465,
+      "grad_norm": 2.346745014190674,
+      "learning_rate": 3.6697806661251016e-06,
+      "loss": 1.8614,
+      "step": 943
+    },
+    {
+      "epoch": 1.0345205479452055,
+      "grad_norm": 2.4765074253082275,
+      "learning_rate": 3.6677497969130793e-06,
+      "loss": 1.8654,
+      "step": 944
+    },
+    {
+      "epoch": 1.0356164383561643,
+      "grad_norm": 2.3023204803466797,
+      "learning_rate": 3.665718927701056e-06,
+      "loss": 1.9901,
+      "step": 945
+    },
+    {
+      "epoch": 1.0367123287671234,
+      "grad_norm": 2.322421073913574,
+      "learning_rate": 3.663688058489034e-06,
+      "loss": 1.8775,
+      "step": 946
+    },
+    {
+      "epoch": 1.0378082191780822,
+      "grad_norm": 2.2324275970458984,
+      "learning_rate": 3.6616571892770107e-06,
+      "loss": 1.9303,
+      "step": 947
+    },
+    {
+      "epoch": 1.038904109589041,
+      "grad_norm": 2.372067928314209,
+      "learning_rate": 3.659626320064988e-06,
+      "loss": 1.8673,
+      "step": 948
+    },
+    {
+      "epoch": 1.04,
+      "grad_norm": 2.2404983043670654,
+      "learning_rate": 3.6575954508529653e-06,
+      "loss": 1.8025,
+      "step": 949
+    },
+    {
+      "epoch": 1.0410958904109588,
+      "grad_norm": 2.180673599243164,
+      "learning_rate": 3.6555645816409426e-06,
+      "loss": 1.8563,
+      "step": 950
+    },
+    {
+      "epoch": 1.0410958904109588,
+      "eval_loss": 2.061262845993042,
+      "eval_runtime": 36.7727,
+      "eval_samples_per_second": 27.194,
+      "eval_steps_per_second": 1.142,
+      "step": 950
+    },
+    {
+      "epoch": 1.0421917808219179,
+      "grad_norm": 2.299043893814087,
+      "learning_rate": 3.65353371242892e-06,
+      "loss": 1.8453,
+      "step": 951
+    },
+    {
+      "epoch": 1.0432876712328767,
+      "grad_norm": 2.4878089427948,
+      "learning_rate": 3.651502843216897e-06,
+      "loss": 1.8901,
+      "step": 952
+    },
+    {
+      "epoch": 1.0443835616438357,
+      "grad_norm": 2.150404930114746,
+      "learning_rate": 3.6494719740048744e-06,
+      "loss": 1.7371,
+      "step": 953
+    },
+    {
+      "epoch": 1.0454794520547945,
+      "grad_norm": 2.177750587463379,
+      "learning_rate": 3.6474411047928517e-06,
+      "loss": 1.8814,
+      "step": 954
+    },
+    {
+      "epoch": 1.0465753424657533,
+      "grad_norm": 2.230797052383423,
+      "learning_rate": 3.645410235580829e-06,
+      "loss": 1.8323,
+      "step": 955
+    },
+    {
+      "epoch": 1.0476712328767124,
+      "grad_norm": 2.2484378814697266,
+      "learning_rate": 3.6433793663688062e-06,
+      "loss": 1.8749,
+      "step": 956
+    },
+    {
+      "epoch": 1.0487671232876712,
+      "grad_norm": 2.201279640197754,
+      "learning_rate": 3.6413484971567835e-06,
+      "loss": 1.8518,
+      "step": 957
+    },
+    {
+      "epoch": 1.0498630136986302,
+      "grad_norm": 2.1125736236572266,
+      "learning_rate": 3.6393176279447608e-06,
+      "loss": 1.846,
+      "step": 958
+    },
+    {
+      "epoch": 1.050958904109589,
+      "grad_norm": 2.3078789710998535,
+      "learning_rate": 3.637286758732738e-06,
+      "loss": 1.8791,
+      "step": 959
+    },
+    {
+      "epoch": 1.0520547945205478,
+      "grad_norm": 2.2911148071289062,
+      "learning_rate": 3.6352558895207153e-06,
+      "loss": 1.8901,
+      "step": 960
+    },
+    {
+      "epoch": 1.0520547945205478,
+      "eval_loss": 2.0583066940307617,
+      "eval_runtime": 36.7535,
+      "eval_samples_per_second": 27.208,
+      "eval_steps_per_second": 1.143,
+      "step": 960
+    },
+    {
+      "epoch": 1.0531506849315069,
+      "grad_norm": 2.3990273475646973,
+      "learning_rate": 3.633225020308692e-06,
+      "loss": 1.9071,
+      "step": 961
+    },
+    {
+      "epoch": 1.0542465753424657,
+      "grad_norm": 2.358685255050659,
+      "learning_rate": 3.63119415109667e-06,
+      "loss": 1.8982,
+      "step": 962
+    },
+    {
+      "epoch": 1.0553424657534247,
+      "grad_norm": 2.1145992279052734,
+      "learning_rate": 3.6291632818846467e-06,
+      "loss": 1.8649,
+      "step": 963
+    },
+    {
+      "epoch": 1.0564383561643835,
+      "grad_norm": 2.2302935123443604,
+      "learning_rate": 3.6271324126726244e-06,
+      "loss": 1.8911,
+      "step": 964
+    },
+    {
+      "epoch": 1.0575342465753426,
+      "grad_norm": 2.344797372817993,
+      "learning_rate": 3.6251015434606013e-06,
+      "loss": 1.8,
+      "step": 965
+    },
+    {
+      "epoch": 1.0586301369863014,
+      "grad_norm": 2.2514073848724365,
+      "learning_rate": 3.623070674248579e-06,
+      "loss": 1.8128,
+      "step": 966
+    },
+    {
+      "epoch": 1.0597260273972602,
+      "grad_norm": 2.541221857070923,
+      "learning_rate": 3.621039805036556e-06,
+      "loss": 1.8504,
+      "step": 967
+    },
+    {
+      "epoch": 1.0608219178082192,
+      "grad_norm": 2.3211770057678223,
+      "learning_rate": 3.6190089358245335e-06,
+      "loss": 1.8329,
+      "step": 968
+    },
+    {
+      "epoch": 1.061917808219178,
+      "grad_norm": 2.1008756160736084,
+      "learning_rate": 3.6169780666125104e-06,
+      "loss": 1.8972,
+      "step": 969
+    },
+    {
+      "epoch": 1.063013698630137,
+      "grad_norm": 2.4148898124694824,
+      "learning_rate": 3.614947197400488e-06,
+      "loss": 1.8748,
+      "step": 970
+    },
+    {
+      "epoch": 1.063013698630137,
+      "eval_loss": 2.0592916011810303,
+      "eval_runtime": 36.7663,
+      "eval_samples_per_second": 27.199,
+      "eval_steps_per_second": 1.142,
+      "step": 970
+    },
+    {
+      "epoch": 1.0641095890410959,
+      "grad_norm": 2.4116220474243164,
+      "learning_rate": 3.612916328188465e-06,
+      "loss": 1.8915,
+      "step": 971
+    },
+    {
+      "epoch": 1.0652054794520547,
+      "grad_norm": 2.257411241531372,
+      "learning_rate": 3.610885458976442e-06,
+      "loss": 1.7901,
+      "step": 972
+    },
+    {
+      "epoch": 1.0663013698630137,
+      "grad_norm": 2.166875123977661,
+      "learning_rate": 3.6088545897644195e-06,
+      "loss": 1.8969,
+      "step": 973
+    },
+    {
+      "epoch": 1.0673972602739725,
+      "grad_norm": 2.464716911315918,
+      "learning_rate": 3.6068237205523963e-06,
+      "loss": 1.8844,
+      "step": 974
+    },
+    {
+      "epoch": 1.0684931506849316,
+      "grad_norm": 2.337334156036377,
+      "learning_rate": 3.604792851340374e-06,
+      "loss": 1.8192,
+      "step": 975
+    },
+    {
+      "epoch": 1.0695890410958904,
+      "grad_norm": 2.117663621902466,
+      "learning_rate": 3.602761982128351e-06,
+      "loss": 1.8233,
+      "step": 976
+    },
+    {
+      "epoch": 1.0706849315068494,
+      "grad_norm": 2.3353586196899414,
+      "learning_rate": 3.6007311129163286e-06,
+      "loss": 1.9133,
+      "step": 977
+    },
+    {
+      "epoch": 1.0717808219178082,
+      "grad_norm": 2.444279909133911,
+      "learning_rate": 3.5987002437043054e-06,
+      "loss": 1.8803,
+      "step": 978
+    },
+    {
+      "epoch": 1.072876712328767,
+      "grad_norm": 2.3458383083343506,
+      "learning_rate": 3.596669374492283e-06,
+      "loss": 1.8108,
+      "step": 979
+    },
+    {
+      "epoch": 1.073972602739726,
+      "grad_norm": 2.207977771759033,
+      "learning_rate": 3.59463850528026e-06,
+      "loss": 1.8423,
+      "step": 980
+    },
+    {
+      "epoch": 1.073972602739726,
+      "eval_loss": 2.0578773021698,
+      "eval_runtime": 36.8123,
+      "eval_samples_per_second": 27.165,
+      "eval_steps_per_second": 1.141,
+      "step": 980
+    },
+    {
+      "epoch": 1.0750684931506849,
+      "grad_norm": 2.378127336502075,
+      "learning_rate": 3.5926076360682377e-06,
+      "loss": 1.9268,
+      "step": 981
+    },
+    {
+      "epoch": 1.076164383561644,
+      "grad_norm": 2.248105525970459,
+      "learning_rate": 3.5905767668562145e-06,
+      "loss": 1.8111,
+      "step": 982
+    },
+    {
+      "epoch": 1.0772602739726027,
+      "grad_norm": 2.275592088699341,
+      "learning_rate": 3.5885458976441922e-06,
+      "loss": 1.9099,
+      "step": 983
+    },
+    {
+      "epoch": 1.0783561643835617,
+      "grad_norm": 3.108377456665039,
+      "learning_rate": 3.586515028432169e-06,
+      "loss": 1.8728,
+      "step": 984
+    },
+    {
+      "epoch": 1.0794520547945206,
+      "grad_norm": 2.2702999114990234,
+      "learning_rate": 3.5844841592201464e-06,
+      "loss": 1.9007,
+      "step": 985
+    },
+    {
+      "epoch": 1.0805479452054794,
+      "grad_norm": 3.218647003173828,
+      "learning_rate": 3.5824532900081237e-06,
+      "loss": 1.9157,
+      "step": 986
+    },
+    {
+      "epoch": 1.0816438356164384,
+      "grad_norm": 2.3393027782440186,
+      "learning_rate": 3.580422420796101e-06,
+      "loss": 1.7936,
+      "step": 987
+    },
+    {
+      "epoch": 1.0827397260273972,
+      "grad_norm": 2.420541286468506,
+      "learning_rate": 3.578391551584078e-06,
+      "loss": 1.8608,
+      "step": 988
+    },
+    {
+      "epoch": 1.0838356164383562,
+      "grad_norm": 2.3354556560516357,
+      "learning_rate": 3.5763606823720555e-06,
+      "loss": 1.8684,
+      "step": 989
+    },
+    {
+      "epoch": 1.084931506849315,
+      "grad_norm": 2.200282573699951,
+      "learning_rate": 3.5743298131600328e-06,
+      "loss": 1.7814,
+      "step": 990
+    },
+    {
+      "epoch": 1.084931506849315,
+      "eval_loss": 2.0565249919891357,
+      "eval_runtime": 36.952,
+      "eval_samples_per_second": 27.062,
+      "eval_steps_per_second": 1.137,
+      "step": 990
+    },
+    {
+      "epoch": 1.086027397260274,
+      "grad_norm": 2.149033308029175,
+      "learning_rate": 3.57229894394801e-06,
+      "loss": 1.8708,
+      "step": 991
+    },
+    {
+      "epoch": 1.087123287671233,
+      "grad_norm": 2.4663679599761963,
+      "learning_rate": 3.5702680747359873e-06,
+      "loss": 1.8947,
+      "step": 992
+    },
+    {
+      "epoch": 1.0882191780821917,
+      "grad_norm": 2.3142261505126953,
+      "learning_rate": 3.5682372055239646e-06,
+      "loss": 1.8727,
+      "step": 993
+    },
+    {
+      "epoch": 1.0893150684931507,
+      "grad_norm": 2.2223033905029297,
+      "learning_rate": 3.566206336311942e-06,
+      "loss": 1.8203,
+      "step": 994
+    },
+    {
+      "epoch": 1.0904109589041096,
+      "grad_norm": 2.423855781555176,
+      "learning_rate": 3.564175467099919e-06,
+      "loss": 1.8687,
+      "step": 995
+    },
+    {
+      "epoch": 1.0915068493150686,
+      "grad_norm": 2.4022302627563477,
+      "learning_rate": 3.5621445978878964e-06,
+      "loss": 1.8215,
+      "step": 996
+    },
+    {
+      "epoch": 1.0926027397260274,
+      "grad_norm": 2.378236770629883,
+      "learning_rate": 3.5601137286758737e-06,
+      "loss": 1.7994,
+      "step": 997
+    },
+    {
+      "epoch": 1.0936986301369862,
+      "grad_norm": 2.4499096870422363,
+      "learning_rate": 3.5580828594638505e-06,
+      "loss": 1.8779,
+      "step": 998
+    },
+    {
+      "epoch": 1.0947945205479452,
+      "grad_norm": 2.3181755542755127,
+      "learning_rate": 3.5560519902518282e-06,
+      "loss": 1.8341,
+      "step": 999
+    },
+    {
+      "epoch": 1.095890410958904,
+      "grad_norm": 2.2406623363494873,
+      "learning_rate": 3.554021121039805e-06,
+      "loss": 1.874,
+      "step": 1000
+    },
+    {
+      "epoch": 1.095890410958904,
+      "eval_loss": 2.0523672103881836,
+      "eval_runtime": 37.0454,
+      "eval_samples_per_second": 26.994,
+      "eval_steps_per_second": 1.134,
+      "step": 1000
+    },
+    {
+      "epoch": 1.096986301369863,
+      "grad_norm": 2.254667043685913,
+      "learning_rate": 3.551990251827783e-06,
+      "loss": 1.9017,
+      "step": 1001
+    },
+    {
+      "epoch": 1.098082191780822,
+      "grad_norm": 2.2846896648406982,
+      "learning_rate": 3.5499593826157596e-06,
+      "loss": 1.8505,
+      "step": 1002
+    },
+    {
+      "epoch": 1.099178082191781,
+      "grad_norm": 2.183469772338867,
+      "learning_rate": 3.5479285134037373e-06,
+      "loss": 1.8266,
+      "step": 1003
+    },
+    {
+      "epoch": 1.1002739726027397,
+      "grad_norm": 2.151798963546753,
+      "learning_rate": 3.545897644191714e-06,
+      "loss": 1.8487,
+      "step": 1004
+    },
+    {
+      "epoch": 1.1013698630136985,
+      "grad_norm": 2.2039146423339844,
+      "learning_rate": 3.543866774979692e-06,
+      "loss": 1.8327,
+      "step": 1005
+    },
+    {
+      "epoch": 1.1024657534246576,
+      "grad_norm": 2.2091665267944336,
+      "learning_rate": 3.5418359057676687e-06,
+      "loss": 1.8214,
+      "step": 1006
+    },
+    {
+      "epoch": 1.1035616438356164,
+      "grad_norm": 2.2821404933929443,
+      "learning_rate": 3.5398050365556464e-06,
+      "loss": 1.8591,
+      "step": 1007
+    },
+    {
+      "epoch": 1.1046575342465754,
+      "grad_norm": 2.1726973056793213,
+      "learning_rate": 3.5377741673436233e-06,
+      "loss": 1.8265,
+      "step": 1008
+    },
+    {
+      "epoch": 1.1057534246575342,
+      "grad_norm": 2.2870233058929443,
+      "learning_rate": 3.535743298131601e-06,
+      "loss": 1.8783,
+      "step": 1009
+    },
+    {
+      "epoch": 1.106849315068493,
+      "grad_norm": 2.148080587387085,
+      "learning_rate": 3.533712428919578e-06,
+      "loss": 1.8272,
+      "step": 1010
+    },
+    {
+      "epoch": 1.106849315068493,
+      "eval_loss": 2.049504041671753,
+      "eval_runtime": 36.8717,
+      "eval_samples_per_second": 27.121,
+      "eval_steps_per_second": 1.139,
+      "step": 1010
+    },
+    {
+      "epoch": 1.107945205479452,
+      "grad_norm": 2.182035446166992,
+      "learning_rate": 3.5316815597075547e-06,
+      "loss": 1.8771,
+      "step": 1011
+    },
+    {
+      "epoch": 1.1090410958904109,
+      "grad_norm": 2.135085344314575,
+      "learning_rate": 3.5296506904955324e-06,
+      "loss": 1.9031,
+      "step": 1012
+    },
+    {
+      "epoch": 1.11013698630137,
+      "grad_norm": 2.1017255783081055,
+      "learning_rate": 3.5276198212835093e-06,
+      "loss": 1.8341,
+      "step": 1013
+    },
+    {
+      "epoch": 1.1112328767123287,
+      "grad_norm": 2.2773115634918213,
+      "learning_rate": 3.525588952071487e-06,
+      "loss": 1.8564,
+      "step": 1014
+    },
+    {
+      "epoch": 1.1123287671232878,
+      "grad_norm": 2.2773115634918213,
+      "learning_rate": 3.525588952071487e-06,
+      "loss": 1.8971,
+      "step": 1015
+    },
+    {
+      "epoch": 1.1134246575342466,
+      "grad_norm": 2.273375988006592,
+      "learning_rate": 3.523558082859464e-06,
+      "loss": 1.9074,
+      "step": 1016
+    },
+    {
+      "epoch": 1.1145205479452054,
+      "grad_norm": 2.2055182456970215,
+      "learning_rate": 3.5215272136474415e-06,
+      "loss": 1.8724,
+      "step": 1017
+    },
+    {
+      "epoch": 1.1156164383561644,
+      "grad_norm": 2.246126651763916,
+      "learning_rate": 3.5194963444354184e-06,
+      "loss": 1.8387,
+      "step": 1018
+    },
+    {
+      "epoch": 1.1167123287671232,
+      "grad_norm": 2.3076183795928955,
+      "learning_rate": 3.517465475223396e-06,
+      "loss": 1.8387,
+      "step": 1019
+    },
+    {
+      "epoch": 1.1178082191780823,
+      "grad_norm": 2.3365488052368164,
+      "learning_rate": 3.515434606011373e-06,
+      "loss": 1.8063,
+      "step": 1020
+    },
+    {
+      "epoch": 1.1178082191780823,
+      "eval_loss": 2.0528781414031982,
+      "eval_runtime": 36.8885,
+      "eval_samples_per_second": 27.109,
+      "eval_steps_per_second": 1.139,
+      "step": 1020
+    },
+    {
+      "epoch": 1.118904109589041,
+      "grad_norm": 2.319324493408203,
+      "learning_rate": 3.5134037367993506e-06,
+      "loss": 1.8843,
+      "step": 1021
+    },
+    {
+      "epoch": 1.12,
+      "grad_norm": 2.238720655441284,
+      "learning_rate": 3.5113728675873275e-06,
+      "loss": 1.9375,
+      "step": 1022
+    },
+    {
+      "epoch": 1.121095890410959,
+      "grad_norm": 2.2824833393096924,
+      "learning_rate": 3.5093419983753047e-06,
+      "loss": 1.8749,
+      "step": 1023
+    },
+    {
+      "epoch": 1.1221917808219177,
+      "grad_norm": 2.5120770931243896,
+      "learning_rate": 3.507311129163282e-06,
+      "loss": 1.8774,
+      "step": 1024
+    },
+    {
+      "epoch": 1.1232876712328768,
+      "grad_norm": 2.1762094497680664,
+      "learning_rate": 3.5052802599512593e-06,
+      "loss": 1.8018,
+      "step": 1025
+    },
+    {
+      "epoch": 1.1243835616438356,
+      "grad_norm": 2.2530343532562256,
+      "learning_rate": 3.5032493907392366e-06,
+      "loss": 1.8451,
+      "step": 1026
+    },
+    {
+      "epoch": 1.1254794520547946,
+      "grad_norm": 2.21649169921875,
+      "learning_rate": 3.501218521527214e-06,
+      "loss": 1.8613,
+      "step": 1027
+    },
+    {
+      "epoch": 1.1265753424657534,
+      "grad_norm": 2.306431293487549,
+      "learning_rate": 3.499187652315191e-06,
+      "loss": 1.7947,
+      "step": 1028
+    },
+    {
+      "epoch": 1.1276712328767124,
+      "grad_norm": 2.203005313873291,
+      "learning_rate": 3.4971567831031684e-06,
+      "loss": 1.8483,
+      "step": 1029
+    },
+    {
+      "epoch": 1.1287671232876713,
+      "grad_norm": 2.1645009517669678,
+      "learning_rate": 3.4951259138911457e-06,
+      "loss": 1.8915,
+      "step": 1030
+    },
+    {
+      "epoch": 1.1287671232876713,
+      "eval_loss": 2.053393602371216,
+      "eval_runtime": 36.855,
+      "eval_samples_per_second": 27.133,
+      "eval_steps_per_second": 1.14,
+      "step": 1030
+    },
+    {
+      "epoch": 1.12986301369863,
+      "grad_norm": 2.2863385677337646,
+      "learning_rate": 3.493095044679123e-06,
+      "loss": 1.8073,
+      "step": 1031
+    },
+    {
+      "epoch": 1.130958904109589,
+      "grad_norm": 2.1439812183380127,
+      "learning_rate": 3.4910641754671002e-06,
+      "loss": 1.9191,
+      "step": 1032
+    },
+    {
+      "epoch": 1.132054794520548,
+      "grad_norm": 2.224228858947754,
+      "learning_rate": 3.4890333062550775e-06,
+      "loss": 1.7478,
+      "step": 1033
+    },
+    {
+      "epoch": 1.133150684931507,
+      "grad_norm": 2.319059133529663,
+      "learning_rate": 3.4870024370430548e-06,
+      "loss": 1.8779,
+      "step": 1034
+    },
+    {
+      "epoch": 1.1342465753424658,
+      "grad_norm": 2.5003139972686768,
+      "learning_rate": 3.484971567831032e-06,
+      "loss": 1.776,
+      "step": 1035
+    },
+    {
+      "epoch": 1.1353424657534246,
+      "grad_norm": 2.2032580375671387,
+      "learning_rate": 3.482940698619009e-06,
+      "loss": 1.8221,
+      "step": 1036
+    },
+    {
+      "epoch": 1.1364383561643836,
+      "grad_norm": 2.5253560543060303,
+      "learning_rate": 3.4809098294069866e-06,
+      "loss": 1.8952,
+      "step": 1037
+    },
+    {
+      "epoch": 1.1375342465753424,
+      "grad_norm": 2.3215854167938232,
+      "learning_rate": 3.4788789601949634e-06,
+      "loss": 1.847,
+      "step": 1038
+    },
+    {
+      "epoch": 1.1386301369863014,
+      "grad_norm": 2.2543280124664307,
+      "learning_rate": 3.476848090982941e-06,
+      "loss": 1.8689,
+      "step": 1039
+    },
+    {
+      "epoch": 1.1397260273972603,
+      "grad_norm": 2.1592042446136475,
+      "learning_rate": 3.474817221770918e-06,
+      "loss": 1.8375,
+      "step": 1040
+    },
+    {
+      "epoch": 1.1397260273972603,
+      "eval_loss": 2.0524044036865234,
+      "eval_runtime": 36.8163,
+      "eval_samples_per_second": 27.162,
+      "eval_steps_per_second": 1.141,
+      "step": 1040
+    },
+    {
+      "epoch": 1.140821917808219,
+      "grad_norm": 2.2874810695648193,
+      "learning_rate": 3.4727863525588957e-06,
+      "loss": 1.9442,
+      "step": 1041
+    },
+    {
+      "epoch": 1.141917808219178,
+      "grad_norm": 2.288395404815674,
+      "learning_rate": 3.4707554833468725e-06,
+      "loss": 1.8881,
+      "step": 1042
+    },
+    {
+      "epoch": 1.143013698630137,
+      "grad_norm": 2.2264153957366943,
+      "learning_rate": 3.4687246141348502e-06,
+      "loss": 1.8518,
+      "step": 1043
+    },
+    {
+      "epoch": 1.144109589041096,
+      "grad_norm": 2.4413652420043945,
+      "learning_rate": 3.466693744922827e-06,
+      "loss": 1.8146,
+      "step": 1044
+    },
+    {
+      "epoch": 1.1452054794520548,
+      "grad_norm": 2.3597898483276367,
+      "learning_rate": 3.464662875710805e-06,
+      "loss": 1.8421,
+      "step": 1045
+    },
+    {
+      "epoch": 1.1463013698630138,
+      "grad_norm": 2.2489402294158936,
+      "learning_rate": 3.4626320064987817e-06,
+      "loss": 1.8643,
+      "step": 1046
+    },
+    {
+      "epoch": 1.1473972602739726,
+      "grad_norm": 2.3958940505981445,
+      "learning_rate": 3.4606011372867594e-06,
+      "loss": 1.8522,
+      "step": 1047
+    },
+    {
+      "epoch": 1.1484931506849314,
+      "grad_norm": 2.295349597930908,
+      "learning_rate": 3.458570268074736e-06,
+      "loss": 1.8677,
+      "step": 1048
+    },
+    {
+      "epoch": 1.1495890410958904,
+      "grad_norm": 2.2571280002593994,
+      "learning_rate": 3.4565393988627135e-06,
+      "loss": 1.8393,
+      "step": 1049
+    },
+    {
+      "epoch": 1.1506849315068493,
+      "grad_norm": 2.173983335494995,
+      "learning_rate": 3.4545085296506908e-06,
+      "loss": 1.8267,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1506849315068493,
+      "eval_loss": 2.0489604473114014,
+      "eval_runtime": 36.7678,
+      "eval_samples_per_second": 27.198,
+      "eval_steps_per_second": 1.142,
+      "step": 1050
+    },
+    {
+      "epoch": 1.1517808219178083,
+      "grad_norm": 2.3395209312438965,
+      "learning_rate": 3.452477660438668e-06,
+      "loss": 1.8032,
+      "step": 1051
+    },
+    {
+      "epoch": 1.152876712328767,
+      "grad_norm": 2.1894757747650146,
+      "learning_rate": 3.4504467912266453e-06,
+      "loss": 1.8834,
+      "step": 1052
+    },
+    {
+      "epoch": 1.1539726027397261,
+      "grad_norm": 2.1978437900543213,
+      "learning_rate": 3.4484159220146226e-06,
+      "loss": 1.7897,
+      "step": 1053
+    },
+    {
+      "epoch": 1.155068493150685,
+      "grad_norm": 2.275792360305786,
+      "learning_rate": 3.4463850528026e-06,
+      "loss": 1.8566,
+      "step": 1054
+    },
+    {
+      "epoch": 1.1561643835616437,
+      "grad_norm": 2.2659800052642822,
+      "learning_rate": 3.444354183590577e-06,
+      "loss": 1.8307,
+      "step": 1055
+    },
+    {
+      "epoch": 1.1572602739726028,
+      "grad_norm": 2.232149600982666,
+      "learning_rate": 3.4423233143785544e-06,
+      "loss": 1.8701,
+      "step": 1056
+    },
+    {
+      "epoch": 1.1583561643835616,
+      "grad_norm": 2.3636281490325928,
+      "learning_rate": 3.4402924451665317e-06,
+      "loss": 1.8008,
+      "step": 1057
+    },
+    {
+      "epoch": 1.1594520547945206,
+      "grad_norm": 2.271533489227295,
+      "learning_rate": 3.438261575954509e-06,
+      "loss": 1.7751,
+      "step": 1058
+    },
+    {
+      "epoch": 1.1605479452054794,
+      "grad_norm": 2.3286566734313965,
+      "learning_rate": 3.4362307067424862e-06,
+      "loss": 1.865,
+      "step": 1059
+    },
+    {
+      "epoch": 1.1616438356164385,
+      "grad_norm": 2.46665096282959,
+      "learning_rate": 3.434199837530463e-06,
+      "loss": 1.9001,
+      "step": 1060
+    },
+    {
+      "epoch": 1.1616438356164385,
+      "eval_loss": 2.0477516651153564,
+      "eval_runtime": 36.7818,
+      "eval_samples_per_second": 27.187,
+      "eval_steps_per_second": 1.142,
+      "step": 1060
+    },
+    {
+      "epoch": 1.1627397260273973,
+      "grad_norm": 2.307744264602661,
+      "learning_rate": 3.432168968318441e-06,
+      "loss": 1.8637,
+      "step": 1061
+    },
+    {
+      "epoch": 1.163835616438356,
+      "grad_norm": 2.3935985565185547,
+      "learning_rate": 3.4301380991064176e-06,
+      "loss": 1.7867,
+      "step": 1062
+    },
+    {
+      "epoch": 1.1649315068493151,
+      "grad_norm": 2.290971040725708,
+      "learning_rate": 3.4281072298943953e-06,
+      "loss": 1.8708,
+      "step": 1063
+    },
+    {
+      "epoch": 1.166027397260274,
+      "grad_norm": 2.191629409790039,
+      "learning_rate": 3.426076360682372e-06,
+      "loss": 1.9031,
+      "step": 1064
+    },
+    {
+      "epoch": 1.167123287671233,
+      "grad_norm": 2.284426689147949,
+      "learning_rate": 3.42404549147035e-06,
+      "loss": 1.9026,
+      "step": 1065
+    },
+    {
+      "epoch": 1.1682191780821918,
+      "grad_norm": 2.2567849159240723,
+      "learning_rate": 3.4220146222583267e-06,
+      "loss": 1.8507,
+      "step": 1066
+    },
+    {
+      "epoch": 1.1693150684931506,
+      "grad_norm": 2.2093160152435303,
+      "learning_rate": 3.4199837530463044e-06,
+      "loss": 1.8001,
+      "step": 1067
+    },
+    {
+      "epoch": 1.1704109589041096,
+      "grad_norm": 2.092560291290283,
+      "learning_rate": 3.4179528838342813e-06,
+      "loss": 1.8393,
+      "step": 1068
+    },
+    {
+      "epoch": 1.1715068493150684,
+      "grad_norm": 2.313749074935913,
+      "learning_rate": 3.415922014622259e-06,
+      "loss": 1.9385,
+      "step": 1069
+    },
+    {
+      "epoch": 1.1726027397260275,
+      "grad_norm": 2.488673210144043,
+      "learning_rate": 3.413891145410236e-06,
+      "loss": 1.8924,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1726027397260275,
+      "eval_loss": 2.0482401847839355,
+      "eval_runtime": 36.7919,
+      "eval_samples_per_second": 27.18,
+      "eval_steps_per_second": 1.142,
+      "step": 1070
+    },
+    {
+      "epoch": 1.1736986301369863,
+      "grad_norm": 2.0804412364959717,
+      "learning_rate": 3.4118602761982135e-06,
+      "loss": 1.881,
+      "step": 1071
+    },
+    {
+      "epoch": 1.174794520547945,
+      "grad_norm": 2.165426254272461,
+      "learning_rate": 3.4098294069861904e-06,
+      "loss": 1.9035,
+      "step": 1072
+    },
+    {
+      "epoch": 1.1758904109589041,
+      "grad_norm": 2.3034675121307373,
+      "learning_rate": 3.4077985377741673e-06,
+      "loss": 1.8554,
+      "step": 1073
+    },
+    {
+      "epoch": 1.176986301369863,
+      "grad_norm": 2.303159713745117,
+      "learning_rate": 3.405767668562145e-06,
+      "loss": 1.8255,
+      "step": 1074
+    },
+    {
+      "epoch": 1.178082191780822,
+      "grad_norm": 2.2402753829956055,
+      "learning_rate": 3.403736799350122e-06,
+      "loss": 1.9087,
+      "step": 1075
+    },
+    {
+      "epoch": 1.1791780821917808,
+      "grad_norm": 2.251638650894165,
+      "learning_rate": 3.4017059301380995e-06,
+      "loss": 1.8878,
+      "step": 1076
+    },
+    {
+      "epoch": 1.1802739726027398,
+      "grad_norm": 2.2661657333374023,
+      "learning_rate": 3.3996750609260764e-06,
+      "loss": 1.8244,
+      "step": 1077
+    },
+    {
+      "epoch": 1.1813698630136986,
+      "grad_norm": 2.3280904293060303,
+      "learning_rate": 3.397644191714054e-06,
+      "loss": 1.9328,
+      "step": 1078
+    },
+    {
+      "epoch": 1.1824657534246574,
+      "grad_norm": 2.315779447555542,
+      "learning_rate": 3.395613322502031e-06,
+      "loss": 1.8289,
+      "step": 1079
+    },
+    {
+      "epoch": 1.1835616438356165,
+      "grad_norm": 2.136967182159424,
+      "learning_rate": 3.3935824532900086e-06,
+      "loss": 1.8307,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1835616438356165,
+      "eval_loss": 2.0473902225494385,
+      "eval_runtime": 36.8094,
+      "eval_samples_per_second": 27.167,
+      "eval_steps_per_second": 1.141,
+      "step": 1080
+    },
+    {
+      "epoch": 1.1846575342465753,
+      "grad_norm": 2.3701720237731934,
+      "learning_rate": 3.3915515840779855e-06,
+      "loss": 1.8123,
+      "step": 1081
+    },
+    {
+      "epoch": 1.1857534246575343,
+      "grad_norm": 2.2662320137023926,
+      "learning_rate": 3.389520714865963e-06,
+      "loss": 1.85,
+      "step": 1082
+    },
+    {
+      "epoch": 1.1868493150684931,
+      "grad_norm": 2.2151782512664795,
+      "learning_rate": 3.38748984565394e-06,
+      "loss": 1.845,
+      "step": 1083
+    },
+    {
+      "epoch": 1.1879452054794521,
+      "grad_norm": 2.2209115028381348,
+      "learning_rate": 3.3854589764419177e-06,
+      "loss": 1.8787,
+      "step": 1084
+    },
+    {
+      "epoch": 1.189041095890411,
+      "grad_norm": 2.4760477542877197,
+      "learning_rate": 3.3834281072298946e-06,
+      "loss": 1.818,
+      "step": 1085
+    },
+    {
+      "epoch": 1.1901369863013698,
+      "grad_norm": 2.309034585952759,
+      "learning_rate": 3.381397238017872e-06,
+      "loss": 1.8387,
+      "step": 1086
+    },
+    {
+      "epoch": 1.1912328767123288,
+      "grad_norm": 2.2174596786499023,
+      "learning_rate": 3.379366368805849e-06,
+      "loss": 1.8135,
+      "step": 1087
+    },
+    {
+      "epoch": 1.1923287671232876,
+      "grad_norm": 2.3043556213378906,
+      "learning_rate": 3.3773354995938264e-06,
+      "loss": 1.9156,
+      "step": 1088
+    },
+    {
+      "epoch": 1.1934246575342466,
+      "grad_norm": 2.2240524291992188,
+      "learning_rate": 3.3753046303818037e-06,
+      "loss": 1.8636,
+      "step": 1089
+    },
+    {
+      "epoch": 1.1945205479452055,
+      "grad_norm": 2.1108691692352295,
+      "learning_rate": 3.373273761169781e-06,
+      "loss": 1.8458,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1945205479452055,
+      "eval_loss": 2.04378604888916,
+      "eval_runtime": 36.9058,
+      "eval_samples_per_second": 27.096,
+      "eval_steps_per_second": 1.138,
+      "step": 1090
+    },
+    {
+      "epoch": 1.1956164383561645,
+      "grad_norm": 2.1893398761749268,
+      "learning_rate": 3.3712428919577582e-06,
+      "loss": 1.8446,
+      "step": 1091
+    },
+    {
+      "epoch": 1.1967123287671233,
+      "grad_norm": 2.305298089981079,
+      "learning_rate": 3.3692120227457355e-06,
+      "loss": 1.8283,
+      "step": 1092
+    },
+    {
+      "epoch": 1.1978082191780821,
+      "grad_norm": 2.364433765411377,
+      "learning_rate": 3.3671811535337128e-06,
+      "loss": 1.8708,
+      "step": 1093
+    },
+    {
+      "epoch": 1.1989041095890411,
+      "grad_norm": 2.259892225265503,
+      "learning_rate": 3.36515028432169e-06,
+      "loss": 1.7966,
+      "step": 1094
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 2.2487263679504395,
+      "learning_rate": 3.3631194151096673e-06,
+      "loss": 1.8238,
+      "step": 1095
+    },
+    {
+      "epoch": 1.201095890410959,
+      "grad_norm": 2.1592557430267334,
+      "learning_rate": 3.3610885458976446e-06,
+      "loss": 1.8651,
+      "step": 1096
+    },
+    {
+      "epoch": 1.2021917808219178,
+      "grad_norm": 2.257737159729004,
+      "learning_rate": 3.359057676685622e-06,
+      "loss": 1.8926,
+      "step": 1097
+    },
+    {
+      "epoch": 1.2032876712328768,
+      "grad_norm": 2.0702707767486572,
+      "learning_rate": 3.357026807473599e-06,
+      "loss": 1.8833,
+      "step": 1098
+    },
+    {
+      "epoch": 1.2043835616438356,
+      "grad_norm": 2.1568455696105957,
+      "learning_rate": 3.354995938261576e-06,
+      "loss": 1.7732,
+      "step": 1099
+    },
+    {
+      "epoch": 1.2054794520547945,
+      "grad_norm": 2.294715404510498,
+      "learning_rate": 3.3529650690495537e-06,
+      "loss": 1.8205,
+      "step": 1100
+    },
+    {
+      "epoch": 1.2054794520547945,
+      "eval_loss": 2.044201612472534,
+      "eval_runtime": 36.9861,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 1100
+    },
+    {
+      "epoch": 1.2065753424657535,
+      "grad_norm": 2.3234570026397705,
+      "learning_rate": 3.3509341998375305e-06,
+      "loss": 1.7935,
+      "step": 1101
+    },
+    {
+      "epoch": 1.2076712328767123,
+      "grad_norm": 2.307224750518799,
+      "learning_rate": 3.3489033306255082e-06,
+      "loss": 1.8254,
+      "step": 1102
+    },
+    {
+      "epoch": 1.208767123287671,
+      "grad_norm": 2.3725457191467285,
+      "learning_rate": 3.346872461413485e-06,
+      "loss": 1.8493,
+      "step": 1103
+    },
+    {
+      "epoch": 1.2098630136986301,
+      "grad_norm": 2.2057435512542725,
+      "learning_rate": 3.344841592201463e-06,
+      "loss": 1.8241,
+      "step": 1104
+    },
+    {
+      "epoch": 1.210958904109589,
+      "grad_norm": 2.232947826385498,
+      "learning_rate": 3.3428107229894397e-06,
+      "loss": 1.8057,
+      "step": 1105
+    },
+    {
+      "epoch": 1.212054794520548,
+      "grad_norm": 2.4268925189971924,
+      "learning_rate": 3.3407798537774174e-06,
+      "loss": 1.8334,
+      "step": 1106
+    },
+    {
+      "epoch": 1.2131506849315068,
+      "grad_norm": 2.3134701251983643,
+      "learning_rate": 3.338748984565394e-06,
+      "loss": 1.8516,
+      "step": 1107
+    },
+    {
+      "epoch": 1.2142465753424658,
+      "grad_norm": 2.375873327255249,
+      "learning_rate": 3.336718115353372e-06,
+      "loss": 1.8942,
+      "step": 1108
+    },
+    {
+      "epoch": 1.2153424657534246,
+      "grad_norm": 2.622528076171875,
+      "learning_rate": 3.3346872461413488e-06,
+      "loss": 1.9027,
+      "step": 1109
+    },
+    {
+      "epoch": 1.2164383561643834,
+      "grad_norm": 2.290865898132324,
+      "learning_rate": 3.3326563769293256e-06,
+      "loss": 1.8258,
+      "step": 1110
+    },
+    {
+      "epoch": 1.2164383561643834,
+      "eval_loss": 2.0408973693847656,
+      "eval_runtime": 36.8596,
+      "eval_samples_per_second": 27.13,
+      "eval_steps_per_second": 1.139,
+      "step": 1110
+    },
+    {
+      "epoch": 1.2175342465753425,
+      "grad_norm": 2.1596908569335938,
+      "learning_rate": 3.3306255077173033e-06,
+      "loss": 1.8066,
+      "step": 1111
+    },
+    {
+      "epoch": 1.2186301369863013,
+      "grad_norm": 2.4652187824249268,
+      "learning_rate": 3.32859463850528e-06,
+      "loss": 1.8372,
+      "step": 1112
+    },
+    {
+      "epoch": 1.2197260273972603,
+      "grad_norm": 2.282440185546875,
+      "learning_rate": 3.326563769293258e-06,
+      "loss": 1.8535,
+      "step": 1113
+    },
+    {
+      "epoch": 1.2208219178082191,
+      "grad_norm": 2.2390308380126953,
+      "learning_rate": 3.3245329000812347e-06,
+      "loss": 1.7719,
+      "step": 1114
+    },
+    {
+      "epoch": 1.2219178082191782,
+      "grad_norm": 2.368565320968628,
+      "learning_rate": 3.3225020308692124e-06,
+      "loss": 1.8207,
+      "step": 1115
+    },
+    {
+      "epoch": 1.223013698630137,
+      "grad_norm": 2.463942766189575,
+      "learning_rate": 3.3204711616571893e-06,
+      "loss": 1.8402,
+      "step": 1116
+    },
+    {
+      "epoch": 1.2241095890410958,
+      "grad_norm": 2.266096830368042,
+      "learning_rate": 3.318440292445167e-06,
+      "loss": 1.7854,
+      "step": 1117
+    },
+    {
+      "epoch": 1.2252054794520548,
+      "grad_norm": 2.3463494777679443,
+      "learning_rate": 3.316409423233144e-06,
+      "loss": 1.8449,
+      "step": 1118
+    },
+    {
+      "epoch": 1.2263013698630136,
+      "grad_norm": 2.381727933883667,
+      "learning_rate": 3.3143785540211215e-06,
+      "loss": 1.8085,
+      "step": 1119
+    },
+    {
+      "epoch": 1.2273972602739727,
+      "grad_norm": 2.3251113891601562,
+      "learning_rate": 3.3123476848090984e-06,
+      "loss": 1.8922,
+      "step": 1120
+    },
+    {
+      "epoch": 1.2273972602739727,
+      "eval_loss": 2.0425429344177246,
+      "eval_runtime": 36.851,
+      "eval_samples_per_second": 27.136,
+      "eval_steps_per_second": 1.14,
+      "step": 1120
+    },
+    {
+      "epoch": 1.2284931506849315,
+      "grad_norm": 2.2314298152923584,
+      "learning_rate": 3.310316815597076e-06,
+      "loss": 1.8474,
+      "step": 1121
+    },
+    {
+      "epoch": 1.2295890410958905,
+      "grad_norm": 2.1847472190856934,
+      "learning_rate": 3.308285946385053e-06,
+      "loss": 1.8454,
+      "step": 1122
+    },
+    {
+      "epoch": 1.2306849315068493,
+      "grad_norm": 2.249422788619995,
+      "learning_rate": 3.30625507717303e-06,
+      "loss": 1.9257,
+      "step": 1123
+    },
+    {
+      "epoch": 1.2317808219178081,
+      "grad_norm": 2.223930835723877,
+      "learning_rate": 3.3042242079610075e-06,
+      "loss": 1.7816,
+      "step": 1124
+    },
+    {
+      "epoch": 1.2328767123287672,
+      "grad_norm": 2.3209826946258545,
+      "learning_rate": 3.3021933387489847e-06,
+      "loss": 1.8773,
+      "step": 1125
+    },
+    {
+      "epoch": 1.233972602739726,
+      "grad_norm": 2.2288055419921875,
+      "learning_rate": 3.300162469536962e-06,
+      "loss": 1.8931,
+      "step": 1126
+    },
+    {
+      "epoch": 1.235068493150685,
+      "grad_norm": 2.201449155807495,
+      "learning_rate": 3.2981316003249393e-06,
+      "loss": 1.8149,
+      "step": 1127
+    },
+    {
+      "epoch": 1.2361643835616438,
+      "grad_norm": 2.215824842453003,
+      "learning_rate": 3.2961007311129166e-06,
+      "loss": 1.8949,
+      "step": 1128
+    },
+    {
+      "epoch": 1.2372602739726029,
+      "grad_norm": 2.133671522140503,
+      "learning_rate": 3.294069861900894e-06,
+      "loss": 1.8669,
+      "step": 1129
+    },
+    {
+      "epoch": 1.2383561643835617,
+      "grad_norm": 2.2195792198181152,
+      "learning_rate": 3.292038992688871e-06,
+      "loss": 1.9245,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2383561643835617,
+      "eval_loss": 2.0442771911621094,
+      "eval_runtime": 36.8077,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 1130
+    },
+    {
+      "epoch": 1.2394520547945205,
+      "grad_norm": 2.310866117477417,
+      "learning_rate": 3.2900081234768484e-06,
+      "loss": 1.8523,
+      "step": 1131
+    },
+    {
+      "epoch": 1.2405479452054795,
+      "grad_norm": 2.318300247192383,
+      "learning_rate": 3.2879772542648257e-06,
+      "loss": 1.8665,
+      "step": 1132
+    },
+    {
+      "epoch": 1.2416438356164383,
+      "grad_norm": 2.082564115524292,
+      "learning_rate": 3.285946385052803e-06,
+      "loss": 1.8962,
+      "step": 1133
+    },
+    {
+      "epoch": 1.2427397260273974,
+      "grad_norm": 2.1982104778289795,
+      "learning_rate": 3.2839155158407802e-06,
+      "loss": 1.7722,
+      "step": 1134
+    },
+    {
+      "epoch": 1.2438356164383562,
+      "grad_norm": 2.330782890319824,
+      "learning_rate": 3.2818846466287575e-06,
+      "loss": 1.8787,
+      "step": 1135
+    },
+    {
+      "epoch": 1.244931506849315,
+      "grad_norm": 2.136747121810913,
+      "learning_rate": 3.2798537774167344e-06,
+      "loss": 1.7739,
+      "step": 1136
+    },
+    {
+      "epoch": 1.246027397260274,
+      "grad_norm": 2.205862283706665,
+      "learning_rate": 3.277822908204712e-06,
+      "loss": 1.9127,
+      "step": 1137
+    },
+    {
+      "epoch": 1.2471232876712328,
+      "grad_norm": 2.2060093879699707,
+      "learning_rate": 3.275792038992689e-06,
+      "loss": 1.8379,
+      "step": 1138
+    },
+    {
+      "epoch": 1.2482191780821918,
+      "grad_norm": 2.255772829055786,
+      "learning_rate": 3.2737611697806666e-06,
+      "loss": 1.7931,
+      "step": 1139
+    },
+    {
+      "epoch": 1.2493150684931507,
+      "grad_norm": 2.240095615386963,
+      "learning_rate": 3.2717303005686435e-06,
+      "loss": 1.8466,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2493150684931507,
+      "eval_loss": 2.0429375171661377,
+      "eval_runtime": 36.7983,
+      "eval_samples_per_second": 27.175,
+      "eval_steps_per_second": 1.141,
+      "step": 1140
+    },
+    {
+      "epoch": 1.2504109589041095,
+      "grad_norm": 2.1491949558258057,
+      "learning_rate": 3.269699431356621e-06,
+      "loss": 1.8137,
+      "step": 1141
+    },
+    {
+      "epoch": 1.2515068493150685,
+      "grad_norm": 2.3854610919952393,
+      "learning_rate": 3.267668562144598e-06,
+      "loss": 1.8486,
+      "step": 1142
+    },
+    {
+      "epoch": 1.2526027397260273,
+      "grad_norm": 2.3498709201812744,
+      "learning_rate": 3.2656376929325757e-06,
+      "loss": 1.837,
+      "step": 1143
+    },
+    {
+      "epoch": 1.2536986301369863,
+      "grad_norm": 2.1651649475097656,
+      "learning_rate": 3.2636068237205526e-06,
+      "loss": 1.8,
+      "step": 1144
+    },
+    {
+      "epoch": 1.2547945205479452,
+      "grad_norm": 2.2500200271606445,
+      "learning_rate": 3.2615759545085303e-06,
+      "loss": 1.8691,
+      "step": 1145
+    },
+    {
+      "epoch": 1.2558904109589042,
+      "grad_norm": 2.2178192138671875,
+      "learning_rate": 3.259545085296507e-06,
+      "loss": 1.8345,
+      "step": 1146
+    },
+    {
+      "epoch": 1.256986301369863,
+      "grad_norm": 2.238114595413208,
+      "learning_rate": 3.257514216084484e-06,
+      "loss": 1.8935,
+      "step": 1147
+    },
+    {
+      "epoch": 1.2580821917808218,
+      "grad_norm": 2.37042498588562,
+      "learning_rate": 3.2554833468724617e-06,
+      "loss": 1.7872,
+      "step": 1148
+    },
+    {
+      "epoch": 1.2591780821917808,
+      "grad_norm": 2.1831908226013184,
+      "learning_rate": 3.2534524776604385e-06,
+      "loss": 1.8913,
+      "step": 1149
+    },
+    {
+      "epoch": 1.2602739726027397,
+      "grad_norm": 2.251382827758789,
+      "learning_rate": 3.2514216084484162e-06,
+      "loss": 1.8611,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2602739726027397,
+      "eval_loss": 2.041306257247925,
+      "eval_runtime": 36.8199,
+      "eval_samples_per_second": 27.159,
+      "eval_steps_per_second": 1.141,
+      "step": 1150
+    },
+    {
+      "epoch": 1.2613698630136987,
+      "grad_norm": 2.2690420150756836,
+      "learning_rate": 3.249390739236393e-06,
+      "loss": 1.8283,
+      "step": 1151
+    },
+    {
+      "epoch": 1.2624657534246575,
+      "grad_norm": 2.255833387374878,
+      "learning_rate": 3.2473598700243708e-06,
+      "loss": 1.8563,
+      "step": 1152
+    },
+    {
+      "epoch": 1.2635616438356165,
+      "grad_norm": 2.1362855434417725,
+      "learning_rate": 3.2453290008123476e-06,
+      "loss": 1.809,
+      "step": 1153
+    },
+    {
+      "epoch": 1.2646575342465753,
+      "grad_norm": 2.1227219104766846,
+      "learning_rate": 3.2432981316003253e-06,
+      "loss": 1.848,
+      "step": 1154
+    },
+    {
+      "epoch": 1.2657534246575342,
+      "grad_norm": 2.3639719486236572,
+      "learning_rate": 3.241267262388302e-06,
+      "loss": 1.8472,
+      "step": 1155
+    },
+    {
+      "epoch": 1.2668493150684932,
+      "grad_norm": 2.334404468536377,
+      "learning_rate": 3.23923639317628e-06,
+      "loss": 1.8073,
+      "step": 1156
+    },
+    {
+      "epoch": 1.267945205479452,
+      "grad_norm": 2.231656312942505,
+      "learning_rate": 3.2372055239642567e-06,
+      "loss": 1.8446,
+      "step": 1157
+    },
+    {
+      "epoch": 1.269041095890411,
+      "grad_norm": 2.371375560760498,
+      "learning_rate": 3.2351746547522344e-06,
+      "loss": 1.8002,
+      "step": 1158
+    },
+    {
+      "epoch": 1.2701369863013698,
+      "grad_norm": 2.5212879180908203,
+      "learning_rate": 3.2331437855402113e-06,
+      "loss": 1.863,
+      "step": 1159
+    },
+    {
+      "epoch": 1.2712328767123289,
+      "grad_norm": 2.3285911083221436,
+      "learning_rate": 3.2311129163281885e-06,
+      "loss": 1.9101,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2712328767123289,
+      "eval_loss": 2.038487195968628,
+      "eval_runtime": 36.8002,
+      "eval_samples_per_second": 27.174,
+      "eval_steps_per_second": 1.141,
+      "step": 1160
+    },
+    {
+      "epoch": 1.2723287671232877,
+      "grad_norm": 2.1316046714782715,
+      "learning_rate": 3.229082047116166e-06,
+      "loss": 1.7952,
+      "step": 1161
+    },
+    {
+      "epoch": 1.2734246575342465,
+      "grad_norm": 2.4078125953674316,
+      "learning_rate": 3.227051177904143e-06,
+      "loss": 1.869,
+      "step": 1162
+    },
+    {
+      "epoch": 1.2745205479452055,
+      "grad_norm": 2.363672971725464,
+      "learning_rate": 3.2250203086921204e-06,
+      "loss": 1.8478,
+      "step": 1163
+    },
+    {
+      "epoch": 1.2756164383561643,
+      "grad_norm": 2.240558385848999,
+      "learning_rate": 3.2229894394800977e-06,
+      "loss": 1.8483,
+      "step": 1164
+    },
+    {
+      "epoch": 1.2767123287671232,
+      "grad_norm": 2.9142000675201416,
+      "learning_rate": 3.220958570268075e-06,
+      "loss": 1.8789,
+      "step": 1165
+    },
+    {
+      "epoch": 1.2778082191780822,
+      "grad_norm": 2.1930558681488037,
+      "learning_rate": 3.218927701056052e-06,
+      "loss": 1.7659,
+      "step": 1166
+    },
+    {
+      "epoch": 1.2789041095890412,
+      "grad_norm": 2.312471866607666,
+      "learning_rate": 3.2168968318440295e-06,
+      "loss": 1.8374,
+      "step": 1167
+    },
+    {
+      "epoch": 1.28,
+      "grad_norm": 2.6222236156463623,
+      "learning_rate": 3.2148659626320068e-06,
+      "loss": 1.7539,
+      "step": 1168
+    },
+    {
+      "epoch": 1.2810958904109588,
+      "grad_norm": 2.3013322353363037,
+      "learning_rate": 3.212835093419984e-06,
+      "loss": 1.9011,
+      "step": 1169
+    },
+    {
+      "epoch": 1.2821917808219179,
+      "grad_norm": 2.215240716934204,
+      "learning_rate": 3.2108042242079613e-06,
+      "loss": 1.8032,
+      "step": 1170
+    },
+    {
+      "epoch": 1.2821917808219179,
+      "eval_loss": 2.0365872383117676,
+      "eval_runtime": 36.8265,
+      "eval_samples_per_second": 27.154,
+      "eval_steps_per_second": 1.14,
+      "step": 1170
+    },
+    {
+      "epoch": 1.2832876712328767,
+      "grad_norm": 2.3620481491088867,
+      "learning_rate": 3.2087733549959386e-06,
+      "loss": 1.8258,
+      "step": 1171
+    },
+    {
+      "epoch": 1.2843835616438355,
+      "grad_norm": 2.2650208473205566,
+      "learning_rate": 3.206742485783916e-06,
+      "loss": 1.8888,
+      "step": 1172
+    },
+    {
+      "epoch": 1.2854794520547945,
+      "grad_norm": 2.2564077377319336,
+      "learning_rate": 3.2047116165718927e-06,
+      "loss": 1.7874,
+      "step": 1173
+    },
+    {
+      "epoch": 1.2865753424657536,
+      "grad_norm": 2.2855281829833984,
+      "learning_rate": 3.2026807473598704e-06,
+      "loss": 1.8003,
+      "step": 1174
+    },
+    {
+      "epoch": 1.2876712328767124,
+      "grad_norm": 2.353893756866455,
+      "learning_rate": 3.2006498781478473e-06,
+      "loss": 1.8466,
+      "step": 1175
+    },
+    {
+      "epoch": 1.2887671232876712,
+      "grad_norm": 2.1893045902252197,
+      "learning_rate": 3.198619008935825e-06,
+      "loss": 1.8652,
+      "step": 1176
+    },
+    {
+      "epoch": 1.2898630136986302,
+      "grad_norm": 2.6629793643951416,
+      "learning_rate": 3.196588139723802e-06,
+      "loss": 1.797,
+      "step": 1177
+    },
+    {
+      "epoch": 1.290958904109589,
+      "grad_norm": 2.2427361011505127,
+      "learning_rate": 3.1945572705117795e-06,
+      "loss": 1.8294,
+      "step": 1178
+    },
+    {
+      "epoch": 1.2920547945205478,
+      "grad_norm": 2.321878671646118,
+      "learning_rate": 3.1925264012997564e-06,
+      "loss": 1.8494,
+      "step": 1179
+    },
+    {
+      "epoch": 1.2931506849315069,
+      "grad_norm": 2.6928281784057617,
+      "learning_rate": 3.190495532087734e-06,
+      "loss": 1.8574,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2931506849315069,
+      "eval_loss": 2.0426652431488037,
+      "eval_runtime": 36.8482,
+      "eval_samples_per_second": 27.138,
+      "eval_steps_per_second": 1.14,
+      "step": 1180
+    },
+    {
+      "epoch": 1.2942465753424657,
+      "grad_norm": 2.382164239883423,
+      "learning_rate": 3.188464662875711e-06,
+      "loss": 1.8518,
+      "step": 1181
+    },
+    {
+      "epoch": 1.2953424657534247,
+      "grad_norm": 2.4109256267547607,
+      "learning_rate": 3.1864337936636886e-06,
+      "loss": 1.9083,
+      "step": 1182
+    },
+    {
+      "epoch": 1.2964383561643835,
+      "grad_norm": 2.468062400817871,
+      "learning_rate": 3.1844029244516655e-06,
+      "loss": 1.8247,
+      "step": 1183
+    },
+    {
+      "epoch": 1.2975342465753426,
+      "grad_norm": 2.459378242492676,
+      "learning_rate": 3.182372055239643e-06,
+      "loss": 1.791,
+      "step": 1184
+    },
+    {
+      "epoch": 1.2986301369863014,
+      "grad_norm": 2.3331072330474854,
+      "learning_rate": 3.18034118602762e-06,
+      "loss": 1.7869,
+      "step": 1185
+    },
+    {
+      "epoch": 1.2997260273972602,
+      "grad_norm": 2.339428424835205,
+      "learning_rate": 3.1783103168155973e-06,
+      "loss": 1.8729,
+      "step": 1186
+    },
+    {
+      "epoch": 1.3008219178082192,
+      "grad_norm": 2.3033406734466553,
+      "learning_rate": 3.1762794476035746e-06,
+      "loss": 1.8455,
+      "step": 1187
+    },
+    {
+      "epoch": 1.301917808219178,
+      "grad_norm": 2.2480714321136475,
+      "learning_rate": 3.174248578391552e-06,
+      "loss": 1.8233,
+      "step": 1188
+    },
+    {
+      "epoch": 1.303013698630137,
+      "grad_norm": 2.2953624725341797,
+      "learning_rate": 3.172217709179529e-06,
+      "loss": 1.8261,
+      "step": 1189
+    },
+    {
+      "epoch": 1.3041095890410959,
+      "grad_norm": 2.494307518005371,
+      "learning_rate": 3.1701868399675064e-06,
+      "loss": 1.8153,
+      "step": 1190
+    },
+    {
+      "epoch": 1.3041095890410959,
+      "eval_loss": 2.037719964981079,
+      "eval_runtime": 36.9788,
+      "eval_samples_per_second": 27.043,
+      "eval_steps_per_second": 1.136,
+      "step": 1190
+    },
+    {
+      "epoch": 1.305205479452055,
+      "grad_norm": 2.314159631729126,
+      "learning_rate": 3.1681559707554837e-06,
+      "loss": 1.8406,
+      "step": 1191
+    },
+    {
+      "epoch": 1.3063013698630137,
+      "grad_norm": 2.4262588024139404,
+      "learning_rate": 3.166125101543461e-06,
+      "loss": 1.7803,
+      "step": 1192
+    },
+    {
+      "epoch": 1.3073972602739725,
+      "grad_norm": 2.3800835609436035,
+      "learning_rate": 3.1640942323314382e-06,
+      "loss": 1.8521,
+      "step": 1193
+    },
+    {
+      "epoch": 1.3084931506849315,
+      "grad_norm": 2.234551429748535,
+      "learning_rate": 3.1620633631194155e-06,
+      "loss": 1.845,
+      "step": 1194
+    },
+    {
+      "epoch": 1.3095890410958904,
+      "grad_norm": 2.167635440826416,
+      "learning_rate": 3.1600324939073928e-06,
+      "loss": 1.789,
+      "step": 1195
+    },
+    {
+      "epoch": 1.3106849315068494,
+      "grad_norm": 2.320247173309326,
+      "learning_rate": 3.15800162469537e-06,
+      "loss": 1.8311,
+      "step": 1196
+    },
+    {
+      "epoch": 1.3117808219178082,
+      "grad_norm": 2.2715561389923096,
+      "learning_rate": 3.155970755483347e-06,
+      "loss": 1.8308,
+      "step": 1197
+    },
+    {
+      "epoch": 1.3128767123287672,
+      "grad_norm": 2.2535455226898193,
+      "learning_rate": 3.1539398862713246e-06,
+      "loss": 1.8358,
+      "step": 1198
+    },
+    {
+      "epoch": 1.313972602739726,
+      "grad_norm": 2.2399163246154785,
+      "learning_rate": 3.1519090170593015e-06,
+      "loss": 1.8513,
+      "step": 1199
+    },
+    {
+      "epoch": 1.3150684931506849,
+      "grad_norm": 2.2984132766723633,
+      "learning_rate": 3.149878147847279e-06,
+      "loss": 1.8377,
+      "step": 1200
+    },
+    {
+      "epoch": 1.3150684931506849,
+      "eval_loss": 2.0334064960479736,
+      "eval_runtime": 37.0346,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1200
+    },
+    {
+      "epoch": 1.316164383561644,
+      "grad_norm": 2.1296892166137695,
+      "learning_rate": 3.147847278635256e-06,
+      "loss": 1.9023,
+      "step": 1201
+    },
+    {
+      "epoch": 1.3172602739726027,
+      "grad_norm": 2.1034350395202637,
+      "learning_rate": 3.1458164094232337e-06,
+      "loss": 1.917,
+      "step": 1202
+    },
+    {
+      "epoch": 1.3183561643835615,
+      "grad_norm": 2.258863687515259,
+      "learning_rate": 3.1437855402112106e-06,
+      "loss": 1.8093,
+      "step": 1203
+    },
+    {
+      "epoch": 1.3194520547945205,
+      "grad_norm": 2.20119309425354,
+      "learning_rate": 3.1417546709991883e-06,
+      "loss": 1.7796,
+      "step": 1204
+    },
+    {
+      "epoch": 1.3205479452054796,
+      "grad_norm": 2.3005869388580322,
+      "learning_rate": 3.139723801787165e-06,
+      "loss": 1.8496,
+      "step": 1205
+    },
+    {
+      "epoch": 1.3216438356164384,
+      "grad_norm": 2.2009031772613525,
+      "learning_rate": 3.137692932575143e-06,
+      "loss": 1.8244,
+      "step": 1206
+    },
+    {
+      "epoch": 1.3227397260273972,
+      "grad_norm": 2.3467633724212646,
+      "learning_rate": 3.1356620633631197e-06,
+      "loss": 1.8803,
+      "step": 1207
+    },
+    {
+      "epoch": 1.3238356164383562,
+      "grad_norm": 2.318711519241333,
+      "learning_rate": 3.1336311941510974e-06,
+      "loss": 1.7999,
+      "step": 1208
+    },
+    {
+      "epoch": 1.324931506849315,
+      "grad_norm": 2.182602882385254,
+      "learning_rate": 3.1316003249390742e-06,
+      "loss": 1.836,
+      "step": 1209
+    },
+    {
+      "epoch": 1.3260273972602739,
+      "grad_norm": 2.217298984527588,
+      "learning_rate": 3.129569455727051e-06,
+      "loss": 1.8776,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3260273972602739,
+      "eval_loss": 2.033573627471924,
+      "eval_runtime": 36.8029,
+      "eval_samples_per_second": 27.172,
+      "eval_steps_per_second": 1.141,
+      "step": 1210
+    },
+    {
+      "epoch": 1.3271232876712329,
+      "grad_norm": 2.2814879417419434,
+      "learning_rate": 3.1275385865150288e-06,
+      "loss": 1.8739,
+      "step": 1211
+    },
+    {
+      "epoch": 1.3282191780821917,
+      "grad_norm": 2.1692471504211426,
+      "learning_rate": 3.1255077173030056e-06,
+      "loss": 1.8539,
+      "step": 1212
+    },
+    {
+      "epoch": 1.3293150684931507,
+      "grad_norm": 2.2472000122070312,
+      "learning_rate": 3.1234768480909833e-06,
+      "loss": 1.8527,
+      "step": 1213
+    },
+    {
+      "epoch": 1.3304109589041095,
+      "grad_norm": 2.1969852447509766,
+      "learning_rate": 3.12144597887896e-06,
+      "loss": 1.8429,
+      "step": 1214
+    },
+    {
+      "epoch": 1.3315068493150686,
+      "grad_norm": 2.141629219055176,
+      "learning_rate": 3.119415109666938e-06,
+      "loss": 1.868,
+      "step": 1215
+    },
+    {
+      "epoch": 1.3326027397260274,
+      "grad_norm": 2.282562494277954,
+      "learning_rate": 3.1173842404549147e-06,
+      "loss": 1.8913,
+      "step": 1216
+    },
+    {
+      "epoch": 1.3336986301369862,
+      "grad_norm": 2.2375850677490234,
+      "learning_rate": 3.1153533712428924e-06,
+      "loss": 1.8462,
+      "step": 1217
+    },
+    {
+      "epoch": 1.3347945205479452,
+      "grad_norm": 2.3529837131500244,
+      "learning_rate": 3.1133225020308693e-06,
+      "loss": 1.8445,
+      "step": 1218
+    },
+    {
+      "epoch": 1.335890410958904,
+      "grad_norm": 2.385857343673706,
+      "learning_rate": 3.111291632818847e-06,
+      "loss": 1.89,
+      "step": 1219
+    },
+    {
+      "epoch": 1.336986301369863,
+      "grad_norm": 2.301253080368042,
+      "learning_rate": 3.109260763606824e-06,
+      "loss": 1.8882,
+      "step": 1220
+    },
+    {
+      "epoch": 1.336986301369863,
+      "eval_loss": 2.040025234222412,
+      "eval_runtime": 36.839,
+      "eval_samples_per_second": 27.145,
+      "eval_steps_per_second": 1.14,
+      "step": 1220
+    },
+    {
+      "epoch": 1.3380821917808219,
+      "grad_norm": 2.3772335052490234,
+      "learning_rate": 3.1072298943948015e-06,
+      "loss": 1.8545,
+      "step": 1221
+    },
+    {
+      "epoch": 1.339178082191781,
+      "grad_norm": 2.387830972671509,
+      "learning_rate": 3.1051990251827784e-06,
+      "loss": 1.8653,
+      "step": 1222
+    },
+    {
+      "epoch": 1.3402739726027397,
+      "grad_norm": 2.4285190105438232,
+      "learning_rate": 3.1031681559707557e-06,
+      "loss": 1.8268,
+      "step": 1223
+    },
+    {
+      "epoch": 1.3413698630136985,
+      "grad_norm": 2.1892075538635254,
+      "learning_rate": 3.101137286758733e-06,
+      "loss": 1.8818,
+      "step": 1224
+    },
+    {
+      "epoch": 1.3424657534246576,
+      "grad_norm": 2.2510695457458496,
+      "learning_rate": 3.09910641754671e-06,
+      "loss": 1.879,
+      "step": 1225
+    },
+    {
+      "epoch": 1.3435616438356164,
+      "grad_norm": 2.505671977996826,
+      "learning_rate": 3.0970755483346875e-06,
+      "loss": 1.872,
+      "step": 1226
+    },
+    {
+      "epoch": 1.3446575342465754,
+      "grad_norm": 2.5315067768096924,
+      "learning_rate": 3.0950446791226648e-06,
+      "loss": 1.8651,
+      "step": 1227
+    },
+    {
+      "epoch": 1.3457534246575342,
+      "grad_norm": 2.223600387573242,
+      "learning_rate": 3.093013809910642e-06,
+      "loss": 1.9167,
+      "step": 1228
+    },
+    {
+      "epoch": 1.3468493150684933,
+      "grad_norm": 2.357802391052246,
+      "learning_rate": 3.0909829406986193e-06,
+      "loss": 1.8593,
+      "step": 1229
+    },
+    {
+      "epoch": 1.347945205479452,
+      "grad_norm": 2.3102364540100098,
+      "learning_rate": 3.0889520714865966e-06,
+      "loss": 1.8994,
+      "step": 1230
+    },
+    {
+      "epoch": 1.347945205479452,
+      "eval_loss": 2.0332930088043213,
+      "eval_runtime": 36.989,
+      "eval_samples_per_second": 27.035,
+      "eval_steps_per_second": 1.135,
+      "step": 1230
+    },
+    {
+      "epoch": 1.3490410958904109,
+      "grad_norm": 2.085080623626709,
+      "learning_rate": 3.086921202274574e-06,
+      "loss": 1.8741,
+      "step": 1231
+    },
+    {
+      "epoch": 1.35013698630137,
+      "grad_norm": 2.267305612564087,
+      "learning_rate": 3.084890333062551e-06,
+      "loss": 1.8426,
+      "step": 1232
+    },
+    {
+      "epoch": 1.3512328767123287,
+      "grad_norm": 2.181033134460449,
+      "learning_rate": 3.0828594638505284e-06,
+      "loss": 1.8199,
+      "step": 1233
+    },
+    {
+      "epoch": 1.3523287671232875,
+      "grad_norm": 2.288386821746826,
+      "learning_rate": 3.0808285946385057e-06,
+      "loss": 1.8469,
+      "step": 1234
+    },
+    {
+      "epoch": 1.3534246575342466,
+      "grad_norm": 2.1516127586364746,
+      "learning_rate": 3.078797725426483e-06,
+      "loss": 1.8056,
+      "step": 1235
+    },
+    {
+      "epoch": 1.3545205479452056,
+      "grad_norm": 2.2699198722839355,
+      "learning_rate": 3.07676685621446e-06,
+      "loss": 1.7535,
+      "step": 1236
+    },
+    {
+      "epoch": 1.3556164383561644,
+      "grad_norm": 2.193251132965088,
+      "learning_rate": 3.0747359870024375e-06,
+      "loss": 1.8407,
+      "step": 1237
+    },
+    {
+      "epoch": 1.3567123287671232,
+      "grad_norm": 2.2217774391174316,
+      "learning_rate": 3.0727051177904144e-06,
+      "loss": 1.8062,
+      "step": 1238
+    },
+    {
+      "epoch": 1.3578082191780823,
+      "grad_norm": 2.2376203536987305,
+      "learning_rate": 3.070674248578392e-06,
+      "loss": 1.829,
+      "step": 1239
+    },
+    {
+      "epoch": 1.358904109589041,
+      "grad_norm": 2.387559413909912,
+      "learning_rate": 3.068643379366369e-06,
+      "loss": 1.8459,
+      "step": 1240
+    },
+    {
+      "epoch": 1.358904109589041,
+      "eval_loss": 2.029298782348633,
+      "eval_runtime": 37.0491,
+      "eval_samples_per_second": 26.991,
+      "eval_steps_per_second": 1.134,
+      "step": 1240
+    },
+    {
+      "epoch": 1.3599999999999999,
+      "grad_norm": 2.1940178871154785,
+      "learning_rate": 3.0666125101543466e-06,
+      "loss": 1.831,
+      "step": 1241
+    },
+    {
+      "epoch": 1.361095890410959,
+      "grad_norm": 2.145514488220215,
+      "learning_rate": 3.0645816409423235e-06,
+      "loss": 1.7919,
+      "step": 1242
+    },
+    {
+      "epoch": 1.362191780821918,
+      "grad_norm": 2.319164276123047,
+      "learning_rate": 3.062550771730301e-06,
+      "loss": 1.8472,
+      "step": 1243
+    },
+    {
+      "epoch": 1.3632876712328768,
+      "grad_norm": 2.218986749649048,
+      "learning_rate": 3.060519902518278e-06,
+      "loss": 1.8184,
+      "step": 1244
+    },
+    {
+      "epoch": 1.3643835616438356,
+      "grad_norm": 2.3755338191986084,
+      "learning_rate": 3.0584890333062557e-06,
+      "loss": 1.8166,
+      "step": 1245
+    },
+    {
+      "epoch": 1.3654794520547946,
+      "grad_norm": 2.3186213970184326,
+      "learning_rate": 3.0564581640942326e-06,
+      "loss": 1.7599,
+      "step": 1246
+    },
+    {
+      "epoch": 1.3665753424657534,
+      "grad_norm": 2.193958282470703,
+      "learning_rate": 3.0544272948822094e-06,
+      "loss": 1.8404,
+      "step": 1247
+    },
+    {
+      "epoch": 1.3676712328767122,
+      "grad_norm": 2.40419602394104,
+      "learning_rate": 3.052396425670187e-06,
+      "loss": 1.821,
+      "step": 1248
+    },
+    {
+      "epoch": 1.3687671232876713,
+      "grad_norm": 2.445542573928833,
+      "learning_rate": 3.050365556458164e-06,
+      "loss": 1.8518,
+      "step": 1249
+    },
+    {
+      "epoch": 1.36986301369863,
+      "grad_norm": 2.2776567935943604,
+      "learning_rate": 3.0483346872461417e-06,
+      "loss": 1.8121,
+      "step": 1250
+    },
+    {
+      "epoch": 1.36986301369863,
+      "eval_loss": 2.030043125152588,
+      "eval_runtime": 37.0721,
+      "eval_samples_per_second": 26.974,
+      "eval_steps_per_second": 1.133,
+      "step": 1250
+    },
+    {
+      "epoch": 1.370958904109589,
+      "grad_norm": 2.420448064804077,
+      "learning_rate": 3.0463038180341185e-06,
+      "loss": 1.8168,
+      "step": 1251
+    },
+    {
+      "epoch": 1.372054794520548,
+      "grad_norm": 2.252065658569336,
+      "learning_rate": 3.0442729488220962e-06,
+      "loss": 1.8757,
+      "step": 1252
+    },
+    {
+      "epoch": 1.373150684931507,
+      "grad_norm": 2.4772560596466064,
+      "learning_rate": 3.042242079610073e-06,
+      "loss": 1.8711,
+      "step": 1253
+    },
+    {
+      "epoch": 1.3742465753424657,
+      "grad_norm": 2.3986129760742188,
+      "learning_rate": 3.0402112103980508e-06,
+      "loss": 1.8636,
+      "step": 1254
+    },
+    {
+      "epoch": 1.3753424657534246,
+      "grad_norm": 2.257274866104126,
+      "learning_rate": 3.0381803411860276e-06,
+      "loss": 1.7858,
+      "step": 1255
+    },
+    {
+      "epoch": 1.3764383561643836,
+      "grad_norm": 2.2373499870300293,
+      "learning_rate": 3.0361494719740053e-06,
+      "loss": 1.817,
+      "step": 1256
+    },
+    {
+      "epoch": 1.3775342465753424,
+      "grad_norm": 2.292388916015625,
+      "learning_rate": 3.034118602761982e-06,
+      "loss": 1.827,
+      "step": 1257
+    },
+    {
+      "epoch": 1.3786301369863014,
+      "grad_norm": 2.4618642330169678,
+      "learning_rate": 3.03208773354996e-06,
+      "loss": 1.8218,
+      "step": 1258
+    },
+    {
+      "epoch": 1.3797260273972602,
+      "grad_norm": 2.271958112716675,
+      "learning_rate": 3.0300568643379367e-06,
+      "loss": 1.8092,
+      "step": 1259
+    },
+    {
+      "epoch": 1.3808219178082193,
+      "grad_norm": 2.3990538120269775,
+      "learning_rate": 3.028025995125914e-06,
+      "loss": 1.8633,
+      "step": 1260
+    },
+    {
+      "epoch": 1.3808219178082193,
+      "eval_loss": 2.028458595275879,
+      "eval_runtime": 37.0659,
+      "eval_samples_per_second": 26.979,
+      "eval_steps_per_second": 1.133,
+      "step": 1260
+    },
+    {
+      "epoch": 1.381917808219178,
+      "grad_norm": 2.208803653717041,
+      "learning_rate": 3.0259951259138913e-06,
+      "loss": 1.8463,
+      "step": 1261
+    },
+    {
+      "epoch": 1.383013698630137,
+      "grad_norm": 2.3269028663635254,
+      "learning_rate": 3.0239642567018686e-06,
+      "loss": 1.7784,
+      "step": 1262
+    },
+    {
+      "epoch": 1.384109589041096,
+      "grad_norm": 2.367910623550415,
+      "learning_rate": 3.021933387489846e-06,
+      "loss": 1.8281,
+      "step": 1263
+    },
+    {
+      "epoch": 1.3852054794520547,
+      "grad_norm": 2.273686647415161,
+      "learning_rate": 3.019902518277823e-06,
+      "loss": 1.7832,
+      "step": 1264
+    },
+    {
+      "epoch": 1.3863013698630138,
+      "grad_norm": 2.421964168548584,
+      "learning_rate": 3.0178716490658004e-06,
+      "loss": 1.8407,
+      "step": 1265
+    },
+    {
+      "epoch": 1.3873972602739726,
+      "grad_norm": 2.2627696990966797,
+      "learning_rate": 3.0158407798537777e-06,
+      "loss": 1.8718,
+      "step": 1266
+    },
+    {
+      "epoch": 1.3884931506849316,
+      "grad_norm": 2.1981618404388428,
+      "learning_rate": 3.013809910641755e-06,
+      "loss": 1.7763,
+      "step": 1267
+    },
+    {
+      "epoch": 1.3895890410958904,
+      "grad_norm": 2.245253562927246,
+      "learning_rate": 3.0117790414297322e-06,
+      "loss": 1.7915,
+      "step": 1268
+    },
+    {
+      "epoch": 1.3906849315068492,
+      "grad_norm": 2.2292444705963135,
+      "learning_rate": 3.0097481722177095e-06,
+      "loss": 1.7909,
+      "step": 1269
+    },
+    {
+      "epoch": 1.3917808219178083,
+      "grad_norm": 2.2290430068969727,
+      "learning_rate": 3.0077173030056868e-06,
+      "loss": 1.8364,
+      "step": 1270
+    },
+    {
+      "epoch": 1.3917808219178083,
+      "eval_loss": 2.0290327072143555,
+      "eval_runtime": 37.0666,
+      "eval_samples_per_second": 26.978,
+      "eval_steps_per_second": 1.133,
+      "step": 1270
+    },
+    {
+      "epoch": 1.392876712328767,
+      "grad_norm": 2.220881700515747,
+      "learning_rate": 3.005686433793664e-06,
+      "loss": 1.7968,
+      "step": 1271
+    },
+    {
+      "epoch": 1.393972602739726,
+      "grad_norm": 2.375332832336426,
+      "learning_rate": 3.0036555645816413e-06,
+      "loss": 1.8502,
+      "step": 1272
+    },
+    {
+      "epoch": 1.395068493150685,
+      "grad_norm": 2.3065083026885986,
+      "learning_rate": 3.001624695369618e-06,
+      "loss": 1.8116,
+      "step": 1273
+    },
+    {
+      "epoch": 1.396164383561644,
+      "grad_norm": 2.7262511253356934,
+      "learning_rate": 2.999593826157596e-06,
+      "loss": 1.8641,
+      "step": 1274
+    },
+    {
+      "epoch": 1.3972602739726028,
+      "grad_norm": 2.3381216526031494,
+      "learning_rate": 2.9975629569455727e-06,
+      "loss": 1.8128,
+      "step": 1275
+    },
+    {
+      "epoch": 1.3983561643835616,
+      "grad_norm": 2.1994192600250244,
+      "learning_rate": 2.9955320877335504e-06,
+      "loss": 1.8563,
+      "step": 1276
+    },
+    {
+      "epoch": 1.3994520547945206,
+      "grad_norm": 2.3810958862304688,
+      "learning_rate": 2.9935012185215273e-06,
+      "loss": 1.8196,
+      "step": 1277
+    },
+    {
+      "epoch": 1.4005479452054794,
+      "grad_norm": 2.4343955516815186,
+      "learning_rate": 2.991470349309505e-06,
+      "loss": 1.8564,
+      "step": 1278
+    },
+    {
+      "epoch": 1.4016438356164382,
+      "grad_norm": 2.1929988861083984,
+      "learning_rate": 2.989439480097482e-06,
+      "loss": 1.8119,
+      "step": 1279
+    },
+    {
+      "epoch": 1.4027397260273973,
+      "grad_norm": 2.2305309772491455,
+      "learning_rate": 2.9874086108854595e-06,
+      "loss": 1.8289,
+      "step": 1280
+    },
+    {
+      "epoch": 1.4027397260273973,
+      "eval_loss": 2.032022476196289,
+      "eval_runtime": 37.084,
+      "eval_samples_per_second": 26.966,
+      "eval_steps_per_second": 1.133,
+      "step": 1280
+    },
+    {
+      "epoch": 1.403835616438356,
+      "grad_norm": 2.6255314350128174,
+      "learning_rate": 2.9853777416734364e-06,
+      "loss": 1.8176,
+      "step": 1281
+    },
+    {
+      "epoch": 1.4049315068493151,
+      "grad_norm": 2.238891363143921,
+      "learning_rate": 2.983346872461414e-06,
+      "loss": 1.8679,
+      "step": 1282
+    },
+    {
+      "epoch": 1.406027397260274,
+      "grad_norm": 2.9972007274627686,
+      "learning_rate": 2.981316003249391e-06,
+      "loss": 1.8839,
+      "step": 1283
+    },
+    {
+      "epoch": 1.407123287671233,
+      "grad_norm": 2.3255932331085205,
+      "learning_rate": 2.9792851340373678e-06,
+      "loss": 1.8451,
+      "step": 1284
+    },
+    {
+      "epoch": 1.4082191780821918,
+      "grad_norm": 2.290395498275757,
+      "learning_rate": 2.9772542648253455e-06,
+      "loss": 1.8244,
+      "step": 1285
+    },
+    {
+      "epoch": 1.4093150684931506,
+      "grad_norm": 2.288728713989258,
+      "learning_rate": 2.9752233956133223e-06,
+      "loss": 1.8628,
+      "step": 1286
+    },
+    {
+      "epoch": 1.4104109589041096,
+      "grad_norm": 2.177381992340088,
+      "learning_rate": 2.9731925264013e-06,
+      "loss": 1.8111,
+      "step": 1287
+    },
+    {
+      "epoch": 1.4115068493150684,
+      "grad_norm": 2.214956760406494,
+      "learning_rate": 2.971161657189277e-06,
+      "loss": 1.8377,
+      "step": 1288
+    },
+    {
+      "epoch": 1.4126027397260275,
+      "grad_norm": 2.3859663009643555,
+      "learning_rate": 2.9691307879772546e-06,
+      "loss": 1.8154,
+      "step": 1289
+    },
+    {
+      "epoch": 1.4136986301369863,
+      "grad_norm": 2.0873706340789795,
+      "learning_rate": 2.9670999187652314e-06,
+      "loss": 1.8098,
+      "step": 1290
+    },
+    {
+      "epoch": 1.4136986301369863,
+      "eval_loss": 2.029996633529663,
+      "eval_runtime": 37.0979,
+      "eval_samples_per_second": 26.956,
+      "eval_steps_per_second": 1.132,
+      "step": 1290
+    },
+    {
+      "epoch": 1.4147945205479453,
+      "grad_norm": 2.2632675170898438,
+      "learning_rate": 2.965069049553209e-06,
+      "loss": 1.7974,
+      "step": 1291
+    },
+    {
+      "epoch": 1.415890410958904,
+      "grad_norm": 2.208515167236328,
+      "learning_rate": 2.963038180341186e-06,
+      "loss": 1.8856,
+      "step": 1292
+    },
+    {
+      "epoch": 1.416986301369863,
+      "grad_norm": 2.176525115966797,
+      "learning_rate": 2.9610073111291637e-06,
+      "loss": 1.7783,
+      "step": 1293
+    },
+    {
+      "epoch": 1.418082191780822,
+      "grad_norm": 2.226682186126709,
+      "learning_rate": 2.9589764419171405e-06,
+      "loss": 1.865,
+      "step": 1294
+    },
+    {
+      "epoch": 1.4191780821917808,
+      "grad_norm": 2.3544883728027344,
+      "learning_rate": 2.9569455727051182e-06,
+      "loss": 1.8808,
+      "step": 1295
+    },
+    {
+      "epoch": 1.4202739726027398,
+      "grad_norm": 2.1729962825775146,
+      "learning_rate": 2.954914703493095e-06,
+      "loss": 1.7927,
+      "step": 1296
+    },
+    {
+      "epoch": 1.4213698630136986,
+      "grad_norm": 2.168168067932129,
+      "learning_rate": 2.9528838342810724e-06,
+      "loss": 1.8253,
+      "step": 1297
+    },
+    {
+      "epoch": 1.4224657534246576,
+      "grad_norm": 2.2479910850524902,
+      "learning_rate": 2.9508529650690496e-06,
+      "loss": 1.7976,
+      "step": 1298
+    },
+    {
+      "epoch": 1.4235616438356165,
+      "grad_norm": 2.1733837127685547,
+      "learning_rate": 2.948822095857027e-06,
+      "loss": 1.8015,
+      "step": 1299
+    },
+    {
+      "epoch": 1.4246575342465753,
+      "grad_norm": 2.2553391456604004,
+      "learning_rate": 2.946791226645004e-06,
+      "loss": 1.8545,
+      "step": 1300
+    },
+    {
+      "epoch": 1.4246575342465753,
+      "eval_loss": 2.0245048999786377,
+      "eval_runtime": 37.0943,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 1300
+    },
+    {
+      "epoch": 1.4257534246575343,
+      "grad_norm": 2.2817769050598145,
+      "learning_rate": 2.9447603574329815e-06,
+      "loss": 1.8017,
+      "step": 1301
+    },
+    {
+      "epoch": 1.426849315068493,
+      "grad_norm": 2.170865297317505,
+      "learning_rate": 2.9427294882209587e-06,
+      "loss": 1.7738,
+      "step": 1302
+    },
+    {
+      "epoch": 1.427945205479452,
+      "grad_norm": 2.296919107437134,
+      "learning_rate": 2.940698619008936e-06,
+      "loss": 1.8098,
+      "step": 1303
+    },
+    {
+      "epoch": 1.429041095890411,
+      "grad_norm": 2.476266622543335,
+      "learning_rate": 2.9386677497969133e-06,
+      "loss": 1.8229,
+      "step": 1304
+    },
+    {
+      "epoch": 1.43013698630137,
+      "grad_norm": 2.400918483734131,
+      "learning_rate": 2.9366368805848906e-06,
+      "loss": 1.8057,
+      "step": 1305
+    },
+    {
+      "epoch": 1.4312328767123288,
+      "grad_norm": 2.393892765045166,
+      "learning_rate": 2.934606011372868e-06,
+      "loss": 1.8568,
+      "step": 1306
+    },
+    {
+      "epoch": 1.4323287671232876,
+      "grad_norm": 2.2551307678222656,
+      "learning_rate": 2.932575142160845e-06,
+      "loss": 1.8493,
+      "step": 1307
+    },
+    {
+      "epoch": 1.4334246575342466,
+      "grad_norm": 2.3091611862182617,
+      "learning_rate": 2.9305442729488224e-06,
+      "loss": 1.8297,
+      "step": 1308
+    },
+    {
+      "epoch": 1.4345205479452054,
+      "grad_norm": 2.231614828109741,
+      "learning_rate": 2.9285134037367997e-06,
+      "loss": 1.8345,
+      "step": 1309
+    },
+    {
+      "epoch": 1.4356164383561643,
+      "grad_norm": 2.24942946434021,
+      "learning_rate": 2.9264825345247765e-06,
+      "loss": 1.8284,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4356164383561643,
+      "eval_loss": 2.024843215942383,
+      "eval_runtime": 36.7795,
+      "eval_samples_per_second": 27.189,
+      "eval_steps_per_second": 1.142,
+      "step": 1310
+    },
+    {
+      "epoch": 1.4367123287671233,
+      "grad_norm": 2.2798643112182617,
+      "learning_rate": 2.9244516653127542e-06,
+      "loss": 1.8554,
+      "step": 1311
+    },
+    {
+      "epoch": 1.4378082191780823,
+      "grad_norm": 2.342576265335083,
+      "learning_rate": 2.922420796100731e-06,
+      "loss": 1.8296,
+      "step": 1312
+    },
+    {
+      "epoch": 1.4389041095890411,
+      "grad_norm": 2.268016815185547,
+      "learning_rate": 2.9203899268887088e-06,
+      "loss": 1.7467,
+      "step": 1313
+    },
+    {
+      "epoch": 1.44,
+      "grad_norm": 2.2520217895507812,
+      "learning_rate": 2.9183590576766856e-06,
+      "loss": 1.8748,
+      "step": 1314
+    },
+    {
+      "epoch": 1.441095890410959,
+      "grad_norm": 2.2091598510742188,
+      "learning_rate": 2.9163281884646633e-06,
+      "loss": 1.7715,
+      "step": 1315
+    },
+    {
+      "epoch": 1.4421917808219178,
+      "grad_norm": 2.197575092315674,
+      "learning_rate": 2.91429731925264e-06,
+      "loss": 1.8265,
+      "step": 1316
+    },
+    {
+      "epoch": 1.4432876712328766,
+      "grad_norm": 2.261122465133667,
+      "learning_rate": 2.912266450040618e-06,
+      "loss": 1.8758,
+      "step": 1317
+    },
+    {
+      "epoch": 1.4443835616438356,
+      "grad_norm": 2.3291778564453125,
+      "learning_rate": 2.9102355808285947e-06,
+      "loss": 1.8213,
+      "step": 1318
+    },
+    {
+      "epoch": 1.4454794520547944,
+      "grad_norm": 2.2182250022888184,
+      "learning_rate": 2.9082047116165724e-06,
+      "loss": 1.822,
+      "step": 1319
+    },
+    {
+      "epoch": 1.4465753424657535,
+      "grad_norm": 2.295978546142578,
+      "learning_rate": 2.9061738424045493e-06,
+      "loss": 1.8008,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4465753424657535,
+      "eval_loss": 2.0230674743652344,
+      "eval_runtime": 37.0013,
+      "eval_samples_per_second": 27.026,
+      "eval_steps_per_second": 1.135,
+      "step": 1320
+    },
+    {
+      "epoch": 1.4476712328767123,
+      "grad_norm": 2.3187007904052734,
+      "learning_rate": 2.904142973192527e-06,
+      "loss": 1.7923,
+      "step": 1321
+    },
+    {
+      "epoch": 1.4487671232876713,
+      "grad_norm": 2.242851972579956,
+      "learning_rate": 2.902112103980504e-06,
+      "loss": 1.8874,
+      "step": 1322
+    },
+    {
+      "epoch": 1.4498630136986301,
+      "grad_norm": 2.2560081481933594,
+      "learning_rate": 2.900081234768481e-06,
+      "loss": 1.8284,
+      "step": 1323
+    },
+    {
+      "epoch": 1.450958904109589,
+      "grad_norm": 2.3017799854278564,
+      "learning_rate": 2.8980503655564584e-06,
+      "loss": 1.8263,
+      "step": 1324
+    },
+    {
+      "epoch": 1.452054794520548,
+      "grad_norm": 2.3016278743743896,
+      "learning_rate": 2.8960194963444357e-06,
+      "loss": 1.8174,
+      "step": 1325
+    },
+    {
+      "epoch": 1.4531506849315068,
+      "grad_norm": 2.1965372562408447,
+      "learning_rate": 2.893988627132413e-06,
+      "loss": 1.7812,
+      "step": 1326
+    },
+    {
+      "epoch": 1.4542465753424658,
+      "grad_norm": 2.302969455718994,
+      "learning_rate": 2.8919577579203902e-06,
+      "loss": 1.8115,
+      "step": 1327
+    },
+    {
+      "epoch": 1.4553424657534246,
+      "grad_norm": 2.338066339492798,
+      "learning_rate": 2.8899268887083675e-06,
+      "loss": 1.7535,
+      "step": 1328
+    },
+    {
+      "epoch": 1.4564383561643837,
+      "grad_norm": 2.446503162384033,
+      "learning_rate": 2.8878960194963448e-06,
+      "loss": 1.8052,
+      "step": 1329
+    },
+    {
+      "epoch": 1.4575342465753425,
+      "grad_norm": 2.2319273948669434,
+      "learning_rate": 2.885865150284322e-06,
+      "loss": 1.8226,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4575342465753425,
+      "eval_loss": 2.022364854812622,
+      "eval_runtime": 37.0901,
+      "eval_samples_per_second": 26.961,
+      "eval_steps_per_second": 1.132,
+      "step": 1330
+    },
+    {
+      "epoch": 1.4586301369863013,
+      "grad_norm": 2.391080856323242,
+      "learning_rate": 2.8838342810722993e-06,
+      "loss": 1.8529,
+      "step": 1331
+    },
+    {
+      "epoch": 1.4597260273972603,
+      "grad_norm": 2.274463415145874,
+      "learning_rate": 2.8818034118602766e-06,
+      "loss": 1.8043,
+      "step": 1332
+    },
+    {
+      "epoch": 1.4608219178082191,
+      "grad_norm": 2.4040963649749756,
+      "learning_rate": 2.879772542648254e-06,
+      "loss": 1.7842,
+      "step": 1333
+    },
+    {
+      "epoch": 1.4619178082191782,
+      "grad_norm": 2.268629312515259,
+      "learning_rate": 2.8777416734362307e-06,
+      "loss": 1.859,
+      "step": 1334
+    },
+    {
+      "epoch": 1.463013698630137,
+      "grad_norm": 2.4727842807769775,
+      "learning_rate": 2.8757108042242084e-06,
+      "loss": 1.7615,
+      "step": 1335
+    },
+    {
+      "epoch": 1.464109589041096,
+      "grad_norm": 2.179358959197998,
+      "learning_rate": 2.8736799350121853e-06,
+      "loss": 1.9085,
+      "step": 1336
+    },
+    {
+      "epoch": 1.4652054794520548,
+      "grad_norm": 2.220917224884033,
+      "learning_rate": 2.871649065800163e-06,
+      "loss": 1.8618,
+      "step": 1337
+    },
+    {
+      "epoch": 1.4663013698630136,
+      "grad_norm": 2.3807291984558105,
+      "learning_rate": 2.86961819658814e-06,
+      "loss": 1.88,
+      "step": 1338
+    },
+    {
+      "epoch": 1.4673972602739727,
+      "grad_norm": 2.272507667541504,
+      "learning_rate": 2.8675873273761175e-06,
+      "loss": 1.8045,
+      "step": 1339
+    },
+    {
+      "epoch": 1.4684931506849315,
+      "grad_norm": 2.1832377910614014,
+      "learning_rate": 2.8655564581640944e-06,
+      "loss": 1.8273,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4684931506849315,
+      "eval_loss": 2.0218358039855957,
+      "eval_runtime": 37.1223,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 1340
+    },
+    {
+      "epoch": 1.4695890410958903,
+      "grad_norm": 2.2655251026153564,
+      "learning_rate": 2.863525588952072e-06,
+      "loss": 1.8162,
+      "step": 1341
+    },
+    {
+      "epoch": 1.4706849315068493,
+      "grad_norm": 2.191852331161499,
+      "learning_rate": 2.861494719740049e-06,
+      "loss": 1.8778,
+      "step": 1342
+    },
+    {
+      "epoch": 1.4717808219178083,
+      "grad_norm": 2.1869187355041504,
+      "learning_rate": 2.8594638505280266e-06,
+      "loss": 1.7695,
+      "step": 1343
+    },
+    {
+      "epoch": 1.4728767123287672,
+      "grad_norm": 2.248108148574829,
+      "learning_rate": 2.8574329813160035e-06,
+      "loss": 1.7759,
+      "step": 1344
+    },
+    {
+      "epoch": 1.473972602739726,
+      "grad_norm": 2.2599704265594482,
+      "learning_rate": 2.855402112103981e-06,
+      "loss": 1.8214,
+      "step": 1345
+    },
+    {
+      "epoch": 1.475068493150685,
+      "grad_norm": 2.2920355796813965,
+      "learning_rate": 2.853371242891958e-06,
+      "loss": 1.8061,
+      "step": 1346
+    },
+    {
+      "epoch": 1.4761643835616438,
+      "grad_norm": 2.26715087890625,
+      "learning_rate": 2.851340373679935e-06,
+      "loss": 1.8244,
+      "step": 1347
+    },
+    {
+      "epoch": 1.4772602739726026,
+      "grad_norm": 2.349169969558716,
+      "learning_rate": 2.8493095044679126e-06,
+      "loss": 1.8401,
+      "step": 1348
+    },
+    {
+      "epoch": 1.4783561643835617,
+      "grad_norm": 2.3571698665618896,
+      "learning_rate": 2.8472786352558894e-06,
+      "loss": 1.8913,
+      "step": 1349
+    },
+    {
+      "epoch": 1.4794520547945205,
+      "grad_norm": 2.334057331085205,
+      "learning_rate": 2.845247766043867e-06,
+      "loss": 1.8434,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4794520547945205,
+      "eval_loss": 2.0251200199127197,
+      "eval_runtime": 37.1298,
+      "eval_samples_per_second": 26.933,
+      "eval_steps_per_second": 1.131,
+      "step": 1350
+    },
+    {
+      "epoch": 1.4805479452054795,
+      "grad_norm": 2.363516330718994,
+      "learning_rate": 2.843216896831844e-06,
+      "loss": 1.7846,
+      "step": 1351
+    },
+    {
+      "epoch": 1.4816438356164383,
+      "grad_norm": 2.372403860092163,
+      "learning_rate": 2.8411860276198217e-06,
+      "loss": 1.7791,
+      "step": 1352
+    },
+    {
+      "epoch": 1.4827397260273973,
+      "grad_norm": 2.2666590213775635,
+      "learning_rate": 2.8391551584077985e-06,
+      "loss": 1.7969,
+      "step": 1353
+    },
+    {
+      "epoch": 1.4838356164383562,
+      "grad_norm": 2.3651468753814697,
+      "learning_rate": 2.8371242891957762e-06,
+      "loss": 1.8565,
+      "step": 1354
+    },
+    {
+      "epoch": 1.484931506849315,
+      "grad_norm": 2.3559513092041016,
+      "learning_rate": 2.835093419983753e-06,
+      "loss": 1.8765,
+      "step": 1355
+    },
+    {
+      "epoch": 1.486027397260274,
+      "grad_norm": 2.232041358947754,
+      "learning_rate": 2.833062550771731e-06,
+      "loss": 1.805,
+      "step": 1356
+    },
+    {
+      "epoch": 1.4871232876712328,
+      "grad_norm": 2.2423202991485596,
+      "learning_rate": 2.8310316815597076e-06,
+      "loss": 1.8137,
+      "step": 1357
+    },
+    {
+      "epoch": 1.4882191780821918,
+      "grad_norm": 2.3454222679138184,
+      "learning_rate": 2.8290008123476853e-06,
+      "loss": 1.7839,
+      "step": 1358
+    },
+    {
+      "epoch": 1.4893150684931507,
+      "grad_norm": 2.351266384124756,
+      "learning_rate": 2.826969943135662e-06,
+      "loss": 1.8208,
+      "step": 1359
+    },
+    {
+      "epoch": 1.4904109589041097,
+      "grad_norm": 2.3025388717651367,
+      "learning_rate": 2.8249390739236395e-06,
+      "loss": 1.8059,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4904109589041097,
+      "eval_loss": 2.0235962867736816,
+      "eval_runtime": 37.1359,
+      "eval_samples_per_second": 26.928,
+      "eval_steps_per_second": 1.131,
+      "step": 1360
+    },
+    {
+      "epoch": 1.4915068493150685,
+      "grad_norm": 2.2277750968933105,
+      "learning_rate": 2.8229082047116167e-06,
+      "loss": 1.8238,
+      "step": 1361
+    },
+    {
+      "epoch": 1.4926027397260273,
+      "grad_norm": 2.318286180496216,
+      "learning_rate": 2.820877335499594e-06,
+      "loss": 1.7867,
+      "step": 1362
+    },
+    {
+      "epoch": 1.4936986301369863,
+      "grad_norm": 2.296259880065918,
+      "learning_rate": 2.8188464662875713e-06,
+      "loss": 1.8068,
+      "step": 1363
+    },
+    {
+      "epoch": 1.4947945205479451,
+      "grad_norm": 2.2617714405059814,
+      "learning_rate": 2.8168155970755486e-06,
+      "loss": 1.7607,
+      "step": 1364
+    },
+    {
+      "epoch": 1.4958904109589042,
+      "grad_norm": 2.318294048309326,
+      "learning_rate": 2.814784727863526e-06,
+      "loss": 1.7872,
+      "step": 1365
+    },
+    {
+      "epoch": 1.496986301369863,
+      "grad_norm": 2.260286808013916,
+      "learning_rate": 2.812753858651503e-06,
+      "loss": 1.8013,
+      "step": 1366
+    },
+    {
+      "epoch": 1.498082191780822,
+      "grad_norm": 2.222074508666992,
+      "learning_rate": 2.8107229894394804e-06,
+      "loss": 1.8338,
+      "step": 1367
+    },
+    {
+      "epoch": 1.4991780821917808,
+      "grad_norm": 2.6403002738952637,
+      "learning_rate": 2.8086921202274577e-06,
+      "loss": 1.8126,
+      "step": 1368
+    },
+    {
+      "epoch": 1.5002739726027396,
+      "grad_norm": 2.2439680099487305,
+      "learning_rate": 2.806661251015435e-06,
+      "loss": 1.7669,
+      "step": 1369
+    },
+    {
+      "epoch": 1.5013698630136987,
+      "grad_norm": 2.3914687633514404,
+      "learning_rate": 2.8046303818034122e-06,
+      "loss": 1.8421,
+      "step": 1370
+    },
+    {
+      "epoch": 1.5013698630136987,
+      "eval_loss": 2.0222702026367188,
+      "eval_runtime": 37.1598,
+      "eval_samples_per_second": 26.911,
+      "eval_steps_per_second": 1.13,
+      "step": 1370
+    },
+    {
+      "epoch": 1.5024657534246575,
+      "grad_norm": 2.3510093688964844,
+      "learning_rate": 2.802599512591389e-06,
+      "loss": 1.8902,
+      "step": 1371
+    },
+    {
+      "epoch": 1.5035616438356163,
+      "grad_norm": 2.4427363872528076,
+      "learning_rate": 2.8005686433793668e-06,
+      "loss": 1.9376,
+      "step": 1372
+    },
+    {
+      "epoch": 1.5046575342465753,
+      "grad_norm": 2.299356460571289,
+      "learning_rate": 2.7985377741673436e-06,
+      "loss": 1.812,
+      "step": 1373
+    },
+    {
+      "epoch": 1.5057534246575344,
+      "grad_norm": 2.3738906383514404,
+      "learning_rate": 2.7965069049553213e-06,
+      "loss": 1.7666,
+      "step": 1374
+    },
+    {
+      "epoch": 1.5068493150684932,
+      "grad_norm": 2.3620493412017822,
+      "learning_rate": 2.794476035743298e-06,
+      "loss": 1.824,
+      "step": 1375
+    },
+    {
+      "epoch": 1.507945205479452,
+      "grad_norm": 2.539048671722412,
+      "learning_rate": 2.792445166531276e-06,
+      "loss": 1.8712,
+      "step": 1376
+    },
+    {
+      "epoch": 1.509041095890411,
+      "grad_norm": 2.2931456565856934,
+      "learning_rate": 2.7904142973192527e-06,
+      "loss": 1.8414,
+      "step": 1377
+    },
+    {
+      "epoch": 1.5101369863013698,
+      "grad_norm": 2.4368224143981934,
+      "learning_rate": 2.7883834281072304e-06,
+      "loss": 1.8419,
+      "step": 1378
+    },
+    {
+      "epoch": 1.5112328767123286,
+      "grad_norm": 2.298302173614502,
+      "learning_rate": 2.7863525588952073e-06,
+      "loss": 1.7846,
+      "step": 1379
+    },
+    {
+      "epoch": 1.5123287671232877,
+      "grad_norm": 2.3726806640625,
+      "learning_rate": 2.784321689683185e-06,
+      "loss": 1.7733,
+      "step": 1380
+    },
+    {
+      "epoch": 1.5123287671232877,
+      "eval_loss": 2.0201807022094727,
+      "eval_runtime": 37.1012,
+      "eval_samples_per_second": 26.953,
+      "eval_steps_per_second": 1.132,
+      "step": 1380
+    },
+    {
+      "epoch": 1.5134246575342467,
+      "grad_norm": 2.6515002250671387,
+      "learning_rate": 2.782290820471162e-06,
+      "loss": 1.7764,
+      "step": 1381
+    },
+    {
+      "epoch": 1.5145205479452055,
+      "grad_norm": 2.273272752761841,
+      "learning_rate": 2.7802599512591395e-06,
+      "loss": 1.7747,
+      "step": 1382
+    },
+    {
+      "epoch": 1.5156164383561643,
+      "grad_norm": 2.570009708404541,
+      "learning_rate": 2.7782290820471164e-06,
+      "loss": 1.8612,
+      "step": 1383
+    },
+    {
+      "epoch": 1.5167123287671234,
+      "grad_norm": 2.567878007888794,
+      "learning_rate": 2.7761982128350932e-06,
+      "loss": 1.8402,
+      "step": 1384
+    },
+    {
+      "epoch": 1.5178082191780822,
+      "grad_norm": 2.3768470287323,
+      "learning_rate": 2.774167343623071e-06,
+      "loss": 1.8473,
+      "step": 1385
+    },
+    {
+      "epoch": 1.518904109589041,
+      "grad_norm": 2.1383397579193115,
+      "learning_rate": 2.772136474411048e-06,
+      "loss": 1.8593,
+      "step": 1386
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 2.4437034130096436,
+      "learning_rate": 2.7701056051990255e-06,
+      "loss": 1.792,
+      "step": 1387
+    },
+    {
+      "epoch": 1.521095890410959,
+      "grad_norm": 2.433933734893799,
+      "learning_rate": 2.7680747359870023e-06,
+      "loss": 1.7312,
+      "step": 1388
+    },
+    {
+      "epoch": 1.5221917808219179,
+      "grad_norm": 2.340994119644165,
+      "learning_rate": 2.76604386677498e-06,
+      "loss": 1.8326,
+      "step": 1389
+    },
+    {
+      "epoch": 1.5232876712328767,
+      "grad_norm": 2.4730165004730225,
+      "learning_rate": 2.764012997562957e-06,
+      "loss": 1.8129,
+      "step": 1390
+    },
+    {
+      "epoch": 1.5232876712328767,
+      "eval_loss": 2.0158324241638184,
+      "eval_runtime": 37.0997,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 1390
+    },
+    {
+      "epoch": 1.5243835616438357,
+      "grad_norm": 2.2898151874542236,
+      "learning_rate": 2.7619821283509346e-06,
+      "loss": 1.8726,
+      "step": 1391
+    },
+    {
+      "epoch": 1.5254794520547945,
+      "grad_norm": 2.255439519882202,
+      "learning_rate": 2.7599512591389114e-06,
+      "loss": 1.7587,
+      "step": 1392
+    },
+    {
+      "epoch": 1.5265753424657533,
+      "grad_norm": 2.4269862174987793,
+      "learning_rate": 2.757920389926889e-06,
+      "loss": 1.8498,
+      "step": 1393
+    },
+    {
+      "epoch": 1.5276712328767124,
+      "grad_norm": 2.3411412239074707,
+      "learning_rate": 2.755889520714866e-06,
+      "loss": 1.8103,
+      "step": 1394
+    },
+    {
+      "epoch": 1.5287671232876714,
+      "grad_norm": 2.2859277725219727,
+      "learning_rate": 2.7538586515028437e-06,
+      "loss": 1.7909,
+      "step": 1395
+    },
+    {
+      "epoch": 1.52986301369863,
+      "grad_norm": 2.2675201892852783,
+      "learning_rate": 2.7518277822908205e-06,
+      "loss": 1.7672,
+      "step": 1396
+    },
+    {
+      "epoch": 1.530958904109589,
+      "grad_norm": 2.240610361099243,
+      "learning_rate": 2.749796913078798e-06,
+      "loss": 1.8162,
+      "step": 1397
+    },
+    {
+      "epoch": 1.532054794520548,
+      "grad_norm": 2.272045373916626,
+      "learning_rate": 2.747766043866775e-06,
+      "loss": 1.8593,
+      "step": 1398
+    },
+    {
+      "epoch": 1.5331506849315069,
+      "grad_norm": 2.1736183166503906,
+      "learning_rate": 2.7457351746547524e-06,
+      "loss": 1.8154,
+      "step": 1399
+    },
+    {
+      "epoch": 1.5342465753424657,
+      "grad_norm": 2.197085380554199,
+      "learning_rate": 2.7437043054427297e-06,
+      "loss": 1.8758,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5342465753424657,
+      "eval_loss": 2.017298698425293,
+      "eval_runtime": 37.1666,
+      "eval_samples_per_second": 26.906,
+      "eval_steps_per_second": 1.13,
+      "step": 1400
+    },
+    {
+      "epoch": 1.5353424657534247,
+      "grad_norm": 2.2274084091186523,
+      "learning_rate": 2.741673436230707e-06,
+      "loss": 1.9241,
+      "step": 1401
+    },
+    {
+      "epoch": 1.5364383561643835,
+      "grad_norm": 2.3320305347442627,
+      "learning_rate": 2.739642567018684e-06,
+      "loss": 1.8209,
+      "step": 1402
+    },
+    {
+      "epoch": 1.5375342465753423,
+      "grad_norm": 2.370651960372925,
+      "learning_rate": 2.7376116978066615e-06,
+      "loss": 1.7897,
+      "step": 1403
+    },
+    {
+      "epoch": 1.5386301369863014,
+      "grad_norm": 2.387906312942505,
+      "learning_rate": 2.7355808285946388e-06,
+      "loss": 1.8385,
+      "step": 1404
+    },
+    {
+      "epoch": 1.5397260273972604,
+      "grad_norm": 2.340228796005249,
+      "learning_rate": 2.733549959382616e-06,
+      "loss": 1.8279,
+      "step": 1405
+    },
+    {
+      "epoch": 1.5408219178082192,
+      "grad_norm": 2.338785409927368,
+      "learning_rate": 2.7315190901705933e-06,
+      "loss": 1.7887,
+      "step": 1406
+    },
+    {
+      "epoch": 1.541917808219178,
+      "grad_norm": 2.350057601928711,
+      "learning_rate": 2.7294882209585706e-06,
+      "loss": 1.8772,
+      "step": 1407
+    },
+    {
+      "epoch": 1.543013698630137,
+      "grad_norm": 2.1129322052001953,
+      "learning_rate": 2.727457351746548e-06,
+      "loss": 1.8393,
+      "step": 1408
+    },
+    {
+      "epoch": 1.5441095890410959,
+      "grad_norm": 2.282553195953369,
+      "learning_rate": 2.725426482534525e-06,
+      "loss": 1.7893,
+      "step": 1409
+    },
+    {
+      "epoch": 1.5452054794520547,
+      "grad_norm": 2.3291871547698975,
+      "learning_rate": 2.723395613322502e-06,
+      "loss": 1.7921,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5452054794520547,
+      "eval_loss": 2.0166046619415283,
+      "eval_runtime": 36.7906,
+      "eval_samples_per_second": 27.181,
+      "eval_steps_per_second": 1.142,
+      "step": 1410
+    },
+    {
+      "epoch": 1.5463013698630137,
+      "grad_norm": 2.3050696849823,
+      "learning_rate": 2.7213647441104797e-06,
+      "loss": 1.8292,
+      "step": 1411
+    },
+    {
+      "epoch": 1.5473972602739727,
+      "grad_norm": 2.2762458324432373,
+      "learning_rate": 2.7193338748984565e-06,
+      "loss": 1.811,
+      "step": 1412
+    },
+    {
+      "epoch": 1.5484931506849315,
+      "grad_norm": 2.5544228553771973,
+      "learning_rate": 2.7173030056864342e-06,
+      "loss": 1.8203,
+      "step": 1413
+    },
+    {
+      "epoch": 1.5495890410958904,
+      "grad_norm": 2.6392276287078857,
+      "learning_rate": 2.715272136474411e-06,
+      "loss": 1.9724,
+      "step": 1414
+    },
+    {
+      "epoch": 1.5506849315068494,
+      "grad_norm": 2.445366382598877,
+      "learning_rate": 2.713241267262389e-06,
+      "loss": 1.7589,
+      "step": 1415
+    },
+    {
+      "epoch": 1.5517808219178082,
+      "grad_norm": 2.5305662155151367,
+      "learning_rate": 2.7112103980503656e-06,
+      "loss": 1.8311,
+      "step": 1416
+    },
+    {
+      "epoch": 1.552876712328767,
+      "grad_norm": 2.405482769012451,
+      "learning_rate": 2.7091795288383433e-06,
+      "loss": 1.7893,
+      "step": 1417
+    },
+    {
+      "epoch": 1.553972602739726,
+      "grad_norm": 2.258361577987671,
+      "learning_rate": 2.70714865962632e-06,
+      "loss": 1.8489,
+      "step": 1418
+    },
+    {
+      "epoch": 1.555068493150685,
+      "grad_norm": 2.1798577308654785,
+      "learning_rate": 2.705117790414298e-06,
+      "loss": 1.7983,
+      "step": 1419
+    },
+    {
+      "epoch": 1.5561643835616439,
+      "grad_norm": 2.1985363960266113,
+      "learning_rate": 2.7030869212022747e-06,
+      "loss": 1.7934,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5561643835616439,
+      "eval_loss": 2.0219907760620117,
+      "eval_runtime": 36.9416,
+      "eval_samples_per_second": 27.07,
+      "eval_steps_per_second": 1.137,
+      "step": 1420
+    },
+    {
+      "epoch": 1.5572602739726027,
+      "grad_norm": 2.273252248764038,
+      "learning_rate": 2.7010560519902516e-06,
+      "loss": 1.8571,
+      "step": 1421
+    },
+    {
+      "epoch": 1.5583561643835617,
+      "grad_norm": 2.22920298576355,
+      "learning_rate": 2.6990251827782293e-06,
+      "loss": 1.7785,
+      "step": 1422
+    },
+    {
+      "epoch": 1.5594520547945205,
+      "grad_norm": 2.1784098148345947,
+      "learning_rate": 2.696994313566206e-06,
+      "loss": 1.8053,
+      "step": 1423
+    },
+    {
+      "epoch": 1.5605479452054793,
+      "grad_norm": 2.223404884338379,
+      "learning_rate": 2.694963444354184e-06,
+      "loss": 1.7965,
+      "step": 1424
+    },
+    {
+      "epoch": 1.5616438356164384,
+      "grad_norm": 2.2004787921905518,
+      "learning_rate": 2.6929325751421607e-06,
+      "loss": 1.8851,
+      "step": 1425
+    },
+    {
+      "epoch": 1.5627397260273974,
+      "grad_norm": 2.202467441558838,
+      "learning_rate": 2.6909017059301384e-06,
+      "loss": 1.7884,
+      "step": 1426
+    },
+    {
+      "epoch": 1.563835616438356,
+      "grad_norm": 2.337430238723755,
+      "learning_rate": 2.6888708367181153e-06,
+      "loss": 1.8543,
+      "step": 1427
+    },
+    {
+      "epoch": 1.564931506849315,
+      "grad_norm": 2.322848081588745,
+      "learning_rate": 2.686839967506093e-06,
+      "loss": 1.7607,
+      "step": 1428
+    },
+    {
+      "epoch": 1.566027397260274,
+      "grad_norm": 2.277765989303589,
+      "learning_rate": 2.68480909829407e-06,
+      "loss": 1.8354,
+      "step": 1429
+    },
+    {
+      "epoch": 1.5671232876712329,
+      "grad_norm": 2.3229448795318604,
+      "learning_rate": 2.6827782290820475e-06,
+      "loss": 1.7277,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5671232876712329,
+      "eval_loss": 2.0220985412597656,
+      "eval_runtime": 37.0054,
+      "eval_samples_per_second": 27.023,
+      "eval_steps_per_second": 1.135,
+      "step": 1430
+    },
+    {
+      "epoch": 1.5682191780821917,
+      "grad_norm": 2.49545955657959,
+      "learning_rate": 2.6807473598700244e-06,
+      "loss": 1.7351,
+      "step": 1431
+    },
+    {
+      "epoch": 1.5693150684931507,
+      "grad_norm": 2.530553102493286,
+      "learning_rate": 2.678716490658002e-06,
+      "loss": 1.8614,
+      "step": 1432
+    },
+    {
+      "epoch": 1.5704109589041098,
+      "grad_norm": 2.3352842330932617,
+      "learning_rate": 2.676685621445979e-06,
+      "loss": 1.8218,
+      "step": 1433
+    },
+    {
+      "epoch": 1.5715068493150683,
+      "grad_norm": 2.3533058166503906,
+      "learning_rate": 2.674654752233956e-06,
+      "loss": 1.8282,
+      "step": 1434
+    },
+    {
+      "epoch": 1.5726027397260274,
+      "grad_norm": 2.5509746074676514,
+      "learning_rate": 2.6726238830219335e-06,
+      "loss": 1.8333,
+      "step": 1435
+    },
+    {
+      "epoch": 1.5736986301369864,
+      "grad_norm": 2.185175895690918,
+      "learning_rate": 2.6705930138099107e-06,
+      "loss": 1.8299,
+      "step": 1436
+    },
+    {
+      "epoch": 1.5747945205479452,
+      "grad_norm": 2.2338390350341797,
+      "learning_rate": 2.668562144597888e-06,
+      "loss": 1.857,
+      "step": 1437
+    },
+    {
+      "epoch": 1.575890410958904,
+      "grad_norm": 2.4094536304473877,
+      "learning_rate": 2.6665312753858653e-06,
+      "loss": 1.7333,
+      "step": 1438
+    },
+    {
+      "epoch": 1.576986301369863,
+      "grad_norm": 2.4650957584381104,
+      "learning_rate": 2.6645004061738426e-06,
+      "loss": 1.8237,
+      "step": 1439
+    },
+    {
+      "epoch": 1.5780821917808219,
+      "grad_norm": 2.427415132522583,
+      "learning_rate": 2.66246953696182e-06,
+      "loss": 1.8206,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5780821917808219,
+      "eval_loss": 2.0222699642181396,
+      "eval_runtime": 37.0252,
+      "eval_samples_per_second": 27.009,
+      "eval_steps_per_second": 1.134,
+      "step": 1440
+    },
+    {
+      "epoch": 1.5791780821917807,
+      "grad_norm": 2.3878679275512695,
+      "learning_rate": 2.660438667749797e-06,
+      "loss": 1.775,
+      "step": 1441
+    },
+    {
+      "epoch": 1.5802739726027397,
+      "grad_norm": 2.317819595336914,
+      "learning_rate": 2.6584077985377744e-06,
+      "loss": 1.7479,
+      "step": 1442
+    },
+    {
+      "epoch": 1.5813698630136988,
+      "grad_norm": 2.34951114654541,
+      "learning_rate": 2.6563769293257517e-06,
+      "loss": 1.8025,
+      "step": 1443
+    },
+    {
+      "epoch": 1.5824657534246576,
+      "grad_norm": 2.2849795818328857,
+      "learning_rate": 2.654346060113729e-06,
+      "loss": 1.8152,
+      "step": 1444
+    },
+    {
+      "epoch": 1.5835616438356164,
+      "grad_norm": 2.211489677429199,
+      "learning_rate": 2.6523151909017062e-06,
+      "loss": 1.7666,
+      "step": 1445
+    },
+    {
+      "epoch": 1.5846575342465754,
+      "grad_norm": 2.1876230239868164,
+      "learning_rate": 2.6502843216896835e-06,
+      "loss": 1.7933,
+      "step": 1446
+    },
+    {
+      "epoch": 1.5857534246575342,
+      "grad_norm": 2.367628335952759,
+      "learning_rate": 2.6482534524776603e-06,
+      "loss": 1.7756,
+      "step": 1447
+    },
+    {
+      "epoch": 1.586849315068493,
+      "grad_norm": 2.1198348999023438,
+      "learning_rate": 2.646222583265638e-06,
+      "loss": 1.8908,
+      "step": 1448
+    },
+    {
+      "epoch": 1.587945205479452,
+      "grad_norm": 2.192788600921631,
+      "learning_rate": 2.644191714053615e-06,
+      "loss": 1.8432,
+      "step": 1449
+    },
+    {
+      "epoch": 1.589041095890411,
+      "grad_norm": 2.4480655193328857,
+      "learning_rate": 2.6421608448415926e-06,
+      "loss": 1.8223,
+      "step": 1450
+    },
+    {
+      "epoch": 1.589041095890411,
+      "eval_loss": 2.019976854324341,
+      "eval_runtime": 37.0117,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1450
+    },
+    {
+      "epoch": 1.59013698630137,
+      "grad_norm": 2.573078155517578,
+      "learning_rate": 2.6401299756295694e-06,
+      "loss": 1.7881,
+      "step": 1451
+    },
+    {
+      "epoch": 1.5912328767123287,
+      "grad_norm": 2.2820355892181396,
+      "learning_rate": 2.638099106417547e-06,
+      "loss": 1.7703,
+      "step": 1452
+    },
+    {
+      "epoch": 1.5923287671232877,
+      "grad_norm": 2.526477575302124,
+      "learning_rate": 2.636068237205524e-06,
+      "loss": 1.8281,
+      "step": 1453
+    },
+    {
+      "epoch": 1.5934246575342466,
+      "grad_norm": 2.873647689819336,
+      "learning_rate": 2.6340373679935017e-06,
+      "loss": 1.7771,
+      "step": 1454
+    },
+    {
+      "epoch": 1.5945205479452054,
+      "grad_norm": 2.3782413005828857,
+      "learning_rate": 2.6320064987814785e-06,
+      "loss": 1.8864,
+      "step": 1455
+    },
+    {
+      "epoch": 1.5956164383561644,
+      "grad_norm": 2.3153343200683594,
+      "learning_rate": 2.6299756295694562e-06,
+      "loss": 1.734,
+      "step": 1456
+    },
+    {
+      "epoch": 1.5967123287671234,
+      "grad_norm": 2.631880283355713,
+      "learning_rate": 2.627944760357433e-06,
+      "loss": 1.8295,
+      "step": 1457
+    },
+    {
+      "epoch": 1.5978082191780822,
+      "grad_norm": 2.5827910900115967,
+      "learning_rate": 2.6259138911454104e-06,
+      "loss": 1.8738,
+      "step": 1458
+    },
+    {
+      "epoch": 1.598904109589041,
+      "grad_norm": 2.3519935607910156,
+      "learning_rate": 2.6238830219333877e-06,
+      "loss": 1.8832,
+      "step": 1459
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 2.5236215591430664,
+      "learning_rate": 2.621852152721365e-06,
+      "loss": 1.8595,
+      "step": 1460
+    },
+    {
+      "epoch": 1.6,
+      "eval_loss": 2.0170512199401855,
+      "eval_runtime": 37.0302,
+      "eval_samples_per_second": 27.005,
+      "eval_steps_per_second": 1.134,
+      "step": 1460
+    },
+    {
+      "epoch": 1.601095890410959,
+      "grad_norm": 2.725083112716675,
+      "learning_rate": 2.619821283509342e-06,
+      "loss": 1.7476,
+      "step": 1461
+    },
+    {
+      "epoch": 1.6021917808219177,
+      "grad_norm": 2.212644100189209,
+      "learning_rate": 2.6177904142973195e-06,
+      "loss": 1.8475,
+      "step": 1462
+    },
+    {
+      "epoch": 1.6032876712328767,
+      "grad_norm": 2.346808910369873,
+      "learning_rate": 2.6157595450852968e-06,
+      "loss": 1.7762,
+      "step": 1463
+    },
+    {
+      "epoch": 1.6043835616438358,
+      "grad_norm": 2.702331304550171,
+      "learning_rate": 2.613728675873274e-06,
+      "loss": 1.8296,
+      "step": 1464
+    },
+    {
+      "epoch": 1.6054794520547944,
+      "grad_norm": 2.4175899028778076,
+      "learning_rate": 2.6116978066612513e-06,
+      "loss": 1.8476,
+      "step": 1465
+    },
+    {
+      "epoch": 1.6065753424657534,
+      "grad_norm": 2.325457811355591,
+      "learning_rate": 2.6096669374492286e-06,
+      "loss": 1.8706,
+      "step": 1466
+    },
+    {
+      "epoch": 1.6076712328767124,
+      "grad_norm": 2.383842945098877,
+      "learning_rate": 2.607636068237206e-06,
+      "loss": 1.8246,
+      "step": 1467
+    },
+    {
+      "epoch": 1.6087671232876712,
+      "grad_norm": 2.289515256881714,
+      "learning_rate": 2.605605199025183e-06,
+      "loss": 1.7728,
+      "step": 1468
+    },
+    {
+      "epoch": 1.60986301369863,
+      "grad_norm": 2.5116689205169678,
+      "learning_rate": 2.6035743298131604e-06,
+      "loss": 1.8359,
+      "step": 1469
+    },
+    {
+      "epoch": 1.610958904109589,
+      "grad_norm": 2.279695749282837,
+      "learning_rate": 2.6015434606011377e-06,
+      "loss": 1.8155,
+      "step": 1470
+    },
+    {
+      "epoch": 1.610958904109589,
+      "eval_loss": 2.0142691135406494,
+      "eval_runtime": 37.0119,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1470
+    },
+    {
+      "epoch": 1.612054794520548,
+      "grad_norm": 2.4397690296173096,
+      "learning_rate": 2.5995125913891145e-06,
+      "loss": 1.8424,
+      "step": 1471
+    },
+    {
+      "epoch": 1.6131506849315067,
+      "grad_norm": 2.433577299118042,
+      "learning_rate": 2.5974817221770922e-06,
+      "loss": 1.8196,
+      "step": 1472
+    },
+    {
+      "epoch": 1.6142465753424657,
+      "grad_norm": 2.365459442138672,
+      "learning_rate": 2.595450852965069e-06,
+      "loss": 1.8124,
+      "step": 1473
+    },
+    {
+      "epoch": 1.6153424657534248,
+      "grad_norm": 2.253133535385132,
+      "learning_rate": 2.593419983753047e-06,
+      "loss": 1.8443,
+      "step": 1474
+    },
+    {
+      "epoch": 1.6164383561643836,
+      "grad_norm": 2.227625608444214,
+      "learning_rate": 2.5913891145410236e-06,
+      "loss": 1.8524,
+      "step": 1475
+    },
+    {
+      "epoch": 1.6175342465753424,
+      "grad_norm": 2.403702974319458,
+      "learning_rate": 2.5893582453290013e-06,
+      "loss": 1.7983,
+      "step": 1476
+    },
+    {
+      "epoch": 1.6186301369863014,
+      "grad_norm": 2.24507999420166,
+      "learning_rate": 2.587327376116978e-06,
+      "loss": 1.8563,
+      "step": 1477
+    },
+    {
+      "epoch": 1.6197260273972602,
+      "grad_norm": 2.284700393676758,
+      "learning_rate": 2.585296506904956e-06,
+      "loss": 1.7958,
+      "step": 1478
+    },
+    {
+      "epoch": 1.620821917808219,
+      "grad_norm": 2.288215398788452,
+      "learning_rate": 2.5832656376929327e-06,
+      "loss": 1.8021,
+      "step": 1479
+    },
+    {
+      "epoch": 1.621917808219178,
+      "grad_norm": 2.2312939167022705,
+      "learning_rate": 2.5812347684809104e-06,
+      "loss": 1.7856,
+      "step": 1480
+    },
+    {
+      "epoch": 1.621917808219178,
+      "eval_loss": 2.01674222946167,
+      "eval_runtime": 36.9943,
+      "eval_samples_per_second": 27.031,
+      "eval_steps_per_second": 1.135,
+      "step": 1480
+    },
+    {
+      "epoch": 1.6230136986301371,
+      "grad_norm": 2.319363832473755,
+      "learning_rate": 2.5792038992688873e-06,
+      "loss": 1.7263,
+      "step": 1481
+    },
+    {
+      "epoch": 1.624109589041096,
+      "grad_norm": 2.4342033863067627,
+      "learning_rate": 2.577173030056865e-06,
+      "loss": 1.897,
+      "step": 1482
+    },
+    {
+      "epoch": 1.6252054794520547,
+      "grad_norm": 2.3052282333374023,
+      "learning_rate": 2.575142160844842e-06,
+      "loss": 1.7165,
+      "step": 1483
+    },
+    {
+      "epoch": 1.6263013698630138,
+      "grad_norm": 2.2476744651794434,
+      "learning_rate": 2.5731112916328187e-06,
+      "loss": 1.8347,
+      "step": 1484
+    },
+    {
+      "epoch": 1.6273972602739726,
+      "grad_norm": 2.3128387928009033,
+      "learning_rate": 2.5710804224207964e-06,
+      "loss": 1.8403,
+      "step": 1485
+    },
+    {
+      "epoch": 1.6284931506849314,
+      "grad_norm": 2.3483974933624268,
+      "learning_rate": 2.5690495532087733e-06,
+      "loss": 1.781,
+      "step": 1486
+    },
+    {
+      "epoch": 1.6295890410958904,
+      "grad_norm": 2.328298568725586,
+      "learning_rate": 2.567018683996751e-06,
+      "loss": 1.7356,
+      "step": 1487
+    },
+    {
+      "epoch": 1.6306849315068495,
+      "grad_norm": 2.2756922245025635,
+      "learning_rate": 2.564987814784728e-06,
+      "loss": 1.8449,
+      "step": 1488
+    },
+    {
+      "epoch": 1.6317808219178083,
+      "grad_norm": 2.363874912261963,
+      "learning_rate": 2.5629569455727055e-06,
+      "loss": 1.7767,
+      "step": 1489
+    },
+    {
+      "epoch": 1.632876712328767,
+      "grad_norm": 2.2872767448425293,
+      "learning_rate": 2.5609260763606824e-06,
+      "loss": 1.8052,
+      "step": 1490
+    },
+    {
+      "epoch": 1.632876712328767,
+      "eval_loss": 2.0134243965148926,
+      "eval_runtime": 37.029,
+      "eval_samples_per_second": 27.006,
+      "eval_steps_per_second": 1.134,
+      "step": 1490
+    },
+    {
+      "epoch": 1.633972602739726,
+      "grad_norm": 2.354851007461548,
+      "learning_rate": 2.55889520714866e-06,
+      "loss": 1.7953,
+      "step": 1491
+    },
+    {
+      "epoch": 1.635068493150685,
+      "grad_norm": 2.4440531730651855,
+      "learning_rate": 2.556864337936637e-06,
+      "loss": 1.9063,
+      "step": 1492
+    },
+    {
+      "epoch": 1.6361643835616437,
+      "grad_norm": 2.296835422515869,
+      "learning_rate": 2.5548334687246146e-06,
+      "loss": 1.7961,
+      "step": 1493
+    },
+    {
+      "epoch": 1.6372602739726028,
+      "grad_norm": 2.7421159744262695,
+      "learning_rate": 2.5528025995125915e-06,
+      "loss": 1.809,
+      "step": 1494
+    },
+    {
+      "epoch": 1.6383561643835618,
+      "grad_norm": 2.819125175476074,
+      "learning_rate": 2.550771730300569e-06,
+      "loss": 1.853,
+      "step": 1495
+    },
+    {
+      "epoch": 1.6394520547945204,
+      "grad_norm": 2.2536680698394775,
+      "learning_rate": 2.548740861088546e-06,
+      "loss": 1.8297,
+      "step": 1496
+    },
+    {
+      "epoch": 1.6405479452054794,
+      "grad_norm": 2.7130825519561768,
+      "learning_rate": 2.5467099918765233e-06,
+      "loss": 1.843,
+      "step": 1497
+    },
+    {
+      "epoch": 1.6416438356164385,
+      "grad_norm": 3.0625038146972656,
+      "learning_rate": 2.5446791226645006e-06,
+      "loss": 1.7462,
+      "step": 1498
+    },
+    {
+      "epoch": 1.6427397260273973,
+      "grad_norm": 2.3095271587371826,
+      "learning_rate": 2.542648253452478e-06,
+      "loss": 1.7363,
+      "step": 1499
+    },
+    {
+      "epoch": 1.643835616438356,
+      "grad_norm": 2.536547899246216,
+      "learning_rate": 2.540617384240455e-06,
+      "loss": 1.8422,
+      "step": 1500
+    },
+    {
+      "epoch": 1.643835616438356,
+      "eval_loss": 2.0126097202301025,
+      "eval_runtime": 37.0469,
+      "eval_samples_per_second": 26.993,
+      "eval_steps_per_second": 1.134,
+      "step": 1500
+    },
+    {
+      "epoch": 1.644931506849315,
+      "grad_norm": 2.7454309463500977,
+      "learning_rate": 2.5385865150284324e-06,
+      "loss": 1.8183,
+      "step": 1501
+    },
+    {
+      "epoch": 1.6460273972602741,
+      "grad_norm": 2.3443901538848877,
+      "learning_rate": 2.5365556458164097e-06,
+      "loss": 1.7827,
+      "step": 1502
+    },
+    {
+      "epoch": 1.6471232876712327,
+      "grad_norm": 2.2412657737731934,
+      "learning_rate": 2.534524776604387e-06,
+      "loss": 1.797,
+      "step": 1503
+    },
+    {
+      "epoch": 1.6482191780821918,
+      "grad_norm": 2.6193699836730957,
+      "learning_rate": 2.5324939073923642e-06,
+      "loss": 1.8103,
+      "step": 1504
+    },
+    {
+      "epoch": 1.6493150684931508,
+      "grad_norm": 2.363919258117676,
+      "learning_rate": 2.5304630381803415e-06,
+      "loss": 1.6987,
+      "step": 1505
+    },
+    {
+      "epoch": 1.6504109589041096,
+      "grad_norm": 2.303164482116699,
+      "learning_rate": 2.5284321689683188e-06,
+      "loss": 1.8005,
+      "step": 1506
+    },
+    {
+      "epoch": 1.6515068493150684,
+      "grad_norm": 2.444174289703369,
+      "learning_rate": 2.526401299756296e-06,
+      "loss": 1.8009,
+      "step": 1507
+    },
+    {
+      "epoch": 1.6526027397260274,
+      "grad_norm": 2.418454647064209,
+      "learning_rate": 2.524370430544273e-06,
+      "loss": 1.782,
+      "step": 1508
+    },
+    {
+      "epoch": 1.6536986301369863,
+      "grad_norm": 2.48840069770813,
+      "learning_rate": 2.5223395613322506e-06,
+      "loss": 1.8143,
+      "step": 1509
+    },
+    {
+      "epoch": 1.654794520547945,
+      "grad_norm": 2.2814347743988037,
+      "learning_rate": 2.5203086921202274e-06,
+      "loss": 1.7687,
+      "step": 1510
+    },
+    {
+      "epoch": 1.654794520547945,
+      "eval_loss": 2.0148253440856934,
+      "eval_runtime": 36.7833,
+      "eval_samples_per_second": 27.186,
+      "eval_steps_per_second": 1.142,
+      "step": 1510
+    },
+    {
+      "epoch": 1.655890410958904,
+      "grad_norm": 2.228933572769165,
+      "learning_rate": 2.518277822908205e-06,
+      "loss": 1.8416,
+      "step": 1511
+    },
+    {
+      "epoch": 1.6569863013698631,
+      "grad_norm": 2.3696742057800293,
+      "learning_rate": 2.516246953696182e-06,
+      "loss": 1.7868,
+      "step": 1512
+    },
+    {
+      "epoch": 1.658082191780822,
+      "grad_norm": 2.216360092163086,
+      "learning_rate": 2.5142160844841597e-06,
+      "loss": 1.8312,
+      "step": 1513
+    },
+    {
+      "epoch": 1.6591780821917808,
+      "grad_norm": 2.454345226287842,
+      "learning_rate": 2.5121852152721365e-06,
+      "loss": 1.8309,
+      "step": 1514
+    },
+    {
+      "epoch": 1.6602739726027398,
+      "grad_norm": 2.2106776237487793,
+      "learning_rate": 2.5101543460601142e-06,
+      "loss": 1.8155,
+      "step": 1515
+    },
+    {
+      "epoch": 1.6613698630136986,
+      "grad_norm": 2.289006233215332,
+      "learning_rate": 2.508123476848091e-06,
+      "loss": 1.8162,
+      "step": 1516
+    },
+    {
+      "epoch": 1.6624657534246574,
+      "grad_norm": 2.297246217727661,
+      "learning_rate": 2.506092607636069e-06,
+      "loss": 1.7559,
+      "step": 1517
+    },
+    {
+      "epoch": 1.6635616438356164,
+      "grad_norm": 2.3909482955932617,
+      "learning_rate": 2.5040617384240457e-06,
+      "loss": 1.8389,
+      "step": 1518
+    },
+    {
+      "epoch": 1.6646575342465755,
+      "grad_norm": 2.398171901702881,
+      "learning_rate": 2.5020308692120234e-06,
+      "loss": 1.7785,
+      "step": 1519
+    },
+    {
+      "epoch": 1.6657534246575343,
+      "grad_norm": 2.377707004547119,
+      "learning_rate": 2.5e-06,
+      "loss": 1.768,
+      "step": 1520
+    },
+    {
+      "epoch": 1.6657534246575343,
+      "eval_loss": 2.012815475463867,
+      "eval_runtime": 36.9603,
+      "eval_samples_per_second": 27.056,
+      "eval_steps_per_second": 1.136,
+      "step": 1520
+    },
+    {
+      "epoch": 1.666849315068493,
+      "grad_norm": 2.336160898208618,
+      "learning_rate": 2.4979691307879775e-06,
+      "loss": 1.8653,
+      "step": 1521
+    },
+    {
+      "epoch": 1.6679452054794521,
+      "grad_norm": 2.288424253463745,
+      "learning_rate": 2.4959382615759548e-06,
+      "loss": 1.7798,
+      "step": 1522
+    },
+    {
+      "epoch": 1.669041095890411,
+      "grad_norm": 2.2599472999572754,
+      "learning_rate": 2.493907392363932e-06,
+      "loss": 1.7736,
+      "step": 1523
+    },
+    {
+      "epoch": 1.6701369863013698,
+      "grad_norm": 2.242276430130005,
+      "learning_rate": 2.4918765231519093e-06,
+      "loss": 1.8753,
+      "step": 1524
+    },
+    {
+      "epoch": 1.6712328767123288,
+      "grad_norm": 2.1897733211517334,
+      "learning_rate": 2.4898456539398866e-06,
+      "loss": 1.8353,
+      "step": 1525
+    },
+    {
+      "epoch": 1.6723287671232878,
+      "grad_norm": 2.254009246826172,
+      "learning_rate": 2.487814784727864e-06,
+      "loss": 1.7825,
+      "step": 1526
+    },
+    {
+      "epoch": 1.6734246575342466,
+      "grad_norm": 2.3704216480255127,
+      "learning_rate": 2.4857839155158407e-06,
+      "loss": 1.8742,
+      "step": 1527
+    },
+    {
+      "epoch": 1.6745205479452054,
+      "grad_norm": 2.1590118408203125,
+      "learning_rate": 2.483753046303818e-06,
+      "loss": 1.7974,
+      "step": 1528
+    },
+    {
+      "epoch": 1.6756164383561645,
+      "grad_norm": 2.2389416694641113,
+      "learning_rate": 2.4817221770917953e-06,
+      "loss": 1.8007,
+      "step": 1529
+    },
+    {
+      "epoch": 1.6767123287671233,
+      "grad_norm": 2.324131488800049,
+      "learning_rate": 2.4796913078797725e-06,
+      "loss": 1.8079,
+      "step": 1530
+    },
+    {
+      "epoch": 1.6767123287671233,
+      "eval_loss": 2.013923168182373,
+      "eval_runtime": 37.0828,
+      "eval_samples_per_second": 26.967,
+      "eval_steps_per_second": 1.133,
+      "step": 1530
+    },
+    {
+      "epoch": 1.677808219178082,
+      "grad_norm": 2.406456708908081,
+      "learning_rate": 2.47766043866775e-06,
+      "loss": 1.8356,
+      "step": 1531
+    },
+    {
+      "epoch": 1.6789041095890411,
+      "grad_norm": 2.289919137954712,
+      "learning_rate": 2.475629569455727e-06,
+      "loss": 1.8239,
+      "step": 1532
+    },
+    {
+      "epoch": 1.6800000000000002,
+      "grad_norm": 2.2617733478546143,
+      "learning_rate": 2.4735987002437044e-06,
+      "loss": 1.8025,
+      "step": 1533
+    },
+    {
+      "epoch": 1.6810958904109587,
+      "grad_norm": 2.133652925491333,
+      "learning_rate": 2.4715678310316816e-06,
+      "loss": 1.7743,
+      "step": 1534
+    },
+    {
+      "epoch": 1.6821917808219178,
+      "grad_norm": 2.215322256088257,
+      "learning_rate": 2.469536961819659e-06,
+      "loss": 1.7979,
+      "step": 1535
+    },
+    {
+      "epoch": 1.6832876712328768,
+      "grad_norm": 2.336021900177002,
+      "learning_rate": 2.467506092607636e-06,
+      "loss": 1.8027,
+      "step": 1536
+    },
+    {
+      "epoch": 1.6843835616438356,
+      "grad_norm": 2.3567142486572266,
+      "learning_rate": 2.4654752233956135e-06,
+      "loss": 1.7877,
+      "step": 1537
+    },
+    {
+      "epoch": 1.6854794520547944,
+      "grad_norm": 2.2642927169799805,
+      "learning_rate": 2.4634443541835907e-06,
+      "loss": 1.7695,
+      "step": 1538
+    },
+    {
+      "epoch": 1.6865753424657535,
+      "grad_norm": 2.353303909301758,
+      "learning_rate": 2.461413484971568e-06,
+      "loss": 1.867,
+      "step": 1539
+    },
+    {
+      "epoch": 1.6876712328767123,
+      "grad_norm": 2.252074956893921,
+      "learning_rate": 2.4593826157595453e-06,
+      "loss": 1.8009,
+      "step": 1540
+    },
+    {
+      "epoch": 1.6876712328767123,
+      "eval_loss": 2.0113608837127686,
+      "eval_runtime": 37.0875,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 1540
+    },
+    {
+      "epoch": 1.688767123287671,
+      "grad_norm": 2.2303407192230225,
+      "learning_rate": 2.4573517465475226e-06,
+      "loss": 1.8881,
+      "step": 1541
+    },
+    {
+      "epoch": 1.6898630136986301,
+      "grad_norm": 2.3359081745147705,
+      "learning_rate": 2.4553208773355e-06,
+      "loss": 1.7388,
+      "step": 1542
+    },
+    {
+      "epoch": 1.6909589041095892,
+      "grad_norm": 2.3783583641052246,
+      "learning_rate": 2.453290008123477e-06,
+      "loss": 1.7993,
+      "step": 1543
+    },
+    {
+      "epoch": 1.692054794520548,
+      "grad_norm": 2.159264326095581,
+      "learning_rate": 2.4512591389114544e-06,
+      "loss": 1.81,
+      "step": 1544
+    },
+    {
+      "epoch": 1.6931506849315068,
+      "grad_norm": 2.2232203483581543,
+      "learning_rate": 2.4492282696994317e-06,
+      "loss": 1.8095,
+      "step": 1545
+    },
+    {
+      "epoch": 1.6942465753424658,
+      "grad_norm": 2.1755056381225586,
+      "learning_rate": 2.447197400487409e-06,
+      "loss": 1.8149,
+      "step": 1546
+    },
+    {
+      "epoch": 1.6953424657534246,
+      "grad_norm": 2.325871467590332,
+      "learning_rate": 2.4451665312753862e-06,
+      "loss": 1.8154,
+      "step": 1547
+    },
+    {
+      "epoch": 1.6964383561643834,
+      "grad_norm": 2.286848545074463,
+      "learning_rate": 2.4431356620633635e-06,
+      "loss": 1.8409,
+      "step": 1548
+    },
+    {
+      "epoch": 1.6975342465753425,
+      "grad_norm": 2.359510898590088,
+      "learning_rate": 2.4411047928513408e-06,
+      "loss": 1.843,
+      "step": 1549
+    },
+    {
+      "epoch": 1.6986301369863015,
+      "grad_norm": 2.474428176879883,
+      "learning_rate": 2.439073923639318e-06,
+      "loss": 1.8089,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6986301369863015,
+      "eval_loss": 2.0077731609344482,
+      "eval_runtime": 37.1092,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1550
+    },
+    {
+      "epoch": 1.6997260273972603,
+      "grad_norm": 2.3662912845611572,
+      "learning_rate": 2.4370430544272953e-06,
+      "loss": 1.8176,
+      "step": 1551
+    },
+    {
+      "epoch": 1.7008219178082191,
+      "grad_norm": 2.448948860168457,
+      "learning_rate": 2.435012185215272e-06,
+      "loss": 1.7776,
+      "step": 1552
+    },
+    {
+      "epoch": 1.7019178082191782,
+      "grad_norm": 2.333739757537842,
+      "learning_rate": 2.4329813160032495e-06,
+      "loss": 1.7552,
+      "step": 1553
+    },
+    {
+      "epoch": 1.703013698630137,
+      "grad_norm": 2.170365571975708,
+      "learning_rate": 2.4309504467912267e-06,
+      "loss": 1.7759,
+      "step": 1554
+    },
+    {
+      "epoch": 1.7041095890410958,
+      "grad_norm": 2.2343010902404785,
+      "learning_rate": 2.428919577579204e-06,
+      "loss": 1.8073,
+      "step": 1555
+    },
+    {
+      "epoch": 1.7052054794520548,
+      "grad_norm": 2.2245287895202637,
+      "learning_rate": 2.4268887083671813e-06,
+      "loss": 1.7826,
+      "step": 1556
+    },
+    {
+      "epoch": 1.7063013698630138,
+      "grad_norm": 2.339942216873169,
+      "learning_rate": 2.4248578391551586e-06,
+      "loss": 1.8766,
+      "step": 1557
+    },
+    {
+      "epoch": 1.7073972602739727,
+      "grad_norm": 2.1585116386413574,
+      "learning_rate": 2.422826969943136e-06,
+      "loss": 1.7605,
+      "step": 1558
+    },
+    {
+      "epoch": 1.7084931506849315,
+      "grad_norm": 2.2108142375946045,
+      "learning_rate": 2.420796100731113e-06,
+      "loss": 1.7855,
+      "step": 1559
+    },
+    {
+      "epoch": 1.7095890410958905,
+      "grad_norm": 2.192457675933838,
+      "learning_rate": 2.4187652315190904e-06,
+      "loss": 1.7489,
+      "step": 1560
+    },
+    {
+      "epoch": 1.7095890410958905,
+      "eval_loss": 2.0085604190826416,
+      "eval_runtime": 37.1105,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1560
+    },
+    {
+      "epoch": 1.7106849315068493,
+      "grad_norm": 2.291106939315796,
+      "learning_rate": 2.4167343623070677e-06,
+      "loss": 1.7981,
+      "step": 1561
+    },
+    {
+      "epoch": 1.7117808219178081,
+      "grad_norm": 2.509237766265869,
+      "learning_rate": 2.414703493095045e-06,
+      "loss": 1.8146,
+      "step": 1562
+    },
+    {
+      "epoch": 1.7128767123287671,
+      "grad_norm": 2.497162342071533,
+      "learning_rate": 2.4126726238830222e-06,
+      "loss": 1.807,
+      "step": 1563
+    },
+    {
+      "epoch": 1.7139726027397262,
+      "grad_norm": 2.256833076477051,
+      "learning_rate": 2.410641754670999e-06,
+      "loss": 1.8304,
+      "step": 1564
+    },
+    {
+      "epoch": 1.7150684931506848,
+      "grad_norm": 2.2523772716522217,
+      "learning_rate": 2.4086108854589763e-06,
+      "loss": 1.7794,
+      "step": 1565
+    },
+    {
+      "epoch": 1.7161643835616438,
+      "grad_norm": 2.289182424545288,
+      "learning_rate": 2.4065800162469536e-06,
+      "loss": 1.8425,
+      "step": 1566
+    },
+    {
+      "epoch": 1.7172602739726028,
+      "grad_norm": 2.734487771987915,
+      "learning_rate": 2.404549147034931e-06,
+      "loss": 1.7791,
+      "step": 1567
+    },
+    {
+      "epoch": 1.7183561643835616,
+      "grad_norm": 2.3307888507843018,
+      "learning_rate": 2.402518277822908e-06,
+      "loss": 1.8064,
+      "step": 1568
+    },
+    {
+      "epoch": 1.7194520547945205,
+      "grad_norm": 2.5016298294067383,
+      "learning_rate": 2.4004874086108854e-06,
+      "loss": 1.8544,
+      "step": 1569
+    },
+    {
+      "epoch": 1.7205479452054795,
+      "grad_norm": 2.142246961593628,
+      "learning_rate": 2.3984565393988627e-06,
+      "loss": 1.7653,
+      "step": 1570
+    },
+    {
+      "epoch": 1.7205479452054795,
+      "eval_loss": 2.0068864822387695,
+      "eval_runtime": 37.1078,
+      "eval_samples_per_second": 26.949,
+      "eval_steps_per_second": 1.132,
+      "step": 1570
+    },
+    {
+      "epoch": 1.7216438356164385,
+      "grad_norm": 2.5692505836486816,
+      "learning_rate": 2.39642567018684e-06,
+      "loss": 1.8444,
+      "step": 1571
+    },
+    {
+      "epoch": 1.7227397260273971,
+      "grad_norm": 2.2596213817596436,
+      "learning_rate": 2.3943948009748173e-06,
+      "loss": 1.7694,
+      "step": 1572
+    },
+    {
+      "epoch": 1.7238356164383561,
+      "grad_norm": 2.230984926223755,
+      "learning_rate": 2.3923639317627945e-06,
+      "loss": 1.786,
+      "step": 1573
+    },
+    {
+      "epoch": 1.7249315068493152,
+      "grad_norm": 2.366178274154663,
+      "learning_rate": 2.390333062550772e-06,
+      "loss": 1.8138,
+      "step": 1574
+    },
+    {
+      "epoch": 1.726027397260274,
+      "grad_norm": 2.23783016204834,
+      "learning_rate": 2.388302193338749e-06,
+      "loss": 1.7615,
+      "step": 1575
+    },
+    {
+      "epoch": 1.7271232876712328,
+      "grad_norm": 2.3766677379608154,
+      "learning_rate": 2.3862713241267264e-06,
+      "loss": 1.6811,
+      "step": 1576
+    },
+    {
+      "epoch": 1.7282191780821918,
+      "grad_norm": 2.3434040546417236,
+      "learning_rate": 2.3842404549147037e-06,
+      "loss": 1.8806,
+      "step": 1577
+    },
+    {
+      "epoch": 1.7293150684931506,
+      "grad_norm": 2.2444276809692383,
+      "learning_rate": 2.382209585702681e-06,
+      "loss": 1.7968,
+      "step": 1578
+    },
+    {
+      "epoch": 1.7304109589041095,
+      "grad_norm": 2.2524919509887695,
+      "learning_rate": 2.380178716490658e-06,
+      "loss": 1.7602,
+      "step": 1579
+    },
+    {
+      "epoch": 1.7315068493150685,
+      "grad_norm": 2.3310587406158447,
+      "learning_rate": 2.3781478472786355e-06,
+      "loss": 1.8414,
+      "step": 1580
+    },
+    {
+      "epoch": 1.7315068493150685,
+      "eval_loss": 1.9981482028961182,
+      "eval_runtime": 37.0827,
+      "eval_samples_per_second": 26.967,
+      "eval_steps_per_second": 1.133,
+      "step": 1580
+    },
+    {
+      "epoch": 1.7326027397260275,
+      "grad_norm": 2.4081802368164062,
+      "learning_rate": 2.3761169780666128e-06,
+      "loss": 1.8002,
+      "step": 1581
+    },
+    {
+      "epoch": 1.7336986301369863,
+      "grad_norm": 2.293506622314453,
+      "learning_rate": 2.37408610885459e-06,
+      "loss": 1.8365,
+      "step": 1582
+    },
+    {
+      "epoch": 1.7347945205479451,
+      "grad_norm": 2.2015790939331055,
+      "learning_rate": 2.3720552396425673e-06,
+      "loss": 1.8571,
+      "step": 1583
+    },
+    {
+      "epoch": 1.7358904109589042,
+      "grad_norm": 2.308809518814087,
+      "learning_rate": 2.3700243704305446e-06,
+      "loss": 1.7821,
+      "step": 1584
+    },
+    {
+      "epoch": 1.736986301369863,
+      "grad_norm": 2.376842975616455,
+      "learning_rate": 2.367993501218522e-06,
+      "loss": 1.8701,
+      "step": 1585
+    },
+    {
+      "epoch": 1.7380821917808218,
+      "grad_norm": 2.4102022647857666,
+      "learning_rate": 2.365962632006499e-06,
+      "loss": 1.8055,
+      "step": 1586
+    },
+    {
+      "epoch": 1.7391780821917808,
+      "grad_norm": 2.2734079360961914,
+      "learning_rate": 2.3639317627944764e-06,
+      "loss": 1.8117,
+      "step": 1587
+    },
+    {
+      "epoch": 1.7402739726027399,
+      "grad_norm": 2.3426599502563477,
+      "learning_rate": 2.3619008935824537e-06,
+      "loss": 1.8337,
+      "step": 1588
+    },
+    {
+      "epoch": 1.7413698630136987,
+      "grad_norm": 2.340383768081665,
+      "learning_rate": 2.3598700243704305e-06,
+      "loss": 1.8247,
+      "step": 1589
+    },
+    {
+      "epoch": 1.7424657534246575,
+      "grad_norm": 2.2561380863189697,
+      "learning_rate": 2.357839155158408e-06,
+      "loss": 1.8249,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7424657534246575,
+      "eval_loss": 1.9979513883590698,
+      "eval_runtime": 37.0948,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 1590
+    },
+    {
+      "epoch": 1.7435616438356165,
+      "grad_norm": 2.2459654808044434,
+      "learning_rate": 2.355808285946385e-06,
+      "loss": 1.8273,
+      "step": 1591
+    },
+    {
+      "epoch": 1.7446575342465753,
+      "grad_norm": 2.2017974853515625,
+      "learning_rate": 2.3537774167343624e-06,
+      "loss": 1.828,
+      "step": 1592
+    },
+    {
+      "epoch": 1.7457534246575341,
+      "grad_norm": 2.286304235458374,
+      "learning_rate": 2.3517465475223396e-06,
+      "loss": 1.8497,
+      "step": 1593
+    },
+    {
+      "epoch": 1.7468493150684932,
+      "grad_norm": 2.296494722366333,
+      "learning_rate": 2.349715678310317e-06,
+      "loss": 1.8155,
+      "step": 1594
+    },
+    {
+      "epoch": 1.7479452054794522,
+      "grad_norm": 2.2989203929901123,
+      "learning_rate": 2.347684809098294e-06,
+      "loss": 1.8214,
+      "step": 1595
+    },
+    {
+      "epoch": 1.749041095890411,
+      "grad_norm": 2.3034965991973877,
+      "learning_rate": 2.3456539398862715e-06,
+      "loss": 1.7792,
+      "step": 1596
+    },
+    {
+      "epoch": 1.7501369863013698,
+      "grad_norm": 2.3823719024658203,
+      "learning_rate": 2.3436230706742487e-06,
+      "loss": 1.8326,
+      "step": 1597
+    },
+    {
+      "epoch": 1.7512328767123289,
+      "grad_norm": 2.231095552444458,
+      "learning_rate": 2.341592201462226e-06,
+      "loss": 1.7713,
+      "step": 1598
+    },
+    {
+      "epoch": 1.7523287671232877,
+      "grad_norm": 2.287524938583374,
+      "learning_rate": 2.3395613322502033e-06,
+      "loss": 1.8143,
+      "step": 1599
+    },
+    {
+      "epoch": 1.7534246575342465,
+      "grad_norm": 2.267552614212036,
+      "learning_rate": 2.3375304630381806e-06,
+      "loss": 1.7883,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7534246575342465,
+      "eval_loss": 1.998821496963501,
+      "eval_runtime": 37.1222,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 1600
+    },
+    {
+      "epoch": 1.7545205479452055,
+      "grad_norm": 2.2986714839935303,
+      "learning_rate": 2.335499593826158e-06,
+      "loss": 1.7615,
+      "step": 1601
+    },
+    {
+      "epoch": 1.7556164383561645,
+      "grad_norm": 2.2913246154785156,
+      "learning_rate": 2.333468724614135e-06,
+      "loss": 1.7584,
+      "step": 1602
+    },
+    {
+      "epoch": 1.7567123287671231,
+      "grad_norm": 2.2486486434936523,
+      "learning_rate": 2.3314378554021124e-06,
+      "loss": 1.8089,
+      "step": 1603
+    },
+    {
+      "epoch": 1.7578082191780822,
+      "grad_norm": 2.319581985473633,
+      "learning_rate": 2.3294069861900897e-06,
+      "loss": 1.8364,
+      "step": 1604
+    },
+    {
+      "epoch": 1.7589041095890412,
+      "grad_norm": 2.2775609493255615,
+      "learning_rate": 2.327376116978067e-06,
+      "loss": 1.7661,
+      "step": 1605
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 2.3445472717285156,
+      "learning_rate": 2.3253452477660442e-06,
+      "loss": 1.7583,
+      "step": 1606
+    },
+    {
+      "epoch": 1.7610958904109588,
+      "grad_norm": 2.420865058898926,
+      "learning_rate": 2.3233143785540215e-06,
+      "loss": 1.8411,
+      "step": 1607
+    },
+    {
+      "epoch": 1.7621917808219179,
+      "grad_norm": 2.294821262359619,
+      "learning_rate": 2.3212835093419988e-06,
+      "loss": 1.7812,
+      "step": 1608
+    },
+    {
+      "epoch": 1.7632876712328767,
+      "grad_norm": 2.332784414291382,
+      "learning_rate": 2.319252640129976e-06,
+      "loss": 1.8252,
+      "step": 1609
+    },
+    {
+      "epoch": 1.7643835616438355,
+      "grad_norm": 2.433619499206543,
+      "learning_rate": 2.3172217709179533e-06,
+      "loss": 1.7709,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7643835616438355,
+      "eval_loss": 2.0010287761688232,
+      "eval_runtime": 36.7858,
+      "eval_samples_per_second": 27.184,
+      "eval_steps_per_second": 1.142,
+      "step": 1610
+    },
+    {
+      "epoch": 1.7654794520547945,
+      "grad_norm": 2.532618522644043,
+      "learning_rate": 2.3151909017059306e-06,
+      "loss": 1.8677,
+      "step": 1611
+    },
+    {
+      "epoch": 1.7665753424657535,
+      "grad_norm": 2.354578733444214,
+      "learning_rate": 2.313160032493908e-06,
+      "loss": 1.8412,
+      "step": 1612
+    },
+    {
+      "epoch": 1.7676712328767124,
+      "grad_norm": 2.382774591445923,
+      "learning_rate": 2.311129163281885e-06,
+      "loss": 1.7543,
+      "step": 1613
+    },
+    {
+      "epoch": 1.7687671232876712,
+      "grad_norm": 2.3443398475646973,
+      "learning_rate": 2.309098294069862e-06,
+      "loss": 1.8267,
+      "step": 1614
+    },
+    {
+      "epoch": 1.7698630136986302,
+      "grad_norm": 2.3804502487182617,
+      "learning_rate": 2.3070674248578393e-06,
+      "loss": 1.7582,
+      "step": 1615
+    },
+    {
+      "epoch": 1.770958904109589,
+      "grad_norm": 2.3163580894470215,
+      "learning_rate": 2.3050365556458166e-06,
+      "loss": 1.7539,
+      "step": 1616
+    },
+    {
+      "epoch": 1.7720547945205478,
+      "grad_norm": 2.605738639831543,
+      "learning_rate": 2.303005686433794e-06,
+      "loss": 1.8124,
+      "step": 1617
+    },
+    {
+      "epoch": 1.7731506849315068,
+      "grad_norm": 2.3859915733337402,
+      "learning_rate": 2.300974817221771e-06,
+      "loss": 1.8271,
+      "step": 1618
+    },
+    {
+      "epoch": 1.7742465753424659,
+      "grad_norm": 2.3715479373931885,
+      "learning_rate": 2.2989439480097484e-06,
+      "loss": 1.7823,
+      "step": 1619
+    },
+    {
+      "epoch": 1.7753424657534247,
+      "grad_norm": 2.441100597381592,
+      "learning_rate": 2.2969130787977257e-06,
+      "loss": 1.7461,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7753424657534247,
+      "eval_loss": 2.002152442932129,
+      "eval_runtime": 37.0246,
+      "eval_samples_per_second": 27.009,
+      "eval_steps_per_second": 1.134,
+      "step": 1620
+    },
+    {
+      "epoch": 1.7764383561643835,
+      "grad_norm": 2.4154021739959717,
+      "learning_rate": 2.294882209585703e-06,
+      "loss": 1.8012,
+      "step": 1621
+    },
+    {
+      "epoch": 1.7775342465753425,
+      "grad_norm": 2.333400249481201,
+      "learning_rate": 2.2928513403736802e-06,
+      "loss": 1.8222,
+      "step": 1622
+    },
+    {
+      "epoch": 1.7786301369863013,
+      "grad_norm": 2.3183937072753906,
+      "learning_rate": 2.2908204711616575e-06,
+      "loss": 1.7561,
+      "step": 1623
+    },
+    {
+      "epoch": 1.7797260273972602,
+      "grad_norm": 2.302823066711426,
+      "learning_rate": 2.2887896019496348e-06,
+      "loss": 1.7975,
+      "step": 1624
+    },
+    {
+      "epoch": 1.7808219178082192,
+      "grad_norm": 2.3702917098999023,
+      "learning_rate": 2.286758732737612e-06,
+      "loss": 1.8478,
+      "step": 1625
+    },
+    {
+      "epoch": 1.7819178082191782,
+      "grad_norm": 2.4997549057006836,
+      "learning_rate": 2.284727863525589e-06,
+      "loss": 1.7855,
+      "step": 1626
+    },
+    {
+      "epoch": 1.783013698630137,
+      "grad_norm": 2.296191930770874,
+      "learning_rate": 2.282696994313566e-06,
+      "loss": 1.7922,
+      "step": 1627
+    },
+    {
+      "epoch": 1.7841095890410958,
+      "grad_norm": 2.2567873001098633,
+      "learning_rate": 2.2806661251015434e-06,
+      "loss": 1.8455,
+      "step": 1628
+    },
+    {
+      "epoch": 1.7852054794520549,
+      "grad_norm": 2.393336534500122,
+      "learning_rate": 2.2786352558895207e-06,
+      "loss": 1.8141,
+      "step": 1629
+    },
+    {
+      "epoch": 1.7863013698630137,
+      "grad_norm": 2.610466480255127,
+      "learning_rate": 2.276604386677498e-06,
+      "loss": 1.7875,
+      "step": 1630
+    },
+    {
+      "epoch": 1.7863013698630137,
+      "eval_loss": 1.9985637664794922,
+      "eval_runtime": 37.1001,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 1630
+    },
+    {
+      "epoch": 1.7873972602739725,
+      "grad_norm": 2.2163283824920654,
+      "learning_rate": 2.2745735174654753e-06,
+      "loss": 1.7479,
+      "step": 1631
+    },
+    {
+      "epoch": 1.7884931506849315,
+      "grad_norm": 2.2641165256500244,
+      "learning_rate": 2.2725426482534526e-06,
+      "loss": 1.7749,
+      "step": 1632
+    },
+    {
+      "epoch": 1.7895890410958906,
+      "grad_norm": 2.2786946296691895,
+      "learning_rate": 2.27051177904143e-06,
+      "loss": 1.8403,
+      "step": 1633
+    },
+    {
+      "epoch": 1.7906849315068492,
+      "grad_norm": 2.660356044769287,
+      "learning_rate": 2.268480909829407e-06,
+      "loss": 1.7994,
+      "step": 1634
+    },
+    {
+      "epoch": 1.7917808219178082,
+      "grad_norm": 2.322181224822998,
+      "learning_rate": 2.2664500406173844e-06,
+      "loss": 1.7917,
+      "step": 1635
+    },
+    {
+      "epoch": 1.7928767123287672,
+      "grad_norm": 2.3020682334899902,
+      "learning_rate": 2.2644191714053617e-06,
+      "loss": 1.7983,
+      "step": 1636
+    },
+    {
+      "epoch": 1.793972602739726,
+      "grad_norm": 2.5280368328094482,
+      "learning_rate": 2.262388302193339e-06,
+      "loss": 1.8378,
+      "step": 1637
+    },
+    {
+      "epoch": 1.7950684931506848,
+      "grad_norm": 2.3975653648376465,
+      "learning_rate": 2.260357432981316e-06,
+      "loss": 1.8081,
+      "step": 1638
+    },
+    {
+      "epoch": 1.7961643835616439,
+      "grad_norm": 2.419806957244873,
+      "learning_rate": 2.2583265637692935e-06,
+      "loss": 1.793,
+      "step": 1639
+    },
+    {
+      "epoch": 1.7972602739726027,
+      "grad_norm": 2.4601471424102783,
+      "learning_rate": 2.2562956945572708e-06,
+      "loss": 1.8116,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7972602739726027,
+      "eval_loss": 1.9951634407043457,
+      "eval_runtime": 37.1095,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 1640
+    },
+    {
+      "epoch": 1.7983561643835615,
+      "grad_norm": 2.17830753326416,
+      "learning_rate": 2.254264825345248e-06,
+      "loss": 1.7845,
+      "step": 1641
+    },
+    {
+      "epoch": 1.7994520547945205,
+      "grad_norm": 2.3332247734069824,
+      "learning_rate": 2.2522339561332253e-06,
+      "loss": 1.7945,
+      "step": 1642
+    },
+    {
+      "epoch": 1.8005479452054796,
+      "grad_norm": 2.3841960430145264,
+      "learning_rate": 2.2502030869212026e-06,
+      "loss": 1.7994,
+      "step": 1643
+    },
+    {
+      "epoch": 1.8016438356164384,
+      "grad_norm": 2.3669073581695557,
+      "learning_rate": 2.24817221770918e-06,
+      "loss": 1.8053,
+      "step": 1644
+    },
+    {
+      "epoch": 1.8027397260273972,
+      "grad_norm": 2.406858205795288,
+      "learning_rate": 2.246141348497157e-06,
+      "loss": 1.8647,
+      "step": 1645
+    },
+    {
+      "epoch": 1.8038356164383562,
+      "grad_norm": 2.2854161262512207,
+      "learning_rate": 2.2441104792851344e-06,
+      "loss": 1.8312,
+      "step": 1646
+    },
+    {
+      "epoch": 1.804931506849315,
+      "grad_norm": 2.252976655960083,
+      "learning_rate": 2.2420796100731117e-06,
+      "loss": 1.7834,
+      "step": 1647
+    },
+    {
+      "epoch": 1.8060273972602738,
+      "grad_norm": 2.4131624698638916,
+      "learning_rate": 2.240048740861089e-06,
+      "loss": 1.8204,
+      "step": 1648
+    },
+    {
+      "epoch": 1.8071232876712329,
+      "grad_norm": 2.4727940559387207,
+      "learning_rate": 2.2380178716490662e-06,
+      "loss": 1.8326,
+      "step": 1649
+    },
+    {
+      "epoch": 1.808219178082192,
+      "grad_norm": 2.4432709217071533,
+      "learning_rate": 2.2359870024370435e-06,
+      "loss": 1.6911,
+      "step": 1650
+    },
+    {
+      "epoch": 1.808219178082192,
+      "eval_loss": 1.996634840965271,
+      "eval_runtime": 37.072,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 1650
+    },
+    {
+      "epoch": 1.8093150684931507,
+      "grad_norm": 2.3810362815856934,
+      "learning_rate": 2.2339561332250204e-06,
+      "loss": 1.8256,
+      "step": 1651
+    },
+    {
+      "epoch": 1.8104109589041095,
+      "grad_norm": 2.4489142894744873,
+      "learning_rate": 2.2319252640129976e-06,
+      "loss": 1.8368,
+      "step": 1652
+    },
+    {
+      "epoch": 1.8115068493150686,
+      "grad_norm": 2.458686113357544,
+      "learning_rate": 2.229894394800975e-06,
+      "loss": 1.8176,
+      "step": 1653
+    },
+    {
+      "epoch": 1.8126027397260274,
+      "grad_norm": 2.3977842330932617,
+      "learning_rate": 2.227863525588952e-06,
+      "loss": 1.7516,
+      "step": 1654
+    },
+    {
+      "epoch": 1.8136986301369862,
+      "grad_norm": 2.3777663707733154,
+      "learning_rate": 2.2258326563769295e-06,
+      "loss": 1.732,
+      "step": 1655
+    },
+    {
+      "epoch": 1.8147945205479452,
+      "grad_norm": 2.2874817848205566,
+      "learning_rate": 2.2238017871649067e-06,
+      "loss": 1.749,
+      "step": 1656
+    },
+    {
+      "epoch": 1.8158904109589042,
+      "grad_norm": 2.288114547729492,
+      "learning_rate": 2.221770917952884e-06,
+      "loss": 1.7357,
+      "step": 1657
+    },
+    {
+      "epoch": 1.816986301369863,
+      "grad_norm": 2.3532187938690186,
+      "learning_rate": 2.2197400487408613e-06,
+      "loss": 1.7953,
+      "step": 1658
+    },
+    {
+      "epoch": 1.8180821917808219,
+      "grad_norm": 2.2829396724700928,
+      "learning_rate": 2.2177091795288386e-06,
+      "loss": 1.8358,
+      "step": 1659
+    },
+    {
+      "epoch": 1.819178082191781,
+      "grad_norm": 2.3282570838928223,
+      "learning_rate": 2.215678310316816e-06,
+      "loss": 1.7448,
+      "step": 1660
+    },
+    {
+      "epoch": 1.819178082191781,
+      "eval_loss": 1.999374270439148,
+      "eval_runtime": 37.0739,
+      "eval_samples_per_second": 26.973,
+      "eval_steps_per_second": 1.133,
+      "step": 1660
+    },
+    {
+      "epoch": 1.8202739726027397,
+      "grad_norm": 2.474106788635254,
+      "learning_rate": 2.213647441104793e-06,
+      "loss": 1.7979,
+      "step": 1661
+    },
+    {
+      "epoch": 1.8213698630136985,
+      "grad_norm": 2.159498453140259,
+      "learning_rate": 2.2116165718927704e-06,
+      "loss": 1.8075,
+      "step": 1662
+    },
+    {
+      "epoch": 1.8224657534246576,
+      "grad_norm": 2.2503573894500732,
+      "learning_rate": 2.2095857026807473e-06,
+      "loss": 1.7742,
+      "step": 1663
+    },
+    {
+      "epoch": 1.8235616438356166,
+      "grad_norm": 2.4465060234069824,
+      "learning_rate": 2.2075548334687245e-06,
+      "loss": 1.8067,
+      "step": 1664
+    },
+    {
+      "epoch": 1.8246575342465754,
+      "grad_norm": 2.462158679962158,
+      "learning_rate": 2.205523964256702e-06,
+      "loss": 1.7456,
+      "step": 1665
+    },
+    {
+      "epoch": 1.8257534246575342,
+      "grad_norm": 2.354051113128662,
+      "learning_rate": 2.203493095044679e-06,
+      "loss": 1.7688,
+      "step": 1666
+    },
+    {
+      "epoch": 1.8268493150684932,
+      "grad_norm": 2.3165128231048584,
+      "learning_rate": 2.2014622258326564e-06,
+      "loss": 1.7986,
+      "step": 1667
+    },
+    {
+      "epoch": 1.827945205479452,
+      "grad_norm": 2.3131113052368164,
+      "learning_rate": 2.1994313566206336e-06,
+      "loss": 1.8348,
+      "step": 1668
+    },
+    {
+      "epoch": 1.8290410958904109,
+      "grad_norm": 2.303086996078491,
+      "learning_rate": 2.197400487408611e-06,
+      "loss": 1.8413,
+      "step": 1669
+    },
+    {
+      "epoch": 1.83013698630137,
+      "grad_norm": 2.392995834350586,
+      "learning_rate": 2.195369618196588e-06,
+      "loss": 1.7663,
+      "step": 1670
+    },
+    {
+      "epoch": 1.83013698630137,
+      "eval_loss": 1.9936721324920654,
+      "eval_runtime": 37.0717,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 1670
+    },
+    {
+      "epoch": 1.831232876712329,
+      "grad_norm": 2.323596477508545,
+      "learning_rate": 2.1933387489845655e-06,
+      "loss": 1.7463,
+      "step": 1671
+    },
+    {
+      "epoch": 1.8323287671232875,
+      "grad_norm": 2.3553171157836914,
+      "learning_rate": 2.1913078797725427e-06,
+      "loss": 1.7844,
+      "step": 1672
+    },
+    {
+      "epoch": 1.8334246575342465,
+      "grad_norm": 2.3724591732025146,
+      "learning_rate": 2.18927701056052e-06,
+      "loss": 1.7955,
+      "step": 1673
+    },
+    {
+      "epoch": 1.8345205479452056,
+      "grad_norm": 2.4300284385681152,
+      "learning_rate": 2.1872461413484973e-06,
+      "loss": 1.8697,
+      "step": 1674
+    },
+    {
+      "epoch": 1.8356164383561644,
+      "grad_norm": 2.4300284385681152,
+      "learning_rate": 2.1872461413484973e-06,
+      "loss": 1.8065,
+      "step": 1675
+    },
+    {
+      "epoch": 1.8367123287671232,
+      "grad_norm": 2.523627519607544,
+      "learning_rate": 2.1852152721364746e-06,
+      "loss": 1.801,
+      "step": 1676
+    },
+    {
+      "epoch": 1.8378082191780822,
+      "grad_norm": 2.2884891033172607,
+      "learning_rate": 2.183184402924452e-06,
+      "loss": 1.7512,
+      "step": 1677
+    },
+    {
+      "epoch": 1.838904109589041,
+      "grad_norm": 2.3114986419677734,
+      "learning_rate": 2.181153533712429e-06,
+      "loss": 1.8245,
+      "step": 1678
+    },
+    {
+      "epoch": 1.8399999999999999,
+      "grad_norm": 2.326855421066284,
+      "learning_rate": 2.1791226645004064e-06,
+      "loss": 1.7921,
+      "step": 1679
+    },
+    {
+      "epoch": 1.841095890410959,
+      "grad_norm": 2.3448753356933594,
+      "learning_rate": 2.1770917952883837e-06,
+      "loss": 1.7709,
+      "step": 1680
+    },
+    {
+      "epoch": 1.841095890410959,
+      "eval_loss": 1.996195673942566,
+      "eval_runtime": 37.0481,
+      "eval_samples_per_second": 26.992,
+      "eval_steps_per_second": 1.134,
+      "step": 1680
+    },
+    {
+      "epoch": 1.842191780821918,
+      "grad_norm": 2.223470687866211,
+      "learning_rate": 2.175060926076361e-06,
+      "loss": 1.8468,
+      "step": 1681
+    },
+    {
+      "epoch": 1.8432876712328767,
+      "grad_norm": 2.29726505279541,
+      "learning_rate": 2.1730300568643382e-06,
+      "loss": 1.8753,
+      "step": 1682
+    },
+    {
+      "epoch": 1.8443835616438355,
+      "grad_norm": 2.3869571685791016,
+      "learning_rate": 2.1709991876523155e-06,
+      "loss": 1.7835,
+      "step": 1683
+    },
+    {
+      "epoch": 1.8454794520547946,
+      "grad_norm": 2.192608594894409,
+      "learning_rate": 2.1689683184402928e-06,
+      "loss": 1.7556,
+      "step": 1684
+    },
+    {
+      "epoch": 1.8465753424657534,
+      "grad_norm": 2.355975866317749,
+      "learning_rate": 2.16693744922827e-06,
+      "loss": 1.8323,
+      "step": 1685
+    },
+    {
+      "epoch": 1.8476712328767122,
+      "grad_norm": 2.243938446044922,
+      "learning_rate": 2.1649065800162473e-06,
+      "loss": 1.7942,
+      "step": 1686
+    },
+    {
+      "epoch": 1.8487671232876712,
+      "grad_norm": 2.337069511413574,
+      "learning_rate": 2.1628757108042246e-06,
+      "loss": 1.7709,
+      "step": 1687
+    },
+    {
+      "epoch": 1.8498630136986303,
+      "grad_norm": 2.257504463195801,
+      "learning_rate": 2.160844841592202e-06,
+      "loss": 1.7496,
+      "step": 1688
+    },
+    {
+      "epoch": 1.850958904109589,
+      "grad_norm": 2.287336826324463,
+      "learning_rate": 2.1588139723801787e-06,
+      "loss": 1.7454,
+      "step": 1689
+    },
+    {
+      "epoch": 1.8520547945205479,
+      "grad_norm": 2.2864668369293213,
+      "learning_rate": 2.156783103168156e-06,
+      "loss": 1.7807,
+      "step": 1690
+    },
+    {
+      "epoch": 1.8520547945205479,
+      "eval_loss": 1.9940879344940186,
+      "eval_runtime": 37.0261,
+      "eval_samples_per_second": 27.008,
+      "eval_steps_per_second": 1.134,
+      "step": 1690
+    },
+    {
+      "epoch": 1.853150684931507,
+      "grad_norm": 2.1782145500183105,
+      "learning_rate": 2.1547522339561333e-06,
+      "loss": 1.7567,
+      "step": 1691
+    },
+    {
+      "epoch": 1.8542465753424657,
+      "grad_norm": 2.252023458480835,
+      "learning_rate": 2.1527213647441106e-06,
+      "loss": 1.811,
+      "step": 1692
+    },
+    {
+      "epoch": 1.8553424657534245,
+      "grad_norm": 2.2532849311828613,
+      "learning_rate": 2.150690495532088e-06,
+      "loss": 1.8491,
+      "step": 1693
+    },
+    {
+      "epoch": 1.8564383561643836,
+      "grad_norm": 2.35648250579834,
+      "learning_rate": 2.148659626320065e-06,
+      "loss": 1.8125,
+      "step": 1694
+    },
+    {
+      "epoch": 1.8575342465753426,
+      "grad_norm": 2.3490521907806396,
+      "learning_rate": 2.1466287571080424e-06,
+      "loss": 1.8035,
+      "step": 1695
+    },
+    {
+      "epoch": 1.8586301369863014,
+      "grad_norm": 2.2851879596710205,
+      "learning_rate": 2.1445978878960197e-06,
+      "loss": 1.7755,
+      "step": 1696
+    },
+    {
+      "epoch": 1.8597260273972602,
+      "grad_norm": 2.166754961013794,
+      "learning_rate": 2.142567018683997e-06,
+      "loss": 1.7861,
+      "step": 1697
+    },
+    {
+      "epoch": 1.8608219178082193,
+      "grad_norm": 2.2356410026550293,
+      "learning_rate": 2.140536149471974e-06,
+      "loss": 1.8412,
+      "step": 1698
+    },
+    {
+      "epoch": 1.861917808219178,
+      "grad_norm": 2.215618133544922,
+      "learning_rate": 2.1385052802599515e-06,
+      "loss": 1.7623,
+      "step": 1699
+    },
+    {
+      "epoch": 1.8630136986301369,
+      "grad_norm": 2.283298969268799,
+      "learning_rate": 2.1364744110479288e-06,
+      "loss": 1.7692,
+      "step": 1700
+    },
+    {
+      "epoch": 1.8630136986301369,
+      "eval_loss": 1.9961501359939575,
+      "eval_runtime": 37.0338,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1700
+    },
+    {
+      "epoch": 1.864109589041096,
+      "grad_norm": 2.292818546295166,
+      "learning_rate": 2.134443541835906e-06,
+      "loss": 1.8251,
+      "step": 1701
+    },
+    {
+      "epoch": 1.865205479452055,
+      "grad_norm": 2.235895872116089,
+      "learning_rate": 2.132412672623883e-06,
+      "loss": 1.7726,
+      "step": 1702
+    },
+    {
+      "epoch": 1.8663013698630135,
+      "grad_norm": 2.239413022994995,
+      "learning_rate": 2.13038180341186e-06,
+      "loss": 1.7509,
+      "step": 1703
+    },
+    {
+      "epoch": 1.8673972602739726,
+      "grad_norm": 2.4125585556030273,
+      "learning_rate": 2.1283509341998374e-06,
+      "loss": 1.7973,
+      "step": 1704
+    },
+    {
+      "epoch": 1.8684931506849316,
+      "grad_norm": 2.3774733543395996,
+      "learning_rate": 2.1263200649878147e-06,
+      "loss": 1.8526,
+      "step": 1705
+    },
+    {
+      "epoch": 1.8695890410958904,
+      "grad_norm": 2.344918966293335,
+      "learning_rate": 2.124289195775792e-06,
+      "loss": 1.7708,
+      "step": 1706
+    },
+    {
+      "epoch": 1.8706849315068492,
+      "grad_norm": 2.1933865547180176,
+      "learning_rate": 2.1222583265637693e-06,
+      "loss": 1.7935,
+      "step": 1707
+    },
+    {
+      "epoch": 1.8717808219178083,
+      "grad_norm": 2.4939582347869873,
+      "learning_rate": 2.1202274573517465e-06,
+      "loss": 1.706,
+      "step": 1708
+    },
+    {
+      "epoch": 1.872876712328767,
+      "grad_norm": 2.407532215118408,
+      "learning_rate": 2.118196588139724e-06,
+      "loss": 1.7522,
+      "step": 1709
+    },
+    {
+      "epoch": 1.8739726027397259,
+      "grad_norm": 2.297900915145874,
+      "learning_rate": 2.116165718927701e-06,
+      "loss": 1.7558,
+      "step": 1710
+    },
+    {
+      "epoch": 1.8739726027397259,
+      "eval_loss": 1.9946404695510864,
+      "eval_runtime": 36.7569,
+      "eval_samples_per_second": 27.206,
+      "eval_steps_per_second": 1.143,
+      "step": 1710
+    },
+    {
+      "epoch": 1.875068493150685,
+      "grad_norm": 2.2298660278320312,
+      "learning_rate": 2.1141348497156784e-06,
+      "loss": 1.7836,
+      "step": 1711
+    },
+    {
+      "epoch": 1.876164383561644,
+      "grad_norm": 2.308060884475708,
+      "learning_rate": 2.1121039805036556e-06,
+      "loss": 1.7616,
+      "step": 1712
+    },
+    {
+      "epoch": 1.8772602739726028,
+      "grad_norm": 2.288381338119507,
+      "learning_rate": 2.110073111291633e-06,
+      "loss": 1.7502,
+      "step": 1713
+    },
+    {
+      "epoch": 1.8783561643835616,
+      "grad_norm": 2.2191097736358643,
+      "learning_rate": 2.10804224207961e-06,
+      "loss": 1.7946,
+      "step": 1714
+    },
+    {
+      "epoch": 1.8794520547945206,
+      "grad_norm": 2.211777925491333,
+      "learning_rate": 2.1060113728675875e-06,
+      "loss": 1.8238,
+      "step": 1715
+    },
+    {
+      "epoch": 1.8805479452054794,
+      "grad_norm": 2.3946406841278076,
+      "learning_rate": 2.1039805036555647e-06,
+      "loss": 1.8146,
+      "step": 1716
+    },
+    {
+      "epoch": 1.8816438356164382,
+      "grad_norm": 2.2949914932250977,
+      "learning_rate": 2.101949634443542e-06,
+      "loss": 1.6877,
+      "step": 1717
+    },
+    {
+      "epoch": 1.8827397260273973,
+      "grad_norm": 2.2905256748199463,
+      "learning_rate": 2.0999187652315193e-06,
+      "loss": 1.7868,
+      "step": 1718
+    },
+    {
+      "epoch": 1.8838356164383563,
+      "grad_norm": 2.289280891418457,
+      "learning_rate": 2.0978878960194966e-06,
+      "loss": 1.7387,
+      "step": 1719
+    },
+    {
+      "epoch": 1.884931506849315,
+      "grad_norm": 2.35497784614563,
+      "learning_rate": 2.095857026807474e-06,
+      "loss": 1.8002,
+      "step": 1720
+    },
+    {
+      "epoch": 1.884931506849315,
+      "eval_loss": 1.9928913116455078,
+      "eval_runtime": 36.9289,
+      "eval_samples_per_second": 27.079,
+      "eval_steps_per_second": 1.137,
+      "step": 1720
+    },
+    {
+      "epoch": 1.886027397260274,
+      "grad_norm": 2.4404468536376953,
+      "learning_rate": 2.093826157595451e-06,
+      "loss": 1.7981,
+      "step": 1721
+    },
+    {
+      "epoch": 1.887123287671233,
+      "grad_norm": 2.315718173980713,
+      "learning_rate": 2.0917952883834284e-06,
+      "loss": 1.7581,
+      "step": 1722
+    },
+    {
+      "epoch": 1.8882191780821918,
+      "grad_norm": 2.2984843254089355,
+      "learning_rate": 2.0897644191714057e-06,
+      "loss": 1.8153,
+      "step": 1723
+    },
+    {
+      "epoch": 1.8893150684931506,
+      "grad_norm": 2.3332557678222656,
+      "learning_rate": 2.087733549959383e-06,
+      "loss": 1.831,
+      "step": 1724
+    },
+    {
+      "epoch": 1.8904109589041096,
+      "grad_norm": 2.3394649028778076,
+      "learning_rate": 2.0857026807473602e-06,
+      "loss": 1.7486,
+      "step": 1725
+    },
+    {
+      "epoch": 1.8915068493150686,
+      "grad_norm": 2.2985246181488037,
+      "learning_rate": 2.0836718115353375e-06,
+      "loss": 1.721,
+      "step": 1726
+    },
+    {
+      "epoch": 1.8926027397260274,
+      "grad_norm": 2.318504810333252,
+      "learning_rate": 2.0816409423233144e-06,
+      "loss": 1.8469,
+      "step": 1727
+    },
+    {
+      "epoch": 1.8936986301369862,
+      "grad_norm": 2.4938831329345703,
+      "learning_rate": 2.0796100731112916e-06,
+      "loss": 1.7525,
+      "step": 1728
+    },
+    {
+      "epoch": 1.8947945205479453,
+      "grad_norm": 2.3391780853271484,
+      "learning_rate": 2.077579203899269e-06,
+      "loss": 1.7733,
+      "step": 1729
+    },
+    {
+      "epoch": 1.895890410958904,
+      "grad_norm": 2.437540054321289,
+      "learning_rate": 2.075548334687246e-06,
+      "loss": 1.7344,
+      "step": 1730
+    },
+    {
+      "epoch": 1.895890410958904,
+      "eval_loss": 1.990493655204773,
+      "eval_runtime": 37.0347,
+      "eval_samples_per_second": 27.002,
+      "eval_steps_per_second": 1.134,
+      "step": 1730
+    },
+    {
+      "epoch": 1.896986301369863,
+      "grad_norm": 2.413628578186035,
+      "learning_rate": 2.0735174654752235e-06,
+      "loss": 1.7978,
+      "step": 1731
+    },
+    {
+      "epoch": 1.898082191780822,
+      "grad_norm": 2.3673973083496094,
+      "learning_rate": 2.0714865962632007e-06,
+      "loss": 1.7414,
+      "step": 1732
+    },
+    {
+      "epoch": 1.899178082191781,
+      "grad_norm": 2.3961968421936035,
+      "learning_rate": 2.069455727051178e-06,
+      "loss": 1.7854,
+      "step": 1733
+    },
+    {
+      "epoch": 1.9002739726027398,
+      "grad_norm": 2.5583736896514893,
+      "learning_rate": 2.0674248578391553e-06,
+      "loss": 1.7629,
+      "step": 1734
+    },
+    {
+      "epoch": 1.9013698630136986,
+      "grad_norm": 2.256418228149414,
+      "learning_rate": 2.0653939886271326e-06,
+      "loss": 1.7845,
+      "step": 1735
+    },
+    {
+      "epoch": 1.9024657534246576,
+      "grad_norm": 2.4235012531280518,
+      "learning_rate": 2.06336311941511e-06,
+      "loss": 1.8151,
+      "step": 1736
+    },
+    {
+      "epoch": 1.9035616438356164,
+      "grad_norm": 2.2945358753204346,
+      "learning_rate": 2.061332250203087e-06,
+      "loss": 1.7665,
+      "step": 1737
+    },
+    {
+      "epoch": 1.9046575342465752,
+      "grad_norm": 2.371809959411621,
+      "learning_rate": 2.0593013809910644e-06,
+      "loss": 1.7355,
+      "step": 1738
+    },
+    {
+      "epoch": 1.9057534246575343,
+      "grad_norm": 2.26947021484375,
+      "learning_rate": 2.0572705117790417e-06,
+      "loss": 1.7702,
+      "step": 1739
+    },
+    {
+      "epoch": 1.9068493150684933,
+      "grad_norm": 2.314237356185913,
+      "learning_rate": 2.055239642567019e-06,
+      "loss": 1.8084,
+      "step": 1740
+    },
+    {
+      "epoch": 1.9068493150684933,
+      "eval_loss": 1.9961000680923462,
+      "eval_runtime": 37.0775,
+      "eval_samples_per_second": 26.971,
+      "eval_steps_per_second": 1.133,
+      "step": 1740
+    },
+    {
+      "epoch": 1.907945205479452,
+      "grad_norm": 2.248680591583252,
+      "learning_rate": 2.0532087733549962e-06,
+      "loss": 1.7898,
+      "step": 1741
+    },
+    {
+      "epoch": 1.909041095890411,
+      "grad_norm": 2.336378812789917,
+      "learning_rate": 2.0511779041429735e-06,
+      "loss": 1.795,
+      "step": 1742
+    },
+    {
+      "epoch": 1.91013698630137,
+      "grad_norm": 2.2420554161071777,
+      "learning_rate": 2.0491470349309508e-06,
+      "loss": 1.76,
+      "step": 1743
+    },
+    {
+      "epoch": 1.9112328767123288,
+      "grad_norm": 2.3305511474609375,
+      "learning_rate": 2.047116165718928e-06,
+      "loss": 1.7959,
+      "step": 1744
+    },
+    {
+      "epoch": 1.9123287671232876,
+      "grad_norm": 2.2677786350250244,
+      "learning_rate": 2.0450852965069053e-06,
+      "loss": 1.7922,
+      "step": 1745
+    },
+    {
+      "epoch": 1.9134246575342466,
+      "grad_norm": 2.224148988723755,
+      "learning_rate": 2.0430544272948826e-06,
+      "loss": 1.841,
+      "step": 1746
+    },
+    {
+      "epoch": 1.9145205479452054,
+      "grad_norm": 2.172091245651245,
+      "learning_rate": 2.04102355808286e-06,
+      "loss": 1.784,
+      "step": 1747
+    },
+    {
+      "epoch": 1.9156164383561642,
+      "grad_norm": 2.319962501525879,
+      "learning_rate": 2.038992688870837e-06,
+      "loss": 1.7238,
+      "step": 1748
+    },
+    {
+      "epoch": 1.9167123287671233,
+      "grad_norm": 2.299760103225708,
+      "learning_rate": 2.0369618196588144e-06,
+      "loss": 1.7501,
+      "step": 1749
+    },
+    {
+      "epoch": 1.9178082191780823,
+      "grad_norm": 2.218407154083252,
+      "learning_rate": 2.0349309504467917e-06,
+      "loss": 1.7628,
+      "step": 1750
+    },
+    {
+      "epoch": 1.9178082191780823,
+      "eval_loss": 1.9934331178665161,
+      "eval_runtime": 37.0986,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 1750
+    },
+    {
+      "epoch": 1.9189041095890411,
+      "grad_norm": 2.4121744632720947,
+      "learning_rate": 2.032900081234769e-06,
+      "loss": 1.8128,
+      "step": 1751
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 2.253225326538086,
+      "learning_rate": 2.030869212022746e-06,
+      "loss": 1.8161,
+      "step": 1752
+    },
+    {
+      "epoch": 1.921095890410959,
+      "grad_norm": 2.28702449798584,
+      "learning_rate": 2.028838342810723e-06,
+      "loss": 1.7769,
+      "step": 1753
+    },
+    {
+      "epoch": 1.9221917808219178,
+      "grad_norm": 2.3004465103149414,
+      "learning_rate": 2.0268074735987004e-06,
+      "loss": 1.7631,
+      "step": 1754
+    },
+    {
+      "epoch": 1.9232876712328766,
+      "grad_norm": 2.2733960151672363,
+      "learning_rate": 2.0247766043866777e-06,
+      "loss": 1.75,
+      "step": 1755
+    },
+    {
+      "epoch": 1.9243835616438356,
+      "grad_norm": 2.274019241333008,
+      "learning_rate": 2.022745735174655e-06,
+      "loss": 1.8317,
+      "step": 1756
+    },
+    {
+      "epoch": 1.9254794520547946,
+      "grad_norm": 2.3094100952148438,
+      "learning_rate": 2.020714865962632e-06,
+      "loss": 1.7841,
+      "step": 1757
+    },
+    {
+      "epoch": 1.9265753424657535,
+      "grad_norm": 2.269321918487549,
+      "learning_rate": 2.0186839967506095e-06,
+      "loss": 1.8052,
+      "step": 1758
+    },
+    {
+      "epoch": 1.9276712328767123,
+      "grad_norm": 2.5324459075927734,
+      "learning_rate": 2.0166531275385868e-06,
+      "loss": 1.7916,
+      "step": 1759
+    },
+    {
+      "epoch": 1.9287671232876713,
+      "grad_norm": 2.302569627761841,
+      "learning_rate": 2.014622258326564e-06,
+      "loss": 1.7499,
+      "step": 1760
+    },
+    {
+      "epoch": 1.9287671232876713,
+      "eval_loss": 1.9879337549209595,
+      "eval_runtime": 37.1186,
+      "eval_samples_per_second": 26.941,
+      "eval_steps_per_second": 1.132,
+      "step": 1760
+    },
+    {
+      "epoch": 1.9298630136986301,
+      "grad_norm": 2.29470157623291,
+      "learning_rate": 2.0125913891145413e-06,
+      "loss": 1.7922,
+      "step": 1761
+    },
+    {
+      "epoch": 1.930958904109589,
+      "grad_norm": 2.4307754039764404,
+      "learning_rate": 2.0105605199025186e-06,
+      "loss": 1.7802,
+      "step": 1762
+    },
+    {
+      "epoch": 1.932054794520548,
+      "grad_norm": 2.179588794708252,
+      "learning_rate": 2.008529650690496e-06,
+      "loss": 1.7391,
+      "step": 1763
+    },
+    {
+      "epoch": 1.933150684931507,
+      "grad_norm": 2.5824451446533203,
+      "learning_rate": 2.0064987814784727e-06,
+      "loss": 1.8106,
+      "step": 1764
+    },
+    {
+      "epoch": 1.9342465753424658,
+      "grad_norm": 2.32308292388916,
+      "learning_rate": 2.00446791226645e-06,
+      "loss": 1.7689,
+      "step": 1765
+    },
+    {
+      "epoch": 1.9353424657534246,
+      "grad_norm": 2.262533664703369,
+      "learning_rate": 2.0024370430544273e-06,
+      "loss": 1.7604,
+      "step": 1766
+    },
+    {
+      "epoch": 1.9364383561643836,
+      "grad_norm": 2.325253963470459,
+      "learning_rate": 2.0004061738424045e-06,
+      "loss": 1.7695,
+      "step": 1767
+    },
+    {
+      "epoch": 1.9375342465753425,
+      "grad_norm": 2.2855470180511475,
+      "learning_rate": 1.998375304630382e-06,
+      "loss": 1.8233,
+      "step": 1768
+    },
+    {
+      "epoch": 1.9386301369863013,
+      "grad_norm": 2.3688673973083496,
+      "learning_rate": 1.996344435418359e-06,
+      "loss": 1.7799,
+      "step": 1769
+    },
+    {
+      "epoch": 1.9397260273972603,
+      "grad_norm": 2.4682323932647705,
+      "learning_rate": 1.9943135662063364e-06,
+      "loss": 1.8036,
+      "step": 1770
+    },
+    {
+      "epoch": 1.9397260273972603,
+      "eval_loss": 1.9906389713287354,
+      "eval_runtime": 37.0959,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 1770
+    },
+    {
+      "epoch": 1.9408219178082193,
+      "grad_norm": 2.369166374206543,
+      "learning_rate": 1.9922826969943136e-06,
+      "loss": 1.7878,
+      "step": 1771
+    },
+    {
+      "epoch": 1.941917808219178,
+      "grad_norm": 2.4669666290283203,
+      "learning_rate": 1.990251827782291e-06,
+      "loss": 1.8228,
+      "step": 1772
+    },
+    {
+      "epoch": 1.943013698630137,
+      "grad_norm": 2.3083250522613525,
+      "learning_rate": 1.988220958570268e-06,
+      "loss": 1.7997,
+      "step": 1773
+    },
+    {
+      "epoch": 1.944109589041096,
+      "grad_norm": 2.4884281158447266,
+      "learning_rate": 1.9861900893582455e-06,
+      "loss": 1.8105,
+      "step": 1774
+    },
+    {
+      "epoch": 1.9452054794520548,
+      "grad_norm": 2.3504858016967773,
+      "learning_rate": 1.9841592201462227e-06,
+      "loss": 1.8159,
+      "step": 1775
+    },
+    {
+      "epoch": 1.9463013698630136,
+      "grad_norm": 2.618431329727173,
+      "learning_rate": 1.9821283509342e-06,
+      "loss": 1.7352,
+      "step": 1776
+    },
+    {
+      "epoch": 1.9473972602739726,
+      "grad_norm": 2.3452653884887695,
+      "learning_rate": 1.9800974817221773e-06,
+      "loss": 1.7451,
+      "step": 1777
+    },
+    {
+      "epoch": 1.9484931506849315,
+      "grad_norm": 2.5966849327087402,
+      "learning_rate": 1.9780666125101546e-06,
+      "loss": 1.7965,
+      "step": 1778
+    },
+    {
+      "epoch": 1.9495890410958903,
+      "grad_norm": 2.3017661571502686,
+      "learning_rate": 1.976035743298132e-06,
+      "loss": 1.7629,
+      "step": 1779
+    },
+    {
+      "epoch": 1.9506849315068493,
+      "grad_norm": 2.3766098022460938,
+      "learning_rate": 1.974004874086109e-06,
+      "loss": 1.8403,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9506849315068493,
+      "eval_loss": 1.9936374425888062,
+      "eval_runtime": 37.1158,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 1780
+    },
+    {
+      "epoch": 1.9517808219178083,
+      "grad_norm": 2.2327704429626465,
+      "learning_rate": 1.9719740048740864e-06,
+      "loss": 1.7096,
+      "step": 1781
+    },
+    {
+      "epoch": 1.9528767123287671,
+      "grad_norm": 2.458707809448242,
+      "learning_rate": 1.9699431356620637e-06,
+      "loss": 1.7636,
+      "step": 1782
+    },
+    {
+      "epoch": 1.953972602739726,
+      "grad_norm": 2.4143803119659424,
+      "learning_rate": 1.967912266450041e-06,
+      "loss": 1.757,
+      "step": 1783
+    },
+    {
+      "epoch": 1.955068493150685,
+      "grad_norm": 2.465148448944092,
+      "learning_rate": 1.9658813972380182e-06,
+      "loss": 1.7526,
+      "step": 1784
+    },
+    {
+      "epoch": 1.9561643835616438,
+      "grad_norm": 2.356210231781006,
+      "learning_rate": 1.9638505280259955e-06,
+      "loss": 1.8705,
+      "step": 1785
+    },
+    {
+      "epoch": 1.9572602739726026,
+      "grad_norm": 2.7456002235412598,
+      "learning_rate": 1.9618196588139728e-06,
+      "loss": 1.7574,
+      "step": 1786
+    },
+    {
+      "epoch": 1.9583561643835616,
+      "grad_norm": 2.3724262714385986,
+      "learning_rate": 1.95978878960195e-06,
+      "loss": 1.7854,
+      "step": 1787
+    },
+    {
+      "epoch": 1.9594520547945207,
+      "grad_norm": 2.275827646255493,
+      "learning_rate": 1.9577579203899273e-06,
+      "loss": 1.8447,
+      "step": 1788
+    },
+    {
+      "epoch": 1.9605479452054795,
+      "grad_norm": 2.369992256164551,
+      "learning_rate": 1.955727051177904e-06,
+      "loss": 1.8299,
+      "step": 1789
+    },
+    {
+      "epoch": 1.9616438356164383,
+      "grad_norm": 2.462372064590454,
+      "learning_rate": 1.9536961819658815e-06,
+      "loss": 1.7596,
+      "step": 1790
+    },
+    {
+      "epoch": 1.9616438356164383,
+      "eval_loss": 1.9947243928909302,
+      "eval_runtime": 37.111,
+      "eval_samples_per_second": 26.946,
+      "eval_steps_per_second": 1.132,
+      "step": 1790
+    },
+    {
+      "epoch": 1.9627397260273973,
+      "grad_norm": 2.3040034770965576,
+      "learning_rate": 1.9516653127538587e-06,
+      "loss": 1.7391,
+      "step": 1791
+    },
+    {
+      "epoch": 1.9638356164383561,
+      "grad_norm": 2.449328899383545,
+      "learning_rate": 1.949634443541836e-06,
+      "loss": 1.7428,
+      "step": 1792
+    },
+    {
+      "epoch": 1.964931506849315,
+      "grad_norm": 2.2679781913757324,
+      "learning_rate": 1.9476035743298133e-06,
+      "loss": 1.738,
+      "step": 1793
+    },
+    {
+      "epoch": 1.966027397260274,
+      "grad_norm": 2.3444032669067383,
+      "learning_rate": 1.9455727051177906e-06,
+      "loss": 1.8089,
+      "step": 1794
+    },
+    {
+      "epoch": 1.967123287671233,
+      "grad_norm": 2.3005306720733643,
+      "learning_rate": 1.943541835905768e-06,
+      "loss": 1.7718,
+      "step": 1795
+    },
+    {
+      "epoch": 1.9682191780821918,
+      "grad_norm": 2.6022629737854004,
+      "learning_rate": 1.941510966693745e-06,
+      "loss": 1.8325,
+      "step": 1796
+    },
+    {
+      "epoch": 1.9693150684931506,
+      "grad_norm": 2.407585382461548,
+      "learning_rate": 1.9394800974817224e-06,
+      "loss": 1.737,
+      "step": 1797
+    },
+    {
+      "epoch": 1.9704109589041097,
+      "grad_norm": 2.349339723587036,
+      "learning_rate": 1.9374492282696997e-06,
+      "loss": 1.81,
+      "step": 1798
+    },
+    {
+      "epoch": 1.9715068493150685,
+      "grad_norm": 2.296192169189453,
+      "learning_rate": 1.935418359057677e-06,
+      "loss": 1.8424,
+      "step": 1799
+    },
+    {
+      "epoch": 1.9726027397260273,
+      "grad_norm": 2.175875663757324,
+      "learning_rate": 1.9333874898456542e-06,
+      "loss": 1.7977,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9726027397260273,
+      "eval_loss": 1.9929841756820679,
+      "eval_runtime": 37.1159,
+      "eval_samples_per_second": 26.943,
+      "eval_steps_per_second": 1.132,
+      "step": 1800
+    },
+    {
+      "epoch": 1.9736986301369863,
+      "grad_norm": 2.37428617477417,
+      "learning_rate": 1.931356620633631e-06,
+      "loss": 1.8248,
+      "step": 1801
+    },
+    {
+      "epoch": 1.9747945205479454,
+      "grad_norm": 2.297621965408325,
+      "learning_rate": 1.9293257514216083e-06,
+      "loss": 1.7215,
+      "step": 1802
+    },
+    {
+      "epoch": 1.9758904109589042,
+      "grad_norm": 2.217275857925415,
+      "learning_rate": 1.9272948822095856e-06,
+      "loss": 1.7828,
+      "step": 1803
+    },
+    {
+      "epoch": 1.976986301369863,
+      "grad_norm": 2.250854015350342,
+      "learning_rate": 1.925264012997563e-06,
+      "loss": 1.7245,
+      "step": 1804
+    },
+    {
+      "epoch": 1.978082191780822,
+      "grad_norm": 2.2556822299957275,
+      "learning_rate": 1.92323314378554e-06,
+      "loss": 1.7715,
+      "step": 1805
+    },
+    {
+      "epoch": 1.9791780821917808,
+      "grad_norm": 2.316389322280884,
+      "learning_rate": 1.9212022745735174e-06,
+      "loss": 1.8063,
+      "step": 1806
+    },
+    {
+      "epoch": 1.9802739726027396,
+      "grad_norm": 2.4358153343200684,
+      "learning_rate": 1.9191714053614947e-06,
+      "loss": 1.747,
+      "step": 1807
+    },
+    {
+      "epoch": 1.9813698630136987,
+      "grad_norm": 2.437631368637085,
+      "learning_rate": 1.917140536149472e-06,
+      "loss": 1.7485,
+      "step": 1808
+    },
+    {
+      "epoch": 1.9824657534246577,
+      "grad_norm": 2.3554601669311523,
+      "learning_rate": 1.9151096669374493e-06,
+      "loss": 1.7195,
+      "step": 1809
+    },
+    {
+      "epoch": 1.9835616438356163,
+      "grad_norm": 2.575559616088867,
+      "learning_rate": 1.9130787977254266e-06,
+      "loss": 1.8368,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9835616438356163,
+      "eval_loss": 1.9903514385223389,
+      "eval_runtime": 36.8082,
+      "eval_samples_per_second": 27.168,
+      "eval_steps_per_second": 1.141,
+      "step": 1810
+    },
+    {
+      "epoch": 1.9846575342465753,
+      "grad_norm": 2.3074066638946533,
+      "learning_rate": 1.911047928513404e-06,
+      "loss": 1.7952,
+      "step": 1811
+    },
+    {
+      "epoch": 1.9857534246575343,
+      "grad_norm": 2.3162307739257812,
+      "learning_rate": 1.909017059301381e-06,
+      "loss": 1.7455,
+      "step": 1812
+    },
+    {
+      "epoch": 1.9868493150684932,
+      "grad_norm": 2.3085129261016846,
+      "learning_rate": 1.9069861900893586e-06,
+      "loss": 1.7513,
+      "step": 1813
+    },
+    {
+      "epoch": 1.987945205479452,
+      "grad_norm": 2.3454785346984863,
+      "learning_rate": 1.9049553208773354e-06,
+      "loss": 1.803,
+      "step": 1814
+    },
+    {
+      "epoch": 1.989041095890411,
+      "grad_norm": 2.4522876739501953,
+      "learning_rate": 1.9029244516653127e-06,
+      "loss": 1.7567,
+      "step": 1815
+    },
+    {
+      "epoch": 1.9901369863013698,
+      "grad_norm": 2.3876044750213623,
+      "learning_rate": 1.90089358245329e-06,
+      "loss": 1.7126,
+      "step": 1816
+    },
+    {
+      "epoch": 1.9912328767123286,
+      "grad_norm": 2.4061570167541504,
+      "learning_rate": 1.8988627132412673e-06,
+      "loss": 1.7618,
+      "step": 1817
+    },
+    {
+      "epoch": 1.9923287671232877,
+      "grad_norm": 2.307445764541626,
+      "learning_rate": 1.8968318440292445e-06,
+      "loss": 1.8014,
+      "step": 1818
+    },
+    {
+      "epoch": 1.9934246575342467,
+      "grad_norm": 2.4414806365966797,
+      "learning_rate": 1.8948009748172218e-06,
+      "loss": 1.7202,
+      "step": 1819
+    },
+    {
+      "epoch": 1.9945205479452055,
+      "grad_norm": 2.3933329582214355,
+      "learning_rate": 1.892770105605199e-06,
+      "loss": 1.8133,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9945205479452055,
+      "eval_loss": 1.991702675819397,
+      "eval_runtime": 37.0127,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 1820
+    },
+    {
+      "epoch": 1.9956164383561643,
+      "grad_norm": 2.237699508666992,
+      "learning_rate": 1.8907392363931764e-06,
+      "loss": 1.8026,
+      "step": 1821
+    },
+    {
+      "epoch": 1.9967123287671233,
+      "grad_norm": 2.290004014968872,
+      "learning_rate": 1.8887083671811536e-06,
+      "loss": 1.7689,
+      "step": 1822
+    },
+    {
+      "epoch": 1.9978082191780822,
+      "grad_norm": 2.2264676094055176,
+      "learning_rate": 1.886677497969131e-06,
+      "loss": 1.7862,
+      "step": 1823
+    },
+    {
+      "epoch": 1.998904109589041,
+      "grad_norm": 2.2764949798583984,
+      "learning_rate": 1.8846466287571082e-06,
+      "loss": 1.6763,
+      "step": 1824
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 2.3024826049804688,
+      "learning_rate": 1.8826157595450855e-06,
+      "loss": 1.8076,
+      "step": 1825
+    },
+    {
+      "epoch": 2.001095890410959,
+      "grad_norm": 2.362882614135742,
+      "learning_rate": 1.8805848903330625e-06,
+      "loss": 1.7009,
+      "step": 1826
+    },
+    {
+      "epoch": 2.0021917808219176,
+      "grad_norm": 2.432370185852051,
+      "learning_rate": 1.8785540211210398e-06,
+      "loss": 1.8712,
+      "step": 1827
+    },
+    {
+      "epoch": 2.0032876712328767,
+      "grad_norm": 2.3584744930267334,
+      "learning_rate": 1.876523151909017e-06,
+      "loss": 1.7869,
+      "step": 1828
+    },
+    {
+      "epoch": 2.0043835616438357,
+      "grad_norm": 2.534205198287964,
+      "learning_rate": 1.8744922826969944e-06,
+      "loss": 1.7376,
+      "step": 1829
+    },
+    {
+      "epoch": 2.0054794520547947,
+      "grad_norm": 2.58241605758667,
+      "learning_rate": 1.8724614134849716e-06,
+      "loss": 1.7204,
+      "step": 1830
+    },
+    {
+      "epoch": 2.0054794520547947,
+      "eval_loss": 1.9912636280059814,
+      "eval_runtime": 37.0969,
+      "eval_samples_per_second": 26.956,
+      "eval_steps_per_second": 1.132,
+      "step": 1830
+    },
+    {
+      "epoch": 2.0065753424657533,
+      "grad_norm": 2.3742504119873047,
+      "learning_rate": 1.870430544272949e-06,
+      "loss": 1.7537,
+      "step": 1831
+    },
+    {
+      "epoch": 2.0076712328767123,
+      "grad_norm": 2.6560916900634766,
+      "learning_rate": 1.8683996750609262e-06,
+      "loss": 1.8524,
+      "step": 1832
+    },
+    {
+      "epoch": 2.0087671232876714,
+      "grad_norm": 2.523041248321533,
+      "learning_rate": 1.8663688058489035e-06,
+      "loss": 1.7405,
+      "step": 1833
+    },
+    {
+      "epoch": 2.00986301369863,
+      "grad_norm": 2.5444304943084717,
+      "learning_rate": 1.8643379366368807e-06,
+      "loss": 1.7588,
+      "step": 1834
+    },
+    {
+      "epoch": 2.010958904109589,
+      "grad_norm": 2.708984613418579,
+      "learning_rate": 1.862307067424858e-06,
+      "loss": 1.7686,
+      "step": 1835
+    },
+    {
+      "epoch": 2.012054794520548,
+      "grad_norm": 2.726691961288452,
+      "learning_rate": 1.8602761982128353e-06,
+      "loss": 1.7733,
+      "step": 1836
+    },
+    {
+      "epoch": 2.013150684931507,
+      "grad_norm": 2.5085904598236084,
+      "learning_rate": 1.8582453290008126e-06,
+      "loss": 1.81,
+      "step": 1837
+    },
+    {
+      "epoch": 2.0142465753424657,
+      "grad_norm": 2.4699466228485107,
+      "learning_rate": 1.8562144597887898e-06,
+      "loss": 1.7369,
+      "step": 1838
+    },
+    {
+      "epoch": 2.0153424657534247,
+      "grad_norm": 2.531027317047119,
+      "learning_rate": 1.854183590576767e-06,
+      "loss": 1.7611,
+      "step": 1839
+    },
+    {
+      "epoch": 2.0164383561643837,
+      "grad_norm": 2.5061161518096924,
+      "learning_rate": 1.8521527213647442e-06,
+      "loss": 1.811,
+      "step": 1840
+    },
+    {
+      "epoch": 2.0164383561643837,
+      "eval_loss": 1.9893144369125366,
+      "eval_runtime": 37.1262,
+      "eval_samples_per_second": 26.935,
+      "eval_steps_per_second": 1.131,
+      "step": 1840
+    },
+    {
+      "epoch": 2.0175342465753423,
+      "grad_norm": 2.5739994049072266,
+      "learning_rate": 1.8501218521527215e-06,
+      "loss": 1.7153,
+      "step": 1841
+    },
+    {
+      "epoch": 2.0186301369863013,
+      "grad_norm": 2.5673015117645264,
+      "learning_rate": 1.8480909829406987e-06,
+      "loss": 1.6879,
+      "step": 1842
+    },
+    {
+      "epoch": 2.0197260273972604,
+      "grad_norm": 2.717430353164673,
+      "learning_rate": 1.846060113728676e-06,
+      "loss": 1.723,
+      "step": 1843
+    },
+    {
+      "epoch": 2.020821917808219,
+      "grad_norm": 2.4905285835266113,
+      "learning_rate": 1.8440292445166533e-06,
+      "loss": 1.7651,
+      "step": 1844
+    },
+    {
+      "epoch": 2.021917808219178,
+      "grad_norm": 2.6798791885375977,
+      "learning_rate": 1.8419983753046306e-06,
+      "loss": 1.7308,
+      "step": 1845
+    },
+    {
+      "epoch": 2.023013698630137,
+      "grad_norm": 2.591614007949829,
+      "learning_rate": 1.8399675060926078e-06,
+      "loss": 1.786,
+      "step": 1846
+    },
+    {
+      "epoch": 2.024109589041096,
+      "grad_norm": 2.5405824184417725,
+      "learning_rate": 1.8379366368805851e-06,
+      "loss": 1.7699,
+      "step": 1847
+    },
+    {
+      "epoch": 2.0252054794520546,
+      "grad_norm": 2.4189934730529785,
+      "learning_rate": 1.8359057676685624e-06,
+      "loss": 1.732,
+      "step": 1848
+    },
+    {
+      "epoch": 2.0263013698630137,
+      "grad_norm": 2.5857982635498047,
+      "learning_rate": 1.8338748984565397e-06,
+      "loss": 1.7123,
+      "step": 1849
+    },
+    {
+      "epoch": 2.0273972602739727,
+      "grad_norm": 2.4710893630981445,
+      "learning_rate": 1.831844029244517e-06,
+      "loss": 1.7419,
+      "step": 1850
+    },
+    {
+      "epoch": 2.0273972602739727,
+      "eval_loss": 1.9874999523162842,
+      "eval_runtime": 37.1588,
+      "eval_samples_per_second": 26.912,
+      "eval_steps_per_second": 1.13,
+      "step": 1850
+    },
+    {
+      "epoch": 2.0284931506849313,
+      "grad_norm": 2.3414645195007324,
+      "learning_rate": 1.829813160032494e-06,
+      "loss": 1.7102,
+      "step": 1851
+    },
+    {
+      "epoch": 2.0295890410958903,
+      "grad_norm": 2.4148900508880615,
+      "learning_rate": 1.8277822908204713e-06,
+      "loss": 1.7287,
+      "step": 1852
+    },
+    {
+      "epoch": 2.0306849315068494,
+      "grad_norm": 2.5501346588134766,
+      "learning_rate": 1.8257514216084486e-06,
+      "loss": 1.7633,
+      "step": 1853
+    },
+    {
+      "epoch": 2.0317808219178084,
+      "grad_norm": 2.52215576171875,
+      "learning_rate": 1.8237205523964258e-06,
+      "loss": 1.7385,
+      "step": 1854
+    },
+    {
+      "epoch": 2.032876712328767,
+      "grad_norm": 2.433335065841675,
+      "learning_rate": 1.8216896831844031e-06,
+      "loss": 1.7566,
+      "step": 1855
+    },
+    {
+      "epoch": 2.033972602739726,
+      "grad_norm": 2.441232442855835,
+      "learning_rate": 1.8196588139723804e-06,
+      "loss": 1.8027,
+      "step": 1856
+    },
+    {
+      "epoch": 2.035068493150685,
+      "grad_norm": 2.4601082801818848,
+      "learning_rate": 1.8176279447603577e-06,
+      "loss": 1.6859,
+      "step": 1857
+    },
+    {
+      "epoch": 2.0361643835616436,
+      "grad_norm": 2.5008108615875244,
+      "learning_rate": 1.815597075548335e-06,
+      "loss": 1.7294,
+      "step": 1858
+    },
+    {
+      "epoch": 2.0372602739726027,
+      "grad_norm": 2.433915615081787,
+      "learning_rate": 1.8135662063363122e-06,
+      "loss": 1.755,
+      "step": 1859
+    },
+    {
+      "epoch": 2.0383561643835617,
+      "grad_norm": 2.4938697814941406,
+      "learning_rate": 1.8115353371242895e-06,
+      "loss": 1.7904,
+      "step": 1860
+    },
+    {
+      "epoch": 2.0383561643835617,
+      "eval_loss": 1.9939266443252563,
+      "eval_runtime": 37.1539,
+      "eval_samples_per_second": 26.915,
+      "eval_steps_per_second": 1.13,
+      "step": 1860
+    },
+    {
+      "epoch": 2.0394520547945207,
+      "grad_norm": 2.382324695587158,
+      "learning_rate": 1.8095044679122668e-06,
+      "loss": 1.6856,
+      "step": 1861
+    },
+    {
+      "epoch": 2.0405479452054793,
+      "grad_norm": 2.5250437259674072,
+      "learning_rate": 1.807473598700244e-06,
+      "loss": 1.7434,
+      "step": 1862
+    },
+    {
+      "epoch": 2.0416438356164384,
+      "grad_norm": 2.63969087600708,
+      "learning_rate": 1.805442729488221e-06,
+      "loss": 1.7343,
+      "step": 1863
+    },
+    {
+      "epoch": 2.0427397260273974,
+      "grad_norm": 2.4852068424224854,
+      "learning_rate": 1.8034118602761982e-06,
+      "loss": 1.7494,
+      "step": 1864
+    },
+    {
+      "epoch": 2.043835616438356,
+      "grad_norm": 2.393967390060425,
+      "learning_rate": 1.8013809910641754e-06,
+      "loss": 1.7503,
+      "step": 1865
+    },
+    {
+      "epoch": 2.044931506849315,
+      "grad_norm": 2.4447648525238037,
+      "learning_rate": 1.7993501218521527e-06,
+      "loss": 1.7539,
+      "step": 1866
+    },
+    {
+      "epoch": 2.046027397260274,
+      "grad_norm": 2.483530044555664,
+      "learning_rate": 1.79731925264013e-06,
+      "loss": 1.7001,
+      "step": 1867
+    },
+    {
+      "epoch": 2.047123287671233,
+      "grad_norm": 2.56669282913208,
+      "learning_rate": 1.7952883834281073e-06,
+      "loss": 1.7828,
+      "step": 1868
+    },
+    {
+      "epoch": 2.0482191780821917,
+      "grad_norm": 2.4894964694976807,
+      "learning_rate": 1.7932575142160846e-06,
+      "loss": 1.7208,
+      "step": 1869
+    },
+    {
+      "epoch": 2.0493150684931507,
+      "grad_norm": 2.4713854789733887,
+      "learning_rate": 1.7912266450040618e-06,
+      "loss": 1.7687,
+      "step": 1870
+    },
+    {
+      "epoch": 2.0493150684931507,
+      "eval_loss": 1.9938212633132935,
+      "eval_runtime": 37.1298,
+      "eval_samples_per_second": 26.933,
+      "eval_steps_per_second": 1.131,
+      "step": 1870
+    },
+    {
+      "epoch": 2.0504109589041097,
+      "grad_norm": 2.6204898357391357,
+      "learning_rate": 1.789195775792039e-06,
+      "loss": 1.7718,
+      "step": 1871
+    },
+    {
+      "epoch": 2.0515068493150683,
+      "grad_norm": 2.5194199085235596,
+      "learning_rate": 1.7871649065800164e-06,
+      "loss": 1.6653,
+      "step": 1872
+    },
+    {
+      "epoch": 2.0526027397260274,
+      "grad_norm": 2.5077965259552,
+      "learning_rate": 1.7851340373679937e-06,
+      "loss": 1.8098,
+      "step": 1873
+    },
+    {
+      "epoch": 2.0536986301369864,
+      "grad_norm": 2.5841965675354004,
+      "learning_rate": 1.783103168155971e-06,
+      "loss": 1.7126,
+      "step": 1874
+    },
+    {
+      "epoch": 2.0547945205479454,
+      "grad_norm": 2.617910623550415,
+      "learning_rate": 1.7810722989439482e-06,
+      "loss": 1.7294,
+      "step": 1875
+    },
+    {
+      "epoch": 2.055890410958904,
+      "grad_norm": 2.519253969192505,
+      "learning_rate": 1.7790414297319253e-06,
+      "loss": 1.7497,
+      "step": 1876
+    },
+    {
+      "epoch": 2.056986301369863,
+      "grad_norm": 2.5518412590026855,
+      "learning_rate": 1.7770105605199025e-06,
+      "loss": 1.7188,
+      "step": 1877
+    },
+    {
+      "epoch": 2.058082191780822,
+      "grad_norm": 2.5372824668884277,
+      "learning_rate": 1.7749796913078798e-06,
+      "loss": 1.6998,
+      "step": 1878
+    },
+    {
+      "epoch": 2.0591780821917807,
+      "grad_norm": 2.6649036407470703,
+      "learning_rate": 1.772948822095857e-06,
+      "loss": 1.7621,
+      "step": 1879
+    },
+    {
+      "epoch": 2.0602739726027397,
+      "grad_norm": 2.5216212272644043,
+      "learning_rate": 1.7709179528838344e-06,
+      "loss": 1.7963,
+      "step": 1880
+    },
+    {
+      "epoch": 2.0602739726027397,
+      "eval_loss": 1.993401050567627,
+      "eval_runtime": 37.0954,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 1880
+    },
+    {
+      "epoch": 2.0613698630136987,
+      "grad_norm": 2.5226478576660156,
+      "learning_rate": 1.7688870836718116e-06,
+      "loss": 1.7841,
+      "step": 1881
+    },
+    {
+      "epoch": 2.0624657534246573,
+      "grad_norm": 2.584319829940796,
+      "learning_rate": 1.766856214459789e-06,
+      "loss": 1.7449,
+      "step": 1882
+    },
+    {
+      "epoch": 2.0635616438356164,
+      "grad_norm": 2.5305304527282715,
+      "learning_rate": 1.7648253452477662e-06,
+      "loss": 1.7373,
+      "step": 1883
+    },
+    {
+      "epoch": 2.0646575342465754,
+      "grad_norm": 2.48797607421875,
+      "learning_rate": 1.7627944760357435e-06,
+      "loss": 1.722,
+      "step": 1884
+    },
+    {
+      "epoch": 2.0657534246575344,
+      "grad_norm": 2.4461028575897217,
+      "learning_rate": 1.7607636068237208e-06,
+      "loss": 1.7983,
+      "step": 1885
+    },
+    {
+      "epoch": 2.066849315068493,
+      "grad_norm": 2.624058961868286,
+      "learning_rate": 1.758732737611698e-06,
+      "loss": 1.7709,
+      "step": 1886
+    },
+    {
+      "epoch": 2.067945205479452,
+      "grad_norm": 2.848062515258789,
+      "learning_rate": 1.7567018683996753e-06,
+      "loss": 1.7513,
+      "step": 1887
+    },
+    {
+      "epoch": 2.069041095890411,
+      "grad_norm": 2.5219593048095703,
+      "learning_rate": 1.7546709991876524e-06,
+      "loss": 1.7387,
+      "step": 1888
+    },
+    {
+      "epoch": 2.0701369863013697,
+      "grad_norm": 2.5296876430511475,
+      "learning_rate": 1.7526401299756296e-06,
+      "loss": 1.7378,
+      "step": 1889
+    },
+    {
+      "epoch": 2.0712328767123287,
+      "grad_norm": 2.5452613830566406,
+      "learning_rate": 1.750609260763607e-06,
+      "loss": 1.7402,
+      "step": 1890
+    },
+    {
+      "epoch": 2.0712328767123287,
+      "eval_loss": 1.9914054870605469,
+      "eval_runtime": 37.0548,
+      "eval_samples_per_second": 26.987,
+      "eval_steps_per_second": 1.133,
+      "step": 1890
+    },
+    {
+      "epoch": 2.0723287671232877,
+      "grad_norm": 2.653596878051758,
+      "learning_rate": 1.7485783915515842e-06,
+      "loss": 1.7369,
+      "step": 1891
+    },
+    {
+      "epoch": 2.0734246575342468,
+      "grad_norm": 2.615581750869751,
+      "learning_rate": 1.7465475223395615e-06,
+      "loss": 1.7317,
+      "step": 1892
+    },
+    {
+      "epoch": 2.0745205479452054,
+      "grad_norm": 2.610304832458496,
+      "learning_rate": 1.7445166531275387e-06,
+      "loss": 1.7121,
+      "step": 1893
+    },
+    {
+      "epoch": 2.0756164383561644,
+      "grad_norm": 2.6463494300842285,
+      "learning_rate": 1.742485783915516e-06,
+      "loss": 1.8523,
+      "step": 1894
+    },
+    {
+      "epoch": 2.0767123287671234,
+      "grad_norm": 2.4814131259918213,
+      "learning_rate": 1.7404549147034933e-06,
+      "loss": 1.8239,
+      "step": 1895
+    },
+    {
+      "epoch": 2.077808219178082,
+      "grad_norm": 2.7416090965270996,
+      "learning_rate": 1.7384240454914706e-06,
+      "loss": 1.669,
+      "step": 1896
+    },
+    {
+      "epoch": 2.078904109589041,
+      "grad_norm": 2.9458305835723877,
+      "learning_rate": 1.7363931762794478e-06,
+      "loss": 1.7504,
+      "step": 1897
+    },
+    {
+      "epoch": 2.08,
+      "grad_norm": 2.5118932723999023,
+      "learning_rate": 1.7343623070674251e-06,
+      "loss": 1.7854,
+      "step": 1898
+    },
+    {
+      "epoch": 2.081095890410959,
+      "grad_norm": 2.479421615600586,
+      "learning_rate": 1.7323314378554024e-06,
+      "loss": 1.7325,
+      "step": 1899
+    },
+    {
+      "epoch": 2.0821917808219177,
+      "grad_norm": 2.7400424480438232,
+      "learning_rate": 1.7303005686433797e-06,
+      "loss": 1.7434,
+      "step": 1900
+    },
+    {
+      "epoch": 2.0821917808219177,
+      "eval_loss": 1.9887213706970215,
+      "eval_runtime": 37.0484,
+      "eval_samples_per_second": 26.992,
+      "eval_steps_per_second": 1.134,
+      "step": 1900
+    },
+    {
+      "epoch": 2.0832876712328767,
+      "grad_norm": 2.587714433670044,
+      "learning_rate": 1.7282696994313567e-06,
+      "loss": 1.7335,
+      "step": 1901
+    },
+    {
+      "epoch": 2.0843835616438358,
+      "grad_norm": 2.526818037033081,
+      "learning_rate": 1.726238830219334e-06,
+      "loss": 1.7538,
+      "step": 1902
+    },
+    {
+      "epoch": 2.0854794520547943,
+      "grad_norm": 2.595221519470215,
+      "learning_rate": 1.7242079610073113e-06,
+      "loss": 1.7331,
+      "step": 1903
+    },
+    {
+      "epoch": 2.0865753424657534,
+      "grad_norm": 2.5018110275268555,
+      "learning_rate": 1.7221770917952886e-06,
+      "loss": 1.8069,
+      "step": 1904
+    },
+    {
+      "epoch": 2.0876712328767124,
+      "grad_norm": 2.4378256797790527,
+      "learning_rate": 1.7201462225832658e-06,
+      "loss": 1.757,
+      "step": 1905
+    },
+    {
+      "epoch": 2.0887671232876714,
+      "grad_norm": 2.651167392730713,
+      "learning_rate": 1.7181153533712431e-06,
+      "loss": 1.7696,
+      "step": 1906
+    },
+    {
+      "epoch": 2.08986301369863,
+      "grad_norm": 2.550239086151123,
+      "learning_rate": 1.7160844841592204e-06,
+      "loss": 1.6551,
+      "step": 1907
+    },
+    {
+      "epoch": 2.090958904109589,
+      "grad_norm": 2.4192144870758057,
+      "learning_rate": 1.7140536149471977e-06,
+      "loss": 1.7031,
+      "step": 1908
+    },
+    {
+      "epoch": 2.092054794520548,
+      "grad_norm": 2.5327398777008057,
+      "learning_rate": 1.712022745735175e-06,
+      "loss": 1.734,
+      "step": 1909
+    },
+    {
+      "epoch": 2.0931506849315067,
+      "grad_norm": 2.5962865352630615,
+      "learning_rate": 1.7099918765231522e-06,
+      "loss": 1.732,
+      "step": 1910
+    },
+    {
+      "epoch": 2.0931506849315067,
+      "eval_loss": 1.9915482997894287,
+      "eval_runtime": 36.7856,
+      "eval_samples_per_second": 27.185,
+      "eval_steps_per_second": 1.142,
+      "step": 1910
+    },
+    {
+      "epoch": 2.0942465753424657,
+      "grad_norm": 2.561943531036377,
+      "learning_rate": 1.7079610073111295e-06,
+      "loss": 1.7556,
+      "step": 1911
+    },
+    {
+      "epoch": 2.0953424657534248,
+      "grad_norm": 2.529829263687134,
+      "learning_rate": 1.7059301380991068e-06,
+      "loss": 1.6839,
+      "step": 1912
+    },
+    {
+      "epoch": 2.0964383561643833,
+      "grad_norm": 2.636559009552002,
+      "learning_rate": 1.7038992688870836e-06,
+      "loss": 1.7393,
+      "step": 1913
+    },
+    {
+      "epoch": 2.0975342465753424,
+      "grad_norm": 2.558166742324829,
+      "learning_rate": 1.701868399675061e-06,
+      "loss": 1.7346,
+      "step": 1914
+    },
+    {
+      "epoch": 2.0986301369863014,
+      "grad_norm": 2.5193750858306885,
+      "learning_rate": 1.6998375304630382e-06,
+      "loss": 1.7666,
+      "step": 1915
+    },
+    {
+      "epoch": 2.0997260273972604,
+      "grad_norm": 2.498244047164917,
+      "learning_rate": 1.6978066612510155e-06,
+      "loss": 1.827,
+      "step": 1916
+    },
+    {
+      "epoch": 2.100821917808219,
+      "grad_norm": 2.799603223800659,
+      "learning_rate": 1.6957757920389927e-06,
+      "loss": 1.788,
+      "step": 1917
+    },
+    {
+      "epoch": 2.101917808219178,
+      "grad_norm": 2.733391761779785,
+      "learning_rate": 1.69374492282697e-06,
+      "loss": 1.7487,
+      "step": 1918
+    },
+    {
+      "epoch": 2.103013698630137,
+      "grad_norm": 2.5135183334350586,
+      "learning_rate": 1.6917140536149473e-06,
+      "loss": 1.7114,
+      "step": 1919
+    },
+    {
+      "epoch": 2.1041095890410957,
+      "grad_norm": 2.5275800228118896,
+      "learning_rate": 1.6896831844029246e-06,
+      "loss": 1.797,
+      "step": 1920
+    },
+    {
+      "epoch": 2.1041095890410957,
+      "eval_loss": 1.9926362037658691,
+      "eval_runtime": 36.9326,
+      "eval_samples_per_second": 27.076,
+      "eval_steps_per_second": 1.137,
+      "step": 1920
+    },
+    {
+      "epoch": 2.1052054794520547,
+      "grad_norm": 2.682774782180786,
+      "learning_rate": 1.6876523151909018e-06,
+      "loss": 1.7581,
+      "step": 1921
+    },
+    {
+      "epoch": 2.1063013698630138,
+      "grad_norm": 2.625620126724243,
+      "learning_rate": 1.6856214459788791e-06,
+      "loss": 1.7606,
+      "step": 1922
+    },
+    {
+      "epoch": 2.107397260273973,
+      "grad_norm": 2.5005173683166504,
+      "learning_rate": 1.6835905767668564e-06,
+      "loss": 1.7139,
+      "step": 1923
+    },
+    {
+      "epoch": 2.1084931506849314,
+      "grad_norm": 2.6731526851654053,
+      "learning_rate": 1.6815597075548337e-06,
+      "loss": 1.7174,
+      "step": 1924
+    },
+    {
+      "epoch": 2.1095890410958904,
+      "grad_norm": 2.5407724380493164,
+      "learning_rate": 1.679528838342811e-06,
+      "loss": 1.8049,
+      "step": 1925
+    },
+    {
+      "epoch": 2.1106849315068494,
+      "grad_norm": 2.502427101135254,
+      "learning_rate": 1.677497969130788e-06,
+      "loss": 1.7068,
+      "step": 1926
+    },
+    {
+      "epoch": 2.111780821917808,
+      "grad_norm": 2.5063414573669434,
+      "learning_rate": 1.6754670999187653e-06,
+      "loss": 1.7174,
+      "step": 1927
+    },
+    {
+      "epoch": 2.112876712328767,
+      "grad_norm": 2.524588108062744,
+      "learning_rate": 1.6734362307067426e-06,
+      "loss": 1.7442,
+      "step": 1928
+    },
+    {
+      "epoch": 2.113972602739726,
+      "grad_norm": 2.531909465789795,
+      "learning_rate": 1.6714053614947198e-06,
+      "loss": 1.692,
+      "step": 1929
+    },
+    {
+      "epoch": 2.115068493150685,
+      "grad_norm": 2.5104269981384277,
+      "learning_rate": 1.669374492282697e-06,
+      "loss": 1.7039,
+      "step": 1930
+    },
+    {
+      "epoch": 2.115068493150685,
+      "eval_loss": 1.9927221536636353,
+      "eval_runtime": 37.004,
+      "eval_samples_per_second": 27.024,
+      "eval_steps_per_second": 1.135,
+      "step": 1930
+    },
+    {
+      "epoch": 2.1161643835616437,
+      "grad_norm": 2.5443670749664307,
+      "learning_rate": 1.6673436230706744e-06,
+      "loss": 1.7896,
+      "step": 1931
+    },
+    {
+      "epoch": 2.1172602739726027,
+      "grad_norm": 2.6495745182037354,
+      "learning_rate": 1.6653127538586517e-06,
+      "loss": 1.7735,
+      "step": 1932
+    },
+    {
+      "epoch": 2.118356164383562,
+      "grad_norm": 2.7233121395111084,
+      "learning_rate": 1.663281884646629e-06,
+      "loss": 1.7186,
+      "step": 1933
+    },
+    {
+      "epoch": 2.1194520547945204,
+      "grad_norm": 2.4411816596984863,
+      "learning_rate": 1.6612510154346062e-06,
+      "loss": 1.7378,
+      "step": 1934
+    },
+    {
+      "epoch": 2.1205479452054794,
+      "grad_norm": 2.6478421688079834,
+      "learning_rate": 1.6592201462225835e-06,
+      "loss": 1.7989,
+      "step": 1935
+    },
+    {
+      "epoch": 2.1216438356164384,
+      "grad_norm": 2.56093168258667,
+      "learning_rate": 1.6571892770105608e-06,
+      "loss": 1.7663,
+      "step": 1936
+    },
+    {
+      "epoch": 2.1227397260273975,
+      "grad_norm": 2.5598859786987305,
+      "learning_rate": 1.655158407798538e-06,
+      "loss": 1.7956,
+      "step": 1937
+    },
+    {
+      "epoch": 2.123835616438356,
+      "grad_norm": 2.4958112239837646,
+      "learning_rate": 1.653127538586515e-06,
+      "loss": 1.7482,
+      "step": 1938
+    },
+    {
+      "epoch": 2.124931506849315,
+      "grad_norm": 2.5133280754089355,
+      "learning_rate": 1.6510966693744924e-06,
+      "loss": 1.7811,
+      "step": 1939
+    },
+    {
+      "epoch": 2.126027397260274,
+      "grad_norm": 2.492892026901245,
+      "learning_rate": 1.6490658001624696e-06,
+      "loss": 1.7883,
+      "step": 1940
+    },
+    {
+      "epoch": 2.126027397260274,
+      "eval_loss": 1.9917834997177124,
+      "eval_runtime": 37.0222,
+      "eval_samples_per_second": 27.011,
+      "eval_steps_per_second": 1.134,
+      "step": 1940
+    },
+    {
+      "epoch": 2.1271232876712327,
+      "grad_norm": 2.560483455657959,
+      "learning_rate": 1.647034930950447e-06,
+      "loss": 1.7871,
+      "step": 1941
+    },
+    {
+      "epoch": 2.1282191780821917,
+      "grad_norm": 2.549100160598755,
+      "learning_rate": 1.6450040617384242e-06,
+      "loss": 1.7097,
+      "step": 1942
+    },
+    {
+      "epoch": 2.1293150684931508,
+      "grad_norm": 2.4832775592803955,
+      "learning_rate": 1.6429731925264015e-06,
+      "loss": 1.736,
+      "step": 1943
+    },
+    {
+      "epoch": 2.1304109589041094,
+      "grad_norm": 2.60006046295166,
+      "learning_rate": 1.6409423233143788e-06,
+      "loss": 1.765,
+      "step": 1944
+    },
+    {
+      "epoch": 2.1315068493150684,
+      "grad_norm": 2.4875144958496094,
+      "learning_rate": 1.638911454102356e-06,
+      "loss": 1.7317,
+      "step": 1945
+    },
+    {
+      "epoch": 2.1326027397260274,
+      "grad_norm": 2.5562970638275146,
+      "learning_rate": 1.6368805848903333e-06,
+      "loss": 1.6982,
+      "step": 1946
+    },
+    {
+      "epoch": 2.1336986301369865,
+      "grad_norm": 2.4928972721099854,
+      "learning_rate": 1.6348497156783106e-06,
+      "loss": 1.7821,
+      "step": 1947
+    },
+    {
+      "epoch": 2.134794520547945,
+      "grad_norm": 2.6159043312072754,
+      "learning_rate": 1.6328188464662879e-06,
+      "loss": 1.7223,
+      "step": 1948
+    },
+    {
+      "epoch": 2.135890410958904,
+      "grad_norm": 2.541520833969116,
+      "learning_rate": 1.6307879772542651e-06,
+      "loss": 1.7321,
+      "step": 1949
+    },
+    {
+      "epoch": 2.136986301369863,
+      "grad_norm": 2.5165696144104004,
+      "learning_rate": 1.628757108042242e-06,
+      "loss": 1.7571,
+      "step": 1950
+    },
+    {
+      "epoch": 2.136986301369863,
+      "eval_loss": 1.9868476390838623,
+      "eval_runtime": 37.0367,
+      "eval_samples_per_second": 27.0,
+      "eval_steps_per_second": 1.134,
+      "step": 1950
+    },
+    {
+      "epoch": 2.138082191780822,
+      "grad_norm": 2.5660274028778076,
+      "learning_rate": 1.6267262388302193e-06,
+      "loss": 1.7855,
+      "step": 1951
+    },
+    {
+      "epoch": 2.1391780821917807,
+      "grad_norm": 2.4606971740722656,
+      "learning_rate": 1.6246953696181965e-06,
+      "loss": 1.7324,
+      "step": 1952
+    },
+    {
+      "epoch": 2.1402739726027398,
+      "grad_norm": 2.6528608798980713,
+      "learning_rate": 1.6226645004061738e-06,
+      "loss": 1.7715,
+      "step": 1953
+    },
+    {
+      "epoch": 2.141369863013699,
+      "grad_norm": 2.7232985496520996,
+      "learning_rate": 1.620633631194151e-06,
+      "loss": 1.837,
+      "step": 1954
+    },
+    {
+      "epoch": 2.1424657534246574,
+      "grad_norm": 2.6264185905456543,
+      "learning_rate": 1.6186027619821284e-06,
+      "loss": 1.7678,
+      "step": 1955
+    },
+    {
+      "epoch": 2.1435616438356164,
+      "grad_norm": 2.6248252391815186,
+      "learning_rate": 1.6165718927701056e-06,
+      "loss": 1.7218,
+      "step": 1956
+    },
+    {
+      "epoch": 2.1446575342465755,
+      "grad_norm": 3.00539493560791,
+      "learning_rate": 1.614541023558083e-06,
+      "loss": 1.7678,
+      "step": 1957
+    },
+    {
+      "epoch": 2.145753424657534,
+      "grad_norm": 2.5368528366088867,
+      "learning_rate": 1.6125101543460602e-06,
+      "loss": 1.7737,
+      "step": 1958
+    },
+    {
+      "epoch": 2.146849315068493,
+      "grad_norm": 2.5805916786193848,
+      "learning_rate": 1.6104792851340375e-06,
+      "loss": 1.7394,
+      "step": 1959
+    },
+    {
+      "epoch": 2.147945205479452,
+      "grad_norm": 2.615856647491455,
+      "learning_rate": 1.6084484159220147e-06,
+      "loss": 1.766,
+      "step": 1960
+    },
+    {
+      "epoch": 2.147945205479452,
+      "eval_loss": 1.987140417098999,
+      "eval_runtime": 37.0593,
+      "eval_samples_per_second": 26.984,
+      "eval_steps_per_second": 1.133,
+      "step": 1960
+    },
+    {
+      "epoch": 2.149041095890411,
+      "grad_norm": 2.6550803184509277,
+      "learning_rate": 1.606417546709992e-06,
+      "loss": 1.7342,
+      "step": 1961
+    },
+    {
+      "epoch": 2.1501369863013697,
+      "grad_norm": 2.5999467372894287,
+      "learning_rate": 1.6043866774979693e-06,
+      "loss": 1.7175,
+      "step": 1962
+    },
+    {
+      "epoch": 2.1512328767123288,
+      "grad_norm": 2.561642646789551,
+      "learning_rate": 1.6023558082859464e-06,
+      "loss": 1.7409,
+      "step": 1963
+    },
+    {
+      "epoch": 2.152328767123288,
+      "grad_norm": 2.6062211990356445,
+      "learning_rate": 1.6003249390739236e-06,
+      "loss": 1.7052,
+      "step": 1964
+    },
+    {
+      "epoch": 2.1534246575342464,
+      "grad_norm": 2.5122230052948,
+      "learning_rate": 1.598294069861901e-06,
+      "loss": 1.7444,
+      "step": 1965
+    },
+    {
+      "epoch": 2.1545205479452054,
+      "grad_norm": 2.5838866233825684,
+      "learning_rate": 1.5962632006498782e-06,
+      "loss": 1.716,
+      "step": 1966
+    },
+    {
+      "epoch": 2.1556164383561645,
+      "grad_norm": 2.5454747676849365,
+      "learning_rate": 1.5942323314378555e-06,
+      "loss": 1.7902,
+      "step": 1967
+    },
+    {
+      "epoch": 2.1567123287671235,
+      "grad_norm": 2.57637619972229,
+      "learning_rate": 1.5922014622258327e-06,
+      "loss": 1.6963,
+      "step": 1968
+    },
+    {
+      "epoch": 2.157808219178082,
+      "grad_norm": 2.5542941093444824,
+      "learning_rate": 1.59017059301381e-06,
+      "loss": 1.7151,
+      "step": 1969
+    },
+    {
+      "epoch": 2.158904109589041,
+      "grad_norm": 2.4912102222442627,
+      "learning_rate": 1.5881397238017873e-06,
+      "loss": 1.7624,
+      "step": 1970
+    },
+    {
+      "epoch": 2.158904109589041,
+      "eval_loss": 1.9855093955993652,
+      "eval_runtime": 37.0796,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 1970
+    },
+    {
+      "epoch": 2.16,
+      "grad_norm": 2.4793057441711426,
+      "learning_rate": 1.5861088545897646e-06,
+      "loss": 1.7617,
+      "step": 1971
+    },
+    {
+      "epoch": 2.1610958904109587,
+      "grad_norm": 2.5231704711914062,
+      "learning_rate": 1.5840779853777418e-06,
+      "loss": 1.6845,
+      "step": 1972
+    },
+    {
+      "epoch": 2.1621917808219178,
+      "grad_norm": 2.5126073360443115,
+      "learning_rate": 1.5820471161657191e-06,
+      "loss": 1.7468,
+      "step": 1973
+    },
+    {
+      "epoch": 2.163287671232877,
+      "grad_norm": 2.5558865070343018,
+      "learning_rate": 1.5800162469536964e-06,
+      "loss": 1.7253,
+      "step": 1974
+    },
+    {
+      "epoch": 2.1643835616438354,
+      "grad_norm": 2.808987855911255,
+      "learning_rate": 1.5779853777416735e-06,
+      "loss": 1.7844,
+      "step": 1975
+    },
+    {
+      "epoch": 2.1654794520547944,
+      "grad_norm": 2.765275001525879,
+      "learning_rate": 1.5759545085296507e-06,
+      "loss": 1.7658,
+      "step": 1976
+    },
+    {
+      "epoch": 2.1665753424657535,
+      "grad_norm": 2.5721709728240967,
+      "learning_rate": 1.573923639317628e-06,
+      "loss": 1.74,
+      "step": 1977
+    },
+    {
+      "epoch": 2.1676712328767125,
+      "grad_norm": 2.4276933670043945,
+      "learning_rate": 1.5718927701056053e-06,
+      "loss": 1.7613,
+      "step": 1978
+    },
+    {
+      "epoch": 2.168767123287671,
+      "grad_norm": 2.5665223598480225,
+      "learning_rate": 1.5698619008935826e-06,
+      "loss": 1.7514,
+      "step": 1979
+    },
+    {
+      "epoch": 2.16986301369863,
+      "grad_norm": 2.5588295459747314,
+      "learning_rate": 1.5678310316815598e-06,
+      "loss": 1.7749,
+      "step": 1980
+    },
+    {
+      "epoch": 2.16986301369863,
+      "eval_loss": 1.9858319759368896,
+      "eval_runtime": 37.0705,
+      "eval_samples_per_second": 26.976,
+      "eval_steps_per_second": 1.133,
+      "step": 1980
+    },
+    {
+      "epoch": 2.170958904109589,
+      "grad_norm": 2.5413172245025635,
+      "learning_rate": 1.5658001624695371e-06,
+      "loss": 1.706,
+      "step": 1981
+    },
+    {
+      "epoch": 2.172054794520548,
+      "grad_norm": 2.5246310234069824,
+      "learning_rate": 1.5637692932575144e-06,
+      "loss": 1.7825,
+      "step": 1982
+    },
+    {
+      "epoch": 2.1731506849315068,
+      "grad_norm": 2.385380506515503,
+      "learning_rate": 1.5617384240454917e-06,
+      "loss": 1.7268,
+      "step": 1983
+    },
+    {
+      "epoch": 2.174246575342466,
+      "grad_norm": 2.4967970848083496,
+      "learning_rate": 1.559707554833469e-06,
+      "loss": 1.7424,
+      "step": 1984
+    },
+    {
+      "epoch": 2.175342465753425,
+      "grad_norm": 2.563683271408081,
+      "learning_rate": 1.5576766856214462e-06,
+      "loss": 1.7904,
+      "step": 1985
+    },
+    {
+      "epoch": 2.1764383561643834,
+      "grad_norm": 2.518751859664917,
+      "learning_rate": 1.5556458164094235e-06,
+      "loss": 1.6616,
+      "step": 1986
+    },
+    {
+      "epoch": 2.1775342465753424,
+      "grad_norm": 2.5789380073547363,
+      "learning_rate": 1.5536149471974008e-06,
+      "loss": 1.7363,
+      "step": 1987
+    },
+    {
+      "epoch": 2.1786301369863015,
+      "grad_norm": 2.567406415939331,
+      "learning_rate": 1.5515840779853778e-06,
+      "loss": 1.6835,
+      "step": 1988
+    },
+    {
+      "epoch": 2.17972602739726,
+      "grad_norm": 2.509427309036255,
+      "learning_rate": 1.549553208773355e-06,
+      "loss": 1.7493,
+      "step": 1989
+    },
+    {
+      "epoch": 2.180821917808219,
+      "grad_norm": 2.6355340480804443,
+      "learning_rate": 1.5475223395613324e-06,
+      "loss": 1.7292,
+      "step": 1990
+    },
+    {
+      "epoch": 2.180821917808219,
+      "eval_loss": 1.9893935918807983,
+      "eval_runtime": 37.0928,
+      "eval_samples_per_second": 26.959,
+      "eval_steps_per_second": 1.132,
+      "step": 1990
+    },
+    {
+      "epoch": 2.181917808219178,
+      "grad_norm": 2.564587116241455,
+      "learning_rate": 1.5454914703493097e-06,
+      "loss": 1.7248,
+      "step": 1991
+    },
+    {
+      "epoch": 2.183013698630137,
+      "grad_norm": 2.6708462238311768,
+      "learning_rate": 1.543460601137287e-06,
+      "loss": 1.8123,
+      "step": 1992
+    },
+    {
+      "epoch": 2.1841095890410958,
+      "grad_norm": 2.6396312713623047,
+      "learning_rate": 1.5414297319252642e-06,
+      "loss": 1.7742,
+      "step": 1993
+    },
+    {
+      "epoch": 2.185205479452055,
+      "grad_norm": 2.553544044494629,
+      "learning_rate": 1.5393988627132415e-06,
+      "loss": 1.7705,
+      "step": 1994
+    },
+    {
+      "epoch": 2.186301369863014,
+      "grad_norm": 2.554997682571411,
+      "learning_rate": 1.5373679935012188e-06,
+      "loss": 1.737,
+      "step": 1995
+    },
+    {
+      "epoch": 2.1873972602739724,
+      "grad_norm": 2.5650475025177,
+      "learning_rate": 1.535337124289196e-06,
+      "loss": 1.6865,
+      "step": 1996
+    },
+    {
+      "epoch": 2.1884931506849314,
+      "grad_norm": 2.8019111156463623,
+      "learning_rate": 1.5333062550771733e-06,
+      "loss": 1.7428,
+      "step": 1997
+    },
+    {
+      "epoch": 2.1895890410958905,
+      "grad_norm": 2.5704092979431152,
+      "learning_rate": 1.5312753858651506e-06,
+      "loss": 1.8291,
+      "step": 1998
+    },
+    {
+      "epoch": 2.1906849315068495,
+      "grad_norm": 2.5332226753234863,
+      "learning_rate": 1.5292445166531279e-06,
+      "loss": 1.7274,
+      "step": 1999
+    },
+    {
+      "epoch": 2.191780821917808,
+      "grad_norm": 2.555288553237915,
+      "learning_rate": 1.5272136474411047e-06,
+      "loss": 1.7497,
+      "step": 2000
+    },
+    {
+      "epoch": 2.191780821917808,
+      "eval_loss": 1.9881768226623535,
+      "eval_runtime": 37.0996,
+      "eval_samples_per_second": 26.954,
+      "eval_steps_per_second": 1.132,
+      "step": 2000
+    },
+    {
+      "epoch": 2.192876712328767,
+      "grad_norm": 2.6068379878997803,
+      "learning_rate": 1.525182778229082e-06,
+      "loss": 1.7955,
+      "step": 2001
+    },
+    {
+      "epoch": 2.193972602739726,
+      "grad_norm": 2.479121208190918,
+      "learning_rate": 1.5231519090170593e-06,
+      "loss": 1.7342,
+      "step": 2002
+    },
+    {
+      "epoch": 2.1950684931506848,
+      "grad_norm": 2.495670795440674,
+      "learning_rate": 1.5211210398050365e-06,
+      "loss": 1.7125,
+      "step": 2003
+    },
+    {
+      "epoch": 2.196164383561644,
+      "grad_norm": 2.4988651275634766,
+      "learning_rate": 1.5190901705930138e-06,
+      "loss": 1.7137,
+      "step": 2004
+    },
+    {
+      "epoch": 2.197260273972603,
+      "grad_norm": 2.5191574096679688,
+      "learning_rate": 1.517059301380991e-06,
+      "loss": 1.7672,
+      "step": 2005
+    },
+    {
+      "epoch": 2.198356164383562,
+      "grad_norm": 2.5469892024993896,
+      "learning_rate": 1.5150284321689684e-06,
+      "loss": 1.7127,
+      "step": 2006
+    },
+    {
+      "epoch": 2.1994520547945204,
+      "grad_norm": 2.490896701812744,
+      "learning_rate": 1.5129975629569456e-06,
+      "loss": 1.7397,
+      "step": 2007
+    },
+    {
+      "epoch": 2.2005479452054795,
+      "grad_norm": 2.5142805576324463,
+      "learning_rate": 1.510966693744923e-06,
+      "loss": 1.7431,
+      "step": 2008
+    },
+    {
+      "epoch": 2.2016438356164385,
+      "grad_norm": 2.5397887229919434,
+      "learning_rate": 1.5089358245329002e-06,
+      "loss": 1.7439,
+      "step": 2009
+    },
+    {
+      "epoch": 2.202739726027397,
+      "grad_norm": 2.5063931941986084,
+      "learning_rate": 1.5069049553208775e-06,
+      "loss": 1.7259,
+      "step": 2010
+    },
+    {
+      "epoch": 2.202739726027397,
+      "eval_loss": 1.9853808879852295,
+      "eval_runtime": 36.8066,
+      "eval_samples_per_second": 27.169,
+      "eval_steps_per_second": 1.141,
+      "step": 2010
+    },
+    {
+      "epoch": 2.203835616438356,
+      "grad_norm": 2.6017589569091797,
+      "learning_rate": 1.5048740861088547e-06,
+      "loss": 1.7714,
+      "step": 2011
+    },
+    {
+      "epoch": 2.204931506849315,
+      "grad_norm": 2.5295040607452393,
+      "learning_rate": 1.502843216896832e-06,
+      "loss": 1.7561,
+      "step": 2012
+    },
+    {
+      "epoch": 2.206027397260274,
+      "grad_norm": 2.5187482833862305,
+      "learning_rate": 1.500812347684809e-06,
+      "loss": 1.6761,
+      "step": 2013
+    },
+    {
+      "epoch": 2.207123287671233,
+      "grad_norm": 2.5163650512695312,
+      "learning_rate": 1.4987814784727864e-06,
+      "loss": 1.7469,
+      "step": 2014
+    },
+    {
+      "epoch": 2.208219178082192,
+      "grad_norm": 2.731238603591919,
+      "learning_rate": 1.4967506092607636e-06,
+      "loss": 1.7527,
+      "step": 2015
+    },
+    {
+      "epoch": 2.209315068493151,
+      "grad_norm": 2.5298311710357666,
+      "learning_rate": 1.494719740048741e-06,
+      "loss": 1.6912,
+      "step": 2016
+    },
+    {
+      "epoch": 2.2104109589041094,
+      "grad_norm": 2.6385974884033203,
+      "learning_rate": 1.4926888708367182e-06,
+      "loss": 1.7664,
+      "step": 2017
+    },
+    {
+      "epoch": 2.2115068493150685,
+      "grad_norm": 2.6150033473968506,
+      "learning_rate": 1.4906580016246955e-06,
+      "loss": 1.7708,
+      "step": 2018
+    },
+    {
+      "epoch": 2.2126027397260275,
+      "grad_norm": 2.589329957962036,
+      "learning_rate": 1.4886271324126727e-06,
+      "loss": 1.7875,
+      "step": 2019
+    },
+    {
+      "epoch": 2.213698630136986,
+      "grad_norm": 2.6769933700561523,
+      "learning_rate": 1.48659626320065e-06,
+      "loss": 1.8136,
+      "step": 2020
+    },
+    {
+      "epoch": 2.213698630136986,
+      "eval_loss": 1.9827086925506592,
+      "eval_runtime": 37.0026,
+      "eval_samples_per_second": 27.025,
+      "eval_steps_per_second": 1.135,
+      "step": 2020
+    },
+    {
+      "epoch": 2.214794520547945,
+      "grad_norm": 2.7220497131347656,
+      "learning_rate": 1.4845653939886273e-06,
+      "loss": 1.6627,
+      "step": 2021
+    },
+    {
+      "epoch": 2.215890410958904,
+      "grad_norm": 2.6115968227386475,
+      "learning_rate": 1.4825345247766046e-06,
+      "loss": 1.6676,
+      "step": 2022
+    },
+    {
+      "epoch": 2.216986301369863,
+      "grad_norm": 2.4612581729888916,
+      "learning_rate": 1.4805036555645818e-06,
+      "loss": 1.6644,
+      "step": 2023
+    },
+    {
+      "epoch": 2.2180821917808218,
+      "grad_norm": 2.5708696842193604,
+      "learning_rate": 1.4784727863525591e-06,
+      "loss": 1.7357,
+      "step": 2024
+    },
+    {
+      "epoch": 2.219178082191781,
+      "grad_norm": 2.578158140182495,
+      "learning_rate": 1.4764419171405362e-06,
+      "loss": 1.7011,
+      "step": 2025
+    },
+    {
+      "epoch": 2.22027397260274,
+      "grad_norm": 2.547666072845459,
+      "learning_rate": 1.4744110479285135e-06,
+      "loss": 1.7787,
+      "step": 2026
+    },
+    {
+      "epoch": 2.2213698630136984,
+      "grad_norm": 2.617551326751709,
+      "learning_rate": 1.4723801787164907e-06,
+      "loss": 1.784,
+      "step": 2027
+    },
+    {
+      "epoch": 2.2224657534246575,
+      "grad_norm": 2.4385478496551514,
+      "learning_rate": 1.470349309504468e-06,
+      "loss": 1.7764,
+      "step": 2028
+    },
+    {
+      "epoch": 2.2235616438356165,
+      "grad_norm": 2.6088693141937256,
+      "learning_rate": 1.4683184402924453e-06,
+      "loss": 1.7477,
+      "step": 2029
+    },
+    {
+      "epoch": 2.2246575342465755,
+      "grad_norm": 2.6165895462036133,
+      "learning_rate": 1.4662875710804226e-06,
+      "loss": 1.7531,
+      "step": 2030
+    },
+    {
+      "epoch": 2.2246575342465755,
+      "eval_loss": 1.9826008081436157,
+      "eval_runtime": 37.085,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 2030
+    },
+    {
+      "epoch": 2.225753424657534,
+      "grad_norm": 2.6285006999969482,
+      "learning_rate": 1.4642567018683998e-06,
+      "loss": 1.8403,
+      "step": 2031
+    },
+    {
+      "epoch": 2.226849315068493,
+      "grad_norm": 2.549326181411743,
+      "learning_rate": 1.4622258326563771e-06,
+      "loss": 1.7304,
+      "step": 2032
+    },
+    {
+      "epoch": 2.227945205479452,
+      "grad_norm": 2.6198782920837402,
+      "learning_rate": 1.4601949634443544e-06,
+      "loss": 1.7147,
+      "step": 2033
+    },
+    {
+      "epoch": 2.2290410958904108,
+      "grad_norm": 2.536947727203369,
+      "learning_rate": 1.4581640942323317e-06,
+      "loss": 1.7639,
+      "step": 2034
+    },
+    {
+      "epoch": 2.23013698630137,
+      "grad_norm": 2.6757264137268066,
+      "learning_rate": 1.456133225020309e-06,
+      "loss": 1.8434,
+      "step": 2035
+    },
+    {
+      "epoch": 2.231232876712329,
+      "grad_norm": 2.598095655441284,
+      "learning_rate": 1.4541023558082862e-06,
+      "loss": 1.7593,
+      "step": 2036
+    },
+    {
+      "epoch": 2.232328767123288,
+      "grad_norm": 2.6104440689086914,
+      "learning_rate": 1.4520714865962635e-06,
+      "loss": 1.7473,
+      "step": 2037
+    },
+    {
+      "epoch": 2.2334246575342465,
+      "grad_norm": 2.6350162029266357,
+      "learning_rate": 1.4500406173842406e-06,
+      "loss": 1.777,
+      "step": 2038
+    },
+    {
+      "epoch": 2.2345205479452055,
+      "grad_norm": 2.593874216079712,
+      "learning_rate": 1.4480097481722178e-06,
+      "loss": 1.7506,
+      "step": 2039
+    },
+    {
+      "epoch": 2.2356164383561645,
+      "grad_norm": 2.6991093158721924,
+      "learning_rate": 1.4459788789601951e-06,
+      "loss": 1.7375,
+      "step": 2040
+    },
+    {
+      "epoch": 2.2356164383561645,
+      "eval_loss": 1.9815349578857422,
+      "eval_runtime": 37.1126,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 2040
+    },
+    {
+      "epoch": 2.236712328767123,
+      "grad_norm": 2.6339058876037598,
+      "learning_rate": 1.4439480097481724e-06,
+      "loss": 1.7045,
+      "step": 2041
+    },
+    {
+      "epoch": 2.237808219178082,
+      "grad_norm": 2.691631555557251,
+      "learning_rate": 1.4419171405361497e-06,
+      "loss": 1.7063,
+      "step": 2042
+    },
+    {
+      "epoch": 2.238904109589041,
+      "grad_norm": 3.021667003631592,
+      "learning_rate": 1.439886271324127e-06,
+      "loss": 1.7607,
+      "step": 2043
+    },
+    {
+      "epoch": 2.24,
+      "grad_norm": 2.690802812576294,
+      "learning_rate": 1.4378554021121042e-06,
+      "loss": 1.7594,
+      "step": 2044
+    },
+    {
+      "epoch": 2.241095890410959,
+      "grad_norm": 2.5371177196502686,
+      "learning_rate": 1.4358245329000815e-06,
+      "loss": 1.7177,
+      "step": 2045
+    },
+    {
+      "epoch": 2.242191780821918,
+      "grad_norm": 2.5586729049682617,
+      "learning_rate": 1.4337936636880588e-06,
+      "loss": 1.7864,
+      "step": 2046
+    },
+    {
+      "epoch": 2.243287671232877,
+      "grad_norm": 2.6621484756469727,
+      "learning_rate": 1.431762794476036e-06,
+      "loss": 1.7616,
+      "step": 2047
+    },
+    {
+      "epoch": 2.2443835616438355,
+      "grad_norm": 2.787508010864258,
+      "learning_rate": 1.4297319252640133e-06,
+      "loss": 1.7056,
+      "step": 2048
+    },
+    {
+      "epoch": 2.2454794520547945,
+      "grad_norm": 2.6896064281463623,
+      "learning_rate": 1.4277010560519906e-06,
+      "loss": 1.7413,
+      "step": 2049
+    },
+    {
+      "epoch": 2.2465753424657535,
+      "grad_norm": 2.540524959564209,
+      "learning_rate": 1.4256701868399674e-06,
+      "loss": 1.7096,
+      "step": 2050
+    },
+    {
+      "epoch": 2.2465753424657535,
+      "eval_loss": 1.9832353591918945,
+      "eval_runtime": 37.1202,
+      "eval_samples_per_second": 26.94,
+      "eval_steps_per_second": 1.131,
+      "step": 2050
+    },
+    {
+      "epoch": 2.247671232876712,
+      "grad_norm": 2.582171678543091,
+      "learning_rate": 1.4236393176279447e-06,
+      "loss": 1.8262,
+      "step": 2051
+    },
+    {
+      "epoch": 2.248767123287671,
+      "grad_norm": 2.659044027328491,
+      "learning_rate": 1.421608448415922e-06,
+      "loss": 1.7458,
+      "step": 2052
+    },
+    {
+      "epoch": 2.24986301369863,
+      "grad_norm": 2.7265775203704834,
+      "learning_rate": 1.4195775792038993e-06,
+      "loss": 1.8341,
+      "step": 2053
+    },
+    {
+      "epoch": 2.250958904109589,
+      "grad_norm": 2.922330379486084,
+      "learning_rate": 1.4175467099918765e-06,
+      "loss": 1.7103,
+      "step": 2054
+    },
+    {
+      "epoch": 2.252054794520548,
+      "grad_norm": 2.4665236473083496,
+      "learning_rate": 1.4155158407798538e-06,
+      "loss": 1.7461,
+      "step": 2055
+    },
+    {
+      "epoch": 2.253150684931507,
+      "grad_norm": 2.639538049697876,
+      "learning_rate": 1.413484971567831e-06,
+      "loss": 1.7683,
+      "step": 2056
+    },
+    {
+      "epoch": 2.254246575342466,
+      "grad_norm": 2.558363676071167,
+      "learning_rate": 1.4114541023558084e-06,
+      "loss": 1.6821,
+      "step": 2057
+    },
+    {
+      "epoch": 2.255342465753425,
+      "grad_norm": 2.6308321952819824,
+      "learning_rate": 1.4094232331437856e-06,
+      "loss": 1.7226,
+      "step": 2058
+    },
+    {
+      "epoch": 2.2564383561643835,
+      "grad_norm": 2.6325299739837646,
+      "learning_rate": 1.407392363931763e-06,
+      "loss": 1.783,
+      "step": 2059
+    },
+    {
+      "epoch": 2.2575342465753425,
+      "grad_norm": 2.6024723052978516,
+      "learning_rate": 1.4053614947197402e-06,
+      "loss": 1.8402,
+      "step": 2060
+    },
+    {
+      "epoch": 2.2575342465753425,
+      "eval_loss": 1.983332633972168,
+      "eval_runtime": 37.1133,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 2060
+    },
+    {
+      "epoch": 2.2586301369863016,
+      "grad_norm": 2.548522710800171,
+      "learning_rate": 1.4033306255077175e-06,
+      "loss": 1.7625,
+      "step": 2061
+    },
+    {
+      "epoch": 2.25972602739726,
+      "grad_norm": 2.5953831672668457,
+      "learning_rate": 1.4012997562956945e-06,
+      "loss": 1.6632,
+      "step": 2062
+    },
+    {
+      "epoch": 2.260821917808219,
+      "grad_norm": 2.5880215167999268,
+      "learning_rate": 1.3992688870836718e-06,
+      "loss": 1.7957,
+      "step": 2063
+    },
+    {
+      "epoch": 2.261917808219178,
+      "grad_norm": 2.439326047897339,
+      "learning_rate": 1.397238017871649e-06,
+      "loss": 1.76,
+      "step": 2064
+    },
+    {
+      "epoch": 2.263013698630137,
+      "grad_norm": 2.5185437202453613,
+      "learning_rate": 1.3952071486596264e-06,
+      "loss": 1.7385,
+      "step": 2065
+    },
+    {
+      "epoch": 2.264109589041096,
+      "grad_norm": 2.5943732261657715,
+      "learning_rate": 1.3931762794476036e-06,
+      "loss": 1.7066,
+      "step": 2066
+    },
+    {
+      "epoch": 2.265205479452055,
+      "grad_norm": 2.6375551223754883,
+      "learning_rate": 1.391145410235581e-06,
+      "loss": 1.7884,
+      "step": 2067
+    },
+    {
+      "epoch": 2.266301369863014,
+      "grad_norm": 2.4962515830993652,
+      "learning_rate": 1.3891145410235582e-06,
+      "loss": 1.7295,
+      "step": 2068
+    },
+    {
+      "epoch": 2.2673972602739725,
+      "grad_norm": 2.5873053073883057,
+      "learning_rate": 1.3870836718115355e-06,
+      "loss": 1.6697,
+      "step": 2069
+    },
+    {
+      "epoch": 2.2684931506849315,
+      "grad_norm": 2.573211908340454,
+      "learning_rate": 1.3850528025995127e-06,
+      "loss": 1.7198,
+      "step": 2070
+    },
+    {
+      "epoch": 2.2684931506849315,
+      "eval_loss": 1.9826557636260986,
+      "eval_runtime": 37.1271,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 2070
+    },
+    {
+      "epoch": 2.2695890410958905,
+      "grad_norm": 2.68015718460083,
+      "learning_rate": 1.38302193338749e-06,
+      "loss": 1.7294,
+      "step": 2071
+    },
+    {
+      "epoch": 2.270684931506849,
+      "grad_norm": 2.4882888793945312,
+      "learning_rate": 1.3809910641754673e-06,
+      "loss": 1.6592,
+      "step": 2072
+    },
+    {
+      "epoch": 2.271780821917808,
+      "grad_norm": 2.587048053741455,
+      "learning_rate": 1.3789601949634446e-06,
+      "loss": 1.7605,
+      "step": 2073
+    },
+    {
+      "epoch": 2.272876712328767,
+      "grad_norm": 2.6452062129974365,
+      "learning_rate": 1.3769293257514218e-06,
+      "loss": 1.6664,
+      "step": 2074
+    },
+    {
+      "epoch": 2.2739726027397262,
+      "grad_norm": 2.5663232803344727,
+      "learning_rate": 1.374898456539399e-06,
+      "loss": 1.771,
+      "step": 2075
+    },
+    {
+      "epoch": 2.275068493150685,
+      "grad_norm": 2.652989149093628,
+      "learning_rate": 1.3728675873273762e-06,
+      "loss": 1.7096,
+      "step": 2076
+    },
+    {
+      "epoch": 2.276164383561644,
+      "grad_norm": 2.538435220718384,
+      "learning_rate": 1.3708367181153535e-06,
+      "loss": 1.798,
+      "step": 2077
+    },
+    {
+      "epoch": 2.277260273972603,
+      "grad_norm": 2.4685096740722656,
+      "learning_rate": 1.3688058489033307e-06,
+      "loss": 1.7904,
+      "step": 2078
+    },
+    {
+      "epoch": 2.2783561643835615,
+      "grad_norm": 2.566723108291626,
+      "learning_rate": 1.366774979691308e-06,
+      "loss": 1.7569,
+      "step": 2079
+    },
+    {
+      "epoch": 2.2794520547945205,
+      "grad_norm": 2.560960531234741,
+      "learning_rate": 1.3647441104792853e-06,
+      "loss": 1.7734,
+      "step": 2080
+    },
+    {
+      "epoch": 2.2794520547945205,
+      "eval_loss": 1.9809880256652832,
+      "eval_runtime": 37.1137,
+      "eval_samples_per_second": 26.944,
+      "eval_steps_per_second": 1.132,
+      "step": 2080
+    },
+    {
+      "epoch": 2.2805479452054795,
+      "grad_norm": 2.576608419418335,
+      "learning_rate": 1.3627132412672626e-06,
+      "loss": 1.7431,
+      "step": 2081
+    },
+    {
+      "epoch": 2.281643835616438,
+      "grad_norm": 2.648301601409912,
+      "learning_rate": 1.3606823720552398e-06,
+      "loss": 1.6959,
+      "step": 2082
+    },
+    {
+      "epoch": 2.282739726027397,
+      "grad_norm": 2.516493082046509,
+      "learning_rate": 1.3586515028432171e-06,
+      "loss": 1.7519,
+      "step": 2083
+    },
+    {
+      "epoch": 2.283835616438356,
+      "grad_norm": 2.474616050720215,
+      "learning_rate": 1.3566206336311944e-06,
+      "loss": 1.6796,
+      "step": 2084
+    },
+    {
+      "epoch": 2.2849315068493152,
+      "grad_norm": 2.482044219970703,
+      "learning_rate": 1.3545897644191717e-06,
+      "loss": 1.7443,
+      "step": 2085
+    },
+    {
+      "epoch": 2.286027397260274,
+      "grad_norm": 2.6574456691741943,
+      "learning_rate": 1.352558895207149e-06,
+      "loss": 1.7145,
+      "step": 2086
+    },
+    {
+      "epoch": 2.287123287671233,
+      "grad_norm": 2.5350115299224854,
+      "learning_rate": 1.3505280259951258e-06,
+      "loss": 1.765,
+      "step": 2087
+    },
+    {
+      "epoch": 2.288219178082192,
+      "grad_norm": 2.6488730907440186,
+      "learning_rate": 1.348497156783103e-06,
+      "loss": 1.7407,
+      "step": 2088
+    },
+    {
+      "epoch": 2.289315068493151,
+      "grad_norm": 2.6020500659942627,
+      "learning_rate": 1.3464662875710804e-06,
+      "loss": 1.779,
+      "step": 2089
+    },
+    {
+      "epoch": 2.2904109589041095,
+      "grad_norm": 2.633013963699341,
+      "learning_rate": 1.3444354183590576e-06,
+      "loss": 1.6824,
+      "step": 2090
+    },
+    {
+      "epoch": 2.2904109589041095,
+      "eval_loss": 1.982380986213684,
+      "eval_runtime": 37.1166,
+      "eval_samples_per_second": 26.942,
+      "eval_steps_per_second": 1.132,
+      "step": 2090
+    },
+    {
+      "epoch": 2.2915068493150685,
+      "grad_norm": 2.6399149894714355,
+      "learning_rate": 1.342404549147035e-06,
+      "loss": 1.7634,
+      "step": 2091
+    },
+    {
+      "epoch": 2.2926027397260276,
+      "grad_norm": 2.580634832382202,
+      "learning_rate": 1.3403736799350122e-06,
+      "loss": 1.7099,
+      "step": 2092
+    },
+    {
+      "epoch": 2.293698630136986,
+      "grad_norm": 2.523400068283081,
+      "learning_rate": 1.3383428107229895e-06,
+      "loss": 1.7705,
+      "step": 2093
+    },
+    {
+      "epoch": 2.294794520547945,
+      "grad_norm": 2.5114550590515137,
+      "learning_rate": 1.3363119415109667e-06,
+      "loss": 1.7961,
+      "step": 2094
+    },
+    {
+      "epoch": 2.2958904109589042,
+      "grad_norm": 2.457369804382324,
+      "learning_rate": 1.334281072298944e-06,
+      "loss": 1.7507,
+      "step": 2095
+    },
+    {
+      "epoch": 2.296986301369863,
+      "grad_norm": 2.5671467781066895,
+      "learning_rate": 1.3322502030869213e-06,
+      "loss": 1.7607,
+      "step": 2096
+    },
+    {
+      "epoch": 2.298082191780822,
+      "grad_norm": 2.459540367126465,
+      "learning_rate": 1.3302193338748986e-06,
+      "loss": 1.7193,
+      "step": 2097
+    },
+    {
+      "epoch": 2.299178082191781,
+      "grad_norm": 2.4934308528900146,
+      "learning_rate": 1.3281884646628758e-06,
+      "loss": 1.7775,
+      "step": 2098
+    },
+    {
+      "epoch": 2.30027397260274,
+      "grad_norm": 2.520641565322876,
+      "learning_rate": 1.3261575954508531e-06,
+      "loss": 1.7847,
+      "step": 2099
+    },
+    {
+      "epoch": 2.3013698630136985,
+      "grad_norm": 2.4607160091400146,
+      "learning_rate": 1.3241267262388302e-06,
+      "loss": 1.7812,
+      "step": 2100
+    },
+    {
+      "epoch": 2.3013698630136985,
+      "eval_loss": 1.9836152791976929,
+      "eval_runtime": 37.1239,
+      "eval_samples_per_second": 26.937,
+      "eval_steps_per_second": 1.131,
+      "step": 2100
+    },
+    {
+      "epoch": 2.3024657534246575,
+      "grad_norm": 2.577761650085449,
+      "learning_rate": 1.3220958570268074e-06,
+      "loss": 1.7452,
+      "step": 2101
+    },
+    {
+      "epoch": 2.3035616438356166,
+      "grad_norm": 2.5331356525421143,
+      "learning_rate": 1.3200649878147847e-06,
+      "loss": 1.6989,
+      "step": 2102
+    },
+    {
+      "epoch": 2.304657534246575,
+      "grad_norm": 2.632581949234009,
+      "learning_rate": 1.318034118602762e-06,
+      "loss": 1.7368,
+      "step": 2103
+    },
+    {
+      "epoch": 2.305753424657534,
+      "grad_norm": 2.55940318107605,
+      "learning_rate": 1.3160032493907393e-06,
+      "loss": 1.7806,
+      "step": 2104
+    },
+    {
+      "epoch": 2.3068493150684932,
+      "grad_norm": 2.5680153369903564,
+      "learning_rate": 1.3139723801787166e-06,
+      "loss": 1.7482,
+      "step": 2105
+    },
+    {
+      "epoch": 2.3079452054794523,
+      "grad_norm": 2.7915756702423096,
+      "learning_rate": 1.3119415109666938e-06,
+      "loss": 1.832,
+      "step": 2106
+    },
+    {
+      "epoch": 2.309041095890411,
+      "grad_norm": 2.518031597137451,
+      "learning_rate": 1.309910641754671e-06,
+      "loss": 1.7401,
+      "step": 2107
+    },
+    {
+      "epoch": 2.31013698630137,
+      "grad_norm": 2.7410221099853516,
+      "learning_rate": 1.3078797725426484e-06,
+      "loss": 1.6992,
+      "step": 2108
+    },
+    {
+      "epoch": 2.311232876712329,
+      "grad_norm": 2.611825466156006,
+      "learning_rate": 1.3058489033306257e-06,
+      "loss": 1.7085,
+      "step": 2109
+    },
+    {
+      "epoch": 2.3123287671232875,
+      "grad_norm": 2.5833330154418945,
+      "learning_rate": 1.303818034118603e-06,
+      "loss": 1.7129,
+      "step": 2110
+    },
+    {
+      "epoch": 2.3123287671232875,
+      "eval_loss": 1.9824191331863403,
+      "eval_runtime": 36.773,
+      "eval_samples_per_second": 27.194,
+      "eval_steps_per_second": 1.142,
+      "step": 2110
+    },
+    {
+      "epoch": 2.3134246575342465,
+      "grad_norm": 2.587327003479004,
+      "learning_rate": 1.3017871649065802e-06,
+      "loss": 1.7439,
+      "step": 2111
+    },
+    {
+      "epoch": 2.3145205479452056,
+      "grad_norm": 2.574925422668457,
+      "learning_rate": 1.2997562956945573e-06,
+      "loss": 1.7173,
+      "step": 2112
+    },
+    {
+      "epoch": 2.315616438356164,
+      "grad_norm": 2.6404287815093994,
+      "learning_rate": 1.2977254264825345e-06,
+      "loss": 1.7052,
+      "step": 2113
+    },
+    {
+      "epoch": 2.316712328767123,
+      "grad_norm": 2.634584903717041,
+      "learning_rate": 1.2956945572705118e-06,
+      "loss": 1.741,
+      "step": 2114
+    },
+    {
+      "epoch": 2.317808219178082,
+      "grad_norm": 2.66159987449646,
+      "learning_rate": 1.293663688058489e-06,
+      "loss": 1.76,
+      "step": 2115
+    },
+    {
+      "epoch": 2.3189041095890413,
+      "grad_norm": 2.6337332725524902,
+      "learning_rate": 1.2916328188464664e-06,
+      "loss": 1.7035,
+      "step": 2116
+    },
+    {
+      "epoch": 2.32,
+      "grad_norm": 2.528524398803711,
+      "learning_rate": 1.2896019496344436e-06,
+      "loss": 1.6439,
+      "step": 2117
+    },
+    {
+      "epoch": 2.321095890410959,
+      "grad_norm": 2.5938384532928467,
+      "learning_rate": 1.287571080422421e-06,
+      "loss": 1.7134,
+      "step": 2118
+    },
+    {
+      "epoch": 2.322191780821918,
+      "grad_norm": 2.606271505355835,
+      "learning_rate": 1.2855402112103982e-06,
+      "loss": 1.7758,
+      "step": 2119
+    },
+    {
+      "epoch": 2.323287671232877,
+      "grad_norm": 2.530071258544922,
+      "learning_rate": 1.2835093419983755e-06,
+      "loss": 1.6927,
+      "step": 2120
+    },
+    {
+      "epoch": 2.323287671232877,
+      "eval_loss": 1.9856306314468384,
+      "eval_runtime": 36.9674,
+      "eval_samples_per_second": 27.051,
+      "eval_steps_per_second": 1.136,
+      "step": 2120
+    },
+    {
+      "epoch": 2.3243835616438355,
+      "grad_norm": 2.8125321865081787,
+      "learning_rate": 1.2814784727863528e-06,
+      "loss": 1.7225,
+      "step": 2121
+    },
+    {
+      "epoch": 2.3254794520547946,
+      "grad_norm": 2.57012939453125,
+      "learning_rate": 1.27944760357433e-06,
+      "loss": 1.7177,
+      "step": 2122
+    },
+    {
+      "epoch": 2.3265753424657536,
+      "grad_norm": 2.6595780849456787,
+      "learning_rate": 1.2774167343623073e-06,
+      "loss": 1.7807,
+      "step": 2123
+    },
+    {
+      "epoch": 2.327671232876712,
+      "grad_norm": 2.557115077972412,
+      "learning_rate": 1.2753858651502846e-06,
+      "loss": 1.7762,
+      "step": 2124
+    },
+    {
+      "epoch": 2.328767123287671,
+      "grad_norm": 2.763218879699707,
+      "learning_rate": 1.2733549959382616e-06,
+      "loss": 1.7638,
+      "step": 2125
+    },
+    {
+      "epoch": 2.3298630136986302,
+      "grad_norm": 2.7284505367279053,
+      "learning_rate": 1.271324126726239e-06,
+      "loss": 1.7387,
+      "step": 2126
+    },
+    {
+      "epoch": 2.330958904109589,
+      "grad_norm": 2.5893704891204834,
+      "learning_rate": 1.2692932575142162e-06,
+      "loss": 1.7322,
+      "step": 2127
+    },
+    {
+      "epoch": 2.332054794520548,
+      "grad_norm": 2.625582456588745,
+      "learning_rate": 1.2672623883021935e-06,
+      "loss": 1.6865,
+      "step": 2128
+    },
+    {
+      "epoch": 2.333150684931507,
+      "grad_norm": 2.5458641052246094,
+      "learning_rate": 1.2652315190901707e-06,
+      "loss": 1.6817,
+      "step": 2129
+    },
+    {
+      "epoch": 2.334246575342466,
+      "grad_norm": 2.5718271732330322,
+      "learning_rate": 1.263200649878148e-06,
+      "loss": 1.8076,
+      "step": 2130
+    },
+    {
+      "epoch": 2.334246575342466,
+      "eval_loss": 1.9806523323059082,
+      "eval_runtime": 36.9843,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 2130
+    },
+    {
+      "epoch": 2.3353424657534245,
+      "grad_norm": 2.4864919185638428,
+      "learning_rate": 1.2611697806661253e-06,
+      "loss": 1.7308,
+      "step": 2131
+    },
+    {
+      "epoch": 2.3364383561643836,
+      "grad_norm": 2.6007416248321533,
+      "learning_rate": 1.2591389114541026e-06,
+      "loss": 1.7523,
+      "step": 2132
+    },
+    {
+      "epoch": 2.3375342465753426,
+      "grad_norm": 2.600177049636841,
+      "learning_rate": 1.2571080422420798e-06,
+      "loss": 1.7422,
+      "step": 2133
+    },
+    {
+      "epoch": 2.338630136986301,
+      "grad_norm": 2.489025831222534,
+      "learning_rate": 1.2550771730300571e-06,
+      "loss": 1.7184,
+      "step": 2134
+    },
+    {
+      "epoch": 2.33972602739726,
+      "grad_norm": 2.5677363872528076,
+      "learning_rate": 1.2530463038180344e-06,
+      "loss": 1.7922,
+      "step": 2135
+    },
+    {
+      "epoch": 2.3408219178082192,
+      "grad_norm": 2.616976737976074,
+      "learning_rate": 1.2510154346060117e-06,
+      "loss": 1.7354,
+      "step": 2136
+    },
+    {
+      "epoch": 2.3419178082191783,
+      "grad_norm": 2.646456241607666,
+      "learning_rate": 1.2489845653939887e-06,
+      "loss": 1.7377,
+      "step": 2137
+    },
+    {
+      "epoch": 2.343013698630137,
+      "grad_norm": 2.550020933151245,
+      "learning_rate": 1.246953696181966e-06,
+      "loss": 1.7751,
+      "step": 2138
+    },
+    {
+      "epoch": 2.344109589041096,
+      "grad_norm": 2.634279251098633,
+      "learning_rate": 1.2449228269699433e-06,
+      "loss": 1.8157,
+      "step": 2139
+    },
+    {
+      "epoch": 2.345205479452055,
+      "grad_norm": 2.5621821880340576,
+      "learning_rate": 1.2428919577579204e-06,
+      "loss": 1.7383,
+      "step": 2140
+    },
+    {
+      "epoch": 2.345205479452055,
+      "eval_loss": 1.9809924364089966,
+      "eval_runtime": 37.004,
+      "eval_samples_per_second": 27.024,
+      "eval_steps_per_second": 1.135,
+      "step": 2140
+    },
+    {
+      "epoch": 2.3463013698630135,
+      "grad_norm": 2.638831615447998,
+      "learning_rate": 1.2408610885458976e-06,
+      "loss": 1.6988,
+      "step": 2141
+    },
+    {
+      "epoch": 2.3473972602739726,
+      "grad_norm": 2.634612798690796,
+      "learning_rate": 1.238830219333875e-06,
+      "loss": 1.7331,
+      "step": 2142
+    },
+    {
+      "epoch": 2.3484931506849316,
+      "grad_norm": 2.552955389022827,
+      "learning_rate": 1.2367993501218522e-06,
+      "loss": 1.7148,
+      "step": 2143
+    },
+    {
+      "epoch": 2.34958904109589,
+      "grad_norm": 2.6475753784179688,
+      "learning_rate": 1.2347684809098295e-06,
+      "loss": 1.7815,
+      "step": 2144
+    },
+    {
+      "epoch": 2.350684931506849,
+      "grad_norm": 2.7666079998016357,
+      "learning_rate": 1.2327376116978067e-06,
+      "loss": 1.7634,
+      "step": 2145
+    },
+    {
+      "epoch": 2.3517808219178082,
+      "grad_norm": 2.6810429096221924,
+      "learning_rate": 1.230706742485784e-06,
+      "loss": 1.6769,
+      "step": 2146
+    },
+    {
+      "epoch": 2.3528767123287673,
+      "grad_norm": 2.671382188796997,
+      "learning_rate": 1.2286758732737613e-06,
+      "loss": 1.6884,
+      "step": 2147
+    },
+    {
+      "epoch": 2.353972602739726,
+      "grad_norm": 2.614684820175171,
+      "learning_rate": 1.2266450040617386e-06,
+      "loss": 1.8226,
+      "step": 2148
+    },
+    {
+      "epoch": 2.355068493150685,
+      "grad_norm": 2.47189998626709,
+      "learning_rate": 1.2246141348497158e-06,
+      "loss": 1.819,
+      "step": 2149
+    },
+    {
+      "epoch": 2.356164383561644,
+      "grad_norm": 2.4641449451446533,
+      "learning_rate": 1.2225832656376931e-06,
+      "loss": 1.7494,
+      "step": 2150
+    },
+    {
+      "epoch": 2.356164383561644,
+      "eval_loss": 1.9806426763534546,
+      "eval_runtime": 37.0153,
+      "eval_samples_per_second": 27.016,
+      "eval_steps_per_second": 1.135,
+      "step": 2150
+    },
+    {
+      "epoch": 2.357260273972603,
+      "grad_norm": 2.7310373783111572,
+      "learning_rate": 1.2205523964256704e-06,
+      "loss": 1.7432,
+      "step": 2151
+    },
+    {
+      "epoch": 2.3583561643835615,
+      "grad_norm": 2.5554895401000977,
+      "learning_rate": 1.2185215272136477e-06,
+      "loss": 1.7213,
+      "step": 2152
+    },
+    {
+      "epoch": 2.3594520547945206,
+      "grad_norm": 2.5401952266693115,
+      "learning_rate": 1.2164906580016247e-06,
+      "loss": 1.7469,
+      "step": 2153
+    },
+    {
+      "epoch": 2.3605479452054796,
+      "grad_norm": 2.5213980674743652,
+      "learning_rate": 1.214459788789602e-06,
+      "loss": 1.7459,
+      "step": 2154
+    },
+    {
+      "epoch": 2.361643835616438,
+      "grad_norm": 2.5585155487060547,
+      "learning_rate": 1.2124289195775793e-06,
+      "loss": 1.6587,
+      "step": 2155
+    },
+    {
+      "epoch": 2.3627397260273972,
+      "grad_norm": 2.6329963207244873,
+      "learning_rate": 1.2103980503655566e-06,
+      "loss": 1.6925,
+      "step": 2156
+    },
+    {
+      "epoch": 2.3638356164383563,
+      "grad_norm": 2.678351640701294,
+      "learning_rate": 1.2083671811535338e-06,
+      "loss": 1.7322,
+      "step": 2157
+    },
+    {
+      "epoch": 2.364931506849315,
+      "grad_norm": 2.595313549041748,
+      "learning_rate": 1.2063363119415111e-06,
+      "loss": 1.7142,
+      "step": 2158
+    },
+    {
+      "epoch": 2.366027397260274,
+      "grad_norm": 2.5650861263275146,
+      "learning_rate": 1.2043054427294882e-06,
+      "loss": 1.7024,
+      "step": 2159
+    },
+    {
+      "epoch": 2.367123287671233,
+      "grad_norm": 2.495866537094116,
+      "learning_rate": 1.2022745735174654e-06,
+      "loss": 1.7205,
+      "step": 2160
+    },
+    {
+      "epoch": 2.367123287671233,
+      "eval_loss": 1.976922631263733,
+      "eval_runtime": 37.012,
+      "eval_samples_per_second": 27.018,
+      "eval_steps_per_second": 1.135,
+      "step": 2160
+    },
+    {
+      "epoch": 2.368219178082192,
+      "grad_norm": 2.570307493209839,
+      "learning_rate": 1.2002437043054427e-06,
+      "loss": 1.7601,
+      "step": 2161
+    },
+    {
+      "epoch": 2.3693150684931505,
+      "grad_norm": 2.444373846054077,
+      "learning_rate": 1.19821283509342e-06,
+      "loss": 1.7857,
+      "step": 2162
+    },
+    {
+      "epoch": 2.3704109589041096,
+      "grad_norm": 2.525505781173706,
+      "learning_rate": 1.1961819658813973e-06,
+      "loss": 1.7875,
+      "step": 2163
+    },
+    {
+      "epoch": 2.3715068493150686,
+      "grad_norm": 2.622260093688965,
+      "learning_rate": 1.1941510966693746e-06,
+      "loss": 1.7596,
+      "step": 2164
+    },
+    {
+      "epoch": 2.3726027397260276,
+      "grad_norm": 2.542076349258423,
+      "learning_rate": 1.1921202274573518e-06,
+      "loss": 1.7248,
+      "step": 2165
+    },
+    {
+      "epoch": 2.3736986301369862,
+      "grad_norm": 2.510263681411743,
+      "learning_rate": 1.190089358245329e-06,
+      "loss": 1.7168,
+      "step": 2166
+    },
+    {
+      "epoch": 2.3747945205479453,
+      "grad_norm": 2.605023145675659,
+      "learning_rate": 1.1880584890333064e-06,
+      "loss": 1.798,
+      "step": 2167
+    },
+    {
+      "epoch": 2.3758904109589043,
+      "grad_norm": 2.593261480331421,
+      "learning_rate": 1.1860276198212837e-06,
+      "loss": 1.7247,
+      "step": 2168
+    },
+    {
+      "epoch": 2.376986301369863,
+      "grad_norm": 2.6010820865631104,
+      "learning_rate": 1.183996750609261e-06,
+      "loss": 1.7607,
+      "step": 2169
+    },
+    {
+      "epoch": 2.378082191780822,
+      "grad_norm": 2.62886118888855,
+      "learning_rate": 1.1819658813972382e-06,
+      "loss": 1.7641,
+      "step": 2170
+    },
+    {
+      "epoch": 2.378082191780822,
+      "eval_loss": 1.9790476560592651,
+      "eval_runtime": 37.0413,
+      "eval_samples_per_second": 26.997,
+      "eval_steps_per_second": 1.134,
+      "step": 2170
+    },
+    {
+      "epoch": 2.379178082191781,
+      "grad_norm": 2.5506582260131836,
+      "learning_rate": 1.1799350121852153e-06,
+      "loss": 1.7407,
+      "step": 2171
+    },
+    {
+      "epoch": 2.3802739726027395,
+      "grad_norm": 2.547445774078369,
+      "learning_rate": 1.1779041429731925e-06,
+      "loss": 1.7808,
+      "step": 2172
+    },
+    {
+      "epoch": 2.3813698630136986,
+      "grad_norm": 2.5443003177642822,
+      "learning_rate": 1.1758732737611698e-06,
+      "loss": 1.7842,
+      "step": 2173
+    },
+    {
+      "epoch": 2.3824657534246576,
+      "grad_norm": 2.5733375549316406,
+      "learning_rate": 1.173842404549147e-06,
+      "loss": 1.6902,
+      "step": 2174
+    },
+    {
+      "epoch": 2.383561643835616,
+      "grad_norm": 2.63702392578125,
+      "learning_rate": 1.1718115353371244e-06,
+      "loss": 1.8049,
+      "step": 2175
+    },
+    {
+      "epoch": 2.3846575342465752,
+      "grad_norm": 2.5577163696289062,
+      "learning_rate": 1.1697806661251016e-06,
+      "loss": 1.7162,
+      "step": 2176
+    },
+    {
+      "epoch": 2.3857534246575343,
+      "grad_norm": 2.4259274005889893,
+      "learning_rate": 1.167749796913079e-06,
+      "loss": 1.7883,
+      "step": 2177
+    },
+    {
+      "epoch": 2.3868493150684933,
+      "grad_norm": 2.537519931793213,
+      "learning_rate": 1.1657189277010562e-06,
+      "loss": 1.7882,
+      "step": 2178
+    },
+    {
+      "epoch": 2.387945205479452,
+      "grad_norm": 2.5480241775512695,
+      "learning_rate": 1.1636880584890335e-06,
+      "loss": 1.7245,
+      "step": 2179
+    },
+    {
+      "epoch": 2.389041095890411,
+      "grad_norm": 2.630337953567505,
+      "learning_rate": 1.1616571892770108e-06,
+      "loss": 1.7165,
+      "step": 2180
+    },
+    {
+      "epoch": 2.389041095890411,
+      "eval_loss": 1.9772392511367798,
+      "eval_runtime": 37.1283,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 2180
+    },
+    {
+      "epoch": 2.39013698630137,
+      "grad_norm": 2.5741171836853027,
+      "learning_rate": 1.159626320064988e-06,
+      "loss": 1.7806,
+      "step": 2181
+    },
+    {
+      "epoch": 2.391232876712329,
+      "grad_norm": 2.6166305541992188,
+      "learning_rate": 1.1575954508529653e-06,
+      "loss": 1.744,
+      "step": 2182
+    },
+    {
+      "epoch": 2.3923287671232876,
+      "grad_norm": 2.6965837478637695,
+      "learning_rate": 1.1555645816409426e-06,
+      "loss": 1.7736,
+      "step": 2183
+    },
+    {
+      "epoch": 2.3934246575342466,
+      "grad_norm": 2.586827516555786,
+      "learning_rate": 1.1535337124289196e-06,
+      "loss": 1.7455,
+      "step": 2184
+    },
+    {
+      "epoch": 2.3945205479452056,
+      "grad_norm": 2.587235689163208,
+      "learning_rate": 1.151502843216897e-06,
+      "loss": 1.6999,
+      "step": 2185
+    },
+    {
+      "epoch": 2.3956164383561642,
+      "grad_norm": 2.656740665435791,
+      "learning_rate": 1.1494719740048742e-06,
+      "loss": 1.6518,
+      "step": 2186
+    },
+    {
+      "epoch": 2.3967123287671233,
+      "grad_norm": 2.5353455543518066,
+      "learning_rate": 1.1474411047928515e-06,
+      "loss": 1.7389,
+      "step": 2187
+    },
+    {
+      "epoch": 2.3978082191780823,
+      "grad_norm": 2.622314453125,
+      "learning_rate": 1.1454102355808287e-06,
+      "loss": 1.7309,
+      "step": 2188
+    },
+    {
+      "epoch": 2.398904109589041,
+      "grad_norm": 2.6286797523498535,
+      "learning_rate": 1.143379366368806e-06,
+      "loss": 1.7094,
+      "step": 2189
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 2.5203921794891357,
+      "learning_rate": 1.141348497156783e-06,
+      "loss": 1.6905,
+      "step": 2190
+    },
+    {
+      "epoch": 2.4,
+      "eval_loss": 1.9822566509246826,
+      "eval_runtime": 37.0581,
+      "eval_samples_per_second": 26.985,
+      "eval_steps_per_second": 1.133,
+      "step": 2190
+    },
+    {
+      "epoch": 2.401095890410959,
+      "grad_norm": 2.6562373638153076,
+      "learning_rate": 1.1393176279447604e-06,
+      "loss": 1.8293,
+      "step": 2191
+    },
+    {
+      "epoch": 2.402191780821918,
+      "grad_norm": 2.5581510066986084,
+      "learning_rate": 1.1372867587327376e-06,
+      "loss": 1.7936,
+      "step": 2192
+    },
+    {
+      "epoch": 2.4032876712328766,
+      "grad_norm": 2.534675121307373,
+      "learning_rate": 1.135255889520715e-06,
+      "loss": 1.6648,
+      "step": 2193
+    },
+    {
+      "epoch": 2.4043835616438356,
+      "grad_norm": 2.7236168384552,
+      "learning_rate": 1.1332250203086922e-06,
+      "loss": 1.7654,
+      "step": 2194
+    },
+    {
+      "epoch": 2.4054794520547946,
+      "grad_norm": 2.5706138610839844,
+      "learning_rate": 1.1311941510966695e-06,
+      "loss": 1.7427,
+      "step": 2195
+    },
+    {
+      "epoch": 2.4065753424657537,
+      "grad_norm": 2.7030224800109863,
+      "learning_rate": 1.1291632818846467e-06,
+      "loss": 1.7216,
+      "step": 2196
+    },
+    {
+      "epoch": 2.4076712328767123,
+      "grad_norm": 2.4936161041259766,
+      "learning_rate": 1.127132412672624e-06,
+      "loss": 1.7033,
+      "step": 2197
+    },
+    {
+      "epoch": 2.4087671232876713,
+      "grad_norm": 2.5384793281555176,
+      "learning_rate": 1.1251015434606013e-06,
+      "loss": 1.7023,
+      "step": 2198
+    },
+    {
+      "epoch": 2.4098630136986303,
+      "grad_norm": 2.5271613597869873,
+      "learning_rate": 1.1230706742485786e-06,
+      "loss": 1.738,
+      "step": 2199
+    },
+    {
+      "epoch": 2.410958904109589,
+      "grad_norm": 2.557610034942627,
+      "learning_rate": 1.1210398050365558e-06,
+      "loss": 1.7927,
+      "step": 2200
+    },
+    {
+      "epoch": 2.410958904109589,
+      "eval_loss": 1.9795563220977783,
+      "eval_runtime": 37.0832,
+      "eval_samples_per_second": 26.966,
+      "eval_steps_per_second": 1.133,
+      "step": 2200
+    },
+    {
+      "epoch": 2.412054794520548,
+      "grad_norm": 2.5497145652770996,
+      "learning_rate": 1.1190089358245331e-06,
+      "loss": 1.7361,
+      "step": 2201
+    },
+    {
+      "epoch": 2.413150684931507,
+      "grad_norm": 2.622793436050415,
+      "learning_rate": 1.1169780666125102e-06,
+      "loss": 1.6624,
+      "step": 2202
+    },
+    {
+      "epoch": 2.4142465753424656,
+      "grad_norm": 2.694563865661621,
+      "learning_rate": 1.1149471974004875e-06,
+      "loss": 1.7755,
+      "step": 2203
+    },
+    {
+      "epoch": 2.4153424657534246,
+      "grad_norm": 2.68461012840271,
+      "learning_rate": 1.1129163281884647e-06,
+      "loss": 1.7132,
+      "step": 2204
+    },
+    {
+      "epoch": 2.4164383561643836,
+      "grad_norm": 2.65791654586792,
+      "learning_rate": 1.110885458976442e-06,
+      "loss": 1.7728,
+      "step": 2205
+    },
+    {
+      "epoch": 2.417534246575342,
+      "grad_norm": 2.46262788772583,
+      "learning_rate": 1.1088545897644193e-06,
+      "loss": 1.6973,
+      "step": 2206
+    },
+    {
+      "epoch": 2.4186301369863012,
+      "grad_norm": 2.5602169036865234,
+      "learning_rate": 1.1068237205523966e-06,
+      "loss": 1.7651,
+      "step": 2207
+    },
+    {
+      "epoch": 2.4197260273972603,
+      "grad_norm": 2.5492701530456543,
+      "learning_rate": 1.1047928513403736e-06,
+      "loss": 1.7674,
+      "step": 2208
+    },
+    {
+      "epoch": 2.4208219178082193,
+      "grad_norm": 2.798940420150757,
+      "learning_rate": 1.102761982128351e-06,
+      "loss": 1.7719,
+      "step": 2209
+    },
+    {
+      "epoch": 2.421917808219178,
+      "grad_norm": 2.6989545822143555,
+      "learning_rate": 1.1007311129163282e-06,
+      "loss": 1.6821,
+      "step": 2210
+    },
+    {
+      "epoch": 2.421917808219178,
+      "eval_loss": 1.9763859510421753,
+      "eval_runtime": 36.7821,
+      "eval_samples_per_second": 27.187,
+      "eval_steps_per_second": 1.142,
+      "step": 2210
+    },
+    {
+      "epoch": 2.423013698630137,
+      "grad_norm": 2.752720594406128,
+      "learning_rate": 1.0987002437043055e-06,
+      "loss": 1.7432,
+      "step": 2211
+    },
+    {
+      "epoch": 2.424109589041096,
+      "grad_norm": 2.6162850856781006,
+      "learning_rate": 1.0966693744922827e-06,
+      "loss": 1.718,
+      "step": 2212
+    },
+    {
+      "epoch": 2.425205479452055,
+      "grad_norm": 2.938995838165283,
+      "learning_rate": 1.09463850528026e-06,
+      "loss": 1.6761,
+      "step": 2213
+    },
+    {
+      "epoch": 2.4263013698630136,
+      "grad_norm": 2.7038822174072266,
+      "learning_rate": 1.0926076360682373e-06,
+      "loss": 1.7267,
+      "step": 2214
+    },
+    {
+      "epoch": 2.4273972602739726,
+      "grad_norm": 2.6025028228759766,
+      "learning_rate": 1.0905767668562146e-06,
+      "loss": 1.7021,
+      "step": 2215
+    },
+    {
+      "epoch": 2.4284931506849317,
+      "grad_norm": 2.5922977924346924,
+      "learning_rate": 1.0885458976441918e-06,
+      "loss": 1.7221,
+      "step": 2216
+    },
+    {
+      "epoch": 2.4295890410958902,
+      "grad_norm": 2.634716033935547,
+      "learning_rate": 1.0865150284321691e-06,
+      "loss": 1.8047,
+      "step": 2217
+    },
+    {
+      "epoch": 2.4306849315068493,
+      "grad_norm": 2.655155658721924,
+      "learning_rate": 1.0844841592201464e-06,
+      "loss": 1.7426,
+      "step": 2218
+    },
+    {
+      "epoch": 2.4317808219178083,
+      "grad_norm": 2.5151286125183105,
+      "learning_rate": 1.0824532900081237e-06,
+      "loss": 1.7627,
+      "step": 2219
+    },
+    {
+      "epoch": 2.432876712328767,
+      "grad_norm": 2.696597099304199,
+      "learning_rate": 1.080422420796101e-06,
+      "loss": 1.7325,
+      "step": 2220
+    },
+    {
+      "epoch": 2.432876712328767,
+      "eval_loss": 1.9818732738494873,
+      "eval_runtime": 36.9854,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 2220
+    },
+    {
+      "epoch": 2.433972602739726,
+      "grad_norm": 2.7444307804107666,
+      "learning_rate": 1.078391551584078e-06,
+      "loss": 1.743,
+      "step": 2221
+    },
+    {
+      "epoch": 2.435068493150685,
+      "grad_norm": 2.503706455230713,
+      "learning_rate": 1.0763606823720553e-06,
+      "loss": 1.7699,
+      "step": 2222
+    },
+    {
+      "epoch": 2.436164383561644,
+      "grad_norm": 2.546041965484619,
+      "learning_rate": 1.0743298131600326e-06,
+      "loss": 1.8077,
+      "step": 2223
+    },
+    {
+      "epoch": 2.4372602739726026,
+      "grad_norm": 2.5515291690826416,
+      "learning_rate": 1.0722989439480098e-06,
+      "loss": 1.7637,
+      "step": 2224
+    },
+    {
+      "epoch": 2.4383561643835616,
+      "grad_norm": 2.564180850982666,
+      "learning_rate": 1.070268074735987e-06,
+      "loss": 1.7299,
+      "step": 2225
+    },
+    {
+      "epoch": 2.4394520547945207,
+      "grad_norm": 2.532708168029785,
+      "learning_rate": 1.0682372055239644e-06,
+      "loss": 1.713,
+      "step": 2226
+    },
+    {
+      "epoch": 2.4405479452054797,
+      "grad_norm": 2.672358512878418,
+      "learning_rate": 1.0662063363119414e-06,
+      "loss": 1.7103,
+      "step": 2227
+    },
+    {
+      "epoch": 2.4416438356164383,
+      "grad_norm": 2.7836480140686035,
+      "learning_rate": 1.0641754670999187e-06,
+      "loss": 1.7129,
+      "step": 2228
+    },
+    {
+      "epoch": 2.4427397260273973,
+      "grad_norm": 2.6253440380096436,
+      "learning_rate": 1.062144597887896e-06,
+      "loss": 1.7389,
+      "step": 2229
+    },
+    {
+      "epoch": 2.4438356164383563,
+      "grad_norm": 2.5863685607910156,
+      "learning_rate": 1.0601137286758733e-06,
+      "loss": 1.7293,
+      "step": 2230
+    },
+    {
+      "epoch": 2.4438356164383563,
+      "eval_loss": 1.9771078824996948,
+      "eval_runtime": 37.0603,
+      "eval_samples_per_second": 26.983,
+      "eval_steps_per_second": 1.133,
+      "step": 2230
+    },
+    {
+      "epoch": 2.444931506849315,
+      "grad_norm": 2.732042074203491,
+      "learning_rate": 1.0580828594638505e-06,
+      "loss": 1.8012,
+      "step": 2231
+    },
+    {
+      "epoch": 2.446027397260274,
+      "grad_norm": 2.5275564193725586,
+      "learning_rate": 1.0560519902518278e-06,
+      "loss": 1.6482,
+      "step": 2232
+    },
+    {
+      "epoch": 2.447123287671233,
+      "grad_norm": 2.7054405212402344,
+      "learning_rate": 1.054021121039805e-06,
+      "loss": 1.7799,
+      "step": 2233
+    },
+    {
+      "epoch": 2.4482191780821916,
+      "grad_norm": 2.587719678878784,
+      "learning_rate": 1.0519902518277824e-06,
+      "loss": 1.6797,
+      "step": 2234
+    },
+    {
+      "epoch": 2.4493150684931506,
+      "grad_norm": 2.600907564163208,
+      "learning_rate": 1.0499593826157596e-06,
+      "loss": 1.6727,
+      "step": 2235
+    },
+    {
+      "epoch": 2.4504109589041096,
+      "grad_norm": 2.6509485244750977,
+      "learning_rate": 1.047928513403737e-06,
+      "loss": 1.7535,
+      "step": 2236
+    },
+    {
+      "epoch": 2.4515068493150687,
+      "grad_norm": 2.6827805042266846,
+      "learning_rate": 1.0458976441917142e-06,
+      "loss": 1.7285,
+      "step": 2237
+    },
+    {
+      "epoch": 2.4526027397260273,
+      "grad_norm": 2.7095749378204346,
+      "learning_rate": 1.0438667749796915e-06,
+      "loss": 1.6546,
+      "step": 2238
+    },
+    {
+      "epoch": 2.4536986301369863,
+      "grad_norm": 2.7757186889648438,
+      "learning_rate": 1.0418359057676688e-06,
+      "loss": 1.6659,
+      "step": 2239
+    },
+    {
+      "epoch": 2.4547945205479453,
+      "grad_norm": 2.7696826457977295,
+      "learning_rate": 1.0398050365556458e-06,
+      "loss": 1.7108,
+      "step": 2240
+    },
+    {
+      "epoch": 2.4547945205479453,
+      "eval_loss": 1.980337142944336,
+      "eval_runtime": 37.0883,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 2240
+    },
+    {
+      "epoch": 2.455890410958904,
+      "grad_norm": 2.96648907661438,
+      "learning_rate": 1.037774167343623e-06,
+      "loss": 1.77,
+      "step": 2241
+    },
+    {
+      "epoch": 2.456986301369863,
+      "grad_norm": 2.561461925506592,
+      "learning_rate": 1.0357432981316004e-06,
+      "loss": 1.6962,
+      "step": 2242
+    },
+    {
+      "epoch": 2.458082191780822,
+      "grad_norm": 2.5870988368988037,
+      "learning_rate": 1.0337124289195776e-06,
+      "loss": 1.7385,
+      "step": 2243
+    },
+    {
+      "epoch": 2.459178082191781,
+      "grad_norm": 2.607558488845825,
+      "learning_rate": 1.031681559707555e-06,
+      "loss": 1.7771,
+      "step": 2244
+    },
+    {
+      "epoch": 2.4602739726027396,
+      "grad_norm": 2.6373603343963623,
+      "learning_rate": 1.0296506904955322e-06,
+      "loss": 1.7845,
+      "step": 2245
+    },
+    {
+      "epoch": 2.4613698630136986,
+      "grad_norm": 2.5508599281311035,
+      "learning_rate": 1.0276198212835095e-06,
+      "loss": 1.7384,
+      "step": 2246
+    },
+    {
+      "epoch": 2.4624657534246577,
+      "grad_norm": 2.730696201324463,
+      "learning_rate": 1.0255889520714867e-06,
+      "loss": 1.7458,
+      "step": 2247
+    },
+    {
+      "epoch": 2.4635616438356163,
+      "grad_norm": 2.4905753135681152,
+      "learning_rate": 1.023558082859464e-06,
+      "loss": 1.7195,
+      "step": 2248
+    },
+    {
+      "epoch": 2.4646575342465753,
+      "grad_norm": 2.664184331893921,
+      "learning_rate": 1.0215272136474413e-06,
+      "loss": 1.6986,
+      "step": 2249
+    },
+    {
+      "epoch": 2.4657534246575343,
+      "grad_norm": 2.683645009994507,
+      "learning_rate": 1.0194963444354186e-06,
+      "loss": 1.7981,
+      "step": 2250
+    },
+    {
+      "epoch": 2.4657534246575343,
+      "eval_loss": 1.9768283367156982,
+      "eval_runtime": 37.1086,
+      "eval_samples_per_second": 26.948,
+      "eval_steps_per_second": 1.132,
+      "step": 2250
+    },
+    {
+      "epoch": 2.466849315068493,
+      "grad_norm": 2.633633852005005,
+      "learning_rate": 1.0174654752233958e-06,
+      "loss": 1.6807,
+      "step": 2251
+    },
+    {
+      "epoch": 2.467945205479452,
+      "grad_norm": 2.655794382095337,
+      "learning_rate": 1.015434606011373e-06,
+      "loss": 1.7705,
+      "step": 2252
+    },
+    {
+      "epoch": 2.469041095890411,
+      "grad_norm": 2.6457812786102295,
+      "learning_rate": 1.0134037367993502e-06,
+      "loss": 1.7201,
+      "step": 2253
+    },
+    {
+      "epoch": 2.47013698630137,
+      "grad_norm": 2.623656749725342,
+      "learning_rate": 1.0113728675873275e-06,
+      "loss": 1.7402,
+      "step": 2254
+    },
+    {
+      "epoch": 2.4712328767123286,
+      "grad_norm": 2.65552020072937,
+      "learning_rate": 1.0093419983753047e-06,
+      "loss": 1.7372,
+      "step": 2255
+    },
+    {
+      "epoch": 2.4723287671232876,
+      "grad_norm": 2.707531213760376,
+      "learning_rate": 1.007311129163282e-06,
+      "loss": 1.7299,
+      "step": 2256
+    },
+    {
+      "epoch": 2.4734246575342467,
+      "grad_norm": 2.504805564880371,
+      "learning_rate": 1.0052802599512593e-06,
+      "loss": 1.6858,
+      "step": 2257
+    },
+    {
+      "epoch": 2.4745205479452057,
+      "grad_norm": 2.584251642227173,
+      "learning_rate": 1.0032493907392364e-06,
+      "loss": 1.6867,
+      "step": 2258
+    },
+    {
+      "epoch": 2.4756164383561643,
+      "grad_norm": 2.5356390476226807,
+      "learning_rate": 1.0012185215272136e-06,
+      "loss": 1.7442,
+      "step": 2259
+    },
+    {
+      "epoch": 2.4767123287671233,
+      "grad_norm": 2.5929319858551025,
+      "learning_rate": 9.99187652315191e-07,
+      "loss": 1.6777,
+      "step": 2260
+    },
+    {
+      "epoch": 2.4767123287671233,
+      "eval_loss": 1.9772632122039795,
+      "eval_runtime": 37.0951,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 2260
+    },
+    {
+      "epoch": 2.4778082191780824,
+      "grad_norm": 2.6396026611328125,
+      "learning_rate": 9.971567831031682e-07,
+      "loss": 1.7638,
+      "step": 2261
+    },
+    {
+      "epoch": 2.478904109589041,
+      "grad_norm": 2.5710740089416504,
+      "learning_rate": 9.951259138911455e-07,
+      "loss": 1.7566,
+      "step": 2262
+    },
+    {
+      "epoch": 2.48,
+      "grad_norm": 2.5611767768859863,
+      "learning_rate": 9.930950446791227e-07,
+      "loss": 1.6639,
+      "step": 2263
+    },
+    {
+      "epoch": 2.481095890410959,
+      "grad_norm": 2.5390372276306152,
+      "learning_rate": 9.910641754671e-07,
+      "loss": 1.7887,
+      "step": 2264
+    },
+    {
+      "epoch": 2.4821917808219176,
+      "grad_norm": 2.6124560832977295,
+      "learning_rate": 9.890333062550773e-07,
+      "loss": 1.8221,
+      "step": 2265
+    },
+    {
+      "epoch": 2.4832876712328766,
+      "grad_norm": 2.6506896018981934,
+      "learning_rate": 9.870024370430546e-07,
+      "loss": 1.725,
+      "step": 2266
+    },
+    {
+      "epoch": 2.4843835616438357,
+      "grad_norm": 2.617889642715454,
+      "learning_rate": 9.849715678310318e-07,
+      "loss": 1.7566,
+      "step": 2267
+    },
+    {
+      "epoch": 2.4854794520547947,
+      "grad_norm": 2.571110486984253,
+      "learning_rate": 9.829406986190091e-07,
+      "loss": 1.7327,
+      "step": 2268
+    },
+    {
+      "epoch": 2.4865753424657533,
+      "grad_norm": 2.657665252685547,
+      "learning_rate": 9.809098294069864e-07,
+      "loss": 1.7471,
+      "step": 2269
+    },
+    {
+      "epoch": 2.4876712328767123,
+      "grad_norm": 2.5402324199676514,
+      "learning_rate": 9.788789601949637e-07,
+      "loss": 1.6597,
+      "step": 2270
+    },
+    {
+      "epoch": 2.4876712328767123,
+      "eval_loss": 1.975923776626587,
+      "eval_runtime": 37.0909,
+      "eval_samples_per_second": 26.961,
+      "eval_steps_per_second": 1.132,
+      "step": 2270
+    },
+    {
+      "epoch": 2.4887671232876714,
+      "grad_norm": 2.619765281677246,
+      "learning_rate": 9.768480909829407e-07,
+      "loss": 1.7434,
+      "step": 2271
+    },
+    {
+      "epoch": 2.48986301369863,
+      "grad_norm": 2.664238214492798,
+      "learning_rate": 9.74817221770918e-07,
+      "loss": 1.7937,
+      "step": 2272
+    },
+    {
+      "epoch": 2.490958904109589,
+      "grad_norm": 2.6743860244750977,
+      "learning_rate": 9.727863525588953e-07,
+      "loss": 1.7203,
+      "step": 2273
+    },
+    {
+      "epoch": 2.492054794520548,
+      "grad_norm": 2.695051670074463,
+      "learning_rate": 9.707554833468726e-07,
+      "loss": 1.7473,
+      "step": 2274
+    },
+    {
+      "epoch": 2.493150684931507,
+      "grad_norm": 2.6115036010742188,
+      "learning_rate": 9.687246141348498e-07,
+      "loss": 1.7358,
+      "step": 2275
+    },
+    {
+      "epoch": 2.4942465753424656,
+      "grad_norm": 2.7951300144195557,
+      "learning_rate": 9.666937449228271e-07,
+      "loss": 1.723,
+      "step": 2276
+    },
+    {
+      "epoch": 2.4953424657534247,
+      "grad_norm": 2.6242685317993164,
+      "learning_rate": 9.646628757108042e-07,
+      "loss": 1.7389,
+      "step": 2277
+    },
+    {
+      "epoch": 2.4964383561643837,
+      "grad_norm": 2.567150831222534,
+      "learning_rate": 9.626320064987814e-07,
+      "loss": 1.6775,
+      "step": 2278
+    },
+    {
+      "epoch": 2.4975342465753423,
+      "grad_norm": 2.6157355308532715,
+      "learning_rate": 9.606011372867587e-07,
+      "loss": 1.7624,
+      "step": 2279
+    },
+    {
+      "epoch": 2.4986301369863013,
+      "grad_norm": 2.5663843154907227,
+      "learning_rate": 9.58570268074736e-07,
+      "loss": 1.6955,
+      "step": 2280
+    },
+    {
+      "epoch": 2.4986301369863013,
+      "eval_loss": 1.9744558334350586,
+      "eval_runtime": 37.1094,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 2280
+    },
+    {
+      "epoch": 2.4997260273972604,
+      "grad_norm": 2.667611598968506,
+      "learning_rate": 9.565393988627133e-07,
+      "loss": 1.7261,
+      "step": 2281
+    },
+    {
+      "epoch": 2.500821917808219,
+      "grad_norm": 2.5702202320098877,
+      "learning_rate": 9.545085296506906e-07,
+      "loss": 1.7252,
+      "step": 2282
+    },
+    {
+      "epoch": 2.501917808219178,
+      "grad_norm": 2.5271430015563965,
+      "learning_rate": 9.524776604386677e-07,
+      "loss": 1.7407,
+      "step": 2283
+    },
+    {
+      "epoch": 2.503013698630137,
+      "grad_norm": 2.5804171562194824,
+      "learning_rate": 9.50446791226645e-07,
+      "loss": 1.739,
+      "step": 2284
+    },
+    {
+      "epoch": 2.504109589041096,
+      "grad_norm": 2.571345090866089,
+      "learning_rate": 9.484159220146223e-07,
+      "loss": 1.668,
+      "step": 2285
+    },
+    {
+      "epoch": 2.5052054794520546,
+      "grad_norm": 2.5697009563446045,
+      "learning_rate": 9.463850528025995e-07,
+      "loss": 1.7677,
+      "step": 2286
+    },
+    {
+      "epoch": 2.5063013698630137,
+      "grad_norm": 2.5846893787384033,
+      "learning_rate": 9.443541835905768e-07,
+      "loss": 1.7589,
+      "step": 2287
+    },
+    {
+      "epoch": 2.5073972602739727,
+      "grad_norm": 2.803380250930786,
+      "learning_rate": 9.423233143785541e-07,
+      "loss": 1.8104,
+      "step": 2288
+    },
+    {
+      "epoch": 2.5084931506849317,
+      "grad_norm": 2.6451611518859863,
+      "learning_rate": 9.402924451665313e-07,
+      "loss": 1.8061,
+      "step": 2289
+    },
+    {
+      "epoch": 2.5095890410958903,
+      "grad_norm": 2.6164097785949707,
+      "learning_rate": 9.382615759545085e-07,
+      "loss": 1.6899,
+      "step": 2290
+    },
+    {
+      "epoch": 2.5095890410958903,
+      "eval_loss": 1.9739093780517578,
+      "eval_runtime": 37.1222,
+      "eval_samples_per_second": 26.938,
+      "eval_steps_per_second": 1.131,
+      "step": 2290
+    },
+    {
+      "epoch": 2.5106849315068493,
+      "grad_norm": 2.5874385833740234,
+      "learning_rate": 9.362307067424858e-07,
+      "loss": 1.7692,
+      "step": 2291
+    },
+    {
+      "epoch": 2.5117808219178084,
+      "grad_norm": 2.5705063343048096,
+      "learning_rate": 9.341998375304631e-07,
+      "loss": 1.7303,
+      "step": 2292
+    },
+    {
+      "epoch": 2.512876712328767,
+      "grad_norm": 2.6130380630493164,
+      "learning_rate": 9.321689683184404e-07,
+      "loss": 1.7892,
+      "step": 2293
+    },
+    {
+      "epoch": 2.513972602739726,
+      "grad_norm": 2.5991110801696777,
+      "learning_rate": 9.301380991064176e-07,
+      "loss": 1.6877,
+      "step": 2294
+    },
+    {
+      "epoch": 2.515068493150685,
+      "grad_norm": 2.6092689037323,
+      "learning_rate": 9.281072298943949e-07,
+      "loss": 1.6729,
+      "step": 2295
+    },
+    {
+      "epoch": 2.5161643835616436,
+      "grad_norm": 2.5563883781433105,
+      "learning_rate": 9.260763606823721e-07,
+      "loss": 1.7181,
+      "step": 2296
+    },
+    {
+      "epoch": 2.5172602739726027,
+      "grad_norm": 2.623948574066162,
+      "learning_rate": 9.240454914703494e-07,
+      "loss": 1.7528,
+      "step": 2297
+    },
+    {
+      "epoch": 2.5183561643835617,
+      "grad_norm": 2.661874294281006,
+      "learning_rate": 9.220146222583266e-07,
+      "loss": 1.7582,
+      "step": 2298
+    },
+    {
+      "epoch": 2.5194520547945203,
+      "grad_norm": 2.5936238765716553,
+      "learning_rate": 9.199837530463039e-07,
+      "loss": 1.6987,
+      "step": 2299
+    },
+    {
+      "epoch": 2.5205479452054793,
+      "grad_norm": 2.5134992599487305,
+      "learning_rate": 9.179528838342812e-07,
+      "loss": 1.7265,
+      "step": 2300
+    },
+    {
+      "epoch": 2.5205479452054793,
+      "eval_loss": 1.9737082719802856,
+      "eval_runtime": 37.1105,
+      "eval_samples_per_second": 26.947,
+      "eval_steps_per_second": 1.132,
+      "step": 2300
+    },
+    {
+      "epoch": 2.5216438356164383,
+      "grad_norm": 2.627927541732788,
+      "learning_rate": 9.159220146222585e-07,
+      "loss": 1.7059,
+      "step": 2301
+    },
+    {
+      "epoch": 2.5227397260273974,
+      "grad_norm": 2.6419436931610107,
+      "learning_rate": 9.138911454102356e-07,
+      "loss": 1.7558,
+      "step": 2302
+    },
+    {
+      "epoch": 2.5238356164383564,
+      "grad_norm": 2.6184327602386475,
+      "learning_rate": 9.118602761982129e-07,
+      "loss": 1.6281,
+      "step": 2303
+    },
+    {
+      "epoch": 2.524931506849315,
+      "grad_norm": 2.6111767292022705,
+      "learning_rate": 9.098294069861902e-07,
+      "loss": 1.7408,
+      "step": 2304
+    },
+    {
+      "epoch": 2.526027397260274,
+      "grad_norm": 2.6416513919830322,
+      "learning_rate": 9.077985377741675e-07,
+      "loss": 1.7397,
+      "step": 2305
+    },
+    {
+      "epoch": 2.527123287671233,
+      "grad_norm": 2.5487682819366455,
+      "learning_rate": 9.057676685621447e-07,
+      "loss": 1.6941,
+      "step": 2306
+    },
+    {
+      "epoch": 2.5282191780821917,
+      "grad_norm": 2.6250264644622803,
+      "learning_rate": 9.03736799350122e-07,
+      "loss": 1.66,
+      "step": 2307
+    },
+    {
+      "epoch": 2.5293150684931507,
+      "grad_norm": 2.6054790019989014,
+      "learning_rate": 9.017059301380991e-07,
+      "loss": 1.7829,
+      "step": 2308
+    },
+    {
+      "epoch": 2.5304109589041097,
+      "grad_norm": 2.751016616821289,
+      "learning_rate": 8.996750609260764e-07,
+      "loss": 1.7693,
+      "step": 2309
+    },
+    {
+      "epoch": 2.5315068493150683,
+      "grad_norm": 2.6321675777435303,
+      "learning_rate": 8.976441917140536e-07,
+      "loss": 1.7971,
+      "step": 2310
+    },
+    {
+      "epoch": 2.5315068493150683,
+      "eval_loss": 1.9757944345474243,
+      "eval_runtime": 36.77,
+      "eval_samples_per_second": 27.196,
+      "eval_steps_per_second": 1.142,
+      "step": 2310
+    },
+    {
+      "epoch": 2.5326027397260273,
+      "grad_norm": 2.642702341079712,
+      "learning_rate": 8.956133225020309e-07,
+      "loss": 1.7703,
+      "step": 2311
+    },
+    {
+      "epoch": 2.5336986301369864,
+      "grad_norm": 2.614771604537964,
+      "learning_rate": 8.935824532900082e-07,
+      "loss": 1.8009,
+      "step": 2312
+    },
+    {
+      "epoch": 2.534794520547945,
+      "grad_norm": 2.514615058898926,
+      "learning_rate": 8.915515840779855e-07,
+      "loss": 1.6954,
+      "step": 2313
+    },
+    {
+      "epoch": 2.535890410958904,
+      "grad_norm": 2.644207715988159,
+      "learning_rate": 8.895207148659626e-07,
+      "loss": 1.6779,
+      "step": 2314
+    },
+    {
+      "epoch": 2.536986301369863,
+      "grad_norm": 2.7844860553741455,
+      "learning_rate": 8.874898456539399e-07,
+      "loss": 1.7595,
+      "step": 2315
+    },
+    {
+      "epoch": 2.538082191780822,
+      "grad_norm": 2.714629650115967,
+      "learning_rate": 8.854589764419172e-07,
+      "loss": 1.7448,
+      "step": 2316
+    },
+    {
+      "epoch": 2.539178082191781,
+      "grad_norm": 2.57271409034729,
+      "learning_rate": 8.834281072298945e-07,
+      "loss": 1.7438,
+      "step": 2317
+    },
+    {
+      "epoch": 2.5402739726027397,
+      "grad_norm": 2.6566264629364014,
+      "learning_rate": 8.813972380178717e-07,
+      "loss": 1.6924,
+      "step": 2318
+    },
+    {
+      "epoch": 2.5413698630136987,
+      "grad_norm": 2.619744300842285,
+      "learning_rate": 8.79366368805849e-07,
+      "loss": 1.7317,
+      "step": 2319
+    },
+    {
+      "epoch": 2.5424657534246577,
+      "grad_norm": 2.6249327659606934,
+      "learning_rate": 8.773354995938262e-07,
+      "loss": 1.7467,
+      "step": 2320
+    },
+    {
+      "epoch": 2.5424657534246577,
+      "eval_loss": 1.9757587909698486,
+      "eval_runtime": 36.9112,
+      "eval_samples_per_second": 27.092,
+      "eval_steps_per_second": 1.138,
+      "step": 2320
+    },
+    {
+      "epoch": 2.5435616438356163,
+      "grad_norm": 2.584317445755005,
+      "learning_rate": 8.753046303818035e-07,
+      "loss": 1.6962,
+      "step": 2321
+    },
+    {
+      "epoch": 2.5446575342465754,
+      "grad_norm": 2.7346856594085693,
+      "learning_rate": 8.732737611697807e-07,
+      "loss": 1.6863,
+      "step": 2322
+    },
+    {
+      "epoch": 2.5457534246575344,
+      "grad_norm": 2.574941635131836,
+      "learning_rate": 8.71242891957758e-07,
+      "loss": 1.7271,
+      "step": 2323
+    },
+    {
+      "epoch": 2.546849315068493,
+      "grad_norm": 2.546217441558838,
+      "learning_rate": 8.692120227457353e-07,
+      "loss": 1.751,
+      "step": 2324
+    },
+    {
+      "epoch": 2.547945205479452,
+      "grad_norm": 2.534414768218994,
+      "learning_rate": 8.671811535337126e-07,
+      "loss": 1.7692,
+      "step": 2325
+    },
+    {
+      "epoch": 2.549041095890411,
+      "grad_norm": 2.625565528869629,
+      "learning_rate": 8.651502843216898e-07,
+      "loss": 1.7135,
+      "step": 2326
+    },
+    {
+      "epoch": 2.5501369863013696,
+      "grad_norm": 2.715276002883911,
+      "learning_rate": 8.63119415109667e-07,
+      "loss": 1.8093,
+      "step": 2327
+    },
+    {
+      "epoch": 2.5512328767123287,
+      "grad_norm": 2.6494171619415283,
+      "learning_rate": 8.610885458976443e-07,
+      "loss": 1.7589,
+      "step": 2328
+    },
+    {
+      "epoch": 2.5523287671232877,
+      "grad_norm": 2.6840505599975586,
+      "learning_rate": 8.590576766856216e-07,
+      "loss": 1.7484,
+      "step": 2329
+    },
+    {
+      "epoch": 2.5534246575342463,
+      "grad_norm": 2.510047435760498,
+      "learning_rate": 8.570268074735988e-07,
+      "loss": 1.6985,
+      "step": 2330
+    },
+    {
+      "epoch": 2.5534246575342463,
+      "eval_loss": 1.9751781225204468,
+      "eval_runtime": 36.9998,
+      "eval_samples_per_second": 27.027,
+      "eval_steps_per_second": 1.135,
+      "step": 2330
+    },
+    {
+      "epoch": 2.5545205479452053,
+      "grad_norm": 2.6174721717834473,
+      "learning_rate": 8.549959382615761e-07,
+      "loss": 1.6873,
+      "step": 2331
+    },
+    {
+      "epoch": 2.5556164383561644,
+      "grad_norm": 2.570403814315796,
+      "learning_rate": 8.529650690495534e-07,
+      "loss": 1.7058,
+      "step": 2332
+    },
+    {
+      "epoch": 2.5567123287671234,
+      "grad_norm": 2.8619518280029297,
+      "learning_rate": 8.509341998375305e-07,
+      "loss": 1.7342,
+      "step": 2333
+    },
+    {
+      "epoch": 2.5578082191780824,
+      "grad_norm": 2.6569507122039795,
+      "learning_rate": 8.489033306255077e-07,
+      "loss": 1.6736,
+      "step": 2334
+    },
+    {
+      "epoch": 2.558904109589041,
+      "grad_norm": 2.5437800884246826,
+      "learning_rate": 8.46872461413485e-07,
+      "loss": 1.648,
+      "step": 2335
+    },
+    {
+      "epoch": 2.56,
+      "grad_norm": 2.652275323867798,
+      "learning_rate": 8.448415922014623e-07,
+      "loss": 1.692,
+      "step": 2336
+    },
+    {
+      "epoch": 2.561095890410959,
+      "grad_norm": 2.6103672981262207,
+      "learning_rate": 8.428107229894396e-07,
+      "loss": 1.7033,
+      "step": 2337
+    },
+    {
+      "epoch": 2.5621917808219177,
+      "grad_norm": 2.6765308380126953,
+      "learning_rate": 8.407798537774168e-07,
+      "loss": 1.7473,
+      "step": 2338
+    },
+    {
+      "epoch": 2.5632876712328767,
+      "grad_norm": 2.708935260772705,
+      "learning_rate": 8.38748984565394e-07,
+      "loss": 1.7346,
+      "step": 2339
+    },
+    {
+      "epoch": 2.5643835616438357,
+      "grad_norm": 2.597233772277832,
+      "learning_rate": 8.367181153533713e-07,
+      "loss": 1.6927,
+      "step": 2340
+    },
+    {
+      "epoch": 2.5643835616438357,
+      "eval_loss": 1.9768768548965454,
+      "eval_runtime": 37.0191,
+      "eval_samples_per_second": 27.013,
+      "eval_steps_per_second": 1.135,
+      "step": 2340
+    },
+    {
+      "epoch": 2.5654794520547943,
+      "grad_norm": 2.624343156814575,
+      "learning_rate": 8.346872461413486e-07,
+      "loss": 1.7219,
+      "step": 2341
+    },
+    {
+      "epoch": 2.5665753424657534,
+      "grad_norm": 2.675095796585083,
+      "learning_rate": 8.326563769293258e-07,
+      "loss": 1.7136,
+      "step": 2342
+    },
+    {
+      "epoch": 2.5676712328767124,
+      "grad_norm": 2.575009822845459,
+      "learning_rate": 8.306255077173031e-07,
+      "loss": 1.7187,
+      "step": 2343
+    },
+    {
+      "epoch": 2.568767123287671,
+      "grad_norm": 2.539958953857422,
+      "learning_rate": 8.285946385052804e-07,
+      "loss": 1.8232,
+      "step": 2344
+    },
+    {
+      "epoch": 2.56986301369863,
+      "grad_norm": 2.613976001739502,
+      "learning_rate": 8.265637692932575e-07,
+      "loss": 1.6746,
+      "step": 2345
+    },
+    {
+      "epoch": 2.570958904109589,
+      "grad_norm": 2.5001742839813232,
+      "learning_rate": 8.245329000812348e-07,
+      "loss": 1.7542,
+      "step": 2346
+    },
+    {
+      "epoch": 2.572054794520548,
+      "grad_norm": 2.6975460052490234,
+      "learning_rate": 8.225020308692121e-07,
+      "loss": 1.7434,
+      "step": 2347
+    },
+    {
+      "epoch": 2.573150684931507,
+      "grad_norm": 2.6900603771209717,
+      "learning_rate": 8.204711616571894e-07,
+      "loss": 1.6843,
+      "step": 2348
+    },
+    {
+      "epoch": 2.5742465753424657,
+      "grad_norm": 2.5895581245422363,
+      "learning_rate": 8.184402924451667e-07,
+      "loss": 1.7279,
+      "step": 2349
+    },
+    {
+      "epoch": 2.5753424657534247,
+      "grad_norm": 2.5448262691497803,
+      "learning_rate": 8.164094232331439e-07,
+      "loss": 1.6743,
+      "step": 2350
+    },
+    {
+      "epoch": 2.5753424657534247,
+      "eval_loss": 1.9762015342712402,
+      "eval_runtime": 37.0186,
+      "eval_samples_per_second": 27.013,
+      "eval_steps_per_second": 1.135,
+      "step": 2350
+    },
+    {
+      "epoch": 2.5764383561643838,
+      "grad_norm": 2.6080751419067383,
+      "learning_rate": 8.14378554021121e-07,
+      "loss": 1.7525,
+      "step": 2351
+    },
+    {
+      "epoch": 2.5775342465753424,
+      "grad_norm": 2.6307425498962402,
+      "learning_rate": 8.123476848090983e-07,
+      "loss": 1.7304,
+      "step": 2352
+    },
+    {
+      "epoch": 2.5786301369863014,
+      "grad_norm": 2.7442290782928467,
+      "learning_rate": 8.103168155970755e-07,
+      "loss": 1.7018,
+      "step": 2353
+    },
+    {
+      "epoch": 2.5797260273972604,
+      "grad_norm": 2.5672149658203125,
+      "learning_rate": 8.082859463850528e-07,
+      "loss": 1.7115,
+      "step": 2354
+    },
+    {
+      "epoch": 2.580821917808219,
+      "grad_norm": 2.7201836109161377,
+      "learning_rate": 8.062550771730301e-07,
+      "loss": 1.7461,
+      "step": 2355
+    },
+    {
+      "epoch": 2.581917808219178,
+      "grad_norm": 2.7145047187805176,
+      "learning_rate": 8.042242079610074e-07,
+      "loss": 1.7169,
+      "step": 2356
+    },
+    {
+      "epoch": 2.583013698630137,
+      "grad_norm": 2.7162413597106934,
+      "learning_rate": 8.021933387489846e-07,
+      "loss": 1.71,
+      "step": 2357
+    },
+    {
+      "epoch": 2.5841095890410957,
+      "grad_norm": 2.515169858932495,
+      "learning_rate": 8.001624695369618e-07,
+      "loss": 1.7016,
+      "step": 2358
+    },
+    {
+      "epoch": 2.5852054794520547,
+      "grad_norm": 2.620922803878784,
+      "learning_rate": 7.981316003249391e-07,
+      "loss": 1.7251,
+      "step": 2359
+    },
+    {
+      "epoch": 2.5863013698630137,
+      "grad_norm": 2.8066515922546387,
+      "learning_rate": 7.961007311129164e-07,
+      "loss": 1.787,
+      "step": 2360
+    },
+    {
+      "epoch": 2.5863013698630137,
+      "eval_loss": 1.9778543710708618,
+      "eval_runtime": 37.0094,
+      "eval_samples_per_second": 27.02,
+      "eval_steps_per_second": 1.135,
+      "step": 2360
+    },
+    {
+      "epoch": 2.5873972602739728,
+      "grad_norm": 2.5266482830047607,
+      "learning_rate": 7.940698619008936e-07,
+      "loss": 1.6654,
+      "step": 2361
+    },
+    {
+      "epoch": 2.5884931506849314,
+      "grad_norm": 2.6038002967834473,
+      "learning_rate": 7.920389926888709e-07,
+      "loss": 1.7532,
+      "step": 2362
+    },
+    {
+      "epoch": 2.5895890410958904,
+      "grad_norm": 2.716871976852417,
+      "learning_rate": 7.900081234768482e-07,
+      "loss": 1.8235,
+      "step": 2363
+    },
+    {
+      "epoch": 2.5906849315068494,
+      "grad_norm": 2.6962411403656006,
+      "learning_rate": 7.879772542648254e-07,
+      "loss": 1.6927,
+      "step": 2364
+    },
+    {
+      "epoch": 2.5917808219178085,
+      "grad_norm": 2.670116662979126,
+      "learning_rate": 7.859463850528026e-07,
+      "loss": 1.7448,
+      "step": 2365
+    },
+    {
+      "epoch": 2.592876712328767,
+      "grad_norm": 2.6886184215545654,
+      "learning_rate": 7.839155158407799e-07,
+      "loss": 1.7369,
+      "step": 2366
+    },
+    {
+      "epoch": 2.593972602739726,
+      "grad_norm": 2.6570565700531006,
+      "learning_rate": 7.818846466287572e-07,
+      "loss": 1.7368,
+      "step": 2367
+    },
+    {
+      "epoch": 2.595068493150685,
+      "grad_norm": 2.551548719406128,
+      "learning_rate": 7.798537774167345e-07,
+      "loss": 1.7357,
+      "step": 2368
+    },
+    {
+      "epoch": 2.5961643835616437,
+      "grad_norm": 2.646280527114868,
+      "learning_rate": 7.778229082047117e-07,
+      "loss": 1.7614,
+      "step": 2369
+    },
+    {
+      "epoch": 2.5972602739726027,
+      "grad_norm": 2.626239776611328,
+      "learning_rate": 7.757920389926889e-07,
+      "loss": 1.694,
+      "step": 2370
+    },
+    {
+      "epoch": 2.5972602739726027,
+      "eval_loss": 1.9757529497146606,
+      "eval_runtime": 36.9858,
+      "eval_samples_per_second": 27.037,
+      "eval_steps_per_second": 1.136,
+      "step": 2370
+    },
+    {
+      "epoch": 2.5983561643835618,
+      "grad_norm": 2.6001148223876953,
+      "learning_rate": 7.737611697806662e-07,
+      "loss": 1.7342,
+      "step": 2371
+    },
+    {
+      "epoch": 2.5994520547945204,
+      "grad_norm": 2.5435168743133545,
+      "learning_rate": 7.717303005686435e-07,
+      "loss": 1.7031,
+      "step": 2372
+    },
+    {
+      "epoch": 2.6005479452054794,
+      "grad_norm": 2.652031183242798,
+      "learning_rate": 7.696994313566207e-07,
+      "loss": 1.7598,
+      "step": 2373
+    },
+    {
+      "epoch": 2.6016438356164384,
+      "grad_norm": 2.635340690612793,
+      "learning_rate": 7.67668562144598e-07,
+      "loss": 1.8008,
+      "step": 2374
+    },
+    {
+      "epoch": 2.602739726027397,
+      "grad_norm": 2.6935770511627197,
+      "learning_rate": 7.656376929325753e-07,
+      "loss": 1.6821,
+      "step": 2375
+    },
+    {
+      "epoch": 2.603835616438356,
+      "grad_norm": 2.621493339538574,
+      "learning_rate": 7.636068237205524e-07,
+      "loss": 1.6844,
+      "step": 2376
+    },
+    {
+      "epoch": 2.604931506849315,
+      "grad_norm": 2.6231918334960938,
+      "learning_rate": 7.615759545085296e-07,
+      "loss": 1.7622,
+      "step": 2377
+    },
+    {
+      "epoch": 2.606027397260274,
+      "grad_norm": 2.608091115951538,
+      "learning_rate": 7.595450852965069e-07,
+      "loss": 1.6772,
+      "step": 2378
+    },
+    {
+      "epoch": 2.607123287671233,
+      "grad_norm": 2.7701456546783447,
+      "learning_rate": 7.575142160844842e-07,
+      "loss": 1.7139,
+      "step": 2379
+    },
+    {
+      "epoch": 2.6082191780821917,
+      "grad_norm": 2.6648287773132324,
+      "learning_rate": 7.554833468724615e-07,
+      "loss": 1.7525,
+      "step": 2380
+    },
+    {
+      "epoch": 2.6082191780821917,
+      "eval_loss": 1.9792429208755493,
+      "eval_runtime": 37.0103,
+      "eval_samples_per_second": 27.02,
+      "eval_steps_per_second": 1.135,
+      "step": 2380
+    },
+    {
+      "epoch": 2.6093150684931508,
+      "grad_norm": 2.690938949584961,
+      "learning_rate": 7.534524776604387e-07,
+      "loss": 1.679,
+      "step": 2381
+    },
+    {
+      "epoch": 2.61041095890411,
+      "grad_norm": 2.658691167831421,
+      "learning_rate": 7.51421608448416e-07,
+      "loss": 1.7033,
+      "step": 2382
+    },
+    {
+      "epoch": 2.6115068493150684,
+      "grad_norm": 2.576317310333252,
+      "learning_rate": 7.493907392363932e-07,
+      "loss": 1.7527,
+      "step": 2383
+    },
+    {
+      "epoch": 2.6126027397260274,
+      "grad_norm": 2.5913753509521484,
+      "learning_rate": 7.473598700243705e-07,
+      "loss": 1.702,
+      "step": 2384
+    },
+    {
+      "epoch": 2.6136986301369864,
+      "grad_norm": 2.61553692817688,
+      "learning_rate": 7.453290008123477e-07,
+      "loss": 1.6568,
+      "step": 2385
+    },
+    {
+      "epoch": 2.614794520547945,
+      "grad_norm": 2.6631391048431396,
+      "learning_rate": 7.43298131600325e-07,
+      "loss": 1.7216,
+      "step": 2386
+    },
+    {
+      "epoch": 2.615890410958904,
+      "grad_norm": 2.6120619773864746,
+      "learning_rate": 7.412672623883023e-07,
+      "loss": 1.7278,
+      "step": 2387
+    },
+    {
+      "epoch": 2.616986301369863,
+      "grad_norm": 2.6166980266571045,
+      "learning_rate": 7.392363931762796e-07,
+      "loss": 1.7275,
+      "step": 2388
+    },
+    {
+      "epoch": 2.6180821917808217,
+      "grad_norm": 2.7203712463378906,
+      "learning_rate": 7.372055239642567e-07,
+      "loss": 1.6657,
+      "step": 2389
+    },
+    {
+      "epoch": 2.6191780821917807,
+      "grad_norm": 2.6785244941711426,
+      "learning_rate": 7.35174654752234e-07,
+      "loss": 1.7206,
+      "step": 2390
+    },
+    {
+      "epoch": 2.6191780821917807,
+      "eval_loss": 1.9763447046279907,
+      "eval_runtime": 37.0352,
+      "eval_samples_per_second": 27.001,
+      "eval_steps_per_second": 1.134,
+      "step": 2390
+    },
+    {
+      "epoch": 2.6202739726027398,
+      "grad_norm": 2.72511887550354,
+      "learning_rate": 7.331437855402113e-07,
+      "loss": 1.7474,
+      "step": 2391
+    },
+    {
+      "epoch": 2.621369863013699,
+      "grad_norm": 2.6428632736206055,
+      "learning_rate": 7.311129163281886e-07,
+      "loss": 1.7694,
+      "step": 2392
+    },
+    {
+      "epoch": 2.6224657534246574,
+      "grad_norm": 2.6593737602233887,
+      "learning_rate": 7.290820471161658e-07,
+      "loss": 1.6883,
+      "step": 2393
+    },
+    {
+      "epoch": 2.6235616438356164,
+      "grad_norm": 2.6048357486724854,
+      "learning_rate": 7.270511779041431e-07,
+      "loss": 1.7457,
+      "step": 2394
+    },
+    {
+      "epoch": 2.6246575342465754,
+      "grad_norm": 2.6435704231262207,
+      "learning_rate": 7.250203086921203e-07,
+      "loss": 1.7508,
+      "step": 2395
+    },
+    {
+      "epoch": 2.6257534246575345,
+      "grad_norm": 2.609815835952759,
+      "learning_rate": 7.229894394800976e-07,
+      "loss": 1.6603,
+      "step": 2396
+    },
+    {
+      "epoch": 2.626849315068493,
+      "grad_norm": 2.5431087017059326,
+      "learning_rate": 7.209585702680748e-07,
+      "loss": 1.725,
+      "step": 2397
+    },
+    {
+      "epoch": 2.627945205479452,
+      "grad_norm": 2.6712050437927246,
+      "learning_rate": 7.189277010560521e-07,
+      "loss": 1.7428,
+      "step": 2398
+    },
+    {
+      "epoch": 2.629041095890411,
+      "grad_norm": 2.6398820877075195,
+      "learning_rate": 7.168968318440294e-07,
+      "loss": 1.6834,
+      "step": 2399
+    },
+    {
+      "epoch": 2.6301369863013697,
+      "grad_norm": 2.5986127853393555,
+      "learning_rate": 7.148659626320067e-07,
+      "loss": 1.6985,
+      "step": 2400
+    },
+    {
+      "epoch": 2.6301369863013697,
+      "eval_loss": 1.9727513790130615,
+      "eval_runtime": 37.0552,
+      "eval_samples_per_second": 26.987,
+      "eval_steps_per_second": 1.133,
+      "step": 2400
+    },
+    {
+      "epoch": 2.6312328767123287,
+      "grad_norm": 2.6422410011291504,
+      "learning_rate": 7.128350934199837e-07,
+      "loss": 1.6732,
+      "step": 2401
+    },
+    {
+      "epoch": 2.632328767123288,
+      "grad_norm": 2.59894061088562,
+      "learning_rate": 7.10804224207961e-07,
+      "loss": 1.7855,
+      "step": 2402
+    },
+    {
+      "epoch": 2.6334246575342464,
+      "grad_norm": 2.6503429412841797,
+      "learning_rate": 7.087733549959383e-07,
+      "loss": 1.7768,
+      "step": 2403
+    },
+    {
+      "epoch": 2.6345205479452054,
+      "grad_norm": 2.5734922885894775,
+      "learning_rate": 7.067424857839155e-07,
+      "loss": 1.7221,
+      "step": 2404
+    },
+    {
+      "epoch": 2.6356164383561644,
+      "grad_norm": 2.5473709106445312,
+      "learning_rate": 7.047116165718928e-07,
+      "loss": 1.7116,
+      "step": 2405
+    },
+    {
+      "epoch": 2.636712328767123,
+      "grad_norm": 2.655651330947876,
+      "learning_rate": 7.026807473598701e-07,
+      "loss": 1.7085,
+      "step": 2406
+    },
+    {
+      "epoch": 2.637808219178082,
+      "grad_norm": 2.615900993347168,
+      "learning_rate": 7.006498781478473e-07,
+      "loss": 1.7189,
+      "step": 2407
+    },
+    {
+      "epoch": 2.638904109589041,
+      "grad_norm": 2.5976428985595703,
+      "learning_rate": 6.986190089358245e-07,
+      "loss": 1.7902,
+      "step": 2408
+    },
+    {
+      "epoch": 2.64,
+      "grad_norm": 2.6213088035583496,
+      "learning_rate": 6.965881397238018e-07,
+      "loss": 1.6842,
+      "step": 2409
+    },
+    {
+      "epoch": 2.641095890410959,
+      "grad_norm": 2.476685047149658,
+      "learning_rate": 6.945572705117791e-07,
+      "loss": 1.7297,
+      "step": 2410
+    },
+    {
+      "epoch": 2.641095890410959,
+      "eval_loss": 1.9702810049057007,
+      "eval_runtime": 36.7953,
+      "eval_samples_per_second": 27.177,
+      "eval_steps_per_second": 1.141,
+      "step": 2410
+    },
+    {
+      "epoch": 2.6421917808219177,
+      "grad_norm": 2.625504732131958,
+      "learning_rate": 6.925264012997564e-07,
+      "loss": 1.7335,
+      "step": 2411
+    },
+    {
+      "epoch": 2.643287671232877,
+      "grad_norm": 2.849334716796875,
+      "learning_rate": 6.904955320877336e-07,
+      "loss": 1.7339,
+      "step": 2412
+    },
+    {
+      "epoch": 2.644383561643836,
+      "grad_norm": 2.7533011436462402,
+      "learning_rate": 6.884646628757109e-07,
+      "loss": 1.7431,
+      "step": 2413
+    },
+    {
+      "epoch": 2.6454794520547944,
+      "grad_norm": 2.6793112754821777,
+      "learning_rate": 6.864337936636881e-07,
+      "loss": 1.7474,
+      "step": 2414
+    },
+    {
+      "epoch": 2.6465753424657534,
+      "grad_norm": 2.68094539642334,
+      "learning_rate": 6.844029244516654e-07,
+      "loss": 1.8092,
+      "step": 2415
+    },
+    {
+      "epoch": 2.6476712328767125,
+      "grad_norm": 2.7346346378326416,
+      "learning_rate": 6.823720552396426e-07,
+      "loss": 1.7838,
+      "step": 2416
+    },
+    {
+      "epoch": 2.648767123287671,
+      "grad_norm": 2.6850287914276123,
+      "learning_rate": 6.803411860276199e-07,
+      "loss": 1.6927,
+      "step": 2417
+    },
+    {
+      "epoch": 2.64986301369863,
+      "grad_norm": 2.5687966346740723,
+      "learning_rate": 6.783103168155972e-07,
+      "loss": 1.7505,
+      "step": 2418
+    },
+    {
+      "epoch": 2.650958904109589,
+      "grad_norm": 2.515760660171509,
+      "learning_rate": 6.762794476035745e-07,
+      "loss": 1.7285,
+      "step": 2419
+    },
+    {
+      "epoch": 2.6520547945205477,
+      "grad_norm": 2.604621171951294,
+      "learning_rate": 6.742485783915515e-07,
+      "loss": 1.7148,
+      "step": 2420
+    },
+    {
+      "epoch": 2.6520547945205477,
+      "eval_loss": 1.9678161144256592,
+      "eval_runtime": 36.9829,
+      "eval_samples_per_second": 27.04,
+      "eval_steps_per_second": 1.136,
+      "step": 2420
+    },
+    {
+      "epoch": 2.6531506849315067,
+      "grad_norm": 2.5340449810028076,
+      "learning_rate": 6.722177091795288e-07,
+      "loss": 1.756,
+      "step": 2421
+    },
+    {
+      "epoch": 2.6542465753424658,
+      "grad_norm": 2.5800235271453857,
+      "learning_rate": 6.701868399675061e-07,
+      "loss": 1.7439,
+      "step": 2422
+    },
+    {
+      "epoch": 2.655342465753425,
+      "grad_norm": 2.6596486568450928,
+      "learning_rate": 6.681559707554834e-07,
+      "loss": 1.7351,
+      "step": 2423
+    },
+    {
+      "epoch": 2.6564383561643834,
+      "grad_norm": 2.55759334564209,
+      "learning_rate": 6.661251015434606e-07,
+      "loss": 1.6978,
+      "step": 2424
+    },
+    {
+      "epoch": 2.6575342465753424,
+      "grad_norm": 2.6761016845703125,
+      "learning_rate": 6.640942323314379e-07,
+      "loss": 1.7011,
+      "step": 2425
+    },
+    {
+      "epoch": 2.6586301369863015,
+      "grad_norm": 2.5663347244262695,
+      "learning_rate": 6.620633631194151e-07,
+      "loss": 1.6891,
+      "step": 2426
+    },
+    {
+      "epoch": 2.6597260273972605,
+      "grad_norm": 2.64680552482605,
+      "learning_rate": 6.600324939073924e-07,
+      "loss": 1.703,
+      "step": 2427
+    },
+    {
+      "epoch": 2.660821917808219,
+      "grad_norm": 2.730487585067749,
+      "learning_rate": 6.580016246953696e-07,
+      "loss": 1.7329,
+      "step": 2428
+    },
+    {
+      "epoch": 2.661917808219178,
+      "grad_norm": 2.6785171031951904,
+      "learning_rate": 6.559707554833469e-07,
+      "loss": 1.7126,
+      "step": 2429
+    },
+    {
+      "epoch": 2.663013698630137,
+      "grad_norm": 2.657883882522583,
+      "learning_rate": 6.539398862713242e-07,
+      "loss": 1.7663,
+      "step": 2430
+    },
+    {
+      "epoch": 2.663013698630137,
+      "eval_loss": 1.968948245048523,
+      "eval_runtime": 37.0622,
+      "eval_samples_per_second": 26.982,
+      "eval_steps_per_second": 1.133,
+      "step": 2430
+    },
+    {
+      "epoch": 2.6641095890410957,
+      "grad_norm": 2.5575273036956787,
+      "learning_rate": 6.519090170593015e-07,
+      "loss": 1.7067,
+      "step": 2431
+    },
+    {
+      "epoch": 2.6652054794520548,
+      "grad_norm": 2.664853811264038,
+      "learning_rate": 6.498781478472786e-07,
+      "loss": 1.7674,
+      "step": 2432
+    },
+    {
+      "epoch": 2.666301369863014,
+      "grad_norm": 2.571324348449707,
+      "learning_rate": 6.478472786352559e-07,
+      "loss": 1.7328,
+      "step": 2433
+    },
+    {
+      "epoch": 2.6673972602739724,
+      "grad_norm": 2.7143940925598145,
+      "learning_rate": 6.458164094232332e-07,
+      "loss": 1.7736,
+      "step": 2434
+    },
+    {
+      "epoch": 2.6684931506849314,
+      "grad_norm": 2.6819112300872803,
+      "learning_rate": 6.437855402112105e-07,
+      "loss": 1.7266,
+      "step": 2435
+    },
+    {
+      "epoch": 2.6695890410958905,
+      "grad_norm": 2.701815128326416,
+      "learning_rate": 6.417546709991877e-07,
+      "loss": 1.6662,
+      "step": 2436
+    },
+    {
+      "epoch": 2.670684931506849,
+      "grad_norm": 2.6434922218322754,
+      "learning_rate": 6.39723801787165e-07,
+      "loss": 1.6896,
+      "step": 2437
+    },
+    {
+      "epoch": 2.671780821917808,
+      "grad_norm": 2.738567590713501,
+      "learning_rate": 6.376929325751423e-07,
+      "loss": 1.7794,
+      "step": 2438
+    },
+    {
+      "epoch": 2.672876712328767,
+      "grad_norm": 2.6100597381591797,
+      "learning_rate": 6.356620633631195e-07,
+      "loss": 1.7159,
+      "step": 2439
+    },
+    {
+      "epoch": 2.673972602739726,
+      "grad_norm": 2.6096298694610596,
+      "learning_rate": 6.336311941510967e-07,
+      "loss": 1.7106,
+      "step": 2440
+    },
+    {
+      "epoch": 2.673972602739726,
+      "eval_loss": 1.967889428138733,
+      "eval_runtime": 37.0763,
+      "eval_samples_per_second": 26.971,
+      "eval_steps_per_second": 1.133,
+      "step": 2440
+    },
+    {
+      "epoch": 2.675068493150685,
+      "grad_norm": 2.671086072921753,
+      "learning_rate": 6.31600324939074e-07,
+      "loss": 1.7106,
+      "step": 2441
+    },
+    {
+      "epoch": 2.6761643835616438,
+      "grad_norm": 2.7339978218078613,
+      "learning_rate": 6.295694557270513e-07,
+      "loss": 1.734,
+      "step": 2442
+    },
+    {
+      "epoch": 2.677260273972603,
+      "grad_norm": 2.626713752746582,
+      "learning_rate": 6.275385865150286e-07,
+      "loss": 1.7677,
+      "step": 2443
+    },
+    {
+      "epoch": 2.678356164383562,
+      "grad_norm": 2.542891502380371,
+      "learning_rate": 6.255077173030058e-07,
+      "loss": 1.7415,
+      "step": 2444
+    },
+    {
+      "epoch": 2.6794520547945204,
+      "grad_norm": 2.657703161239624,
+      "learning_rate": 6.23476848090983e-07,
+      "loss": 1.724,
+      "step": 2445
+    },
+    {
+      "epoch": 2.6805479452054795,
+      "grad_norm": 2.6566591262817383,
+      "learning_rate": 6.214459788789602e-07,
+      "loss": 1.7319,
+      "step": 2446
+    },
+    {
+      "epoch": 2.6816438356164385,
+      "grad_norm": 2.684117078781128,
+      "learning_rate": 6.194151096669375e-07,
+      "loss": 1.7148,
+      "step": 2447
+    },
+    {
+      "epoch": 2.682739726027397,
+      "grad_norm": 2.6989500522613525,
+      "learning_rate": 6.173842404549147e-07,
+      "loss": 1.7515,
+      "step": 2448
+    },
+    {
+      "epoch": 2.683835616438356,
+      "grad_norm": 2.731233835220337,
+      "learning_rate": 6.15353371242892e-07,
+      "loss": 1.7195,
+      "step": 2449
+    },
+    {
+      "epoch": 2.684931506849315,
+      "grad_norm": 2.6298322677612305,
+      "learning_rate": 6.133225020308693e-07,
+      "loss": 1.7311,
+      "step": 2450
+    },
+    {
+      "epoch": 2.684931506849315,
+      "eval_loss": 1.9676637649536133,
+      "eval_runtime": 37.0926,
+      "eval_samples_per_second": 26.96,
+      "eval_steps_per_second": 1.132,
+      "step": 2450
+    },
+    {
+      "epoch": 2.6860273972602737,
+      "grad_norm": 2.57230806350708,
+      "learning_rate": 6.112916328188466e-07,
+      "loss": 1.6388,
+      "step": 2451
+    },
+    {
+      "epoch": 2.6871232876712328,
+      "grad_norm": 2.5881690979003906,
+      "learning_rate": 6.092607636068238e-07,
+      "loss": 1.7108,
+      "step": 2452
+    },
+    {
+      "epoch": 2.688219178082192,
+      "grad_norm": 2.732182264328003,
+      "learning_rate": 6.07229894394801e-07,
+      "loss": 1.7814,
+      "step": 2453
+    },
+    {
+      "epoch": 2.689315068493151,
+      "grad_norm": 2.6008358001708984,
+      "learning_rate": 6.051990251827783e-07,
+      "loss": 1.6969,
+      "step": 2454
+    },
+    {
+      "epoch": 2.69041095890411,
+      "grad_norm": 2.638119697570801,
+      "learning_rate": 6.031681559707556e-07,
+      "loss": 1.6973,
+      "step": 2455
+    },
+    {
+      "epoch": 2.6915068493150685,
+      "grad_norm": 2.5765764713287354,
+      "learning_rate": 6.011372867587327e-07,
+      "loss": 1.7385,
+      "step": 2456
+    },
+    {
+      "epoch": 2.6926027397260275,
+      "grad_norm": 2.6819193363189697,
+      "learning_rate": 5.9910641754671e-07,
+      "loss": 1.7311,
+      "step": 2457
+    },
+    {
+      "epoch": 2.6936986301369865,
+      "grad_norm": 2.562638521194458,
+      "learning_rate": 5.970755483346873e-07,
+      "loss": 1.6935,
+      "step": 2458
+    },
+    {
+      "epoch": 2.694794520547945,
+      "grad_norm": 2.6333518028259277,
+      "learning_rate": 5.950446791226646e-07,
+      "loss": 1.7083,
+      "step": 2459
+    },
+    {
+      "epoch": 2.695890410958904,
+      "grad_norm": 2.6318745613098145,
+      "learning_rate": 5.930138099106418e-07,
+      "loss": 1.7082,
+      "step": 2460
+    },
+    {
+      "epoch": 2.695890410958904,
+      "eval_loss": 1.9667798280715942,
+      "eval_runtime": 37.0761,
+      "eval_samples_per_second": 26.972,
+      "eval_steps_per_second": 1.133,
+      "step": 2460
+    },
+    {
+      "epoch": 2.696986301369863,
+      "grad_norm": 2.7109830379486084,
+      "learning_rate": 5.909829406986191e-07,
+      "loss": 1.7251,
+      "step": 2461
+    },
+    {
+      "epoch": 2.6980821917808218,
+      "grad_norm": 2.5945379734039307,
+      "learning_rate": 5.889520714865963e-07,
+      "loss": 1.7057,
+      "step": 2462
+    },
+    {
+      "epoch": 2.699178082191781,
+      "grad_norm": 2.658888816833496,
+      "learning_rate": 5.869212022745735e-07,
+      "loss": 1.7358,
+      "step": 2463
+    },
+    {
+      "epoch": 2.70027397260274,
+      "grad_norm": 2.6142160892486572,
+      "learning_rate": 5.848903330625508e-07,
+      "loss": 1.6777,
+      "step": 2464
+    },
+    {
+      "epoch": 2.7013698630136984,
+      "grad_norm": 2.6961355209350586,
+      "learning_rate": 5.828594638505281e-07,
+      "loss": 1.7342,
+      "step": 2465
+    },
+    {
+      "epoch": 2.7024657534246574,
+      "grad_norm": 2.527519464492798,
+      "learning_rate": 5.808285946385054e-07,
+      "loss": 1.7343,
+      "step": 2466
+    },
+    {
+      "epoch": 2.7035616438356165,
+      "grad_norm": 2.661269187927246,
+      "learning_rate": 5.787977254264827e-07,
+      "loss": 1.7084,
+      "step": 2467
+    },
+    {
+      "epoch": 2.704657534246575,
+      "grad_norm": 2.648587703704834,
+      "learning_rate": 5.767668562144598e-07,
+      "loss": 1.7417,
+      "step": 2468
+    },
+    {
+      "epoch": 2.705753424657534,
+      "grad_norm": 2.690467596054077,
+      "learning_rate": 5.747359870024371e-07,
+      "loss": 1.7651,
+      "step": 2469
+    },
+    {
+      "epoch": 2.706849315068493,
+      "grad_norm": 2.6351239681243896,
+      "learning_rate": 5.727051177904144e-07,
+      "loss": 1.7156,
+      "step": 2470
+    },
+    {
+      "epoch": 2.706849315068493,
+      "eval_loss": 1.9673000574111938,
+      "eval_runtime": 37.0884,
+      "eval_samples_per_second": 26.963,
+      "eval_steps_per_second": 1.132,
+      "step": 2470
+    },
+    {
+      "epoch": 2.707945205479452,
+      "grad_norm": 2.674419403076172,
+      "learning_rate": 5.706742485783915e-07,
+      "loss": 1.7276,
+      "step": 2471
+    },
+    {
+      "epoch": 2.709041095890411,
+      "grad_norm": 2.626304864883423,
+      "learning_rate": 5.686433793663688e-07,
+      "loss": 1.7578,
+      "step": 2472
+    },
+    {
+      "epoch": 2.71013698630137,
+      "grad_norm": 2.6923720836639404,
+      "learning_rate": 5.666125101543461e-07,
+      "loss": 1.6909,
+      "step": 2473
+    },
+    {
+      "epoch": 2.711232876712329,
+      "grad_norm": 2.6496634483337402,
+      "learning_rate": 5.645816409423234e-07,
+      "loss": 1.6691,
+      "step": 2474
+    },
+    {
+      "epoch": 2.712328767123288,
+      "grad_norm": 2.640418529510498,
+      "learning_rate": 5.625507717303006e-07,
+      "loss": 1.7705,
+      "step": 2475
+    },
+    {
+      "epoch": 2.7134246575342464,
+      "grad_norm": 2.5379440784454346,
+      "learning_rate": 5.605199025182779e-07,
+      "loss": 1.7356,
+      "step": 2476
+    },
+    {
+      "epoch": 2.7145205479452055,
+      "grad_norm": 2.6212170124053955,
+      "learning_rate": 5.584890333062551e-07,
+      "loss": 1.7128,
+      "step": 2477
+    },
+    {
+      "epoch": 2.7156164383561645,
+      "grad_norm": 2.676152467727661,
+      "learning_rate": 5.564581640942324e-07,
+      "loss": 1.6933,
+      "step": 2478
+    },
+    {
+      "epoch": 2.716712328767123,
+      "grad_norm": 2.684361457824707,
+      "learning_rate": 5.544272948822096e-07,
+      "loss": 1.7651,
+      "step": 2479
+    },
+    {
+      "epoch": 2.717808219178082,
+      "grad_norm": 2.630239486694336,
+      "learning_rate": 5.523964256701868e-07,
+      "loss": 1.7054,
+      "step": 2480
+    },
+    {
+      "epoch": 2.717808219178082,
+      "eval_loss": 1.967214822769165,
+      "eval_runtime": 37.0987,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 2480
+    },
+    {
+      "epoch": 2.718904109589041,
+      "grad_norm": 2.5909178256988525,
+      "learning_rate": 5.503655564581641e-07,
+      "loss": 1.7152,
+      "step": 2481
+    },
+    {
+      "epoch": 2.7199999999999998,
+      "grad_norm": 2.6395678520202637,
+      "learning_rate": 5.483346872461414e-07,
+      "loss": 1.7383,
+      "step": 2482
+    },
+    {
+      "epoch": 2.721095890410959,
+      "grad_norm": 2.5965633392333984,
+      "learning_rate": 5.463038180341186e-07,
+      "loss": 1.7111,
+      "step": 2483
+    },
+    {
+      "epoch": 2.722191780821918,
+      "grad_norm": 2.5583207607269287,
+      "learning_rate": 5.442729488220959e-07,
+      "loss": 1.7126,
+      "step": 2484
+    },
+    {
+      "epoch": 2.723287671232877,
+      "grad_norm": 2.691272258758545,
+      "learning_rate": 5.422420796100732e-07,
+      "loss": 1.698,
+      "step": 2485
+    },
+    {
+      "epoch": 2.724383561643836,
+      "grad_norm": 2.5674681663513184,
+      "learning_rate": 5.402112103980505e-07,
+      "loss": 1.7397,
+      "step": 2486
+    },
+    {
+      "epoch": 2.7254794520547945,
+      "grad_norm": 2.626338243484497,
+      "learning_rate": 5.381803411860276e-07,
+      "loss": 1.7095,
+      "step": 2487
+    },
+    {
+      "epoch": 2.7265753424657535,
+      "grad_norm": 2.636225700378418,
+      "learning_rate": 5.361494719740049e-07,
+      "loss": 1.758,
+      "step": 2488
+    },
+    {
+      "epoch": 2.7276712328767125,
+      "grad_norm": 2.7304928302764893,
+      "learning_rate": 5.341186027619822e-07,
+      "loss": 1.7969,
+      "step": 2489
+    },
+    {
+      "epoch": 2.728767123287671,
+      "grad_norm": 2.660935640335083,
+      "learning_rate": 5.320877335499594e-07,
+      "loss": 1.6876,
+      "step": 2490
+    },
+    {
+      "epoch": 2.728767123287671,
+      "eval_loss": 1.9665592908859253,
+      "eval_runtime": 37.0848,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 2490
+    },
+    {
+      "epoch": 2.72986301369863,
+      "grad_norm": 2.623068332672119,
+      "learning_rate": 5.300568643379366e-07,
+      "loss": 1.699,
+      "step": 2491
+    },
+    {
+      "epoch": 2.730958904109589,
+      "grad_norm": 2.528489112854004,
+      "learning_rate": 5.280259951259139e-07,
+      "loss": 1.7789,
+      "step": 2492
+    },
+    {
+      "epoch": 2.732054794520548,
+      "grad_norm": 2.6441330909729004,
+      "learning_rate": 5.259951259138912e-07,
+      "loss": 1.7358,
+      "step": 2493
+    },
+    {
+      "epoch": 2.733150684931507,
+      "grad_norm": 2.6011250019073486,
+      "learning_rate": 5.239642567018685e-07,
+      "loss": 1.6946,
+      "step": 2494
+    },
+    {
+      "epoch": 2.734246575342466,
+      "grad_norm": 2.669063091278076,
+      "learning_rate": 5.219333874898457e-07,
+      "loss": 1.6922,
+      "step": 2495
+    },
+    {
+      "epoch": 2.7353424657534244,
+      "grad_norm": 2.7551729679107666,
+      "learning_rate": 5.199025182778229e-07,
+      "loss": 1.7191,
+      "step": 2496
+    },
+    {
+      "epoch": 2.7364383561643835,
+      "grad_norm": 2.730877637863159,
+      "learning_rate": 5.178716490658002e-07,
+      "loss": 1.7319,
+      "step": 2497
+    },
+    {
+      "epoch": 2.7375342465753425,
+      "grad_norm": 2.6159300804138184,
+      "learning_rate": 5.158407798537775e-07,
+      "loss": 1.6891,
+      "step": 2498
+    },
+    {
+      "epoch": 2.7386301369863015,
+      "grad_norm": 2.7436540126800537,
+      "learning_rate": 5.138099106417547e-07,
+      "loss": 1.651,
+      "step": 2499
+    },
+    {
+      "epoch": 2.73972602739726,
+      "grad_norm": 2.6271872520446777,
+      "learning_rate": 5.11779041429732e-07,
+      "loss": 1.6241,
+      "step": 2500
+    },
+    {
+      "epoch": 2.73972602739726,
+      "eval_loss": 1.9660766124725342,
+      "eval_runtime": 37.0963,
+      "eval_samples_per_second": 26.957,
+      "eval_steps_per_second": 1.132,
+      "step": 2500
+    },
+    {
+      "epoch": 2.740821917808219,
+      "grad_norm": 2.651845693588257,
+      "learning_rate": 5.097481722177093e-07,
+      "loss": 1.71,
+      "step": 2501
+    },
+    {
+      "epoch": 2.741917808219178,
+      "grad_norm": 2.63710880279541,
+      "learning_rate": 5.077173030056865e-07,
+      "loss": 1.6902,
+      "step": 2502
+    },
+    {
+      "epoch": 2.743013698630137,
+      "grad_norm": 2.6147260665893555,
+      "learning_rate": 5.056864337936637e-07,
+      "loss": 1.7641,
+      "step": 2503
+    },
+    {
+      "epoch": 2.744109589041096,
+      "grad_norm": 2.715195655822754,
+      "learning_rate": 5.03655564581641e-07,
+      "loss": 1.7655,
+      "step": 2504
+    },
+    {
+      "epoch": 2.745205479452055,
+      "grad_norm": 2.5544769763946533,
+      "learning_rate": 5.016246953696182e-07,
+      "loss": 1.7217,
+      "step": 2505
+    },
+    {
+      "epoch": 2.746301369863014,
+      "grad_norm": 2.703354835510254,
+      "learning_rate": 4.995938261575955e-07,
+      "loss": 1.8411,
+      "step": 2506
+    },
+    {
+      "epoch": 2.7473972602739725,
+      "grad_norm": 2.7403740882873535,
+      "learning_rate": 4.975629569455727e-07,
+      "loss": 1.7204,
+      "step": 2507
+    },
+    {
+      "epoch": 2.7484931506849315,
+      "grad_norm": 2.595554828643799,
+      "learning_rate": 4.9553208773355e-07,
+      "loss": 1.7404,
+      "step": 2508
+    },
+    {
+      "epoch": 2.7495890410958905,
+      "grad_norm": 2.5391628742218018,
+      "learning_rate": 4.935012185215273e-07,
+      "loss": 1.7964,
+      "step": 2509
+    },
+    {
+      "epoch": 2.750684931506849,
+      "grad_norm": 2.547213315963745,
+      "learning_rate": 4.914703493095046e-07,
+      "loss": 1.7033,
+      "step": 2510
+    },
+    {
+      "epoch": 2.750684931506849,
+      "eval_loss": 1.9653823375701904,
+      "eval_runtime": 36.7942,
+      "eval_samples_per_second": 27.178,
+      "eval_steps_per_second": 1.141,
+      "step": 2510
+    },
+    {
+      "epoch": 2.751780821917808,
+      "grad_norm": 2.6644885540008545,
+      "learning_rate": 4.894394800974818e-07,
+      "loss": 1.6756,
+      "step": 2511
+    },
+    {
+      "epoch": 2.752876712328767,
+      "grad_norm": 2.617520570755005,
+      "learning_rate": 4.87408610885459e-07,
+      "loss": 1.6903,
+      "step": 2512
+    },
+    {
+      "epoch": 2.7539726027397258,
+      "grad_norm": 2.7279889583587646,
+      "learning_rate": 4.853777416734363e-07,
+      "loss": 1.8086,
+      "step": 2513
+    },
+    {
+      "epoch": 2.755068493150685,
+      "grad_norm": 2.6223795413970947,
+      "learning_rate": 4.833468724614136e-07,
+      "loss": 1.758,
+      "step": 2514
+    },
+    {
+      "epoch": 2.756164383561644,
+      "grad_norm": 2.6278305053710938,
+      "learning_rate": 4.813160032493907e-07,
+      "loss": 1.7157,
+      "step": 2515
+    },
+    {
+      "epoch": 2.757260273972603,
+      "grad_norm": 2.575812339782715,
+      "learning_rate": 4.79285134037368e-07,
+      "loss": 1.6986,
+      "step": 2516
+    },
+    {
+      "epoch": 2.758356164383562,
+      "grad_norm": 2.67112398147583,
+      "learning_rate": 4.772542648253453e-07,
+      "loss": 1.6844,
+      "step": 2517
+    },
+    {
+      "epoch": 2.7594520547945205,
+      "grad_norm": 2.826032876968384,
+      "learning_rate": 4.752233956133225e-07,
+      "loss": 1.6937,
+      "step": 2518
+    },
+    {
+      "epoch": 2.7605479452054795,
+      "grad_norm": 2.6157286167144775,
+      "learning_rate": 4.731925264012998e-07,
+      "loss": 1.7048,
+      "step": 2519
+    },
+    {
+      "epoch": 2.7616438356164386,
+      "grad_norm": 2.7270114421844482,
+      "learning_rate": 4.7116165718927705e-07,
+      "loss": 1.7749,
+      "step": 2520
+    },
+    {
+      "epoch": 2.7616438356164386,
+      "eval_loss": 1.9666064977645874,
+      "eval_runtime": 36.9714,
+      "eval_samples_per_second": 27.048,
+      "eval_steps_per_second": 1.136,
+      "step": 2520
+    },
+    {
+      "epoch": 2.762739726027397,
+      "grad_norm": 2.5923123359680176,
+      "learning_rate": 4.6913078797725427e-07,
+      "loss": 1.73,
+      "step": 2521
+    },
+    {
+      "epoch": 2.763835616438356,
+      "grad_norm": 2.6962082386016846,
+      "learning_rate": 4.6709991876523155e-07,
+      "loss": 1.7124,
+      "step": 2522
+    },
+    {
+      "epoch": 2.764931506849315,
+      "grad_norm": 2.6244304180145264,
+      "learning_rate": 4.650690495532088e-07,
+      "loss": 1.7052,
+      "step": 2523
+    },
+    {
+      "epoch": 2.766027397260274,
+      "grad_norm": 2.58933162689209,
+      "learning_rate": 4.6303818034118605e-07,
+      "loss": 1.7873,
+      "step": 2524
+    },
+    {
+      "epoch": 2.767123287671233,
+      "grad_norm": 2.7308239936828613,
+      "learning_rate": 4.610073111291633e-07,
+      "loss": 1.7642,
+      "step": 2525
+    },
+    {
+      "epoch": 2.768219178082192,
+      "grad_norm": 2.666278600692749,
+      "learning_rate": 4.589764419171406e-07,
+      "loss": 1.6554,
+      "step": 2526
+    },
+    {
+      "epoch": 2.7693150684931505,
+      "grad_norm": 2.591625452041626,
+      "learning_rate": 4.569455727051178e-07,
+      "loss": 1.7196,
+      "step": 2527
+    },
+    {
+      "epoch": 2.7704109589041095,
+      "grad_norm": 2.6249170303344727,
+      "learning_rate": 4.549147034930951e-07,
+      "loss": 1.6805,
+      "step": 2528
+    },
+    {
+      "epoch": 2.7715068493150685,
+      "grad_norm": 2.6104607582092285,
+      "learning_rate": 4.5288383428107237e-07,
+      "loss": 1.7684,
+      "step": 2529
+    },
+    {
+      "epoch": 2.7726027397260276,
+      "grad_norm": 2.6448006629943848,
+      "learning_rate": 4.5085296506904954e-07,
+      "loss": 1.747,
+      "step": 2530
+    },
+    {
+      "epoch": 2.7726027397260276,
+      "eval_loss": 1.9661206007003784,
+      "eval_runtime": 37.0394,
+      "eval_samples_per_second": 26.998,
+      "eval_steps_per_second": 1.134,
+      "step": 2530
+    },
+    {
+      "epoch": 2.773698630136986,
+      "grad_norm": 2.6639955043792725,
+      "learning_rate": 4.488220958570268e-07,
+      "loss": 1.7146,
+      "step": 2531
+    },
+    {
+      "epoch": 2.774794520547945,
+      "grad_norm": 2.6226794719696045,
+      "learning_rate": 4.467912266450041e-07,
+      "loss": 1.6874,
+      "step": 2532
+    },
+    {
+      "epoch": 2.775890410958904,
+      "grad_norm": 2.6525795459747314,
+      "learning_rate": 4.447603574329813e-07,
+      "loss": 1.6824,
+      "step": 2533
+    },
+    {
+      "epoch": 2.7769863013698632,
+      "grad_norm": 2.596771240234375,
+      "learning_rate": 4.427294882209586e-07,
+      "loss": 1.753,
+      "step": 2534
+    },
+    {
+      "epoch": 2.778082191780822,
+      "grad_norm": 2.4983580112457275,
+      "learning_rate": 4.4069861900893587e-07,
+      "loss": 1.6772,
+      "step": 2535
+    },
+    {
+      "epoch": 2.779178082191781,
+      "grad_norm": 2.6461708545684814,
+      "learning_rate": 4.386677497969131e-07,
+      "loss": 1.7303,
+      "step": 2536
+    },
+    {
+      "epoch": 2.78027397260274,
+      "grad_norm": 2.532766103744507,
+      "learning_rate": 4.3663688058489037e-07,
+      "loss": 1.8144,
+      "step": 2537
+    },
+    {
+      "epoch": 2.7813698630136985,
+      "grad_norm": 2.6262009143829346,
+      "learning_rate": 4.3460601137286764e-07,
+      "loss": 1.7293,
+      "step": 2538
+    },
+    {
+      "epoch": 2.7824657534246575,
+      "grad_norm": 2.6355018615722656,
+      "learning_rate": 4.325751421608449e-07,
+      "loss": 1.7619,
+      "step": 2539
+    },
+    {
+      "epoch": 2.7835616438356166,
+      "grad_norm": 2.586331844329834,
+      "learning_rate": 4.3054427294882214e-07,
+      "loss": 1.7309,
+      "step": 2540
+    },
+    {
+      "epoch": 2.7835616438356166,
+      "eval_loss": 1.9656602144241333,
+      "eval_runtime": 37.0374,
+      "eval_samples_per_second": 27.0,
+      "eval_steps_per_second": 1.134,
+      "step": 2540
+    },
+    {
+      "epoch": 2.784657534246575,
+      "grad_norm": 2.564220666885376,
+      "learning_rate": 4.285134037367994e-07,
+      "loss": 1.7478,
+      "step": 2541
+    },
+    {
+      "epoch": 2.785753424657534,
+      "grad_norm": 2.5705366134643555,
+      "learning_rate": 4.264825345247767e-07,
+      "loss": 1.7131,
+      "step": 2542
+    },
+    {
+      "epoch": 2.786849315068493,
+      "grad_norm": 2.5889837741851807,
+      "learning_rate": 4.2445166531275386e-07,
+      "loss": 1.6439,
+      "step": 2543
+    },
+    {
+      "epoch": 2.787945205479452,
+      "grad_norm": 2.632688045501709,
+      "learning_rate": 4.2242079610073114e-07,
+      "loss": 1.6999,
+      "step": 2544
+    },
+    {
+      "epoch": 2.789041095890411,
+      "grad_norm": 2.690004587173462,
+      "learning_rate": 4.203899268887084e-07,
+      "loss": 1.6499,
+      "step": 2545
+    },
+    {
+      "epoch": 2.79013698630137,
+      "grad_norm": 2.723090171813965,
+      "learning_rate": 4.1835905767668564e-07,
+      "loss": 1.7666,
+      "step": 2546
+    },
+    {
+      "epoch": 2.791232876712329,
+      "grad_norm": 2.657174587249756,
+      "learning_rate": 4.163281884646629e-07,
+      "loss": 1.7831,
+      "step": 2547
+    },
+    {
+      "epoch": 2.792328767123288,
+      "grad_norm": 2.6086323261260986,
+      "learning_rate": 4.142973192526402e-07,
+      "loss": 1.7211,
+      "step": 2548
+    },
+    {
+      "epoch": 2.7934246575342465,
+      "grad_norm": 2.59486722946167,
+      "learning_rate": 4.122664500406174e-07,
+      "loss": 1.7582,
+      "step": 2549
+    },
+    {
+      "epoch": 2.7945205479452055,
+      "grad_norm": 2.6308753490448,
+      "learning_rate": 4.102355808285947e-07,
+      "loss": 1.7339,
+      "step": 2550
+    },
+    {
+      "epoch": 2.7945205479452055,
+      "eval_loss": 1.9652767181396484,
+      "eval_runtime": 37.0143,
+      "eval_samples_per_second": 27.017,
+      "eval_steps_per_second": 1.135,
+      "step": 2550
+    },
+    {
+      "epoch": 2.7956164383561646,
+      "grad_norm": 3.7471933364868164,
+      "learning_rate": 4.0820471161657196e-07,
+      "loss": 1.7125,
+      "step": 2551
+    },
+    {
+      "epoch": 2.796712328767123,
+      "grad_norm": 2.7045977115631104,
+      "learning_rate": 4.0617384240454913e-07,
+      "loss": 1.7468,
+      "step": 2552
+    },
+    {
+      "epoch": 2.797808219178082,
+      "grad_norm": 2.595890998840332,
+      "learning_rate": 4.041429731925264e-07,
+      "loss": 1.7574,
+      "step": 2553
+    },
+    {
+      "epoch": 2.7989041095890412,
+      "grad_norm": 2.6221795082092285,
+      "learning_rate": 4.021121039805037e-07,
+      "loss": 1.6704,
+      "step": 2554
+    },
+    {
+      "epoch": 2.8,
+      "grad_norm": 2.5565948486328125,
+      "learning_rate": 4.000812347684809e-07,
+      "loss": 1.6908,
+      "step": 2555
+    },
+    {
+      "epoch": 2.801095890410959,
+      "grad_norm": 2.6208608150482178,
+      "learning_rate": 3.980503655564582e-07,
+      "loss": 1.7007,
+      "step": 2556
+    },
+    {
+      "epoch": 2.802191780821918,
+      "grad_norm": 2.6080801486968994,
+      "learning_rate": 3.9601949634443546e-07,
+      "loss": 1.7264,
+      "step": 2557
+    },
+    {
+      "epoch": 2.8032876712328765,
+      "grad_norm": 2.5089752674102783,
+      "learning_rate": 3.939886271324127e-07,
+      "loss": 1.7196,
+      "step": 2558
+    },
+    {
+      "epoch": 2.8043835616438355,
+      "grad_norm": 2.6230075359344482,
+      "learning_rate": 3.9195775792038996e-07,
+      "loss": 1.6998,
+      "step": 2559
+    },
+    {
+      "epoch": 2.8054794520547945,
+      "grad_norm": 2.5645101070404053,
+      "learning_rate": 3.8992688870836723e-07,
+      "loss": 1.7544,
+      "step": 2560
+    },
+    {
+      "epoch": 2.8054794520547945,
+      "eval_loss": 1.9631915092468262,
+      "eval_runtime": 37.0012,
+      "eval_samples_per_second": 27.026,
+      "eval_steps_per_second": 1.135,
+      "step": 2560
+    },
+    {
+      "epoch": 2.8065753424657536,
+      "grad_norm": 2.5155181884765625,
+      "learning_rate": 3.8789601949634446e-07,
+      "loss": 1.7466,
+      "step": 2561
+    },
+    {
+      "epoch": 2.807671232876712,
+      "grad_norm": 2.6017398834228516,
+      "learning_rate": 3.8586515028432173e-07,
+      "loss": 1.7114,
+      "step": 2562
+    },
+    {
+      "epoch": 2.808767123287671,
+      "grad_norm": 2.6371278762817383,
+      "learning_rate": 3.83834281072299e-07,
+      "loss": 1.6367,
+      "step": 2563
+    },
+    {
+      "epoch": 2.8098630136986302,
+      "grad_norm": 2.644526481628418,
+      "learning_rate": 3.818034118602762e-07,
+      "loss": 1.7213,
+      "step": 2564
+    },
+    {
+      "epoch": 2.8109589041095893,
+      "grad_norm": 2.571929931640625,
+      "learning_rate": 3.7977254264825345e-07,
+      "loss": 1.6771,
+      "step": 2565
+    },
+    {
+      "epoch": 2.812054794520548,
+      "grad_norm": 2.7080020904541016,
+      "learning_rate": 3.7774167343623073e-07,
+      "loss": 1.7317,
+      "step": 2566
+    },
+    {
+      "epoch": 2.813150684931507,
+      "grad_norm": 2.6303861141204834,
+      "learning_rate": 3.75710804224208e-07,
+      "loss": 1.738,
+      "step": 2567
+    },
+    {
+      "epoch": 2.814246575342466,
+      "grad_norm": 2.5958845615386963,
+      "learning_rate": 3.7367993501218523e-07,
+      "loss": 1.7589,
+      "step": 2568
+    },
+    {
+      "epoch": 2.8153424657534245,
+      "grad_norm": 2.6587975025177,
+      "learning_rate": 3.716490658001625e-07,
+      "loss": 1.6922,
+      "step": 2569
+    },
+    {
+      "epoch": 2.8164383561643835,
+      "grad_norm": 2.678011178970337,
+      "learning_rate": 3.696181965881398e-07,
+      "loss": 1.7082,
+      "step": 2570
+    },
+    {
+      "epoch": 2.8164383561643835,
+      "eval_loss": 1.963266372680664,
+      "eval_runtime": 37.0097,
+      "eval_samples_per_second": 27.02,
+      "eval_steps_per_second": 1.135,
+      "step": 2570
+    },
+    {
+      "epoch": 2.8175342465753426,
+      "grad_norm": 2.630423069000244,
+      "learning_rate": 3.67587327376117e-07,
+      "loss": 1.6702,
+      "step": 2571
+    },
+    {
+      "epoch": 2.818630136986301,
+      "grad_norm": 2.6961770057678223,
+      "learning_rate": 3.655564581640943e-07,
+      "loss": 1.7208,
+      "step": 2572
+    },
+    {
+      "epoch": 2.81972602739726,
+      "grad_norm": 2.5735878944396973,
+      "learning_rate": 3.6352558895207155e-07,
+      "loss": 1.7199,
+      "step": 2573
+    },
+    {
+      "epoch": 2.8208219178082192,
+      "grad_norm": 2.754194974899292,
+      "learning_rate": 3.614947197400488e-07,
+      "loss": 1.6768,
+      "step": 2574
+    },
+    {
+      "epoch": 2.821917808219178,
+      "grad_norm": 2.5502395629882812,
+      "learning_rate": 3.5946385052802605e-07,
+      "loss": 1.7271,
+      "step": 2575
+    },
+    {
+      "epoch": 2.823013698630137,
+      "grad_norm": 2.6345295906066895,
+      "learning_rate": 3.5743298131600333e-07,
+      "loss": 1.7188,
+      "step": 2576
+    },
+    {
+      "epoch": 2.824109589041096,
+      "grad_norm": 2.6169989109039307,
+      "learning_rate": 3.554021121039805e-07,
+      "loss": 1.6977,
+      "step": 2577
+    },
+    {
+      "epoch": 2.825205479452055,
+      "grad_norm": 2.6612536907196045,
+      "learning_rate": 3.533712428919578e-07,
+      "loss": 1.7722,
+      "step": 2578
+    },
+    {
+      "epoch": 2.826301369863014,
+      "grad_norm": 2.58573579788208,
+      "learning_rate": 3.5134037367993505e-07,
+      "loss": 1.7119,
+      "step": 2579
+    },
+    {
+      "epoch": 2.8273972602739725,
+      "grad_norm": 2.648927688598633,
+      "learning_rate": 3.4930950446791227e-07,
+      "loss": 1.711,
+      "step": 2580
+    },
+    {
+      "epoch": 2.8273972602739725,
+      "eval_loss": 1.9636057615280151,
+      "eval_runtime": 37.0115,
+      "eval_samples_per_second": 27.019,
+      "eval_steps_per_second": 1.135,
+      "step": 2580
+    },
+    {
+      "epoch": 2.8284931506849316,
+      "grad_norm": 2.648447036743164,
+      "learning_rate": 3.4727863525588955e-07,
+      "loss": 1.7076,
+      "step": 2581
+    },
+    {
+      "epoch": 2.8295890410958906,
+      "grad_norm": 2.633486747741699,
+      "learning_rate": 3.452477660438668e-07,
+      "loss": 1.7107,
+      "step": 2582
+    },
+    {
+      "epoch": 2.830684931506849,
+      "grad_norm": 2.8554606437683105,
+      "learning_rate": 3.4321689683184405e-07,
+      "loss": 1.7511,
+      "step": 2583
+    },
+    {
+      "epoch": 2.831780821917808,
+      "grad_norm": 2.6638591289520264,
+      "learning_rate": 3.411860276198213e-07,
+      "loss": 1.7144,
+      "step": 2584
+    },
+    {
+      "epoch": 2.8328767123287673,
+      "grad_norm": 2.530543804168701,
+      "learning_rate": 3.391551584077986e-07,
+      "loss": 1.6839,
+      "step": 2585
+    },
+    {
+      "epoch": 2.833972602739726,
+      "grad_norm": 2.653428792953491,
+      "learning_rate": 3.3712428919577577e-07,
+      "loss": 1.7418,
+      "step": 2586
+    },
+    {
+      "epoch": 2.835068493150685,
+      "grad_norm": 2.74092960357666,
+      "learning_rate": 3.3509341998375304e-07,
+      "loss": 1.7314,
+      "step": 2587
+    },
+    {
+      "epoch": 2.836164383561644,
+      "grad_norm": 2.632279872894287,
+      "learning_rate": 3.330625507717303e-07,
+      "loss": 1.6719,
+      "step": 2588
+    },
+    {
+      "epoch": 2.8372602739726025,
+      "grad_norm": 2.7040152549743652,
+      "learning_rate": 3.3103168155970754e-07,
+      "loss": 1.7074,
+      "step": 2589
+    },
+    {
+      "epoch": 2.8383561643835615,
+      "grad_norm": 2.6040985584259033,
+      "learning_rate": 3.290008123476848e-07,
+      "loss": 1.782,
+      "step": 2590
+    },
+    {
+      "epoch": 2.8383561643835615,
+      "eval_loss": 1.9626216888427734,
+      "eval_runtime": 36.9933,
+      "eval_samples_per_second": 27.032,
+      "eval_steps_per_second": 1.135,
+      "step": 2590
+    },
+    {
+      "epoch": 2.8394520547945206,
+      "grad_norm": 2.7316091060638428,
+      "learning_rate": 3.269699431356621e-07,
+      "loss": 1.7994,
+      "step": 2591
+    },
+    {
+      "epoch": 2.8405479452054796,
+      "grad_norm": 2.657702922821045,
+      "learning_rate": 3.249390739236393e-07,
+      "loss": 1.6929,
+      "step": 2592
+    },
+    {
+      "epoch": 2.8416438356164386,
+      "grad_norm": 2.6281511783599854,
+      "learning_rate": 3.229082047116166e-07,
+      "loss": 1.6652,
+      "step": 2593
+    },
+    {
+      "epoch": 2.842739726027397,
+      "grad_norm": 2.5952417850494385,
+      "learning_rate": 3.2087733549959387e-07,
+      "loss": 1.7714,
+      "step": 2594
+    },
+    {
+      "epoch": 2.8438356164383563,
+      "grad_norm": 2.7457919120788574,
+      "learning_rate": 3.1884646628757114e-07,
+      "loss": 1.7546,
+      "step": 2595
+    },
+    {
+      "epoch": 2.8449315068493153,
+      "grad_norm": 2.580235481262207,
+      "learning_rate": 3.1681559707554837e-07,
+      "loss": 1.7424,
+      "step": 2596
+    },
+    {
+      "epoch": 2.846027397260274,
+      "grad_norm": 2.6735477447509766,
+      "learning_rate": 3.1478472786352564e-07,
+      "loss": 1.7333,
+      "step": 2597
+    },
+    {
+      "epoch": 2.847123287671233,
+      "grad_norm": 2.6106553077697754,
+      "learning_rate": 3.127538586515029e-07,
+      "loss": 1.6817,
+      "step": 2598
+    },
+    {
+      "epoch": 2.848219178082192,
+      "grad_norm": 2.6753358840942383,
+      "learning_rate": 3.107229894394801e-07,
+      "loss": 1.7129,
+      "step": 2599
+    },
+    {
+      "epoch": 2.8493150684931505,
+      "grad_norm": 2.6023521423339844,
+      "learning_rate": 3.0869212022745736e-07,
+      "loss": 1.7113,
+      "step": 2600
+    },
+    {
+      "epoch": 2.8493150684931505,
+      "eval_loss": 1.9635719060897827,
+      "eval_runtime": 37.0071,
+      "eval_samples_per_second": 27.022,
+      "eval_steps_per_second": 1.135,
+      "step": 2600
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2736,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 8.105004464340992e+17,
+  "train_batch_size": 3,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/training_args.bin b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..b7adbe5a2dbf27f05befec6cc65fbc93e90eefd0
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aaa882279512a8c278376c0c601206933f3c5662ddcd3c13c53bf32ae3680d14
+size 6011
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2600/zero_to_fp32.py
@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2736/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2736/model-00002-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..41fe9ea8a65aa239976bacae40255c0933205f2e
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-2736/model-00002-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9cd736f3a4674554357d2f18f5d2abdb8a9c2ae24f4ed293650f2f1cad8b289
+size 2247734920
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..fb0f9a53a507ffcdb60410deb71d60ef801bf350
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/config.json
@@ -0,0 +1,36 @@
+{
+  "_name_or_path": "meta-llama/Llama-3.2-3B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": 128001,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "torch_dtype": "float16",
+  "transformers_version": "4.45.1",
+  "use_cache": true,
+  "vocab_size": 128256
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/generation_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/generation_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9c389b841a9a29ddd904e02d1eb0e08dcce82ad9
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/generation_config.json
@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": 128001,
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.45.1"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/latest b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/latest
new file mode 100644
index 0000000000000000000000000000000000000000..6761b575fffac7f1984044dcb6446b3a51da04c8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/latest
@@ -0,0 +1 @@
+global_step300
\ No newline at end of file
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/model-00002-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..2c41830082cca19bbd15a6cfc05a8437f4045457
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/model-00002-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:65938a6fb92e91e37c2f860fed2c9bc3e001154475d9596166235d6436e0ce2f
+size 2247734920
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/model.safetensors.index.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/model.safetensors.index.json
new file mode 100644
index 0000000000000000000000000000000000000000..ed64de846d720b9a7859dc20575fea8e8ca51940
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/model.safetensors.index.json
@@ -0,0 +1,262 @@
+{
+  "metadata": {
+    "total_size": 7213504512
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00002-of-00002.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00002-of-00002.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00002.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00002.safetensors",
+    "model.norm.weight": "model-00002-of-00002.safetensors"
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/rng_state_0.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/rng_state_0.pth
new file mode 100644
index 0000000000000000000000000000000000000000..9015a2b291768ac1cbe0fb417354be7e82849ff8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/rng_state_0.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c47a2aec04adfb2b774bed1d02a5453693feb66e4636bdf7cd908de748732842
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/rng_state_1.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/rng_state_1.pth
new file mode 100644
index 0000000000000000000000000000000000000000..a17e6802aa777eb35577f53d24a38a0381fecbb7
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/rng_state_1.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e375d2646839f0fc8836323cd9f6c6fec1f1316a3f21b7366755c68ab3c469c
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/rng_state_2.pth b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/rng_state_2.pth
new file mode 100644
index 0000000000000000000000000000000000000000..8aab09805997622aec2edae7b565ed82010cfd53
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/rng_state_2.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6802d0e782c931221303c74054cdedbab70d850d6310dc553b393a4f3f848807
+size 16567
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/scheduler.pt b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..b16ede372cbc4e0471ac73758e212cfab98b63f6
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b064b83ca0c5dc2bee7765078ca54598951057b21249ec46aeee7763bc8efc7
+size 627
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/special_tokens_map.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..6949c5975ee0e961ef61cf31010dce04df0a03f8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/special_tokens_map.json
@@ -0,0 +1,23 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|end_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/tokenizer.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..309789d5ccd06f418fbcfb7472a6a9affda68101
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d422bc7d90212ffb8070a743cdbe8ff2dfc5634b1fce8f2e3fa671ee441989eb
+size 17210374
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/tokenizer_config.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..5684b14e2756b56aa397c1324e239b442162b34a
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/tokenizer_config.json
@@ -0,0 +1,2070 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128256": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|end_of_text|>",
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "[PAD]",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/trainer_state.json b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..6b60ee231e895b10808b10b40d5fd6f4bda3ed1d
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/trainer_state.json
@@ -0,0 +1,2373 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.3287671232876712,
+  "eval_steps": 10,
+  "global_step": 300,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001095890410958904,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0636,
+      "step": 1
+    },
+    {
+      "epoch": 0.002191780821917808,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0501,
+      "step": 2
+    },
+    {
+      "epoch": 0.0032876712328767125,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0238,
+      "step": 3
+    },
+    {
+      "epoch": 0.004383561643835616,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0346,
+      "step": 4
+    },
+    {
+      "epoch": 0.005479452054794521,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0108,
+      "step": 5
+    },
+    {
+      "epoch": 0.006575342465753425,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.131,
+      "step": 6
+    },
+    {
+      "epoch": 0.007671232876712329,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0488,
+      "step": 7
+    },
+    {
+      "epoch": 0.008767123287671232,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0905,
+      "step": 8
+    },
+    {
+      "epoch": 0.009863013698630137,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0493,
+      "step": 9
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.1038,
+      "step": 10
+    },
+    {
+      "epoch": 0.010958904109589041,
+      "eval_loss": 3.100078821182251,
+      "eval_runtime": 37.5838,
+      "eval_samples_per_second": 26.607,
+      "eval_steps_per_second": 1.118,
+      "step": 10
+    },
+    {
+      "epoch": 0.012054794520547946,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0804,
+      "step": 11
+    },
+    {
+      "epoch": 0.01315068493150685,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0053,
+      "step": 12
+    },
+    {
+      "epoch": 0.014246575342465753,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0607,
+      "step": 13
+    },
+    {
+      "epoch": 0.015342465753424657,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 3.0539,
+      "step": 14
+    },
+    {
+      "epoch": 0.01643835616438356,
+      "grad_norm": 5.420984745025635,
+      "learning_rate": 1.8248175182481753e-08,
+      "loss": 3.0941,
+      "step": 15
+    },
+    {
+      "epoch": 0.017534246575342465,
+      "grad_norm": 5.76840877532959,
+      "learning_rate": 3.6496350364963505e-08,
+      "loss": 3.036,
+      "step": 16
+    },
+    {
+      "epoch": 0.01863013698630137,
+      "grad_norm": 5.905357360839844,
+      "learning_rate": 5.4744525547445264e-08,
+      "loss": 3.0491,
+      "step": 17
+    },
+    {
+      "epoch": 0.019726027397260273,
+      "grad_norm": 5.981553554534912,
+      "learning_rate": 7.299270072992701e-08,
+      "loss": 3.0775,
+      "step": 18
+    },
+    {
+      "epoch": 0.020821917808219178,
+      "grad_norm": 6.240172863006592,
+      "learning_rate": 9.124087591240876e-08,
+      "loss": 3.0345,
+      "step": 19
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "grad_norm": 5.59375,
+      "learning_rate": 1.0948905109489053e-07,
+      "loss": 3.1063,
+      "step": 20
+    },
+    {
+      "epoch": 0.021917808219178082,
+      "eval_loss": 3.098820686340332,
+      "eval_runtime": 37.2695,
+      "eval_samples_per_second": 26.832,
+      "eval_steps_per_second": 1.127,
+      "step": 20
+    },
+    {
+      "epoch": 0.023013698630136987,
+      "grad_norm": 5.819795608520508,
+      "learning_rate": 1.2773722627737227e-07,
+      "loss": 3.0808,
+      "step": 21
+    },
+    {
+      "epoch": 0.02410958904109589,
+      "grad_norm": 5.7714762687683105,
+      "learning_rate": 1.4598540145985402e-07,
+      "loss": 3.0055,
+      "step": 22
+    },
+    {
+      "epoch": 0.025205479452054796,
+      "grad_norm": 5.51552152633667,
+      "learning_rate": 1.642335766423358e-07,
+      "loss": 3.0396,
+      "step": 23
+    },
+    {
+      "epoch": 0.0263013698630137,
+      "grad_norm": 5.594012260437012,
+      "learning_rate": 1.8248175182481753e-07,
+      "loss": 3.0824,
+      "step": 24
+    },
+    {
+      "epoch": 0.0273972602739726,
+      "grad_norm": 5.1974029541015625,
+      "learning_rate": 2.0072992700729928e-07,
+      "loss": 3.08,
+      "step": 25
+    },
+    {
+      "epoch": 0.028493150684931506,
+      "grad_norm": 4.733968257904053,
+      "learning_rate": 2.1897810218978106e-07,
+      "loss": 3.0597,
+      "step": 26
+    },
+    {
+      "epoch": 0.02958904109589041,
+      "grad_norm": 4.717319011688232,
+      "learning_rate": 2.3722627737226278e-07,
+      "loss": 3.0696,
+      "step": 27
+    },
+    {
+      "epoch": 0.030684931506849315,
+      "grad_norm": 4.6303534507751465,
+      "learning_rate": 2.5547445255474454e-07,
+      "loss": 3.0273,
+      "step": 28
+    },
+    {
+      "epoch": 0.03178082191780822,
+      "grad_norm": 4.413999557495117,
+      "learning_rate": 2.737226277372263e-07,
+      "loss": 3.0202,
+      "step": 29
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "grad_norm": 4.516043186187744,
+      "learning_rate": 2.9197080291970804e-07,
+      "loss": 2.9926,
+      "step": 30
+    },
+    {
+      "epoch": 0.03287671232876712,
+      "eval_loss": 3.0561797618865967,
+      "eval_runtime": 37.0351,
+      "eval_samples_per_second": 27.001,
+      "eval_steps_per_second": 1.134,
+      "step": 30
+    },
+    {
+      "epoch": 0.03397260273972603,
+      "grad_norm": 3.7179207801818848,
+      "learning_rate": 3.102189781021898e-07,
+      "loss": 3.0086,
+      "step": 31
+    },
+    {
+      "epoch": 0.03506849315068493,
+      "grad_norm": 3.7840425968170166,
+      "learning_rate": 3.284671532846716e-07,
+      "loss": 2.9986,
+      "step": 32
+    },
+    {
+      "epoch": 0.03616438356164384,
+      "grad_norm": 3.843116521835327,
+      "learning_rate": 3.467153284671533e-07,
+      "loss": 2.9677,
+      "step": 33
+    },
+    {
+      "epoch": 0.03726027397260274,
+      "grad_norm": 3.8528714179992676,
+      "learning_rate": 3.6496350364963505e-07,
+      "loss": 2.9732,
+      "step": 34
+    },
+    {
+      "epoch": 0.038356164383561646,
+      "grad_norm": 3.76963210105896,
+      "learning_rate": 3.8321167883211686e-07,
+      "loss": 2.9855,
+      "step": 35
+    },
+    {
+      "epoch": 0.03945205479452055,
+      "grad_norm": 3.6240015029907227,
+      "learning_rate": 4.0145985401459856e-07,
+      "loss": 3.0117,
+      "step": 36
+    },
+    {
+      "epoch": 0.040547945205479455,
+      "grad_norm": 3.1380624771118164,
+      "learning_rate": 4.197080291970803e-07,
+      "loss": 2.9481,
+      "step": 37
+    },
+    {
+      "epoch": 0.041643835616438356,
+      "grad_norm": 3.565329074859619,
+      "learning_rate": 4.379562043795621e-07,
+      "loss": 2.9175,
+      "step": 38
+    },
+    {
+      "epoch": 0.042739726027397264,
+      "grad_norm": 3.8062448501586914,
+      "learning_rate": 4.562043795620438e-07,
+      "loss": 3.0074,
+      "step": 39
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "grad_norm": 3.412046432495117,
+      "learning_rate": 4.7445255474452557e-07,
+      "loss": 2.9119,
+      "step": 40
+    },
+    {
+      "epoch": 0.043835616438356165,
+      "eval_loss": 2.9874370098114014,
+      "eval_runtime": 37.0514,
+      "eval_samples_per_second": 26.99,
+      "eval_steps_per_second": 1.134,
+      "step": 40
+    },
+    {
+      "epoch": 0.044931506849315066,
+      "grad_norm": 3.227437973022461,
+      "learning_rate": 4.927007299270073e-07,
+      "loss": 2.9436,
+      "step": 41
+    },
+    {
+      "epoch": 0.046027397260273974,
+      "grad_norm": 2.9633443355560303,
+      "learning_rate": 5.109489051094891e-07,
+      "loss": 2.967,
+      "step": 42
+    },
+    {
+      "epoch": 0.047123287671232875,
+      "grad_norm": 2.942678928375244,
+      "learning_rate": 5.291970802919708e-07,
+      "loss": 2.9375,
+      "step": 43
+    },
+    {
+      "epoch": 0.04821917808219178,
+      "grad_norm": 2.9806888103485107,
+      "learning_rate": 5.474452554744526e-07,
+      "loss": 2.9474,
+      "step": 44
+    },
+    {
+      "epoch": 0.049315068493150684,
+      "grad_norm": 3.000063419342041,
+      "learning_rate": 5.656934306569343e-07,
+      "loss": 2.9289,
+      "step": 45
+    },
+    {
+      "epoch": 0.05041095890410959,
+      "grad_norm": 3.1481194496154785,
+      "learning_rate": 5.839416058394161e-07,
+      "loss": 2.8869,
+      "step": 46
+    },
+    {
+      "epoch": 0.05150684931506849,
+      "grad_norm": 2.949202060699463,
+      "learning_rate": 6.021897810218979e-07,
+      "loss": 2.8783,
+      "step": 47
+    },
+    {
+      "epoch": 0.0526027397260274,
+      "grad_norm": 2.706526279449463,
+      "learning_rate": 6.204379562043796e-07,
+      "loss": 2.9024,
+      "step": 48
+    },
+    {
+      "epoch": 0.0536986301369863,
+      "grad_norm": 2.7317655086517334,
+      "learning_rate": 6.386861313868614e-07,
+      "loss": 2.8325,
+      "step": 49
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "grad_norm": 2.25644588470459,
+      "learning_rate": 6.569343065693432e-07,
+      "loss": 2.9409,
+      "step": 50
+    },
+    {
+      "epoch": 0.0547945205479452,
+      "eval_loss": 2.909320116043091,
+      "eval_runtime": 37.0714,
+      "eval_samples_per_second": 26.975,
+      "eval_steps_per_second": 1.133,
+      "step": 50
+    },
+    {
+      "epoch": 0.05589041095890411,
+      "grad_norm": 2.8789613246917725,
+      "learning_rate": 6.751824817518248e-07,
+      "loss": 2.8459,
+      "step": 51
+    },
+    {
+      "epoch": 0.05698630136986301,
+      "grad_norm": 2.686873435974121,
+      "learning_rate": 6.934306569343066e-07,
+      "loss": 2.8653,
+      "step": 52
+    },
+    {
+      "epoch": 0.05808219178082192,
+      "grad_norm": 2.54382586479187,
+      "learning_rate": 7.116788321167883e-07,
+      "loss": 2.7973,
+      "step": 53
+    },
+    {
+      "epoch": 0.05917808219178082,
+      "grad_norm": 2.1871724128723145,
+      "learning_rate": 7.299270072992701e-07,
+      "loss": 2.7982,
+      "step": 54
+    },
+    {
+      "epoch": 0.06027397260273973,
+      "grad_norm": 2.145303249359131,
+      "learning_rate": 7.48175182481752e-07,
+      "loss": 2.8195,
+      "step": 55
+    },
+    {
+      "epoch": 0.06136986301369863,
+      "grad_norm": 2.1635069847106934,
+      "learning_rate": 7.664233576642337e-07,
+      "loss": 2.825,
+      "step": 56
+    },
+    {
+      "epoch": 0.06246575342465754,
+      "grad_norm": 2.3530752658843994,
+      "learning_rate": 7.846715328467154e-07,
+      "loss": 2.8407,
+      "step": 57
+    },
+    {
+      "epoch": 0.06356164383561644,
+      "grad_norm": 2.2455942630767822,
+      "learning_rate": 8.029197080291971e-07,
+      "loss": 2.8182,
+      "step": 58
+    },
+    {
+      "epoch": 0.06465753424657535,
+      "grad_norm": 1.8985673189163208,
+      "learning_rate": 8.211678832116789e-07,
+      "loss": 2.772,
+      "step": 59
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "grad_norm": 1.9699499607086182,
+      "learning_rate": 8.394160583941606e-07,
+      "loss": 2.7367,
+      "step": 60
+    },
+    {
+      "epoch": 0.06575342465753424,
+      "eval_loss": 2.8451859951019287,
+      "eval_runtime": 37.0793,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 60
+    },
+    {
+      "epoch": 0.06684931506849315,
+      "grad_norm": 2.159547805786133,
+      "learning_rate": 8.576642335766425e-07,
+      "loss": 2.7635,
+      "step": 61
+    },
+    {
+      "epoch": 0.06794520547945206,
+      "grad_norm": 2.12705659866333,
+      "learning_rate": 8.759124087591242e-07,
+      "loss": 2.7624,
+      "step": 62
+    },
+    {
+      "epoch": 0.06904109589041096,
+      "grad_norm": 1.8393241167068481,
+      "learning_rate": 8.94160583941606e-07,
+      "loss": 2.7763,
+      "step": 63
+    },
+    {
+      "epoch": 0.07013698630136986,
+      "grad_norm": 1.683522343635559,
+      "learning_rate": 9.124087591240876e-07,
+      "loss": 2.7762,
+      "step": 64
+    },
+    {
+      "epoch": 0.07123287671232877,
+      "grad_norm": 1.9260376691818237,
+      "learning_rate": 9.306569343065694e-07,
+      "loss": 2.7418,
+      "step": 65
+    },
+    {
+      "epoch": 0.07232876712328767,
+      "grad_norm": 2.09322190284729,
+      "learning_rate": 9.489051094890511e-07,
+      "loss": 2.7625,
+      "step": 66
+    },
+    {
+      "epoch": 0.07342465753424658,
+      "grad_norm": 1.9183429479599,
+      "learning_rate": 9.67153284671533e-07,
+      "loss": 2.7453,
+      "step": 67
+    },
+    {
+      "epoch": 0.07452054794520548,
+      "grad_norm": 1.9792778491973877,
+      "learning_rate": 9.854014598540146e-07,
+      "loss": 2.7858,
+      "step": 68
+    },
+    {
+      "epoch": 0.07561643835616438,
+      "grad_norm": 2.1397311687469482,
+      "learning_rate": 1.0036496350364965e-06,
+      "loss": 2.7242,
+      "step": 69
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "grad_norm": 1.8556463718414307,
+      "learning_rate": 1.0218978102189781e-06,
+      "loss": 2.6853,
+      "step": 70
+    },
+    {
+      "epoch": 0.07671232876712329,
+      "eval_loss": 2.7787046432495117,
+      "eval_runtime": 37.1017,
+      "eval_samples_per_second": 26.953,
+      "eval_steps_per_second": 1.132,
+      "step": 70
+    },
+    {
+      "epoch": 0.07780821917808219,
+      "grad_norm": 1.9776800870895386,
+      "learning_rate": 1.04014598540146e-06,
+      "loss": 2.749,
+      "step": 71
+    },
+    {
+      "epoch": 0.0789041095890411,
+      "grad_norm": 2.3972630500793457,
+      "learning_rate": 1.0583941605839416e-06,
+      "loss": 2.6457,
+      "step": 72
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 2.090210199356079,
+      "learning_rate": 1.0766423357664233e-06,
+      "loss": 2.6401,
+      "step": 73
+    },
+    {
+      "epoch": 0.08109589041095891,
+      "grad_norm": 2.4117701053619385,
+      "learning_rate": 1.0948905109489052e-06,
+      "loss": 2.6649,
+      "step": 74
+    },
+    {
+      "epoch": 0.0821917808219178,
+      "grad_norm": 2.4845173358917236,
+      "learning_rate": 1.113138686131387e-06,
+      "loss": 2.6855,
+      "step": 75
+    },
+    {
+      "epoch": 0.08328767123287671,
+      "grad_norm": 1.7779242992401123,
+      "learning_rate": 1.1313868613138687e-06,
+      "loss": 2.6926,
+      "step": 76
+    },
+    {
+      "epoch": 0.08438356164383562,
+      "grad_norm": 2.3163535594940186,
+      "learning_rate": 1.1496350364963505e-06,
+      "loss": 2.6302,
+      "step": 77
+    },
+    {
+      "epoch": 0.08547945205479453,
+      "grad_norm": 2.2296528816223145,
+      "learning_rate": 1.1678832116788322e-06,
+      "loss": 2.6054,
+      "step": 78
+    },
+    {
+      "epoch": 0.08657534246575342,
+      "grad_norm": 2.253431797027588,
+      "learning_rate": 1.1861313868613138e-06,
+      "loss": 2.6104,
+      "step": 79
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "grad_norm": 2.490609645843506,
+      "learning_rate": 1.2043795620437959e-06,
+      "loss": 2.6078,
+      "step": 80
+    },
+    {
+      "epoch": 0.08767123287671233,
+      "eval_loss": 2.706103563308716,
+      "eval_runtime": 37.0892,
+      "eval_samples_per_second": 26.962,
+      "eval_steps_per_second": 1.132,
+      "step": 80
+    },
+    {
+      "epoch": 0.08876712328767124,
+      "grad_norm": 2.3048248291015625,
+      "learning_rate": 1.2226277372262775e-06,
+      "loss": 2.6096,
+      "step": 81
+    },
+    {
+      "epoch": 0.08986301369863013,
+      "grad_norm": 2.440666913986206,
+      "learning_rate": 1.2408759124087592e-06,
+      "loss": 2.6412,
+      "step": 82
+    },
+    {
+      "epoch": 0.09095890410958904,
+      "grad_norm": 1.9456374645233154,
+      "learning_rate": 1.259124087591241e-06,
+      "loss": 2.591,
+      "step": 83
+    },
+    {
+      "epoch": 0.09205479452054795,
+      "grad_norm": 2.598172664642334,
+      "learning_rate": 1.2773722627737229e-06,
+      "loss": 2.6093,
+      "step": 84
+    },
+    {
+      "epoch": 0.09315068493150686,
+      "grad_norm": 2.2587032318115234,
+      "learning_rate": 1.2956204379562045e-06,
+      "loss": 2.5949,
+      "step": 85
+    },
+    {
+      "epoch": 0.09424657534246575,
+      "grad_norm": 2.2927448749542236,
+      "learning_rate": 1.3138686131386864e-06,
+      "loss": 2.5464,
+      "step": 86
+    },
+    {
+      "epoch": 0.09534246575342466,
+      "grad_norm": 2.5940468311309814,
+      "learning_rate": 1.332116788321168e-06,
+      "loss": 2.5613,
+      "step": 87
+    },
+    {
+      "epoch": 0.09643835616438357,
+      "grad_norm": 2.073207378387451,
+      "learning_rate": 1.3503649635036497e-06,
+      "loss": 2.4955,
+      "step": 88
+    },
+    {
+      "epoch": 0.09753424657534246,
+      "grad_norm": 3.211766242980957,
+      "learning_rate": 1.3686131386861315e-06,
+      "loss": 2.5315,
+      "step": 89
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "grad_norm": 2.2590038776397705,
+      "learning_rate": 1.3868613138686132e-06,
+      "loss": 2.4641,
+      "step": 90
+    },
+    {
+      "epoch": 0.09863013698630137,
+      "eval_loss": 2.625087261199951,
+      "eval_runtime": 37.0852,
+      "eval_samples_per_second": 26.965,
+      "eval_steps_per_second": 1.133,
+      "step": 90
+    },
+    {
+      "epoch": 0.09972602739726028,
+      "grad_norm": 2.3959836959838867,
+      "learning_rate": 1.405109489051095e-06,
+      "loss": 2.5537,
+      "step": 91
+    },
+    {
+      "epoch": 0.10082191780821918,
+      "grad_norm": 2.596761465072632,
+      "learning_rate": 1.4233576642335767e-06,
+      "loss": 2.5135,
+      "step": 92
+    },
+    {
+      "epoch": 0.10191780821917808,
+      "grad_norm": 2.475759267807007,
+      "learning_rate": 1.4416058394160583e-06,
+      "loss": 2.5436,
+      "step": 93
+    },
+    {
+      "epoch": 0.10301369863013699,
+      "grad_norm": 3.0613298416137695,
+      "learning_rate": 1.4598540145985402e-06,
+      "loss": 2.5066,
+      "step": 94
+    },
+    {
+      "epoch": 0.10410958904109589,
+      "grad_norm": 2.406097650527954,
+      "learning_rate": 1.478102189781022e-06,
+      "loss": 2.4646,
+      "step": 95
+    },
+    {
+      "epoch": 0.1052054794520548,
+      "grad_norm": 2.5872504711151123,
+      "learning_rate": 1.496350364963504e-06,
+      "loss": 2.5351,
+      "step": 96
+    },
+    {
+      "epoch": 0.1063013698630137,
+      "grad_norm": 2.4294471740722656,
+      "learning_rate": 1.5145985401459856e-06,
+      "loss": 2.4953,
+      "step": 97
+    },
+    {
+      "epoch": 0.1073972602739726,
+      "grad_norm": 2.674884080886841,
+      "learning_rate": 1.5328467153284674e-06,
+      "loss": 2.5133,
+      "step": 98
+    },
+    {
+      "epoch": 0.10849315068493151,
+      "grad_norm": 2.5096852779388428,
+      "learning_rate": 1.551094890510949e-06,
+      "loss": 2.4693,
+      "step": 99
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "grad_norm": 2.7390334606170654,
+      "learning_rate": 1.5693430656934307e-06,
+      "loss": 2.4195,
+      "step": 100
+    },
+    {
+      "epoch": 0.1095890410958904,
+      "eval_loss": 2.55997371673584,
+      "eval_runtime": 37.0732,
+      "eval_samples_per_second": 26.974,
+      "eval_steps_per_second": 1.133,
+      "step": 100
+    },
+    {
+      "epoch": 0.11068493150684931,
+      "grad_norm": 2.8887312412261963,
+      "learning_rate": 1.5875912408759126e-06,
+      "loss": 2.4193,
+      "step": 101
+    },
+    {
+      "epoch": 0.11178082191780822,
+      "grad_norm": 3.0721421241760254,
+      "learning_rate": 1.6058394160583942e-06,
+      "loss": 2.4709,
+      "step": 102
+    },
+    {
+      "epoch": 0.11287671232876713,
+      "grad_norm": 2.728156566619873,
+      "learning_rate": 1.624087591240876e-06,
+      "loss": 2.5105,
+      "step": 103
+    },
+    {
+      "epoch": 0.11397260273972602,
+      "grad_norm": 3.1214346885681152,
+      "learning_rate": 1.6423357664233577e-06,
+      "loss": 2.4316,
+      "step": 104
+    },
+    {
+      "epoch": 0.11506849315068493,
+      "grad_norm": 2.9249517917633057,
+      "learning_rate": 1.6605839416058394e-06,
+      "loss": 2.3737,
+      "step": 105
+    },
+    {
+      "epoch": 0.11616438356164384,
+      "grad_norm": 2.6646173000335693,
+      "learning_rate": 1.6788321167883212e-06,
+      "loss": 2.4165,
+      "step": 106
+    },
+    {
+      "epoch": 0.11726027397260275,
+      "grad_norm": 3.108067750930786,
+      "learning_rate": 1.6970802919708029e-06,
+      "loss": 2.4217,
+      "step": 107
+    },
+    {
+      "epoch": 0.11835616438356164,
+      "grad_norm": 3.142338991165161,
+      "learning_rate": 1.715328467153285e-06,
+      "loss": 2.3527,
+      "step": 108
+    },
+    {
+      "epoch": 0.11945205479452055,
+      "grad_norm": 2.9078011512756348,
+      "learning_rate": 1.7335766423357666e-06,
+      "loss": 2.3927,
+      "step": 109
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "grad_norm": 3.40766978263855,
+      "learning_rate": 1.7518248175182485e-06,
+      "loss": 2.4021,
+      "step": 110
+    },
+    {
+      "epoch": 0.12054794520547946,
+      "eval_loss": 2.5031485557556152,
+      "eval_runtime": 36.8147,
+      "eval_samples_per_second": 27.163,
+      "eval_steps_per_second": 1.141,
+      "step": 110
+    },
+    {
+      "epoch": 0.12164383561643835,
+      "grad_norm": 4.396062850952148,
+      "learning_rate": 1.77007299270073e-06,
+      "loss": 2.4373,
+      "step": 111
+    },
+    {
+      "epoch": 0.12273972602739726,
+      "grad_norm": 4.0441365242004395,
+      "learning_rate": 1.788321167883212e-06,
+      "loss": 2.4162,
+      "step": 112
+    },
+    {
+      "epoch": 0.12383561643835617,
+      "grad_norm": 2.8869969844818115,
+      "learning_rate": 1.8065693430656936e-06,
+      "loss": 2.4515,
+      "step": 113
+    },
+    {
+      "epoch": 0.12493150684931507,
+      "grad_norm": 3.4833669662475586,
+      "learning_rate": 1.8248175182481753e-06,
+      "loss": 2.3965,
+      "step": 114
+    },
+    {
+      "epoch": 0.12602739726027398,
+      "grad_norm": 2.9996068477630615,
+      "learning_rate": 1.8430656934306571e-06,
+      "loss": 2.3374,
+      "step": 115
+    },
+    {
+      "epoch": 0.12712328767123288,
+      "grad_norm": 2.701371192932129,
+      "learning_rate": 1.8613138686131388e-06,
+      "loss": 2.3608,
+      "step": 116
+    },
+    {
+      "epoch": 0.12821917808219177,
+      "grad_norm": 2.9491288661956787,
+      "learning_rate": 1.8795620437956206e-06,
+      "loss": 2.355,
+      "step": 117
+    },
+    {
+      "epoch": 0.1293150684931507,
+      "grad_norm": 2.9266560077667236,
+      "learning_rate": 1.8978102189781023e-06,
+      "loss": 2.3475,
+      "step": 118
+    },
+    {
+      "epoch": 0.1304109589041096,
+      "grad_norm": 3.0419797897338867,
+      "learning_rate": 1.916058394160584e-06,
+      "loss": 2.3475,
+      "step": 119
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "grad_norm": 2.4531214237213135,
+      "learning_rate": 1.934306569343066e-06,
+      "loss": 2.3724,
+      "step": 120
+    },
+    {
+      "epoch": 0.13150684931506848,
+      "eval_loss": 2.455423593521118,
+      "eval_runtime": 37.0304,
+      "eval_samples_per_second": 27.005,
+      "eval_steps_per_second": 1.134,
+      "step": 120
+    },
+    {
+      "epoch": 0.1326027397260274,
+      "grad_norm": 3.572502851486206,
+      "learning_rate": 1.952554744525548e-06,
+      "loss": 2.2671,
+      "step": 121
+    },
+    {
+      "epoch": 0.1336986301369863,
+      "grad_norm": 3.5408504009246826,
+      "learning_rate": 1.9708029197080293e-06,
+      "loss": 2.3133,
+      "step": 122
+    },
+    {
+      "epoch": 0.13479452054794522,
+      "grad_norm": 2.488725423812866,
+      "learning_rate": 1.989051094890511e-06,
+      "loss": 2.3141,
+      "step": 123
+    },
+    {
+      "epoch": 0.1358904109589041,
+      "grad_norm": 3.241276264190674,
+      "learning_rate": 2.007299270072993e-06,
+      "loss": 2.318,
+      "step": 124
+    },
+    {
+      "epoch": 0.136986301369863,
+      "grad_norm": 2.853959798812866,
+      "learning_rate": 2.025547445255475e-06,
+      "loss": 2.3712,
+      "step": 125
+    },
+    {
+      "epoch": 0.13808219178082193,
+      "grad_norm": 2.9798169136047363,
+      "learning_rate": 2.0437956204379563e-06,
+      "loss": 2.3072,
+      "step": 126
+    },
+    {
+      "epoch": 0.13917808219178082,
+      "grad_norm": 2.8529601097106934,
+      "learning_rate": 2.062043795620438e-06,
+      "loss": 2.318,
+      "step": 127
+    },
+    {
+      "epoch": 0.14027397260273972,
+      "grad_norm": 3.475337028503418,
+      "learning_rate": 2.08029197080292e-06,
+      "loss": 2.31,
+      "step": 128
+    },
+    {
+      "epoch": 0.14136986301369864,
+      "grad_norm": 4.213955879211426,
+      "learning_rate": 2.0985401459854014e-06,
+      "loss": 2.3734,
+      "step": 129
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "grad_norm": 3.0833330154418945,
+      "learning_rate": 2.1167883211678833e-06,
+      "loss": 2.2602,
+      "step": 130
+    },
+    {
+      "epoch": 0.14246575342465753,
+      "eval_loss": 2.4205682277679443,
+      "eval_runtime": 37.0798,
+      "eval_samples_per_second": 26.969,
+      "eval_steps_per_second": 1.133,
+      "step": 130
+    },
+    {
+      "epoch": 0.14356164383561643,
+      "grad_norm": 3.411235809326172,
+      "learning_rate": 2.135036496350365e-06,
+      "loss": 2.3255,
+      "step": 131
+    },
+    {
+      "epoch": 0.14465753424657535,
+      "grad_norm": 2.882053852081299,
+      "learning_rate": 2.1532846715328466e-06,
+      "loss": 2.3483,
+      "step": 132
+    },
+    {
+      "epoch": 0.14575342465753424,
+      "grad_norm": 3.2436420917510986,
+      "learning_rate": 2.171532846715329e-06,
+      "loss": 2.2492,
+      "step": 133
+    },
+    {
+      "epoch": 0.14684931506849316,
+      "grad_norm": 2.929466962814331,
+      "learning_rate": 2.1897810218978103e-06,
+      "loss": 2.2647,
+      "step": 134
+    },
+    {
+      "epoch": 0.14794520547945206,
+      "grad_norm": 2.9344966411590576,
+      "learning_rate": 2.208029197080292e-06,
+      "loss": 2.3061,
+      "step": 135
+    },
+    {
+      "epoch": 0.14904109589041095,
+      "grad_norm": 3.2567203044891357,
+      "learning_rate": 2.226277372262774e-06,
+      "loss": 2.2746,
+      "step": 136
+    },
+    {
+      "epoch": 0.15013698630136987,
+      "grad_norm": 3.1494054794311523,
+      "learning_rate": 2.244525547445256e-06,
+      "loss": 2.2866,
+      "step": 137
+    },
+    {
+      "epoch": 0.15123287671232877,
+      "grad_norm": 2.723390817642212,
+      "learning_rate": 2.2627737226277373e-06,
+      "loss": 2.3109,
+      "step": 138
+    },
+    {
+      "epoch": 0.15232876712328766,
+      "grad_norm": 2.8904967308044434,
+      "learning_rate": 2.281021897810219e-06,
+      "loss": 2.301,
+      "step": 139
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "grad_norm": 2.8899483680725098,
+      "learning_rate": 2.299270072992701e-06,
+      "loss": 2.2637,
+      "step": 140
+    },
+    {
+      "epoch": 0.15342465753424658,
+      "eval_loss": 2.3952724933624268,
+      "eval_runtime": 37.1272,
+      "eval_samples_per_second": 26.934,
+      "eval_steps_per_second": 1.131,
+      "step": 140
+    },
+    {
+      "epoch": 0.15452054794520548,
+      "grad_norm": 3.5309066772460938,
+      "learning_rate": 2.3175182481751825e-06,
+      "loss": 2.2572,
+      "step": 141
+    },
+    {
+      "epoch": 0.15561643835616437,
+      "grad_norm": 2.8369667530059814,
+      "learning_rate": 2.3357664233576643e-06,
+      "loss": 2.2712,
+      "step": 142
+    },
+    {
+      "epoch": 0.1567123287671233,
+      "grad_norm": 3.629161834716797,
+      "learning_rate": 2.354014598540146e-06,
+      "loss": 2.2301,
+      "step": 143
+    },
+    {
+      "epoch": 0.1578082191780822,
+      "grad_norm": 3.5313000679016113,
+      "learning_rate": 2.3722627737226276e-06,
+      "loss": 2.2712,
+      "step": 144
+    },
+    {
+      "epoch": 0.1589041095890411,
+      "grad_norm": 3.1328771114349365,
+      "learning_rate": 2.39051094890511e-06,
+      "loss": 2.2706,
+      "step": 145
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 4.6992950439453125,
+      "learning_rate": 2.4087591240875918e-06,
+      "loss": 2.219,
+      "step": 146
+    },
+    {
+      "epoch": 0.1610958904109589,
+      "grad_norm": 5.230221748352051,
+      "learning_rate": 2.427007299270073e-06,
+      "loss": 2.2396,
+      "step": 147
+    },
+    {
+      "epoch": 0.16219178082191782,
+      "grad_norm": 3.5680761337280273,
+      "learning_rate": 2.445255474452555e-06,
+      "loss": 2.2434,
+      "step": 148
+    },
+    {
+      "epoch": 0.1632876712328767,
+      "grad_norm": 3.1642441749572754,
+      "learning_rate": 2.463503649635037e-06,
+      "loss": 2.2856,
+      "step": 149
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "grad_norm": 2.694085121154785,
+      "learning_rate": 2.4817518248175183e-06,
+      "loss": 2.2006,
+      "step": 150
+    },
+    {
+      "epoch": 0.1643835616438356,
+      "eval_loss": 2.37199068069458,
+      "eval_runtime": 37.1342,
+      "eval_samples_per_second": 26.929,
+      "eval_steps_per_second": 1.131,
+      "step": 150
+    },
+    {
+      "epoch": 0.16547945205479453,
+      "grad_norm": 2.9868853092193604,
+      "learning_rate": 2.5e-06,
+      "loss": 2.1694,
+      "step": 151
+    },
+    {
+      "epoch": 0.16657534246575342,
+      "grad_norm": 2.793532371520996,
+      "learning_rate": 2.518248175182482e-06,
+      "loss": 2.2965,
+      "step": 152
+    },
+    {
+      "epoch": 0.16767123287671232,
+      "grad_norm": 2.60152530670166,
+      "learning_rate": 2.5364963503649635e-06,
+      "loss": 2.1954,
+      "step": 153
+    },
+    {
+      "epoch": 0.16876712328767124,
+      "grad_norm": 2.6192219257354736,
+      "learning_rate": 2.5547445255474458e-06,
+      "loss": 2.1909,
+      "step": 154
+    },
+    {
+      "epoch": 0.16986301369863013,
+      "grad_norm": 2.467344045639038,
+      "learning_rate": 2.5729927007299272e-06,
+      "loss": 2.2438,
+      "step": 155
+    },
+    {
+      "epoch": 0.17095890410958905,
+      "grad_norm": 2.772012948989868,
+      "learning_rate": 2.591240875912409e-06,
+      "loss": 2.2213,
+      "step": 156
+    },
+    {
+      "epoch": 0.17205479452054795,
+      "grad_norm": 2.613675355911255,
+      "learning_rate": 2.6094890510948905e-06,
+      "loss": 2.2213,
+      "step": 157
+    },
+    {
+      "epoch": 0.17315068493150684,
+      "grad_norm": 2.849245071411133,
+      "learning_rate": 2.627737226277373e-06,
+      "loss": 2.255,
+      "step": 158
+    },
+    {
+      "epoch": 0.17424657534246576,
+      "grad_norm": 2.8388969898223877,
+      "learning_rate": 2.6459854014598542e-06,
+      "loss": 2.2572,
+      "step": 159
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "grad_norm": 2.5219321250915527,
+      "learning_rate": 2.664233576642336e-06,
+      "loss": 2.2253,
+      "step": 160
+    },
+    {
+      "epoch": 0.17534246575342466,
+      "eval_loss": 2.3509061336517334,
+      "eval_runtime": 37.0953,
+      "eval_samples_per_second": 26.958,
+      "eval_steps_per_second": 1.132,
+      "step": 160
+    },
+    {
+      "epoch": 0.17643835616438355,
+      "grad_norm": 2.8460891246795654,
+      "learning_rate": 2.6824817518248175e-06,
+      "loss": 2.2009,
+      "step": 161
+    },
+    {
+      "epoch": 0.17753424657534247,
+      "grad_norm": 2.673377275466919,
+      "learning_rate": 2.7007299270072994e-06,
+      "loss": 2.2211,
+      "step": 162
+    },
+    {
+      "epoch": 0.17863013698630137,
+      "grad_norm": 2.8514182567596436,
+      "learning_rate": 2.718978102189781e-06,
+      "loss": 2.1955,
+      "step": 163
+    },
+    {
+      "epoch": 0.17972602739726026,
+      "grad_norm": 3.530385971069336,
+      "learning_rate": 2.737226277372263e-06,
+      "loss": 2.2536,
+      "step": 164
+    },
+    {
+      "epoch": 0.18082191780821918,
+      "grad_norm": 2.733394145965576,
+      "learning_rate": 2.755474452554745e-06,
+      "loss": 2.2367,
+      "step": 165
+    },
+    {
+      "epoch": 0.18191780821917808,
+      "grad_norm": 3.871150016784668,
+      "learning_rate": 2.7737226277372264e-06,
+      "loss": 2.1949,
+      "step": 166
+    },
+    {
+      "epoch": 0.18301369863013697,
+      "grad_norm": 3.7782530784606934,
+      "learning_rate": 2.7919708029197082e-06,
+      "loss": 2.2979,
+      "step": 167
+    },
+    {
+      "epoch": 0.1841095890410959,
+      "grad_norm": 3.094599485397339,
+      "learning_rate": 2.81021897810219e-06,
+      "loss": 2.2512,
+      "step": 168
+    },
+    {
+      "epoch": 0.1852054794520548,
+      "grad_norm": 3.2166848182678223,
+      "learning_rate": 2.828467153284672e-06,
+      "loss": 2.1784,
+      "step": 169
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "grad_norm": 3.4127073287963867,
+      "learning_rate": 2.8467153284671534e-06,
+      "loss": 2.2241,
+      "step": 170
+    },
+    {
+      "epoch": 0.1863013698630137,
+      "eval_loss": 2.333245038986206,
+      "eval_runtime": 37.0402,
+      "eval_samples_per_second": 26.998,
+      "eval_steps_per_second": 1.134,
+      "step": 170
+    },
+    {
+      "epoch": 0.1873972602739726,
+      "grad_norm": 2.677579879760742,
+      "learning_rate": 2.8649635036496353e-06,
+      "loss": 2.135,
+      "step": 171
+    },
+    {
+      "epoch": 0.1884931506849315,
+      "grad_norm": 2.9602150917053223,
+      "learning_rate": 2.8832116788321167e-06,
+      "loss": 2.1916,
+      "step": 172
+    },
+    {
+      "epoch": 0.18958904109589042,
+      "grad_norm": 2.580122232437134,
+      "learning_rate": 2.901459854014599e-06,
+      "loss": 2.1477,
+      "step": 173
+    },
+    {
+      "epoch": 0.19068493150684931,
+      "grad_norm": 2.652217388153076,
+      "learning_rate": 2.9197080291970804e-06,
+      "loss": 2.2394,
+      "step": 174
+    },
+    {
+      "epoch": 0.1917808219178082,
+      "grad_norm": 2.6887567043304443,
+      "learning_rate": 2.9379562043795623e-06,
+      "loss": 2.2245,
+      "step": 175
+    },
+    {
+      "epoch": 0.19287671232876713,
+      "grad_norm": 2.768165111541748,
+      "learning_rate": 2.956204379562044e-06,
+      "loss": 2.2017,
+      "step": 176
+    },
+    {
+      "epoch": 0.19397260273972602,
+      "grad_norm": 2.9375691413879395,
+      "learning_rate": 2.9744525547445256e-06,
+      "loss": 2.2464,
+      "step": 177
+    },
+    {
+      "epoch": 0.19506849315068492,
+      "grad_norm": 2.8364248275756836,
+      "learning_rate": 2.992700729927008e-06,
+      "loss": 2.2155,
+      "step": 178
+    },
+    {
+      "epoch": 0.19616438356164384,
+      "grad_norm": 2.843487024307251,
+      "learning_rate": 3.0109489051094893e-06,
+      "loss": 2.1577,
+      "step": 179
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "grad_norm": 3.005230665206909,
+      "learning_rate": 3.029197080291971e-06,
+      "loss": 2.1549,
+      "step": 180
+    },
+    {
+      "epoch": 0.19726027397260273,
+      "eval_loss": 2.3209409713745117,
+      "eval_runtime": 37.022,
+      "eval_samples_per_second": 27.011,
+      "eval_steps_per_second": 1.134,
+      "step": 180
+    },
+    {
+      "epoch": 0.19835616438356166,
+      "grad_norm": 2.490260601043701,
+      "learning_rate": 3.0474452554744526e-06,
+      "loss": 2.2313,
+      "step": 181
+    },
+    {
+      "epoch": 0.19945205479452055,
+      "grad_norm": 2.973118543624878,
+      "learning_rate": 3.065693430656935e-06,
+      "loss": 2.2316,
+      "step": 182
+    },
+    {
+      "epoch": 0.20054794520547944,
+      "grad_norm": 3.3308680057525635,
+      "learning_rate": 3.0839416058394163e-06,
+      "loss": 2.2087,
+      "step": 183
+    },
+    {
+      "epoch": 0.20164383561643837,
+      "grad_norm": 3.1440727710723877,
+      "learning_rate": 3.102189781021898e-06,
+      "loss": 2.2256,
+      "step": 184
+    },
+    {
+      "epoch": 0.20273972602739726,
+      "grad_norm": 2.5696325302124023,
+      "learning_rate": 3.1204379562043796e-06,
+      "loss": 2.2093,
+      "step": 185
+    },
+    {
+      "epoch": 0.20383561643835615,
+      "grad_norm": 3.1225037574768066,
+      "learning_rate": 3.1386861313868614e-06,
+      "loss": 2.1425,
+      "step": 186
+    },
+    {
+      "epoch": 0.20493150684931508,
+      "grad_norm": 3.650743246078491,
+      "learning_rate": 3.156934306569343e-06,
+      "loss": 2.1339,
+      "step": 187
+    },
+    {
+      "epoch": 0.20602739726027397,
+      "grad_norm": 2.924072027206421,
+      "learning_rate": 3.175182481751825e-06,
+      "loss": 2.2363,
+      "step": 188
+    },
+    {
+      "epoch": 0.20712328767123286,
+      "grad_norm": 2.9561989307403564,
+      "learning_rate": 3.193430656934307e-06,
+      "loss": 2.1934,
+      "step": 189
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "grad_norm": 2.731393814086914,
+      "learning_rate": 3.2116788321167884e-06,
+      "loss": 2.1718,
+      "step": 190
+    },
+    {
+      "epoch": 0.20821917808219179,
+      "eval_loss": 2.3087782859802246,
+      "eval_runtime": 37.0161,
+      "eval_samples_per_second": 27.015,
+      "eval_steps_per_second": 1.135,
+      "step": 190
+    },
+    {
+      "epoch": 0.20931506849315068,
+      "grad_norm": 2.774442434310913,
+      "learning_rate": 3.2299270072992707e-06,
+      "loss": 2.1841,
+      "step": 191
+    },
+    {
+      "epoch": 0.2104109589041096,
+      "grad_norm": 3.0579135417938232,
+      "learning_rate": 3.248175182481752e-06,
+      "loss": 2.1677,
+      "step": 192
+    },
+    {
+      "epoch": 0.2115068493150685,
+      "grad_norm": 2.732984781265259,
+      "learning_rate": 3.266423357664234e-06,
+      "loss": 2.0975,
+      "step": 193
+    },
+    {
+      "epoch": 0.2126027397260274,
+      "grad_norm": 3.442659854888916,
+      "learning_rate": 3.2846715328467155e-06,
+      "loss": 2.1923,
+      "step": 194
+    },
+    {
+      "epoch": 0.2136986301369863,
+      "grad_norm": 3.064223527908325,
+      "learning_rate": 3.3029197080291973e-06,
+      "loss": 2.1707,
+      "step": 195
+    },
+    {
+      "epoch": 0.2147945205479452,
+      "grad_norm": 2.5138468742370605,
+      "learning_rate": 3.3211678832116788e-06,
+      "loss": 2.1933,
+      "step": 196
+    },
+    {
+      "epoch": 0.2158904109589041,
+      "grad_norm": 3.1076464653015137,
+      "learning_rate": 3.339416058394161e-06,
+      "loss": 2.1725,
+      "step": 197
+    },
+    {
+      "epoch": 0.21698630136986302,
+      "grad_norm": 2.7291388511657715,
+      "learning_rate": 3.3576642335766425e-06,
+      "loss": 2.1722,
+      "step": 198
+    },
+    {
+      "epoch": 0.21808219178082192,
+      "grad_norm": 2.805506467819214,
+      "learning_rate": 3.3759124087591243e-06,
+      "loss": 2.1688,
+      "step": 199
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "grad_norm": 2.6190085411071777,
+      "learning_rate": 3.3941605839416058e-06,
+      "loss": 2.1708,
+      "step": 200
+    },
+    {
+      "epoch": 0.2191780821917808,
+      "eval_loss": 2.295546054840088,
+      "eval_runtime": 37.0133,
+      "eval_samples_per_second": 27.017,
+      "eval_steps_per_second": 1.135,
+      "step": 200
+    },
+    {
+      "epoch": 0.22027397260273973,
+      "grad_norm": 3.1270673274993896,
+      "learning_rate": 3.412408759124088e-06,
+      "loss": 2.1759,
+      "step": 201
+    },
+    {
+      "epoch": 0.22136986301369863,
+      "grad_norm": 3.494178533554077,
+      "learning_rate": 3.43065693430657e-06,
+      "loss": 2.1795,
+      "step": 202
+    },
+    {
+      "epoch": 0.22246575342465755,
+      "grad_norm": 2.7398736476898193,
+      "learning_rate": 3.4489051094890513e-06,
+      "loss": 2.1201,
+      "step": 203
+    },
+    {
+      "epoch": 0.22356164383561644,
+      "grad_norm": 2.981506824493408,
+      "learning_rate": 3.467153284671533e-06,
+      "loss": 2.1693,
+      "step": 204
+    },
+    {
+      "epoch": 0.22465753424657534,
+      "grad_norm": 2.4129178524017334,
+      "learning_rate": 3.4854014598540146e-06,
+      "loss": 2.1125,
+      "step": 205
+    },
+    {
+      "epoch": 0.22575342465753426,
+      "grad_norm": 2.701197385787964,
+      "learning_rate": 3.503649635036497e-06,
+      "loss": 2.1565,
+      "step": 206
+    },
+    {
+      "epoch": 0.22684931506849315,
+      "grad_norm": 2.533306837081909,
+      "learning_rate": 3.5218978102189783e-06,
+      "loss": 2.1487,
+      "step": 207
+    },
+    {
+      "epoch": 0.22794520547945205,
+      "grad_norm": 2.8030788898468018,
+      "learning_rate": 3.54014598540146e-06,
+      "loss": 2.1801,
+      "step": 208
+    },
+    {
+      "epoch": 0.22904109589041097,
+      "grad_norm": 2.895414352416992,
+      "learning_rate": 3.5583941605839416e-06,
+      "loss": 2.189,
+      "step": 209
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "grad_norm": 2.508390426635742,
+      "learning_rate": 3.576642335766424e-06,
+      "loss": 2.2376,
+      "step": 210
+    },
+    {
+      "epoch": 0.23013698630136986,
+      "eval_loss": 2.2900516986846924,
+      "eval_runtime": 36.7645,
+      "eval_samples_per_second": 27.2,
+      "eval_steps_per_second": 1.142,
+      "step": 210
+    },
+    {
+      "epoch": 0.23123287671232876,
+      "grad_norm": 2.6393816471099854,
+      "learning_rate": 3.5948905109489054e-06,
+      "loss": 2.2267,
+      "step": 211
+    },
+    {
+      "epoch": 0.23232876712328768,
+      "grad_norm": 3.0687549114227295,
+      "learning_rate": 3.6131386861313872e-06,
+      "loss": 2.0744,
+      "step": 212
+    },
+    {
+      "epoch": 0.23342465753424657,
+      "grad_norm": 2.6342105865478516,
+      "learning_rate": 3.6313868613138687e-06,
+      "loss": 2.0737,
+      "step": 213
+    },
+    {
+      "epoch": 0.2345205479452055,
+      "grad_norm": 3.5662875175476074,
+      "learning_rate": 3.6496350364963505e-06,
+      "loss": 2.1147,
+      "step": 214
+    },
+    {
+      "epoch": 0.2356164383561644,
+      "grad_norm": 2.5357561111450195,
+      "learning_rate": 3.667883211678833e-06,
+      "loss": 2.1268,
+      "step": 215
+    },
+    {
+      "epoch": 0.23671232876712328,
+      "grad_norm": 3.439711093902588,
+      "learning_rate": 3.6861313868613142e-06,
+      "loss": 2.1118,
+      "step": 216
+    },
+    {
+      "epoch": 0.2378082191780822,
+      "grad_norm": 2.9410204887390137,
+      "learning_rate": 3.704379562043796e-06,
+      "loss": 2.1087,
+      "step": 217
+    },
+    {
+      "epoch": 0.2389041095890411,
+      "grad_norm": 3.3182294368743896,
+      "learning_rate": 3.7226277372262775e-06,
+      "loss": 2.196,
+      "step": 218
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 3.3834011554718018,
+      "learning_rate": 3.7408759124087594e-06,
+      "loss": 2.1633,
+      "step": 219
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "grad_norm": 3.114952325820923,
+      "learning_rate": 3.7591240875912412e-06,
+      "loss": 2.1107,
+      "step": 220
+    },
+    {
+      "epoch": 0.2410958904109589,
+      "eval_loss": 2.277580976486206,
+      "eval_runtime": 36.9846,
+      "eval_samples_per_second": 27.038,
+      "eval_steps_per_second": 1.136,
+      "step": 220
+    },
+    {
+      "epoch": 0.2421917808219178,
+      "grad_norm": 3.1151392459869385,
+      "learning_rate": 3.777372262773723e-06,
+      "loss": 2.1471,
+      "step": 221
+    },
+    {
+      "epoch": 0.2432876712328767,
+      "grad_norm": 3.798450469970703,
+      "learning_rate": 3.7956204379562045e-06,
+      "loss": 2.1324,
+      "step": 222
+    },
+    {
+      "epoch": 0.24438356164383562,
+      "grad_norm": 2.6187098026275635,
+      "learning_rate": 3.8138686131386864e-06,
+      "loss": 2.1061,
+      "step": 223
+    },
+    {
+      "epoch": 0.24547945205479452,
+      "grad_norm": 4.014018535614014,
+      "learning_rate": 3.832116788321168e-06,
+      "loss": 2.166,
+      "step": 224
+    },
+    {
+      "epoch": 0.2465753424657534,
+      "grad_norm": 3.0043938159942627,
+      "learning_rate": 3.85036496350365e-06,
+      "loss": 2.1197,
+      "step": 225
+    },
+    {
+      "epoch": 0.24767123287671233,
+      "grad_norm": 3.507066249847412,
+      "learning_rate": 3.868613138686132e-06,
+      "loss": 2.0848,
+      "step": 226
+    },
+    {
+      "epoch": 0.24876712328767123,
+      "grad_norm": 3.5713281631469727,
+      "learning_rate": 3.886861313868613e-06,
+      "loss": 2.1239,
+      "step": 227
+    },
+    {
+      "epoch": 0.24986301369863015,
+      "grad_norm": 3.961082696914673,
+      "learning_rate": 3.905109489051096e-06,
+      "loss": 2.1406,
+      "step": 228
+    },
+    {
+      "epoch": 0.25095890410958904,
+      "grad_norm": 3.1514859199523926,
+      "learning_rate": 3.923357664233577e-06,
+      "loss": 2.1689,
+      "step": 229
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "grad_norm": 3.8486833572387695,
+      "learning_rate": 3.9416058394160585e-06,
+      "loss": 2.1321,
+      "step": 230
+    },
+    {
+      "epoch": 0.25205479452054796,
+      "eval_loss": 2.2620601654052734,
+      "eval_runtime": 37.058,
+      "eval_samples_per_second": 26.985,
+      "eval_steps_per_second": 1.133,
+      "step": 230
+    },
+    {
+      "epoch": 0.25315068493150683,
+      "grad_norm": 2.5626907348632812,
+      "learning_rate": 3.95985401459854e-06,
+      "loss": 2.1777,
+      "step": 231
+    },
+    {
+      "epoch": 0.25424657534246575,
+      "grad_norm": 4.105278015136719,
+      "learning_rate": 3.978102189781022e-06,
+      "loss": 2.2099,
+      "step": 232
+    },
+    {
+      "epoch": 0.2553424657534247,
+      "grad_norm": 2.9818739891052246,
+      "learning_rate": 3.996350364963504e-06,
+      "loss": 2.0705,
+      "step": 233
+    },
+    {
+      "epoch": 0.25643835616438354,
+      "grad_norm": 2.87030291557312,
+      "learning_rate": 4.014598540145986e-06,
+      "loss": 2.1453,
+      "step": 234
+    },
+    {
+      "epoch": 0.25753424657534246,
+      "grad_norm": 2.4847915172576904,
+      "learning_rate": 4.032846715328467e-06,
+      "loss": 2.0461,
+      "step": 235
+    },
+    {
+      "epoch": 0.2586301369863014,
+      "grad_norm": 3.451420783996582,
+      "learning_rate": 4.05109489051095e-06,
+      "loss": 2.1133,
+      "step": 236
+    },
+    {
+      "epoch": 0.25972602739726025,
+      "grad_norm": 2.954723358154297,
+      "learning_rate": 4.069343065693431e-06,
+      "loss": 2.1316,
+      "step": 237
+    },
+    {
+      "epoch": 0.2608219178082192,
+      "grad_norm": 2.5295896530151367,
+      "learning_rate": 4.0875912408759126e-06,
+      "loss": 2.0901,
+      "step": 238
+    },
+    {
+      "epoch": 0.2619178082191781,
+      "grad_norm": 2.6359453201293945,
+      "learning_rate": 4.105839416058395e-06,
+      "loss": 2.117,
+      "step": 239
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "grad_norm": 2.5236105918884277,
+      "learning_rate": 4.124087591240876e-06,
+      "loss": 2.1958,
+      "step": 240
+    },
+    {
+      "epoch": 0.26301369863013696,
+      "eval_loss": 2.250295400619507,
+      "eval_runtime": 37.0811,
+      "eval_samples_per_second": 26.968,
+      "eval_steps_per_second": 1.133,
+      "step": 240
+    },
+    {
+      "epoch": 0.2641095890410959,
+      "grad_norm": 2.581918954849243,
+      "learning_rate": 4.1423357664233586e-06,
+      "loss": 2.1054,
+      "step": 241
+    },
+    {
+      "epoch": 0.2652054794520548,
+      "grad_norm": 2.501332998275757,
+      "learning_rate": 4.16058394160584e-06,
+      "loss": 2.0878,
+      "step": 242
+    },
+    {
+      "epoch": 0.2663013698630137,
+      "grad_norm": 2.6477770805358887,
+      "learning_rate": 4.1788321167883214e-06,
+      "loss": 2.1656,
+      "step": 243
+    },
+    {
+      "epoch": 0.2673972602739726,
+      "grad_norm": 2.7252438068389893,
+      "learning_rate": 4.197080291970803e-06,
+      "loss": 2.0497,
+      "step": 244
+    },
+    {
+      "epoch": 0.2684931506849315,
+      "grad_norm": 2.6056301593780518,
+      "learning_rate": 4.215328467153285e-06,
+      "loss": 2.1411,
+      "step": 245
+    },
+    {
+      "epoch": 0.26958904109589044,
+      "grad_norm": 2.6581931114196777,
+      "learning_rate": 4.233576642335767e-06,
+      "loss": 2.0829,
+      "step": 246
+    },
+    {
+      "epoch": 0.2706849315068493,
+      "grad_norm": 2.6316752433776855,
+      "learning_rate": 4.251824817518249e-06,
+      "loss": 2.0882,
+      "step": 247
+    },
+    {
+      "epoch": 0.2717808219178082,
+      "grad_norm": 2.803823947906494,
+      "learning_rate": 4.27007299270073e-06,
+      "loss": 2.1203,
+      "step": 248
+    },
+    {
+      "epoch": 0.27287671232876715,
+      "grad_norm": 2.9551174640655518,
+      "learning_rate": 4.288321167883212e-06,
+      "loss": 2.0897,
+      "step": 249
+    },
+    {
+      "epoch": 0.273972602739726,
+      "grad_norm": 2.4847161769866943,
+      "learning_rate": 4.306569343065693e-06,
+      "loss": 2.0845,
+      "step": 250
+    },
+    {
+      "epoch": 0.273972602739726,
+      "eval_loss": 2.2491259574890137,
+      "eval_runtime": 37.1024,
+      "eval_samples_per_second": 26.952,
+      "eval_steps_per_second": 1.132,
+      "step": 250
+    },
+    {
+      "epoch": 0.27506849315068493,
+      "grad_norm": 2.6398351192474365,
+      "learning_rate": 4.3248175182481755e-06,
+      "loss": 2.0479,
+      "step": 251
+    },
+    {
+      "epoch": 0.27616438356164386,
+      "grad_norm": 3.0948562622070312,
+      "learning_rate": 4.343065693430658e-06,
+      "loss": 2.1184,
+      "step": 252
+    },
+    {
+      "epoch": 0.2772602739726027,
+      "grad_norm": 2.5378012657165527,
+      "learning_rate": 4.361313868613139e-06,
+      "loss": 2.1099,
+      "step": 253
+    },
+    {
+      "epoch": 0.27835616438356164,
+      "grad_norm": 3.0282225608825684,
+      "learning_rate": 4.379562043795621e-06,
+      "loss": 2.0982,
+      "step": 254
+    },
+    {
+      "epoch": 0.27945205479452057,
+      "grad_norm": 2.6205334663391113,
+      "learning_rate": 4.397810218978102e-06,
+      "loss": 2.1285,
+      "step": 255
+    },
+    {
+      "epoch": 0.28054794520547943,
+      "grad_norm": 3.2056095600128174,
+      "learning_rate": 4.416058394160584e-06,
+      "loss": 2.1221,
+      "step": 256
+    },
+    {
+      "epoch": 0.28164383561643835,
+      "grad_norm": 3.1370127201080322,
+      "learning_rate": 4.434306569343066e-06,
+      "loss": 2.1016,
+      "step": 257
+    },
+    {
+      "epoch": 0.2827397260273973,
+      "grad_norm": 2.494168519973755,
+      "learning_rate": 4.452554744525548e-06,
+      "loss": 2.0754,
+      "step": 258
+    },
+    {
+      "epoch": 0.28383561643835614,
+      "grad_norm": 2.7979183197021484,
+      "learning_rate": 4.4708029197080295e-06,
+      "loss": 2.1218,
+      "step": 259
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "grad_norm": 2.755682945251465,
+      "learning_rate": 4.489051094890512e-06,
+      "loss": 2.1136,
+      "step": 260
+    },
+    {
+      "epoch": 0.28493150684931506,
+      "eval_loss": 2.2433817386627197,
+      "eval_runtime": 37.1121,
+      "eval_samples_per_second": 26.945,
+      "eval_steps_per_second": 1.132,
+      "step": 260
+    },
+    {
+      "epoch": 0.286027397260274,
+      "grad_norm": 2.3641891479492188,
+      "learning_rate": 4.507299270072993e-06,
+      "loss": 2.1488,
+      "step": 261
+    },
+    {
+      "epoch": 0.28712328767123285,
+      "grad_norm": 2.741292953491211,
+      "learning_rate": 4.525547445255475e-06,
+      "loss": 2.1048,
+      "step": 262
+    },
+    {
+      "epoch": 0.2882191780821918,
+      "grad_norm": 2.811483383178711,
+      "learning_rate": 4.543795620437956e-06,
+      "loss": 2.0996,
+      "step": 263
+    },
+    {
+      "epoch": 0.2893150684931507,
+      "grad_norm": 2.6705479621887207,
+      "learning_rate": 4.562043795620438e-06,
+      "loss": 2.0652,
+      "step": 264
+    },
+    {
+      "epoch": 0.29041095890410956,
+      "grad_norm": 2.9383747577667236,
+      "learning_rate": 4.580291970802921e-06,
+      "loss": 2.1574,
+      "step": 265
+    },
+    {
+      "epoch": 0.2915068493150685,
+      "grad_norm": 2.591373920440674,
+      "learning_rate": 4.598540145985402e-06,
+      "loss": 2.1655,
+      "step": 266
+    },
+    {
+      "epoch": 0.2926027397260274,
+      "grad_norm": 3.16955828666687,
+      "learning_rate": 4.6167883211678835e-06,
+      "loss": 2.1285,
+      "step": 267
+    },
+    {
+      "epoch": 0.2936986301369863,
+      "grad_norm": 2.66941499710083,
+      "learning_rate": 4.635036496350365e-06,
+      "loss": 2.0777,
+      "step": 268
+    },
+    {
+      "epoch": 0.2947945205479452,
+      "grad_norm": 2.768139600753784,
+      "learning_rate": 4.653284671532847e-06,
+      "loss": 2.0748,
+      "step": 269
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "grad_norm": 2.5951924324035645,
+      "learning_rate": 4.671532846715329e-06,
+      "loss": 2.1289,
+      "step": 270
+    },
+    {
+      "epoch": 0.2958904109589041,
+      "eval_loss": 2.2393550872802734,
+      "eval_runtime": 37.0993,
+      "eval_samples_per_second": 26.955,
+      "eval_steps_per_second": 1.132,
+      "step": 270
+    },
+    {
+      "epoch": 0.29698630136986304,
+      "grad_norm": 2.5422160625457764,
+      "learning_rate": 4.689781021897811e-06,
+      "loss": 2.1342,
+      "step": 271
+    },
+    {
+      "epoch": 0.2980821917808219,
+      "grad_norm": 2.540555477142334,
+      "learning_rate": 4.708029197080292e-06,
+      "loss": 2.097,
+      "step": 272
+    },
+    {
+      "epoch": 0.2991780821917808,
+      "grad_norm": 2.995366334915161,
+      "learning_rate": 4.726277372262774e-06,
+      "loss": 2.1452,
+      "step": 273
+    },
+    {
+      "epoch": 0.30027397260273975,
+      "grad_norm": 2.2561020851135254,
+      "learning_rate": 4.744525547445255e-06,
+      "loss": 2.1069,
+      "step": 274
+    },
+    {
+      "epoch": 0.3013698630136986,
+      "grad_norm": 2.53476619720459,
+      "learning_rate": 4.7627737226277375e-06,
+      "loss": 2.0381,
+      "step": 275
+    },
+    {
+      "epoch": 0.30246575342465754,
+      "grad_norm": 2.3415749073028564,
+      "learning_rate": 4.78102189781022e-06,
+      "loss": 2.1027,
+      "step": 276
+    },
+    {
+      "epoch": 0.30356164383561646,
+      "grad_norm": 2.9756431579589844,
+      "learning_rate": 4.799270072992701e-06,
+      "loss": 2.1136,
+      "step": 277
+    },
+    {
+      "epoch": 0.3046575342465753,
+      "grad_norm": 2.606208324432373,
+      "learning_rate": 4.8175182481751835e-06,
+      "loss": 2.0797,
+      "step": 278
+    },
+    {
+      "epoch": 0.30575342465753425,
+      "grad_norm": 2.5089313983917236,
+      "learning_rate": 4.835766423357665e-06,
+      "loss": 2.1026,
+      "step": 279
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "grad_norm": 2.5003838539123535,
+      "learning_rate": 4.854014598540146e-06,
+      "loss": 2.0923,
+      "step": 280
+    },
+    {
+      "epoch": 0.30684931506849317,
+      "eval_loss": 2.2351036071777344,
+      "eval_runtime": 37.111,
+      "eval_samples_per_second": 26.946,
+      "eval_steps_per_second": 1.132,
+      "step": 280
+    },
+    {
+      "epoch": 0.30794520547945203,
+      "grad_norm": 2.403602361679077,
+      "learning_rate": 4.872262773722628e-06,
+      "loss": 2.0388,
+      "step": 281
+    },
+    {
+      "epoch": 0.30904109589041096,
+      "grad_norm": 2.7775959968566895,
+      "learning_rate": 4.89051094890511e-06,
+      "loss": 2.0561,
+      "step": 282
+    },
+    {
+      "epoch": 0.3101369863013699,
+      "grad_norm": 2.7128653526306152,
+      "learning_rate": 4.9087591240875915e-06,
+      "loss": 2.042,
+      "step": 283
+    },
+    {
+      "epoch": 0.31123287671232874,
+      "grad_norm": 2.7029991149902344,
+      "learning_rate": 4.927007299270074e-06,
+      "loss": 2.0814,
+      "step": 284
+    },
+    {
+      "epoch": 0.31232876712328766,
+      "grad_norm": 2.3942933082580566,
+      "learning_rate": 4.945255474452555e-06,
+      "loss": 2.0906,
+      "step": 285
+    },
+    {
+      "epoch": 0.3134246575342466,
+      "grad_norm": 2.5086541175842285,
+      "learning_rate": 4.963503649635037e-06,
+      "loss": 2.0625,
+      "step": 286
+    },
+    {
+      "epoch": 0.31452054794520545,
+      "grad_norm": 2.6668760776519775,
+      "learning_rate": 4.981751824817518e-06,
+      "loss": 2.1538,
+      "step": 287
+    },
+    {
+      "epoch": 0.3156164383561644,
+      "grad_norm": 2.5066750049591064,
+      "learning_rate": 5e-06,
+      "loss": 2.1423,
+      "step": 288
+    },
+    {
+      "epoch": 0.3167123287671233,
+      "grad_norm": 2.84967041015625,
+      "learning_rate": 4.997969130787978e-06,
+      "loss": 2.0943,
+      "step": 289
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "grad_norm": 2.151946544647217,
+      "learning_rate": 4.995938261575955e-06,
+      "loss": 2.1098,
+      "step": 290
+    },
+    {
+      "epoch": 0.3178082191780822,
+      "eval_loss": 2.223247528076172,
+      "eval_runtime": 37.1755,
+      "eval_samples_per_second": 26.899,
+      "eval_steps_per_second": 1.13,
+      "step": 290
+    },
+    {
+      "epoch": 0.3189041095890411,
+      "grad_norm": 2.887166976928711,
+      "learning_rate": 4.993907392363932e-06,
+      "loss": 2.0281,
+      "step": 291
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 2.5930211544036865,
+      "learning_rate": 4.9918765231519095e-06,
+      "loss": 2.1063,
+      "step": 292
+    },
+    {
+      "epoch": 0.32109589041095893,
+      "grad_norm": 2.5568413734436035,
+      "learning_rate": 4.989845653939887e-06,
+      "loss": 2.0313,
+      "step": 293
+    },
+    {
+      "epoch": 0.3221917808219178,
+      "grad_norm": 2.536095380783081,
+      "learning_rate": 4.987814784727864e-06,
+      "loss": 2.0749,
+      "step": 294
+    },
+    {
+      "epoch": 0.3232876712328767,
+      "grad_norm": 2.3385651111602783,
+      "learning_rate": 4.985783915515841e-06,
+      "loss": 2.0666,
+      "step": 295
+    },
+    {
+      "epoch": 0.32438356164383564,
+      "grad_norm": 2.504826068878174,
+      "learning_rate": 4.983753046303819e-06,
+      "loss": 2.0677,
+      "step": 296
+    },
+    {
+      "epoch": 0.3254794520547945,
+      "grad_norm": 2.5289409160614014,
+      "learning_rate": 4.981722177091796e-06,
+      "loss": 2.0828,
+      "step": 297
+    },
+    {
+      "epoch": 0.3265753424657534,
+      "grad_norm": 2.803795099258423,
+      "learning_rate": 4.979691307879773e-06,
+      "loss": 2.0733,
+      "step": 298
+    },
+    {
+      "epoch": 0.32767123287671235,
+      "grad_norm": 2.651273012161255,
+      "learning_rate": 4.9776604386677504e-06,
+      "loss": 2.0987,
+      "step": 299
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "grad_norm": 2.4715864658355713,
+      "learning_rate": 4.975629569455728e-06,
+      "loss": 2.0733,
+      "step": 300
+    },
+    {
+      "epoch": 0.3287671232876712,
+      "eval_loss": 2.214858293533325,
+      "eval_runtime": 37.0559,
+      "eval_samples_per_second": 26.986,
+      "eval_steps_per_second": 1.133,
+      "step": 300
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 2736,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 9.35192822808576e+16,
+  "train_batch_size": 3,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/training_args.bin b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..b7adbe5a2dbf27f05befec6cc65fbc93e90eefd0
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aaa882279512a8c278376c0c601206933f3c5662ddcd3c13c53bf32ae3680d14
+size 6011
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/zero_to_fp32.py b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/zero_to_fp32.py
new file mode 100644
index 0000000000000000000000000000000000000000..24cc342e78d1a006c782b3a4cd68d9ce786d8fd8
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-300/zero_to_fp32.py
@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-500/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-500/model-00002-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..00b3aa29cb6d87c61d88afe0d8bf7d2d05e18f19
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-500/model-00002-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ebc7519827f5e3ad8e8016bc7d558e0f0ff7846653ddac979ba6b4874e7ffe1c
+size 2247734920
diff --git a/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-700/model-00002-of-00002.safetensors b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-700/model-00002-of-00002.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..207a11a44b4b69d8c006254fd06424e8c163dd02
--- /dev/null
+++ b/checkpoints/Llama-3.2-3B/babylm_shuffle_deterministic84_10M_seed0/runs/checkpoint-700/model-00002-of-00002.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:916c33b29e73f3ddc4812a70d63282ffe05f3ec1ceae2b8f8596193a184212f4
+size 2247734920