Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "/home/ubuntu/tmp/models/tempesthenno-nuslerp-001",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 5120,
+  "initializer_range": 0.02,
+  "intermediate_size": 13824,
+  "max_position_embeddings": 131072,
+  "max_window_layers": 48,
+  "model_type": "qwen2",
+  "num_attention_heads": 40,
+  "num_hidden_layers": 48,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.46.1",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151665
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "transformers_version": "4.46.1"
+}

latest

@@ -0,0 +1 @@
+global_step400

model-00001-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcc08b2fc727b1196be1e317996ce1fe276cf108d3250058a41a558ed98e7793
+size 4899283440

model-00002-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:205481265ead77a0130d89f9b27014bc1e8aff5ff07e78e821c49342493be8de
+size 4954847384

model-00003-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c5e574e69ca270084d6a281d849732ff5f00b7e0acd89c7495fc36971572e6c
+size 4954847376

model-00004-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3360e7803596dcdce9900aae4b779a3416f652ba99a1ce2810787925e66c2659
+size 4954847376

model-00005-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8dfd29c16488d8a3edca37ccffd1115cc709ebd1a40131db1d923946a36f5fa2
+size 4954847376

model-00006-of-00006.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd52f323b90b7064b4d11cbbfa0acfa4e6cdf859679f3b79f076a9db860eca68
+size 4813289432

model.safetensors.index.json

@@ -0,0 +1,586 @@
+{
+  "metadata": {
+    "total_size": 29531895808
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00001-of-00006.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.10.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00001-of-00006.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00001-of-00006.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.11.self_attn.k_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.11.self_attn.q_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.11.self_attn.v_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.12.self_attn.k_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.12.self_attn.q_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.12.self_attn.v_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.13.self_attn.k_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.13.self_attn.q_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.13.self_attn.v_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.14.self_attn.k_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.14.self_attn.q_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.14.self_attn.v_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.15.self_attn.k_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.15.self_attn.q_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.15.self_attn.v_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.16.self_attn.k_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.16.self_attn.q_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.16.self_attn.v_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.17.self_attn.k_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.17.self_attn.q_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.17.self_attn.v_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.18.self_attn.k_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.18.self_attn.q_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.18.self_attn.v_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.19.self_attn.k_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.19.self_attn.q_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.19.self_attn.v_proj.bias": "model-00002-of-00006.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00002-of-00006.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.2.self_attn.k_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.2.self_attn.q_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.2.self_attn.v_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.20.self_attn.k_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.20.self_attn.q_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.20.self_attn.v_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.21.self_attn.k_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.21.self_attn.q_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.21.self_attn.v_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.22.self_attn.k_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.22.self_attn.q_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.22.self_attn.v_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.23.self_attn.k_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.23.self_attn.q_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.23.self_attn.v_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.24.self_attn.k_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.24.self_attn.q_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.24.self_attn.v_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.25.self_attn.k_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.25.self_attn.q_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.25.self_attn.v_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.26.self_attn.k_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.26.self_attn.q_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.26.self_attn.v_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.27.self_attn.k_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.27.self_attn.q_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.27.self_attn.v_proj.bias": "model-00003-of-00006.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00003-of-00006.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.28.self_attn.k_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.28.self_attn.q_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.28.self_attn.v_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.29.self_attn.k_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.29.self_attn.q_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.29.self_attn.v_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.3.self_attn.k_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.3.self_attn.q_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.3.self_attn.v_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.30.self_attn.k_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.30.self_attn.q_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.30.self_attn.v_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.31.self_attn.k_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.31.self_attn.q_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.31.self_attn.v_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.32.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.32.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.32.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.32.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.32.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.32.self_attn.k_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.32.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.32.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.32.self_attn.q_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.32.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.32.self_attn.v_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.32.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.33.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.33.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.33.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.33.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.33.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.33.self_attn.k_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.33.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.33.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.33.self_attn.q_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.33.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.33.self_attn.v_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.33.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.34.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.34.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.34.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.34.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.34.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.34.self_attn.k_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.34.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.34.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.34.self_attn.q_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.34.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.34.self_attn.v_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.34.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.35.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.35.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.35.mlp.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.35.mlp.up_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.35.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.35.self_attn.k_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.35.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.35.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.35.self_attn.q_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.35.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.35.self_attn.v_proj.bias": "model-00004-of-00006.safetensors",
+    "model.layers.35.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.36.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "model.layers.36.mlp.down_proj.weight": "model-00004-of-00006.safetensors",
+    "model.layers.36.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.36.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.36.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.36.self_attn.k_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.36.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.36.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.36.self_attn.q_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.36.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.36.self_attn.v_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.36.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.37.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.37.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.37.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.37.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.37.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.37.self_attn.k_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.37.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.37.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.37.self_attn.q_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.37.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.37.self_attn.v_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.37.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.38.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.38.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.38.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.38.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.38.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.38.self_attn.k_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.38.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.38.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.38.self_attn.q_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.38.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.38.self_attn.v_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.38.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.39.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.39.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.39.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.39.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.39.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.39.self_attn.k_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.39.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.39.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.39.self_attn.q_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.39.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.39.self_attn.v_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.39.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.4.self_attn.k_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.4.self_attn.q_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.4.self_attn.v_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.40.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.40.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.40.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.40.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.40.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.40.self_attn.k_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.40.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.40.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.40.self_attn.q_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.40.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.40.self_attn.v_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.40.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.41.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.41.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.41.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.41.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.41.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.41.self_attn.k_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.41.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.41.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.41.self_attn.q_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.41.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.41.self_attn.v_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.41.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.42.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.42.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.42.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.42.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.42.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.42.self_attn.k_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.42.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.42.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.42.self_attn.q_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.42.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.42.self_attn.v_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.42.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.43.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.43.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.43.mlp.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.43.mlp.up_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.43.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.43.self_attn.k_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.43.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.43.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.43.self_attn.q_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.43.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.43.self_attn.v_proj.bias": "model-00005-of-00006.safetensors",
+    "model.layers.43.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.44.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "model.layers.44.mlp.down_proj.weight": "model-00005-of-00006.safetensors",
+    "model.layers.44.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.44.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.44.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.44.self_attn.k_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.44.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.44.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.44.self_attn.q_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.44.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.44.self_attn.v_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.44.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.45.input_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.45.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.45.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.45.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.45.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.45.self_attn.k_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.45.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.45.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.45.self_attn.q_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.45.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.45.self_attn.v_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.45.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.46.input_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.46.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.46.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.46.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.46.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.46.self_attn.k_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.46.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.46.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.46.self_attn.q_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.46.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.46.self_attn.v_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.46.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.47.input_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.47.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.47.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.47.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.47.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.47.self_attn.k_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.47.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.47.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.47.self_attn.q_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.47.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.47.self_attn.v_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.47.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.5.self_attn.k_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.5.self_attn.q_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.5.self_attn.v_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.6.self_attn.k_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.6.self_attn.q_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.6.self_attn.v_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.7.self_attn.k_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.7.self_attn.q_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.7.self_attn.v_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.8.self_attn.k_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.8.self_attn.q_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.8.self_attn.v_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "model.layers.9.self_attn.k_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.9.self_attn.q_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00006-of-00006.safetensors",
+    "model.layers.9.self_attn.v_proj.bias": "model-00006-of-00006.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00006-of-00006.safetensors",
+    "model.norm.weight": "model-00006-of-00006.safetensors"
+  }
+}

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b29cea827e2ef52791356c5b22c98701f253e4d2ddceeda317c7f482e7f56e54
+size 1064

special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

tokenizer_config.json

@@ -0,0 +1,209 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% endif %}{% if system_message is defined %}{{ '<|im_start|>system\n' + system_message + '<|im_end|>\n' }}{% endif %}{% for message in loop_messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|im_start|>user\n' + content + '<|im_end|>\n<|im_start|>assistant\n' }}{% elif message['role'] == 'assistant' %}{{ content + '<|im_end|>\n' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 2048,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

trainer_state.json

@@ -0,0 +1,3013 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.4036326942482341,
+  "eval_steps": 20,
+  "global_step": 400,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0010090817356205853,
+      "grad_norm": 48.721710205078125,
+      "learning_rate": 8e-08,
+      "loss": 0.8681,
+      "step": 1
+    },
+    {
+      "epoch": 0.0020181634712411706,
+      "grad_norm": 44.42015075683594,
+      "learning_rate": 1.6e-07,
+      "loss": 0.6615,
+      "step": 2
+    },
+    {
+      "epoch": 0.0030272452068617556,
+      "grad_norm": 42.49191665649414,
+      "learning_rate": 2.4e-07,
+      "loss": 0.6742,
+      "step": 3
+    },
+    {
+      "epoch": 0.004036326942482341,
+      "grad_norm": 40.36923599243164,
+      "learning_rate": 3.2e-07,
+      "loss": 0.7824,
+      "step": 4
+    },
+    {
+      "epoch": 0.005045408678102927,
+      "grad_norm": 57.19468688964844,
+      "learning_rate": 4e-07,
+      "loss": 1.0009,
+      "step": 5
+    },
+    {
+      "epoch": 0.006054490413723511,
+      "grad_norm": 48.78745651245117,
+      "learning_rate": 4.8e-07,
+      "loss": 0.7613,
+      "step": 6
+    },
+    {
+      "epoch": 0.007063572149344097,
+      "grad_norm": 37.92594528198242,
+      "learning_rate": 5.6e-07,
+      "loss": 0.6615,
+      "step": 7
+    },
+    {
+      "epoch": 0.008072653884964682,
+      "grad_norm": 37.87478256225586,
+      "learning_rate": 6.4e-07,
+      "loss": 0.6779,
+      "step": 8
+    },
+    {
+      "epoch": 0.009081735620585268,
+      "grad_norm": 43.225589752197266,
+      "learning_rate": 7.2e-07,
+      "loss": 0.796,
+      "step": 9
+    },
+    {
+      "epoch": 0.010090817356205853,
+      "grad_norm": 43.1926383972168,
+      "learning_rate": 8e-07,
+      "loss": 0.8526,
+      "step": 10
+    },
+    {
+      "epoch": 0.011099899091826439,
+      "grad_norm": 42.08029556274414,
+      "learning_rate": 8.799999999999999e-07,
+      "loss": 0.6605,
+      "step": 11
+    },
+    {
+      "epoch": 0.012108980827447022,
+      "grad_norm": 49.14653396606445,
+      "learning_rate": 9.6e-07,
+      "loss": 0.6774,
+      "step": 12
+    },
+    {
+      "epoch": 0.013118062563067608,
+      "grad_norm": 47.238338470458984,
+      "learning_rate": 1.04e-06,
+      "loss": 0.7313,
+      "step": 13
+    },
+    {
+      "epoch": 0.014127144298688193,
+      "grad_norm": 37.998931884765625,
+      "learning_rate": 1.12e-06,
+      "loss": 0.6244,
+      "step": 14
+    },
+    {
+      "epoch": 0.015136226034308779,
+      "grad_norm": 37.945064544677734,
+      "learning_rate": 1.2e-06,
+      "loss": 0.5214,
+      "step": 15
+    },
+    {
+      "epoch": 0.016145307769929364,
+      "grad_norm": 33.03001403808594,
+      "learning_rate": 1.28e-06,
+      "loss": 0.5042,
+      "step": 16
+    },
+    {
+      "epoch": 0.017154389505549948,
+      "grad_norm": 32.11518096923828,
+      "learning_rate": 1.3600000000000001e-06,
+      "loss": 0.3594,
+      "step": 17
+    },
+    {
+      "epoch": 0.018163471241170535,
+      "grad_norm": 30.64830207824707,
+      "learning_rate": 1.44e-06,
+      "loss": 0.3533,
+      "step": 18
+    },
+    {
+      "epoch": 0.01917255297679112,
+      "grad_norm": 41.34456253051758,
+      "learning_rate": 1.5199999999999998e-06,
+      "loss": 0.6093,
+      "step": 19
+    },
+    {
+      "epoch": 0.020181634712411706,
+      "grad_norm": 35.51686477661133,
+      "learning_rate": 1.6e-06,
+      "loss": 0.3241,
+      "step": 20
+    },
+    {
+      "epoch": 0.020181634712411706,
+      "eval_accuracy": 0.7758229284903518,
+      "eval_loss": 0.3856516480445862,
+      "eval_runtime": 63.0132,
+      "eval_samples_per_second": 27.962,
+      "eval_steps_per_second": 3.507,
+      "step": 20
+    },
+    {
+      "epoch": 0.02119071644803229,
+      "grad_norm": 36.574764251708984,
+      "learning_rate": 1.6799999999999998e-06,
+      "loss": 0.3611,
+      "step": 21
+    },
+    {
+      "epoch": 0.022199798183652877,
+      "grad_norm": 42.98014831542969,
+      "learning_rate": 1.7599999999999999e-06,
+      "loss": 0.5187,
+      "step": 22
+    },
+    {
+      "epoch": 0.02320887991927346,
+      "grad_norm": 32.25895690917969,
+      "learning_rate": 1.84e-06,
+      "loss": 0.3201,
+      "step": 23
+    },
+    {
+      "epoch": 0.024217961654894045,
+      "grad_norm": 45.763553619384766,
+      "learning_rate": 1.92e-06,
+      "loss": 0.3977,
+      "step": 24
+    },
+    {
+      "epoch": 0.025227043390514632,
+      "grad_norm": 41.3906135559082,
+      "learning_rate": 2e-06,
+      "loss": 0.3863,
+      "step": 25
+    },
+    {
+      "epoch": 0.026236125126135216,
+      "grad_norm": 55.49371337890625,
+      "learning_rate": 2.08e-06,
+      "loss": 0.4371,
+      "step": 26
+    },
+    {
+      "epoch": 0.027245206861755803,
+      "grad_norm": 49.18912124633789,
+      "learning_rate": 2.16e-06,
+      "loss": 0.4072,
+      "step": 27
+    },
+    {
+      "epoch": 0.028254288597376387,
+      "grad_norm": 53.16490173339844,
+      "learning_rate": 2.24e-06,
+      "loss": 0.5366,
+      "step": 28
+    },
+    {
+      "epoch": 0.029263370332996974,
+      "grad_norm": 58.15668869018555,
+      "learning_rate": 2.32e-06,
+      "loss": 0.5429,
+      "step": 29
+    },
+    {
+      "epoch": 0.030272452068617558,
+      "grad_norm": 32.979671478271484,
+      "learning_rate": 2.4e-06,
+      "loss": 0.2946,
+      "step": 30
+    },
+    {
+      "epoch": 0.03128153380423814,
+      "grad_norm": 40.131385803222656,
+      "learning_rate": 2.48e-06,
+      "loss": 0.298,
+      "step": 31
+    },
+    {
+      "epoch": 0.03229061553985873,
+      "grad_norm": 45.999542236328125,
+      "learning_rate": 2.56e-06,
+      "loss": 0.27,
+      "step": 32
+    },
+    {
+      "epoch": 0.033299697275479316,
+      "grad_norm": 32.2443962097168,
+      "learning_rate": 2.64e-06,
+      "loss": 0.2337,
+      "step": 33
+    },
+    {
+      "epoch": 0.034308779011099896,
+      "grad_norm": 28.031023025512695,
+      "learning_rate": 2.7200000000000002e-06,
+      "loss": 0.1835,
+      "step": 34
+    },
+    {
+      "epoch": 0.035317860746720484,
+      "grad_norm": 22.68515396118164,
+      "learning_rate": 2.8e-06,
+      "loss": 0.1522,
+      "step": 35
+    },
+    {
+      "epoch": 0.03632694248234107,
+      "grad_norm": 38.576019287109375,
+      "learning_rate": 2.88e-06,
+      "loss": 0.3371,
+      "step": 36
+    },
+    {
+      "epoch": 0.03733602421796166,
+      "grad_norm": 23.268003463745117,
+      "learning_rate": 2.96e-06,
+      "loss": 0.161,
+      "step": 37
+    },
+    {
+      "epoch": 0.03834510595358224,
+      "grad_norm": 43.69157409667969,
+      "learning_rate": 3.0399999999999997e-06,
+      "loss": 0.4692,
+      "step": 38
+    },
+    {
+      "epoch": 0.039354187689202826,
+      "grad_norm": 33.641780853271484,
+      "learning_rate": 3.1199999999999998e-06,
+      "loss": 0.3815,
+      "step": 39
+    },
+    {
+      "epoch": 0.04036326942482341,
+      "grad_norm": 32.81157302856445,
+      "learning_rate": 3.2e-06,
+      "loss": 0.3342,
+      "step": 40
+    },
+    {
+      "epoch": 0.04036326942482341,
+      "eval_accuracy": 0.8359818388195233,
+      "eval_loss": 0.30243048071861267,
+      "eval_runtime": 62.0758,
+      "eval_samples_per_second": 28.385,
+      "eval_steps_per_second": 3.56,
+      "step": 40
+    },
+    {
+      "epoch": 0.04137235116044399,
+      "grad_norm": 38.70158004760742,
+      "learning_rate": 3.2799999999999995e-06,
+      "loss": 0.3523,
+      "step": 41
+    },
+    {
+      "epoch": 0.04238143289606458,
+      "grad_norm": 36.827938079833984,
+      "learning_rate": 3.3599999999999996e-06,
+      "loss": 0.3339,
+      "step": 42
+    },
+    {
+      "epoch": 0.04339051463168517,
+      "grad_norm": 40.072757720947266,
+      "learning_rate": 3.4399999999999997e-06,
+      "loss": 0.3718,
+      "step": 43
+    },
+    {
+      "epoch": 0.044399596367305755,
+      "grad_norm": 17.283546447753906,
+      "learning_rate": 3.5199999999999998e-06,
+      "loss": 0.1344,
+      "step": 44
+    },
+    {
+      "epoch": 0.045408678102926335,
+      "grad_norm": 38.25215148925781,
+      "learning_rate": 3.6e-06,
+      "loss": 0.3276,
+      "step": 45
+    },
+    {
+      "epoch": 0.04641775983854692,
+      "grad_norm": 31.07087516784668,
+      "learning_rate": 3.68e-06,
+      "loss": 0.3091,
+      "step": 46
+    },
+    {
+      "epoch": 0.04742684157416751,
+      "grad_norm": 29.617158889770508,
+      "learning_rate": 3.7599999999999996e-06,
+      "loss": 0.2113,
+      "step": 47
+    },
+    {
+      "epoch": 0.04843592330978809,
+      "grad_norm": 37.85402297973633,
+      "learning_rate": 3.84e-06,
+      "loss": 0.3061,
+      "step": 48
+    },
+    {
+      "epoch": 0.04944500504540868,
+      "grad_norm": 92.3378677368164,
+      "learning_rate": 3.92e-06,
+      "loss": 0.8413,
+      "step": 49
+    },
+    {
+      "epoch": 0.050454086781029264,
+      "grad_norm": 25.68213653564453,
+      "learning_rate": 4e-06,
+      "loss": 0.2466,
+      "step": 50
+    },
+    {
+      "epoch": 0.05146316851664985,
+      "grad_norm": 22.958641052246094,
+      "learning_rate": 4.08e-06,
+      "loss": 0.2127,
+      "step": 51
+    },
+    {
+      "epoch": 0.05247225025227043,
+      "grad_norm": 26.674156188964844,
+      "learning_rate": 4.16e-06,
+      "loss": 0.5098,
+      "step": 52
+    },
+    {
+      "epoch": 0.05348133198789102,
+      "grad_norm": 18.660629272460938,
+      "learning_rate": 4.24e-06,
+      "loss": 0.2416,
+      "step": 53
+    },
+    {
+      "epoch": 0.054490413723511606,
+      "grad_norm": 20.310880661010742,
+      "learning_rate": 4.32e-06,
+      "loss": 0.3086,
+      "step": 54
+    },
+    {
+      "epoch": 0.055499495459132187,
+      "grad_norm": 20.308359146118164,
+      "learning_rate": 4.4e-06,
+      "loss": 0.3001,
+      "step": 55
+    },
+    {
+      "epoch": 0.056508577194752774,
+      "grad_norm": 13.419134140014648,
+      "learning_rate": 4.48e-06,
+      "loss": 0.1779,
+      "step": 56
+    },
+    {
+      "epoch": 0.05751765893037336,
+      "grad_norm": 14.848164558410645,
+      "learning_rate": 4.5599999999999995e-06,
+      "loss": 0.1883,
+      "step": 57
+    },
+    {
+      "epoch": 0.05852674066599395,
+      "grad_norm": 23.216663360595703,
+      "learning_rate": 4.64e-06,
+      "loss": 0.3184,
+      "step": 58
+    },
+    {
+      "epoch": 0.05953582240161453,
+      "grad_norm": 65.48796081542969,
+      "learning_rate": 4.72e-06,
+      "loss": 0.3397,
+      "step": 59
+    },
+    {
+      "epoch": 0.060544904137235116,
+      "grad_norm": 97.63284301757812,
+      "learning_rate": 4.8e-06,
+      "loss": 0.3936,
+      "step": 60
+    },
+    {
+      "epoch": 0.060544904137235116,
+      "eval_accuracy": 0.7474460839954598,
+      "eval_loss": 0.4690244793891907,
+      "eval_runtime": 62.2233,
+      "eval_samples_per_second": 28.317,
+      "eval_steps_per_second": 3.552,
+      "step": 60
+    },
+    {
+      "epoch": 0.0615539858728557,
+      "grad_norm": 210.90390014648438,
+      "learning_rate": 4.88e-06,
+      "loss": 0.6895,
+      "step": 61
+    },
+    {
+      "epoch": 0.06256306760847628,
+      "grad_norm": 46.58690643310547,
+      "learning_rate": 4.96e-06,
+      "loss": 0.29,
+      "step": 62
+    },
+    {
+      "epoch": 0.06357214934409687,
+      "grad_norm": 31.453752517700195,
+      "learning_rate": 5.04e-06,
+      "loss": 0.3071,
+      "step": 63
+    },
+    {
+      "epoch": 0.06458123107971746,
+      "grad_norm": 20.283885955810547,
+      "learning_rate": 5.12e-06,
+      "loss": 0.3571,
+      "step": 64
+    },
+    {
+      "epoch": 0.06559031281533804,
+      "grad_norm": 42.37957763671875,
+      "learning_rate": 5.2e-06,
+      "loss": 0.5532,
+      "step": 65
+    },
+    {
+      "epoch": 0.06659939455095863,
+      "grad_norm": 21.637910842895508,
+      "learning_rate": 5.28e-06,
+      "loss": 0.262,
+      "step": 66
+    },
+    {
+      "epoch": 0.06760847628657922,
+      "grad_norm": 18.54781723022461,
+      "learning_rate": 5.36e-06,
+      "loss": 0.2257,
+      "step": 67
+    },
+    {
+      "epoch": 0.06861755802219979,
+      "grad_norm": 20.521879196166992,
+      "learning_rate": 5.4400000000000004e-06,
+      "loss": 0.3857,
+      "step": 68
+    },
+    {
+      "epoch": 0.06962663975782038,
+      "grad_norm": 29.46757698059082,
+      "learning_rate": 5.52e-06,
+      "loss": 0.1029,
+      "step": 69
+    },
+    {
+      "epoch": 0.07063572149344097,
+      "grad_norm": 29.944223403930664,
+      "learning_rate": 5.6e-06,
+      "loss": 0.2712,
+      "step": 70
+    },
+    {
+      "epoch": 0.07164480322906155,
+      "grad_norm": 24.391925811767578,
+      "learning_rate": 5.68e-06,
+      "loss": 0.3976,
+      "step": 71
+    },
+    {
+      "epoch": 0.07265388496468214,
+      "grad_norm": 18.12702751159668,
+      "learning_rate": 5.76e-06,
+      "loss": 0.3472,
+      "step": 72
+    },
+    {
+      "epoch": 0.07366296670030273,
+      "grad_norm": 10.216832160949707,
+      "learning_rate": 5.84e-06,
+      "loss": 0.2769,
+      "step": 73
+    },
+    {
+      "epoch": 0.07467204843592332,
+      "grad_norm": 11.782180786132812,
+      "learning_rate": 5.92e-06,
+      "loss": 0.3737,
+      "step": 74
+    },
+    {
+      "epoch": 0.07568113017154389,
+      "grad_norm": 33.01165771484375,
+      "learning_rate": 6e-06,
+      "loss": 0.2062,
+      "step": 75
+    },
+    {
+      "epoch": 0.07669021190716448,
+      "grad_norm": 16.44459342956543,
+      "learning_rate": 6.079999999999999e-06,
+      "loss": 0.4451,
+      "step": 76
+    },
+    {
+      "epoch": 0.07769929364278506,
+      "grad_norm": 8.65665340423584,
+      "learning_rate": 6.1599999999999995e-06,
+      "loss": 0.2799,
+      "step": 77
+    },
+    {
+      "epoch": 0.07870837537840565,
+      "grad_norm": 9.738832473754883,
+      "learning_rate": 6.2399999999999995e-06,
+      "loss": 0.2986,
+      "step": 78
+    },
+    {
+      "epoch": 0.07971745711402624,
+      "grad_norm": 5.850776195526123,
+      "learning_rate": 6.32e-06,
+      "loss": 0.2548,
+      "step": 79
+    },
+    {
+      "epoch": 0.08072653884964683,
+      "grad_norm": 11.245881080627441,
+      "learning_rate": 6.4e-06,
+      "loss": 0.3249,
+      "step": 80
+    },
+    {
+      "epoch": 0.08072653884964683,
+      "eval_accuracy": 0.8569807037457434,
+      "eval_loss": 0.2707608640193939,
+      "eval_runtime": 62.6893,
+      "eval_samples_per_second": 28.107,
+      "eval_steps_per_second": 3.525,
+      "step": 80
+    },
+    {
+      "epoch": 0.08173562058526741,
+      "grad_norm": 9.138853073120117,
+      "learning_rate": 6.48e-06,
+      "loss": 0.2931,
+      "step": 81
+    },
+    {
+      "epoch": 0.08274470232088799,
+      "grad_norm": 7.408812999725342,
+      "learning_rate": 6.559999999999999e-06,
+      "loss": 0.276,
+      "step": 82
+    },
+    {
+      "epoch": 0.08375378405650857,
+      "grad_norm": 5.13024377822876,
+      "learning_rate": 6.639999999999999e-06,
+      "loss": 0.1842,
+      "step": 83
+    },
+    {
+      "epoch": 0.08476286579212916,
+      "grad_norm": 5.102560043334961,
+      "learning_rate": 6.719999999999999e-06,
+      "loss": 0.2041,
+      "step": 84
+    },
+    {
+      "epoch": 0.08577194752774975,
+      "grad_norm": 8.902843475341797,
+      "learning_rate": 6.799999999999999e-06,
+      "loss": 0.3015,
+      "step": 85
+    },
+    {
+      "epoch": 0.08678102926337034,
+      "grad_norm": 8.0729398727417,
+      "learning_rate": 6.879999999999999e-06,
+      "loss": 0.3637,
+      "step": 86
+    },
+    {
+      "epoch": 0.08779011099899092,
+      "grad_norm": 7.424192428588867,
+      "learning_rate": 6.9599999999999994e-06,
+      "loss": 0.1691,
+      "step": 87
+    },
+    {
+      "epoch": 0.08879919273461151,
+      "grad_norm": 19.509435653686523,
+      "learning_rate": 7.0399999999999995e-06,
+      "loss": 0.3888,
+      "step": 88
+    },
+    {
+      "epoch": 0.08980827447023208,
+      "grad_norm": 13.97599983215332,
+      "learning_rate": 7.12e-06,
+      "loss": 0.2415,
+      "step": 89
+    },
+    {
+      "epoch": 0.09081735620585267,
+      "grad_norm": 14.023358345031738,
+      "learning_rate": 7.2e-06,
+      "loss": 0.2889,
+      "step": 90
+    },
+    {
+      "epoch": 0.09182643794147326,
+      "grad_norm": 15.243273735046387,
+      "learning_rate": 7.28e-06,
+      "loss": 0.4061,
+      "step": 91
+    },
+    {
+      "epoch": 0.09283551967709384,
+      "grad_norm": 9.57010555267334,
+      "learning_rate": 7.36e-06,
+      "loss": 0.2501,
+      "step": 92
+    },
+    {
+      "epoch": 0.09384460141271443,
+      "grad_norm": 20.639694213867188,
+      "learning_rate": 7.44e-06,
+      "loss": 0.2877,
+      "step": 93
+    },
+    {
+      "epoch": 0.09485368314833502,
+      "grad_norm": 75.29905700683594,
+      "learning_rate": 7.519999999999999e-06,
+      "loss": 0.4536,
+      "step": 94
+    },
+    {
+      "epoch": 0.0958627648839556,
+      "grad_norm": 340.7889709472656,
+      "learning_rate": 7.599999999999999e-06,
+      "loss": 0.4443,
+      "step": 95
+    },
+    {
+      "epoch": 0.09687184661957618,
+      "grad_norm": 16.866931915283203,
+      "learning_rate": 7.68e-06,
+      "loss": 0.292,
+      "step": 96
+    },
+    {
+      "epoch": 0.09788092835519677,
+      "grad_norm": 12.300322532653809,
+      "learning_rate": 7.76e-06,
+      "loss": 0.2191,
+      "step": 97
+    },
+    {
+      "epoch": 0.09889001009081735,
+      "grad_norm": 18.15094566345215,
+      "learning_rate": 7.84e-06,
+      "loss": 0.1821,
+      "step": 98
+    },
+    {
+      "epoch": 0.09989909182643794,
+      "grad_norm": 31.187374114990234,
+      "learning_rate": 7.92e-06,
+      "loss": 0.1823,
+      "step": 99
+    },
+    {
+      "epoch": 0.10090817356205853,
+      "grad_norm": 64.1741943359375,
+      "learning_rate": 8e-06,
+      "loss": 0.5276,
+      "step": 100
+    },
+    {
+      "epoch": 0.10090817356205853,
+      "eval_accuracy": 0.8467650397275823,
+      "eval_loss": 0.28612375259399414,
+      "eval_runtime": 62.505,
+      "eval_samples_per_second": 28.19,
+      "eval_steps_per_second": 3.536,
+      "step": 100
+    },
+    {
+      "epoch": 0.10191725529767912,
+      "grad_norm": 12.569557189941406,
+      "learning_rate": 7.999975135834775e-06,
+      "loss": 0.3077,
+      "step": 101
+    },
+    {
+      "epoch": 0.1029263370332997,
+      "grad_norm": 7.0805277824401855,
+      "learning_rate": 7.999900543648217e-06,
+      "loss": 0.1759,
+      "step": 102
+    },
+    {
+      "epoch": 0.10393541876892028,
+      "grad_norm": 9.746481895446777,
+      "learning_rate": 7.999776224367659e-06,
+      "loss": 0.2064,
+      "step": 103
+    },
+    {
+      "epoch": 0.10494450050454086,
+      "grad_norm": 6.783945560455322,
+      "learning_rate": 7.999602179538651e-06,
+      "loss": 0.3028,
+      "step": 104
+    },
+    {
+      "epoch": 0.10595358224016145,
+      "grad_norm": 4.683195114135742,
+      "learning_rate": 7.999378411324933e-06,
+      "loss": 0.1767,
+      "step": 105
+    },
+    {
+      "epoch": 0.10696266397578204,
+      "grad_norm": 6.255279064178467,
+      "learning_rate": 7.999104922508408e-06,
+      "loss": 0.2984,
+      "step": 106
+    },
+    {
+      "epoch": 0.10797174571140263,
+      "grad_norm": 5.866703987121582,
+      "learning_rate": 7.99878171648911e-06,
+      "loss": 0.2472,
+      "step": 107
+    },
+    {
+      "epoch": 0.10898082744702321,
+      "grad_norm": 5.773608207702637,
+      "learning_rate": 7.998408797285167e-06,
+      "loss": 0.225,
+      "step": 108
+    },
+    {
+      "epoch": 0.1099899091826438,
+      "grad_norm": 10.58280086517334,
+      "learning_rate": 7.99798616953274e-06,
+      "loss": 0.3962,
+      "step": 109
+    },
+    {
+      "epoch": 0.11099899091826437,
+      "grad_norm": 5.545925140380859,
+      "learning_rate": 7.997513838485971e-06,
+      "loss": 0.1597,
+      "step": 110
+    },
+    {
+      "epoch": 0.11200807265388496,
+      "grad_norm": 7.84637975692749,
+      "learning_rate": 7.99699181001692e-06,
+      "loss": 0.346,
+      "step": 111
+    },
+    {
+      "epoch": 0.11301715438950555,
+      "grad_norm": 5.827946662902832,
+      "learning_rate": 7.996420090615486e-06,
+      "loss": 0.2507,
+      "step": 112
+    },
+    {
+      "epoch": 0.11402623612512613,
+      "grad_norm": 8.125126838684082,
+      "learning_rate": 7.995798687389334e-06,
+      "loss": 0.2496,
+      "step": 113
+    },
+    {
+      "epoch": 0.11503531786074672,
+      "grad_norm": 6.233455181121826,
+      "learning_rate": 7.9951276080638e-06,
+      "loss": 0.2129,
+      "step": 114
+    },
+    {
+      "epoch": 0.11604439959636731,
+      "grad_norm": 8.677657127380371,
+      "learning_rate": 7.994406860981797e-06,
+      "loss": 0.3926,
+      "step": 115
+    },
+    {
+      "epoch": 0.1170534813319879,
+      "grad_norm": 6.164268493652344,
+      "learning_rate": 7.99363645510371e-06,
+      "loss": 0.2335,
+      "step": 116
+    },
+    {
+      "epoch": 0.11806256306760847,
+      "grad_norm": 5.605584144592285,
+      "learning_rate": 7.992816400007294e-06,
+      "loss": 0.2357,
+      "step": 117
+    },
+    {
+      "epoch": 0.11907164480322906,
+      "grad_norm": 8.481727600097656,
+      "learning_rate": 7.991946705887537e-06,
+      "loss": 0.4145,
+      "step": 118
+    },
+    {
+      "epoch": 0.12008072653884964,
+      "grad_norm": 7.141845226287842,
+      "learning_rate": 7.99102738355655e-06,
+      "loss": 0.3807,
+      "step": 119
+    },
+    {
+      "epoch": 0.12108980827447023,
+      "grad_norm": 22.80964469909668,
+      "learning_rate": 7.990058444443424e-06,
+      "loss": 0.259,
+      "step": 120
+    },
+    {
+      "epoch": 0.12108980827447023,
+      "eval_accuracy": 0.7729852440408627,
+      "eval_loss": 0.44569578766822815,
+      "eval_runtime": 62.4809,
+      "eval_samples_per_second": 28.201,
+      "eval_steps_per_second": 3.537,
+      "step": 120
+    },
+    {
+      "epoch": 0.12209889001009082,
+      "grad_norm": 56.96179962158203,
+      "learning_rate": 7.989039900594089e-06,
+      "loss": 0.4721,
+      "step": 121
+    },
+    {
+      "epoch": 0.1231079717457114,
+      "grad_norm": 9.661979675292969,
+      "learning_rate": 7.987971764671168e-06,
+      "loss": 0.2308,
+      "step": 122
+    },
+    {
+      "epoch": 0.124117053481332,
+      "grad_norm": 20.324970245361328,
+      "learning_rate": 7.986854049953814e-06,
+      "loss": 0.4019,
+      "step": 123
+    },
+    {
+      "epoch": 0.12512613521695257,
+      "grad_norm": 7.678678512573242,
+      "learning_rate": 7.98568677033755e-06,
+      "loss": 0.1287,
+      "step": 124
+    },
+    {
+      "epoch": 0.12613521695257315,
+      "grad_norm": 5.97120475769043,
+      "learning_rate": 7.984469940334089e-06,
+      "loss": 0.3056,
+      "step": 125
+    },
+    {
+      "epoch": 0.12714429868819374,
+      "grad_norm": 4.4789509773254395,
+      "learning_rate": 7.983203575071166e-06,
+      "loss": 0.156,
+      "step": 126
+    },
+    {
+      "epoch": 0.12815338042381433,
+      "grad_norm": 7.833422660827637,
+      "learning_rate": 7.981887690292338e-06,
+      "loss": 0.3012,
+      "step": 127
+    },
+    {
+      "epoch": 0.12916246215943492,
+      "grad_norm": 4.627663612365723,
+      "learning_rate": 7.980522302356792e-06,
+      "loss": 0.1701,
+      "step": 128
+    },
+    {
+      "epoch": 0.1301715438950555,
+      "grad_norm": 23.75330924987793,
+      "learning_rate": 7.979107428239143e-06,
+      "loss": 0.5042,
+      "step": 129
+    },
+    {
+      "epoch": 0.1311806256306761,
+      "grad_norm": 8.093461036682129,
+      "learning_rate": 7.977643085529227e-06,
+      "loss": 0.3362,
+      "step": 130
+    },
+    {
+      "epoch": 0.13218970736629668,
+      "grad_norm": 8.479804992675781,
+      "learning_rate": 7.97612929243187e-06,
+      "loss": 0.2005,
+      "step": 131
+    },
+    {
+      "epoch": 0.13319878910191726,
+      "grad_norm": 8.889948844909668,
+      "learning_rate": 7.974566067766671e-06,
+      "loss": 0.4108,
+      "step": 132
+    },
+    {
+      "epoch": 0.13420787083753785,
+      "grad_norm": 6.473036289215088,
+      "learning_rate": 7.972953430967771e-06,
+      "loss": 0.3489,
+      "step": 133
+    },
+    {
+      "epoch": 0.13521695257315844,
+      "grad_norm": 5.9677863121032715,
+      "learning_rate": 7.971291402083606e-06,
+      "loss": 0.2707,
+      "step": 134
+    },
+    {
+      "epoch": 0.136226034308779,
+      "grad_norm": 9.331769943237305,
+      "learning_rate": 7.969580001776653e-06,
+      "loss": 0.4528,
+      "step": 135
+    },
+    {
+      "epoch": 0.13723511604439959,
+      "grad_norm": 6.988079071044922,
+      "learning_rate": 7.96781925132318e-06,
+      "loss": 0.3895,
+      "step": 136
+    },
+    {
+      "epoch": 0.13824419778002017,
+      "grad_norm": 5.6570281982421875,
+      "learning_rate": 7.966009172612988e-06,
+      "loss": 0.2324,
+      "step": 137
+    },
+    {
+      "epoch": 0.13925327951564076,
+      "grad_norm": 7.507450103759766,
+      "learning_rate": 7.964149788149122e-06,
+      "loss": 0.3709,
+      "step": 138
+    },
+    {
+      "epoch": 0.14026236125126135,
+      "grad_norm": 6.912793159484863,
+      "learning_rate": 7.962241121047602e-06,
+      "loss": 0.2004,
+      "step": 139
+    },
+    {
+      "epoch": 0.14127144298688193,
+      "grad_norm": 7.456505298614502,
+      "learning_rate": 7.960283195037138e-06,
+      "loss": 0.1852,
+      "step": 140
+    },
+    {
+      "epoch": 0.14127144298688193,
+      "eval_accuracy": 0.8382519863791147,
+      "eval_loss": 0.30151110887527466,
+      "eval_runtime": 59.0753,
+      "eval_samples_per_second": 29.826,
+      "eval_steps_per_second": 3.741,
+      "step": 140
+    },
+    {
+      "epoch": 0.14228052472250252,
+      "grad_norm": 5.073808670043945,
+      "learning_rate": 7.958276034458826e-06,
+      "loss": 0.2379,
+      "step": 141
+    },
+    {
+      "epoch": 0.1432896064581231,
+      "grad_norm": 10.782764434814453,
+      "learning_rate": 7.956219664265852e-06,
+      "loss": 0.1299,
+      "step": 142
+    },
+    {
+      "epoch": 0.1442986881937437,
+      "grad_norm": 5.916919708251953,
+      "learning_rate": 7.95411411002318e-06,
+      "loss": 0.2939,
+      "step": 143
+    },
+    {
+      "epoch": 0.14530776992936428,
+      "grad_norm": 4.290775775909424,
+      "learning_rate": 7.951959397907236e-06,
+      "loss": 0.2175,
+      "step": 144
+    },
+    {
+      "epoch": 0.14631685166498487,
+      "grad_norm": 6.01127290725708,
+      "learning_rate": 7.949755554705577e-06,
+      "loss": 0.2962,
+      "step": 145
+    },
+    {
+      "epoch": 0.14732593340060546,
+      "grad_norm": 5.430018901824951,
+      "learning_rate": 7.947502607816566e-06,
+      "loss": 0.3358,
+      "step": 146
+    },
+    {
+      "epoch": 0.14833501513622604,
+      "grad_norm": 2.893122911453247,
+      "learning_rate": 7.945200585249022e-06,
+      "loss": 0.0949,
+      "step": 147
+    },
+    {
+      "epoch": 0.14934409687184663,
+      "grad_norm": 5.385425567626953,
+      "learning_rate": 7.942849515621881e-06,
+      "loss": 0.314,
+      "step": 148
+    },
+    {
+      "epoch": 0.1503531786074672,
+      "grad_norm": 4.384002208709717,
+      "learning_rate": 7.940449428163837e-06,
+      "loss": 0.2312,
+      "step": 149
+    },
+    {
+      "epoch": 0.15136226034308778,
+      "grad_norm": 5.955286979675293,
+      "learning_rate": 7.938000352712972e-06,
+      "loss": 0.1949,
+      "step": 150
+    },
+    {
+      "epoch": 0.15237134207870837,
+      "grad_norm": 6.1054487228393555,
+      "learning_rate": 7.935502319716397e-06,
+      "loss": 0.3697,
+      "step": 151
+    },
+    {
+      "epoch": 0.15338042381432895,
+      "grad_norm": 3.2285423278808594,
+      "learning_rate": 7.932955360229862e-06,
+      "loss": 0.1118,
+      "step": 152
+    },
+    {
+      "epoch": 0.15438950554994954,
+      "grad_norm": 7.409282684326172,
+      "learning_rate": 7.930359505917381e-06,
+      "loss": 0.4898,
+      "step": 153
+    },
+    {
+      "epoch": 0.15539858728557013,
+      "grad_norm": 4.591885566711426,
+      "learning_rate": 7.927714789050827e-06,
+      "loss": 0.2288,
+      "step": 154
+    },
+    {
+      "epoch": 0.15640766902119072,
+      "grad_norm": 5.779257297515869,
+      "learning_rate": 7.925021242509538e-06,
+      "loss": 0.2563,
+      "step": 155
+    },
+    {
+      "epoch": 0.1574167507568113,
+      "grad_norm": 4.496645927429199,
+      "learning_rate": 7.92227889977991e-06,
+      "loss": 0.2096,
+      "step": 156
+    },
+    {
+      "epoch": 0.1584258324924319,
+      "grad_norm": 6.980607509613037,
+      "learning_rate": 7.919487794954972e-06,
+      "loss": 0.3658,
+      "step": 157
+    },
+    {
+      "epoch": 0.15943491422805248,
+      "grad_norm": 5.738483905792236,
+      "learning_rate": 7.91664796273397e-06,
+      "loss": 0.3123,
+      "step": 158
+    },
+    {
+      "epoch": 0.16044399596367306,
+      "grad_norm": 6.03507137298584,
+      "learning_rate": 7.913759438421932e-06,
+      "loss": 0.3345,
+      "step": 159
+    },
+    {
+      "epoch": 0.16145307769929365,
+      "grad_norm": 31.2845458984375,
+      "learning_rate": 7.910822257929234e-06,
+      "loss": 0.1122,
+      "step": 160
+    },
+    {
+      "epoch": 0.16145307769929365,
+      "eval_accuracy": 0.8518728717366629,
+      "eval_loss": 0.38947635889053345,
+      "eval_runtime": 62.3612,
+      "eval_samples_per_second": 28.255,
+      "eval_steps_per_second": 3.544,
+      "step": 160
+    },
+    {
+      "epoch": 0.16246215943491424,
+      "grad_norm": 6.148342609405518,
+      "learning_rate": 7.907836457771143e-06,
+      "loss": 0.2418,
+      "step": 161
+    },
+    {
+      "epoch": 0.16347124117053483,
+      "grad_norm": 38.332550048828125,
+      "learning_rate": 7.904802075067377e-06,
+      "loss": 0.6998,
+      "step": 162
+    },
+    {
+      "epoch": 0.16448032290615539,
+      "grad_norm": 8.149801254272461,
+      "learning_rate": 7.901719147541628e-06,
+      "loss": 0.1255,
+      "step": 163
+    },
+    {
+      "epoch": 0.16548940464177597,
+      "grad_norm": 19.77474021911621,
+      "learning_rate": 7.898587713521109e-06,
+      "loss": 0.6065,
+      "step": 164
+    },
+    {
+      "epoch": 0.16649848637739656,
+      "grad_norm": 5.291502952575684,
+      "learning_rate": 7.895407811936064e-06,
+      "loss": 0.1659,
+      "step": 165
+    },
+    {
+      "epoch": 0.16750756811301715,
+      "grad_norm": 5.293869495391846,
+      "learning_rate": 7.892179482319294e-06,
+      "loss": 0.2073,
+      "step": 166
+    },
+    {
+      "epoch": 0.16851664984863773,
+      "grad_norm": 6.291193962097168,
+      "learning_rate": 7.88890276480566e-06,
+      "loss": 0.3597,
+      "step": 167
+    },
+    {
+      "epoch": 0.16952573158425832,
+      "grad_norm": 5.216683864593506,
+      "learning_rate": 7.885577700131584e-06,
+      "loss": 0.2888,
+      "step": 168
+    },
+    {
+      "epoch": 0.1705348133198789,
+      "grad_norm": 3.573047637939453,
+      "learning_rate": 7.882204329634543e-06,
+      "loss": 0.1343,
+      "step": 169
+    },
+    {
+      "epoch": 0.1715438950554995,
+      "grad_norm": 5.618640899658203,
+      "learning_rate": 7.878782695252562e-06,
+      "loss": 0.2477,
+      "step": 170
+    },
+    {
+      "epoch": 0.17255297679112008,
+      "grad_norm": 6.469369411468506,
+      "learning_rate": 7.875312839523677e-06,
+      "loss": 0.3563,
+      "step": 171
+    },
+    {
+      "epoch": 0.17356205852674067,
+      "grad_norm": 7.982197284698486,
+      "learning_rate": 7.871794805585425e-06,
+      "loss": 0.2748,
+      "step": 172
+    },
+    {
+      "epoch": 0.17457114026236126,
+      "grad_norm": 8.727740287780762,
+      "learning_rate": 7.868228637174292e-06,
+      "loss": 0.2539,
+      "step": 173
+    },
+    {
+      "epoch": 0.17558022199798184,
+      "grad_norm": 9.52997875213623,
+      "learning_rate": 7.86461437862518e-06,
+      "loss": 0.2352,
+      "step": 174
+    },
+    {
+      "epoch": 0.17658930373360243,
+      "grad_norm": 8.447443008422852,
+      "learning_rate": 7.86095207487085e-06,
+      "loss": 0.3129,
+      "step": 175
+    },
+    {
+      "epoch": 0.17759838546922302,
+      "grad_norm": 7.929843425750732,
+      "learning_rate": 7.857241771441364e-06,
+      "loss": 0.3138,
+      "step": 176
+    },
+    {
+      "epoch": 0.17860746720484358,
+      "grad_norm": 7.068841934204102,
+      "learning_rate": 7.853483514463521e-06,
+      "loss": 0.1881,
+      "step": 177
+    },
+    {
+      "epoch": 0.17961654894046417,
+      "grad_norm": 179.09567260742188,
+      "learning_rate": 7.849677350660282e-06,
+      "loss": 1.6187,
+      "step": 178
+    },
+    {
+      "epoch": 0.18062563067608475,
+      "grad_norm": 14.378893852233887,
+      "learning_rate": 7.84582332735019e-06,
+      "loss": 0.44,
+      "step": 179
+    },
+    {
+      "epoch": 0.18163471241170534,
+      "grad_norm": 10.581100463867188,
+      "learning_rate": 7.841921492446781e-06,
+      "loss": 0.1962,
+      "step": 180
+    },
+    {
+      "epoch": 0.18163471241170534,
+      "eval_accuracy": 0.8229284903518729,
+      "eval_loss": 0.2842705547809601,
+      "eval_runtime": 63.0695,
+      "eval_samples_per_second": 27.937,
+      "eval_steps_per_second": 3.504,
+      "step": 180
+    },
+    {
+      "epoch": 0.18264379414732593,
+      "grad_norm": 13.833659172058105,
+      "learning_rate": 7.837971894457989e-06,
+      "loss": 0.4059,
+      "step": 181
+    },
+    {
+      "epoch": 0.18365287588294651,
+      "grad_norm": 6.910711288452148,
+      "learning_rate": 7.833974582485544e-06,
+      "loss": 0.3115,
+      "step": 182
+    },
+    {
+      "epoch": 0.1846619576185671,
+      "grad_norm": 4.3872599601745605,
+      "learning_rate": 7.829929606224356e-06,
+      "loss": 0.1654,
+      "step": 183
+    },
+    {
+      "epoch": 0.1856710393541877,
+      "grad_norm": 8.301304817199707,
+      "learning_rate": 7.825837015961904e-06,
+      "loss": 0.3223,
+      "step": 184
+    },
+    {
+      "epoch": 0.18668012108980828,
+      "grad_norm": 6.772916316986084,
+      "learning_rate": 7.82169686257761e-06,
+      "loss": 0.207,
+      "step": 185
+    },
+    {
+      "epoch": 0.18768920282542886,
+      "grad_norm": 10.749194145202637,
+      "learning_rate": 7.817509197542204e-06,
+      "loss": 0.3322,
+      "step": 186
+    },
+    {
+      "epoch": 0.18869828456104945,
+      "grad_norm": 18.030078887939453,
+      "learning_rate": 7.813274072917081e-06,
+      "loss": 0.3675,
+      "step": 187
+    },
+    {
+      "epoch": 0.18970736629667004,
+      "grad_norm": 15.416492462158203,
+      "learning_rate": 7.80899154135366e-06,
+      "loss": 0.2049,
+      "step": 188
+    },
+    {
+      "epoch": 0.19071644803229063,
+      "grad_norm": 24.75461769104004,
+      "learning_rate": 7.80466165609273e-06,
+      "loss": 0.36,
+      "step": 189
+    },
+    {
+      "epoch": 0.1917255297679112,
+      "grad_norm": 11.484724044799805,
+      "learning_rate": 7.800284470963781e-06,
+      "loss": 0.2014,
+      "step": 190
+    },
+    {
+      "epoch": 0.1927346115035318,
+      "grad_norm": 19.571962356567383,
+      "learning_rate": 7.795860040384339e-06,
+      "loss": 0.5687,
+      "step": 191
+    },
+    {
+      "epoch": 0.19374369323915236,
+      "grad_norm": 6.390493392944336,
+      "learning_rate": 7.791388419359292e-06,
+      "loss": 0.2563,
+      "step": 192
+    },
+    {
+      "epoch": 0.19475277497477295,
+      "grad_norm": 9.572249412536621,
+      "learning_rate": 7.786869663480201e-06,
+      "loss": 0.3565,
+      "step": 193
+    },
+    {
+      "epoch": 0.19576185671039353,
+      "grad_norm": 8.426647186279297,
+      "learning_rate": 7.782303828924613e-06,
+      "loss": 0.5456,
+      "step": 194
+    },
+    {
+      "epoch": 0.19677093844601412,
+      "grad_norm": 14.882564544677734,
+      "learning_rate": 7.77769097245536e-06,
+      "loss": 0.2976,
+      "step": 195
+    },
+    {
+      "epoch": 0.1977800201816347,
+      "grad_norm": 10.191393852233887,
+      "learning_rate": 7.773031151419853e-06,
+      "loss": 0.4641,
+      "step": 196
+    },
+    {
+      "epoch": 0.1987891019172553,
+      "grad_norm": 7.604254245758057,
+      "learning_rate": 7.768324423749376e-06,
+      "loss": 0.2364,
+      "step": 197
+    },
+    {
+      "epoch": 0.19979818365287588,
+      "grad_norm": 9.080754280090332,
+      "learning_rate": 7.763570847958354e-06,
+      "loss": 0.1754,
+      "step": 198
+    },
+    {
+      "epoch": 0.20080726538849647,
+      "grad_norm": 10.90381908416748,
+      "learning_rate": 7.758770483143633e-06,
+      "loss": 0.1799,
+      "step": 199
+    },
+    {
+      "epoch": 0.20181634712411706,
+      "grad_norm": 12.508880615234375,
+      "learning_rate": 7.753923388983747e-06,
+      "loss": 0.2855,
+      "step": 200
+    },
+    {
+      "epoch": 0.20181634712411706,
+      "eval_accuracy": 0.840522133938706,
+      "eval_loss": 0.29473429918289185,
+      "eval_runtime": 62.4643,
+      "eval_samples_per_second": 28.208,
+      "eval_steps_per_second": 3.538,
+      "step": 200
+    },
+    {
+      "epoch": 0.20282542885973764,
+      "grad_norm": 20.691980361938477,
+      "learning_rate": 7.749029625738169e-06,
+      "loss": 0.1942,
+      "step": 201
+    },
+    {
+      "epoch": 0.20383451059535823,
+      "grad_norm": 10.143714904785156,
+      "learning_rate": 7.744089254246569e-06,
+      "loss": 0.2001,
+      "step": 202
+    },
+    {
+      "epoch": 0.20484359233097882,
+      "grad_norm": 17.92678451538086,
+      "learning_rate": 7.739102335928053e-06,
+      "loss": 0.3823,
+      "step": 203
+    },
+    {
+      "epoch": 0.2058526740665994,
+      "grad_norm": 12.694001197814941,
+      "learning_rate": 7.734068932780405e-06,
+      "loss": 0.274,
+      "step": 204
+    },
+    {
+      "epoch": 0.20686175580222,
+      "grad_norm": 12.30357837677002,
+      "learning_rate": 7.728989107379303e-06,
+      "loss": 0.3017,
+      "step": 205
+    },
+    {
+      "epoch": 0.20787083753784055,
+      "grad_norm": 7.952809810638428,
+      "learning_rate": 7.72386292287756e-06,
+      "loss": 0.14,
+      "step": 206
+    },
+    {
+      "epoch": 0.20887991927346114,
+      "grad_norm": 69.02944946289062,
+      "learning_rate": 7.718690443004324e-06,
+      "loss": 0.4633,
+      "step": 207
+    },
+    {
+      "epoch": 0.20988900100908173,
+      "grad_norm": 92.96024322509766,
+      "learning_rate": 7.71347173206429e-06,
+      "loss": 1.1202,
+      "step": 208
+    },
+    {
+      "epoch": 0.21089808274470231,
+      "grad_norm": 83.07585906982422,
+      "learning_rate": 7.708206854936908e-06,
+      "loss": 0.3752,
+      "step": 209
+    },
+    {
+      "epoch": 0.2119071644803229,
+      "grad_norm": 10.343903541564941,
+      "learning_rate": 7.702895877075563e-06,
+      "loss": 0.3543,
+      "step": 210
+    },
+    {
+      "epoch": 0.2129162462159435,
+      "grad_norm": 34.449188232421875,
+      "learning_rate": 7.697538864506767e-06,
+      "loss": 0.5488,
+      "step": 211
+    },
+    {
+      "epoch": 0.21392532795156408,
+      "grad_norm": 5.609644412994385,
+      "learning_rate": 7.692135883829349e-06,
+      "loss": 0.342,
+      "step": 212
+    },
+    {
+      "epoch": 0.21493440968718466,
+      "grad_norm": 6.129890441894531,
+      "learning_rate": 7.686687002213609e-06,
+      "loss": 0.1932,
+      "step": 213
+    },
+    {
+      "epoch": 0.21594349142280525,
+      "grad_norm": 17.425050735473633,
+      "learning_rate": 7.681192287400491e-06,
+      "loss": 0.4084,
+      "step": 214
+    },
+    {
+      "epoch": 0.21695257315842584,
+      "grad_norm": 24.733633041381836,
+      "learning_rate": 7.675651807700748e-06,
+      "loss": 0.2141,
+      "step": 215
+    },
+    {
+      "epoch": 0.21796165489404642,
+      "grad_norm": 33.499794006347656,
+      "learning_rate": 7.670065631994078e-06,
+      "loss": 0.5658,
+      "step": 216
+    },
+    {
+      "epoch": 0.218970736629667,
+      "grad_norm": 9.948688507080078,
+      "learning_rate": 7.664433829728277e-06,
+      "loss": 0.2564,
+      "step": 217
+    },
+    {
+      "epoch": 0.2199798183652876,
+      "grad_norm": 9.255337715148926,
+      "learning_rate": 7.658756470918382e-06,
+      "loss": 0.2294,
+      "step": 218
+    },
+    {
+      "epoch": 0.2209889001009082,
+      "grad_norm": 30.673681259155273,
+      "learning_rate": 7.65303362614578e-06,
+      "loss": 0.4334,
+      "step": 219
+    },
+    {
+      "epoch": 0.22199798183652875,
+      "grad_norm": 5.19248628616333,
+      "learning_rate": 7.647265366557355e-06,
+      "loss": 0.1593,
+      "step": 220
+    },
+    {
+      "epoch": 0.22199798183652875,
+      "eval_accuracy": 0.8382519863791147,
+      "eval_loss": 0.29325050115585327,
+      "eval_runtime": 62.9477,
+      "eval_samples_per_second": 27.991,
+      "eval_steps_per_second": 3.511,
+      "step": 220
+    },
+    {
+      "epoch": 0.22300706357214933,
+      "grad_norm": 5.615384101867676,
+      "learning_rate": 7.641451763864587e-06,
+      "loss": 0.225,
+      "step": 221
+    },
+    {
+      "epoch": 0.22401614530776992,
+      "grad_norm": 6.7938032150268555,
+      "learning_rate": 7.63559289034266e-06,
+      "loss": 0.2151,
+      "step": 222
+    },
+    {
+      "epoch": 0.2250252270433905,
+      "grad_norm": 2.8786470890045166,
+      "learning_rate": 7.629688818829577e-06,
+      "loss": 0.1089,
+      "step": 223
+    },
+    {
+      "epoch": 0.2260343087790111,
+      "grad_norm": 7.671812057495117,
+      "learning_rate": 7.623739622725244e-06,
+      "loss": 0.289,
+      "step": 224
+    },
+    {
+      "epoch": 0.22704339051463168,
+      "grad_norm": 5.496854782104492,
+      "learning_rate": 7.617745375990556e-06,
+      "loss": 0.24,
+      "step": 225
+    },
+    {
+      "epoch": 0.22805247225025227,
+      "grad_norm": 6.458908557891846,
+      "learning_rate": 7.611706153146485e-06,
+      "loss": 0.3702,
+      "step": 226
+    },
+    {
+      "epoch": 0.22906155398587286,
+      "grad_norm": 10.48363971710205,
+      "learning_rate": 7.605622029273148e-06,
+      "loss": 0.1634,
+      "step": 227
+    },
+    {
+      "epoch": 0.23007063572149344,
+      "grad_norm": 32.09809875488281,
+      "learning_rate": 7.599493080008873e-06,
+      "loss": 0.7119,
+      "step": 228
+    },
+    {
+      "epoch": 0.23107971745711403,
+      "grad_norm": 43.372257232666016,
+      "learning_rate": 7.5933193815492675e-06,
+      "loss": 0.4775,
+      "step": 229
+    },
+    {
+      "epoch": 0.23208879919273462,
+      "grad_norm": 21.043750762939453,
+      "learning_rate": 7.587101010646259e-06,
+      "loss": 0.3401,
+      "step": 230
+    },
+    {
+      "epoch": 0.2330978809283552,
+      "grad_norm": 13.805304527282715,
+      "learning_rate": 7.58083804460715e-06,
+      "loss": 0.1592,
+      "step": 231
+    },
+    {
+      "epoch": 0.2341069626639758,
+      "grad_norm": 6.074002265930176,
+      "learning_rate": 7.574530561293649e-06,
+      "loss": 0.3836,
+      "step": 232
+    },
+    {
+      "epoch": 0.23511604439959638,
+      "grad_norm": 4.921712875366211,
+      "learning_rate": 7.5681786391209105e-06,
+      "loss": 0.2796,
+      "step": 233
+    },
+    {
+      "epoch": 0.23612512613521694,
+      "grad_norm": 5.469928741455078,
+      "learning_rate": 7.561782357056557e-06,
+      "loss": 0.2877,
+      "step": 234
+    },
+    {
+      "epoch": 0.23713420787083753,
+      "grad_norm": 5.309055328369141,
+      "learning_rate": 7.555341794619694e-06,
+      "loss": 0.2664,
+      "step": 235
+    },
+    {
+      "epoch": 0.23814328960645811,
+      "grad_norm": 55.794979095458984,
+      "learning_rate": 7.548857031879926e-06,
+      "loss": 0.1148,
+      "step": 236
+    },
+    {
+      "epoch": 0.2391523713420787,
+      "grad_norm": 207.49688720703125,
+      "learning_rate": 7.5423281494563595e-06,
+      "loss": 0.3024,
+      "step": 237
+    },
+    {
+      "epoch": 0.2401614530776993,
+      "grad_norm": 13.331132888793945,
+      "learning_rate": 7.535755228516601e-06,
+      "loss": 0.353,
+      "step": 238
+    },
+    {
+      "epoch": 0.24117053481331988,
+      "grad_norm": 9.264041900634766,
+      "learning_rate": 7.529138350775745e-06,
+      "loss": 0.3563,
+      "step": 239
+    },
+    {
+      "epoch": 0.24217961654894046,
+      "grad_norm": 4.584170818328857,
+      "learning_rate": 7.522477598495363e-06,
+      "loss": 0.2153,
+      "step": 240
+    },
+    {
+      "epoch": 0.24217961654894046,
+      "eval_accuracy": 0.8376844494892168,
+      "eval_loss": 0.2747989892959595,
+      "eval_runtime": 62.479,
+      "eval_samples_per_second": 28.201,
+      "eval_steps_per_second": 3.537,
+      "step": 240
+    },
+    {
+      "epoch": 0.24318869828456105,
+      "grad_norm": 4.465677738189697,
+      "learning_rate": 7.515773054482478e-06,
+      "loss": 0.2503,
+      "step": 241
+    },
+    {
+      "epoch": 0.24419778002018164,
+      "grad_norm": 4.346131324768066,
+      "learning_rate": 7.509024802088534e-06,
+      "loss": 0.2141,
+      "step": 242
+    },
+    {
+      "epoch": 0.24520686175580222,
+      "grad_norm": 7.381392002105713,
+      "learning_rate": 7.502232925208363e-06,
+      "loss": 0.2984,
+      "step": 243
+    },
+    {
+      "epoch": 0.2462159434914228,
+      "grad_norm": 4.441753387451172,
+      "learning_rate": 7.49539750827914e-06,
+      "loss": 0.2058,
+      "step": 244
+    },
+    {
+      "epoch": 0.2472250252270434,
+      "grad_norm": 7.497750282287598,
+      "learning_rate": 7.488518636279331e-06,
+      "loss": 0.4537,
+      "step": 245
+    },
+    {
+      "epoch": 0.248234106962664,
+      "grad_norm": 9.178194046020508,
+      "learning_rate": 7.4815963947276436e-06,
+      "loss": 0.2963,
+      "step": 246
+    },
+    {
+      "epoch": 0.24924318869828457,
+      "grad_norm": 7.65831184387207,
+      "learning_rate": 7.474630869681954e-06,
+      "loss": 0.3374,
+      "step": 247
+    },
+    {
+      "epoch": 0.25025227043390513,
+      "grad_norm": 4.4074482917785645,
+      "learning_rate": 7.467622147738246e-06,
+      "loss": 0.176,
+      "step": 248
+    },
+    {
+      "epoch": 0.2512613521695257,
+      "grad_norm": 3.8703436851501465,
+      "learning_rate": 7.4605703160295315e-06,
+      "loss": 0.1294,
+      "step": 249
+    },
+    {
+      "epoch": 0.2522704339051463,
+      "grad_norm": 14.600114822387695,
+      "learning_rate": 7.453475462224763e-06,
+      "loss": 0.2959,
+      "step": 250
+    },
+    {
+      "epoch": 0.2532795156407669,
+      "grad_norm": 7.458588600158691,
+      "learning_rate": 7.44633767452775e-06,
+      "loss": 0.3291,
+      "step": 251
+    },
+    {
+      "epoch": 0.2542885973763875,
+      "grad_norm": 13.63289737701416,
+      "learning_rate": 7.439157041676058e-06,
+      "loss": 0.4166,
+      "step": 252
+    },
+    {
+      "epoch": 0.25529767911200807,
+      "grad_norm": 8.358410835266113,
+      "learning_rate": 7.431933652939908e-06,
+      "loss": 0.257,
+      "step": 253
+    },
+    {
+      "epoch": 0.25630676084762866,
+      "grad_norm": 5.803023338317871,
+      "learning_rate": 7.424667598121065e-06,
+      "loss": 0.2334,
+      "step": 254
+    },
+    {
+      "epoch": 0.25731584258324924,
+      "grad_norm": 5.000588893890381,
+      "learning_rate": 7.4173589675517245e-06,
+      "loss": 0.2193,
+      "step": 255
+    },
+    {
+      "epoch": 0.25832492431886983,
+      "grad_norm": 5.990160942077637,
+      "learning_rate": 7.410007852093384e-06,
+      "loss": 0.31,
+      "step": 256
+    },
+    {
+      "epoch": 0.2593340060544904,
+      "grad_norm": 6.175328731536865,
+      "learning_rate": 7.40261434313572e-06,
+      "loss": 0.307,
+      "step": 257
+    },
+    {
+      "epoch": 0.260343087790111,
+      "grad_norm": 6.047603130340576,
+      "learning_rate": 7.395178532595444e-06,
+      "loss": 0.223,
+      "step": 258
+    },
+    {
+      "epoch": 0.2613521695257316,
+      "grad_norm": 5.447187423706055,
+      "learning_rate": 7.387700512915168e-06,
+      "loss": 0.0597,
+      "step": 259
+    },
+    {
+      "epoch": 0.2623612512613522,
+      "grad_norm": 6.375972747802734,
+      "learning_rate": 7.380180377062251e-06,
+      "loss": 0.398,
+      "step": 260
+    },
+    {
+      "epoch": 0.2623612512613522,
+      "eval_accuracy": 0.8320090805902384,
+      "eval_loss": 0.2793748378753662,
+      "eval_runtime": 62.5384,
+      "eval_samples_per_second": 28.175,
+      "eval_steps_per_second": 3.534,
+      "step": 260
+    },
+    {
+      "epoch": 0.26337033299697277,
+      "grad_norm": 5.560431003570557,
+      "learning_rate": 7.372618218527644e-06,
+      "loss": 0.2861,
+      "step": 261
+    },
+    {
+      "epoch": 0.26437941473259335,
+      "grad_norm": 5.246089935302734,
+      "learning_rate": 7.365014131324725e-06,
+      "loss": 0.1859,
+      "step": 262
+    },
+    {
+      "epoch": 0.26538849646821394,
+      "grad_norm": 7.6310625076293945,
+      "learning_rate": 7.3573682099881345e-06,
+      "loss": 0.3164,
+      "step": 263
+    },
+    {
+      "epoch": 0.26639757820383453,
+      "grad_norm": 6.964735507965088,
+      "learning_rate": 7.349680549572598e-06,
+      "loss": 0.3655,
+      "step": 264
+    },
+    {
+      "epoch": 0.2674066599394551,
+      "grad_norm": 4.680076599121094,
+      "learning_rate": 7.3419512456517455e-06,
+      "loss": 0.1872,
+      "step": 265
+    },
+    {
+      "epoch": 0.2684157416750757,
+      "grad_norm": 3.2980797290802,
+      "learning_rate": 7.3341803943169214e-06,
+      "loss": 0.0983,
+      "step": 266
+    },
+    {
+      "epoch": 0.2694248234106963,
+      "grad_norm": 9.682284355163574,
+      "learning_rate": 7.326368092175993e-06,
+      "loss": 0.1637,
+      "step": 267
+    },
+    {
+      "epoch": 0.2704339051463169,
+      "grad_norm": 58.10804748535156,
+      "learning_rate": 7.3185144363521435e-06,
+      "loss": 1.2893,
+      "step": 268
+    },
+    {
+      "epoch": 0.2714429868819374,
+      "grad_norm": 3.9088146686553955,
+      "learning_rate": 7.310619524482673e-06,
+      "loss": 0.1103,
+      "step": 269
+    },
+    {
+      "epoch": 0.272452068617558,
+      "grad_norm": 8.596792221069336,
+      "learning_rate": 7.302683454717778e-06,
+      "loss": 0.4524,
+      "step": 270
+    },
+    {
+      "epoch": 0.2734611503531786,
+      "grad_norm": 18.94976234436035,
+      "learning_rate": 7.294706325719331e-06,
+      "loss": 0.3868,
+      "step": 271
+    },
+    {
+      "epoch": 0.27447023208879917,
+      "grad_norm": 28.618101119995117,
+      "learning_rate": 7.28668823665966e-06,
+      "loss": 1.2269,
+      "step": 272
+    },
+    {
+      "epoch": 0.27547931382441976,
+      "grad_norm": 7.027772903442383,
+      "learning_rate": 7.2786292872203125e-06,
+      "loss": 0.2546,
+      "step": 273
+    },
+    {
+      "epoch": 0.27648839556004035,
+      "grad_norm": 4.971069812774658,
+      "learning_rate": 7.270529577590812e-06,
+      "loss": 0.2229,
+      "step": 274
+    },
+    {
+      "epoch": 0.27749747729566093,
+      "grad_norm": 7.468736171722412,
+      "learning_rate": 7.262389208467417e-06,
+      "loss": 0.3655,
+      "step": 275
+    },
+    {
+      "epoch": 0.2785065590312815,
+      "grad_norm": 7.4164509773254395,
+      "learning_rate": 7.2542082810518696e-06,
+      "loss": 0.3047,
+      "step": 276
+    },
+    {
+      "epoch": 0.2795156407669021,
+      "grad_norm": 6.213995933532715,
+      "learning_rate": 7.245986897050137e-06,
+      "loss": 0.2879,
+      "step": 277
+    },
+    {
+      "epoch": 0.2805247225025227,
+      "grad_norm": 6.1668548583984375,
+      "learning_rate": 7.237725158671141e-06,
+      "loss": 0.3405,
+      "step": 278
+    },
+    {
+      "epoch": 0.2815338042381433,
+      "grad_norm": 11.742606163024902,
+      "learning_rate": 7.229423168625498e-06,
+      "loss": 0.3894,
+      "step": 279
+    },
+    {
+      "epoch": 0.28254288597376387,
+      "grad_norm": 5.1043477058410645,
+      "learning_rate": 7.2210810301242345e-06,
+      "loss": 0.199,
+      "step": 280
+    },
+    {
+      "epoch": 0.28254288597376387,
+      "eval_accuracy": 0.8382519863791147,
+      "eval_loss": 0.28534775972366333,
+      "eval_runtime": 62.6946,
+      "eval_samples_per_second": 28.104,
+      "eval_steps_per_second": 3.525,
+      "step": 280
+    },
+    {
+      "epoch": 0.28355196770938446,
+      "grad_norm": 5.588554859161377,
+      "learning_rate": 7.212698846877503e-06,
+      "loss": 0.347,
+      "step": 281
+    },
+    {
+      "epoch": 0.28456104944500504,
+      "grad_norm": 4.220617294311523,
+      "learning_rate": 7.204276723093301e-06,
+      "loss": 0.2267,
+      "step": 282
+    },
+    {
+      "epoch": 0.28557013118062563,
+      "grad_norm": 36.09522247314453,
+      "learning_rate": 7.195814763476164e-06,
+      "loss": 0.6504,
+      "step": 283
+    },
+    {
+      "epoch": 0.2865792129162462,
+      "grad_norm": 5.161518096923828,
+      "learning_rate": 7.187313073225876e-06,
+      "loss": 0.2736,
+      "step": 284
+    },
+    {
+      "epoch": 0.2875882946518668,
+      "grad_norm": 5.858831882476807,
+      "learning_rate": 7.178771758036152e-06,
+      "loss": 0.3758,
+      "step": 285
+    },
+    {
+      "epoch": 0.2885973763874874,
+      "grad_norm": 14.579549789428711,
+      "learning_rate": 7.170190924093326e-06,
+      "loss": 0.3338,
+      "step": 286
+    },
+    {
+      "epoch": 0.289606458123108,
+      "grad_norm": 6.289809703826904,
+      "learning_rate": 7.161570678075037e-06,
+      "loss": 0.4051,
+      "step": 287
+    },
+    {
+      "epoch": 0.29061553985872857,
+      "grad_norm": 4.26041841506958,
+      "learning_rate": 7.152911127148893e-06,
+      "loss": 0.2365,
+      "step": 288
+    },
+    {
+      "epoch": 0.29162462159434915,
+      "grad_norm": 3.567761182785034,
+      "learning_rate": 7.1442123789711495e-06,
+      "loss": 0.1612,
+      "step": 289
+    },
+    {
+      "epoch": 0.29263370332996974,
+      "grad_norm": 4.908886432647705,
+      "learning_rate": 7.135474541685359e-06,
+      "loss": 0.2345,
+      "step": 290
+    },
+    {
+      "epoch": 0.29364278506559033,
+      "grad_norm": 5.828519821166992,
+      "learning_rate": 7.126697723921041e-06,
+      "loss": 0.3646,
+      "step": 291
+    },
+    {
+      "epoch": 0.2946518668012109,
+      "grad_norm": 5.6283674240112305,
+      "learning_rate": 7.117882034792315e-06,
+      "loss": 0.3539,
+      "step": 292
+    },
+    {
+      "epoch": 0.2956609485368315,
+      "grad_norm": 2.5755410194396973,
+      "learning_rate": 7.109027583896559e-06,
+      "loss": 0.072,
+      "step": 293
+    },
+    {
+      "epoch": 0.2966700302724521,
+      "grad_norm": 5.203614234924316,
+      "learning_rate": 7.1001344813130355e-06,
+      "loss": 0.2994,
+      "step": 294
+    },
+    {
+      "epoch": 0.2976791120080727,
+      "grad_norm": 4.4349822998046875,
+      "learning_rate": 7.0912028376015315e-06,
+      "loss": 0.1816,
+      "step": 295
+    },
+    {
+      "epoch": 0.29868819374369326,
+      "grad_norm": 3.131836175918579,
+      "learning_rate": 7.082232763800982e-06,
+      "loss": 0.1107,
+      "step": 296
+    },
+    {
+      "epoch": 0.29969727547931385,
+      "grad_norm": 7.76261568069458,
+      "learning_rate": 7.073224371428083e-06,
+      "loss": 0.3865,
+      "step": 297
+    },
+    {
+      "epoch": 0.3007063572149344,
+      "grad_norm": 8.260970115661621,
+      "learning_rate": 7.064177772475912e-06,
+      "loss": 0.3415,
+      "step": 298
+    },
+    {
+      "epoch": 0.30171543895055497,
+      "grad_norm": 7.274144172668457,
+      "learning_rate": 7.055093079412536e-06,
+      "loss": 0.406,
+      "step": 299
+    },
+    {
+      "epoch": 0.30272452068617556,
+      "grad_norm": 6.949483394622803,
+      "learning_rate": 7.04597040517961e-06,
+      "loss": 0.2363,
+      "step": 300
+    },
+    {
+      "epoch": 0.30272452068617556,
+      "eval_accuracy": 0.8422247446083996,
+      "eval_loss": 0.29199618101119995,
+      "eval_runtime": 62.2656,
+      "eval_samples_per_second": 28.298,
+      "eval_steps_per_second": 3.549,
+      "step": 300
+    },
+    {
+      "epoch": 0.30373360242179614,
+      "grad_norm": 11.153295516967773,
+      "learning_rate": 7.036809863190972e-06,
+      "loss": 0.367,
+      "step": 301
+    },
+    {
+      "epoch": 0.30474268415741673,
+      "grad_norm": 7.238802909851074,
+      "learning_rate": 7.027611567331239e-06,
+      "loss": 0.3071,
+      "step": 302
+    },
+    {
+      "epoch": 0.3057517658930373,
+      "grad_norm": 5.204061031341553,
+      "learning_rate": 7.018375631954384e-06,
+      "loss": 0.1991,
+      "step": 303
+    },
+    {
+      "epoch": 0.3067608476286579,
+      "grad_norm": 4.6721930503845215,
+      "learning_rate": 7.0091021718823185e-06,
+      "loss": 0.1833,
+      "step": 304
+    },
+    {
+      "epoch": 0.3077699293642785,
+      "grad_norm": 4.843674182891846,
+      "learning_rate": 6.999791302403463e-06,
+      "loss": 0.2464,
+      "step": 305
+    },
+    {
+      "epoch": 0.3087790110998991,
+      "grad_norm": 5.238533973693848,
+      "learning_rate": 6.990443139271317e-06,
+      "loss": 0.2551,
+      "step": 306
+    },
+    {
+      "epoch": 0.30978809283551967,
+      "grad_norm": 20.923328399658203,
+      "learning_rate": 6.981057798703019e-06,
+      "loss": 0.7206,
+      "step": 307
+    },
+    {
+      "epoch": 0.31079717457114026,
+      "grad_norm": 4.751791954040527,
+      "learning_rate": 6.971635397377895e-06,
+      "loss": 0.2031,
+      "step": 308
+    },
+    {
+      "epoch": 0.31180625630676084,
+      "grad_norm": 9.50137996673584,
+      "learning_rate": 6.962176052436019e-06,
+      "loss": 0.4901,
+      "step": 309
+    },
+    {
+      "epoch": 0.31281533804238143,
+      "grad_norm": 10.486673355102539,
+      "learning_rate": 6.952679881476746e-06,
+      "loss": 0.5644,
+      "step": 310
+    },
+    {
+      "epoch": 0.313824419778002,
+      "grad_norm": 8.85893726348877,
+      "learning_rate": 6.94314700255726e-06,
+      "loss": 0.5932,
+      "step": 311
+    },
+    {
+      "epoch": 0.3148335015136226,
+      "grad_norm": 33.33445358276367,
+      "learning_rate": 6.933577534191101e-06,
+      "loss": 1.3013,
+      "step": 312
+    },
+    {
+      "epoch": 0.3158425832492432,
+      "grad_norm": 29.183631896972656,
+      "learning_rate": 6.923971595346686e-06,
+      "loss": 0.4947,
+      "step": 313
+    },
+    {
+      "epoch": 0.3168516649848638,
+      "grad_norm": 7.223900318145752,
+      "learning_rate": 6.914329305445844e-06,
+      "loss": 0.3102,
+      "step": 314
+    },
+    {
+      "epoch": 0.31786074672048437,
+      "grad_norm": 4.909074306488037,
+      "learning_rate": 6.904650784362317e-06,
+      "loss": 0.223,
+      "step": 315
+    },
+    {
+      "epoch": 0.31886982845610495,
+      "grad_norm": 5.134761333465576,
+      "learning_rate": 6.89493615242028e-06,
+      "loss": 0.2572,
+      "step": 316
+    },
+    {
+      "epoch": 0.31987891019172554,
+      "grad_norm": 4.783396244049072,
+      "learning_rate": 6.885185530392841e-06,
+      "loss": 0.2743,
+      "step": 317
+    },
+    {
+      "epoch": 0.32088799192734613,
+      "grad_norm": 6.382033824920654,
+      "learning_rate": 6.875399039500535e-06,
+      "loss": 0.3093,
+      "step": 318
+    },
+    {
+      "epoch": 0.3218970736629667,
+      "grad_norm": 7.4557318687438965,
+      "learning_rate": 6.865576801409828e-06,
+      "loss": 0.4611,
+      "step": 319
+    },
+    {
+      "epoch": 0.3229061553985873,
+      "grad_norm": 9.1113862991333,
+      "learning_rate": 6.855718938231597e-06,
+      "loss": 0.5509,
+      "step": 320
+    },
+    {
+      "epoch": 0.3229061553985873,
+      "eval_accuracy": 0.8422247446083996,
+      "eval_loss": 0.2692955434322357,
+      "eval_runtime": 62.6677,
+      "eval_samples_per_second": 28.117,
+      "eval_steps_per_second": 3.527,
+      "step": 320
+    },
+    {
+      "epoch": 0.3239152371342079,
+      "grad_norm": 4.367307186126709,
+      "learning_rate": 6.845825572519606e-06,
+      "loss": 0.2284,
+      "step": 321
+    },
+    {
+      "epoch": 0.3249243188698285,
+      "grad_norm": 4.980352878570557,
+      "learning_rate": 6.8358968272689995e-06,
+      "loss": 0.2197,
+      "step": 322
+    },
+    {
+      "epoch": 0.32593340060544906,
+      "grad_norm": 3.6899354457855225,
+      "learning_rate": 6.825932825914758e-06,
+      "loss": 0.1665,
+      "step": 323
+    },
+    {
+      "epoch": 0.32694248234106965,
+      "grad_norm": 5.287685394287109,
+      "learning_rate": 6.815933692330168e-06,
+      "loss": 0.2254,
+      "step": 324
+    },
+    {
+      "epoch": 0.32795156407669024,
+      "grad_norm": 4.298030376434326,
+      "learning_rate": 6.805899550825285e-06,
+      "loss": 0.1869,
+      "step": 325
+    },
+    {
+      "epoch": 0.32896064581231077,
+      "grad_norm": 6.648251533508301,
+      "learning_rate": 6.795830526145385e-06,
+      "loss": 0.438,
+      "step": 326
+    },
+    {
+      "epoch": 0.32996972754793136,
+      "grad_norm": 8.279413223266602,
+      "learning_rate": 6.785726743469415e-06,
+      "loss": 0.3674,
+      "step": 327
+    },
+    {
+      "epoch": 0.33097880928355194,
+      "grad_norm": 8.667623519897461,
+      "learning_rate": 6.775588328408435e-06,
+      "loss": 0.2876,
+      "step": 328
+    },
+    {
+      "epoch": 0.33198789101917253,
+      "grad_norm": 4.858876705169678,
+      "learning_rate": 6.765415407004061e-06,
+      "loss": 0.2051,
+      "step": 329
+    },
+    {
+      "epoch": 0.3329969727547931,
+      "grad_norm": 5.615779876708984,
+      "learning_rate": 6.75520810572689e-06,
+      "loss": 0.2162,
+      "step": 330
+    },
+    {
+      "epoch": 0.3340060544904137,
+      "grad_norm": 5.255247116088867,
+      "learning_rate": 6.744966551474935e-06,
+      "loss": 0.2493,
+      "step": 331
+    },
+    {
+      "epoch": 0.3350151362260343,
+      "grad_norm": 4.313990116119385,
+      "learning_rate": 6.734690871572044e-06,
+      "loss": 0.1736,
+      "step": 332
+    },
+    {
+      "epoch": 0.3360242179616549,
+      "grad_norm": 6.205388069152832,
+      "learning_rate": 6.72438119376632e-06,
+      "loss": 0.1372,
+      "step": 333
+    },
+    {
+      "epoch": 0.33703329969727547,
+      "grad_norm": 4.961912631988525,
+      "learning_rate": 6.714037646228529e-06,
+      "loss": 0.1083,
+      "step": 334
+    },
+    {
+      "epoch": 0.33804238143289606,
+      "grad_norm": 11.366336822509766,
+      "learning_rate": 6.703660357550507e-06,
+      "loss": 0.1794,
+      "step": 335
+    },
+    {
+      "epoch": 0.33905146316851664,
+      "grad_norm": 36.39216613769531,
+      "learning_rate": 6.693249456743565e-06,
+      "loss": 0.8015,
+      "step": 336
+    },
+    {
+      "epoch": 0.34006054490413723,
+      "grad_norm": 26.968290328979492,
+      "learning_rate": 6.682805073236883e-06,
+      "loss": 0.5109,
+      "step": 337
+    },
+    {
+      "epoch": 0.3410696266397578,
+      "grad_norm": 12.663142204284668,
+      "learning_rate": 6.672327336875899e-06,
+      "loss": 0.5224,
+      "step": 338
+    },
+    {
+      "epoch": 0.3420787083753784,
+      "grad_norm": 7.804121017456055,
+      "learning_rate": 6.661816377920695e-06,
+      "loss": 0.2021,
+      "step": 339
+    },
+    {
+      "epoch": 0.343087790110999,
+      "grad_norm": 8.033084869384766,
+      "learning_rate": 6.651272327044385e-06,
+      "loss": 0.2942,
+      "step": 340
+    },
+    {
+      "epoch": 0.343087790110999,
+      "eval_accuracy": 0.840522133938706,
+      "eval_loss": 0.2765989303588867,
+      "eval_runtime": 62.3633,
+      "eval_samples_per_second": 28.254,
+      "eval_steps_per_second": 3.544,
+      "step": 340
+    },
+    {
+      "epoch": 0.3440968718466196,
+      "grad_norm": 6.994205951690674,
+      "learning_rate": 6.640695315331476e-06,
+      "loss": 0.422,
+      "step": 341
+    },
+    {
+      "epoch": 0.34510595358224017,
+      "grad_norm": 6.707622051239014,
+      "learning_rate": 6.630085474276255e-06,
+      "loss": 0.4127,
+      "step": 342
+    },
+    {
+      "epoch": 0.34611503531786075,
+      "grad_norm": 2.8467178344726562,
+      "learning_rate": 6.619442935781141e-06,
+      "loss": 0.0992,
+      "step": 343
+    },
+    {
+      "epoch": 0.34712411705348134,
+      "grad_norm": 3.68339467048645,
+      "learning_rate": 6.608767832155051e-06,
+      "loss": 0.1661,
+      "step": 344
+    },
+    {
+      "epoch": 0.3481331987891019,
+      "grad_norm": 4.811255931854248,
+      "learning_rate": 6.598060296111755e-06,
+      "loss": 0.1805,
+      "step": 345
+    },
+    {
+      "epoch": 0.3491422805247225,
+      "grad_norm": 7.682583332061768,
+      "learning_rate": 6.58732046076823e-06,
+      "loss": 0.4899,
+      "step": 346
+    },
+    {
+      "epoch": 0.3501513622603431,
+      "grad_norm": 4.097416400909424,
+      "learning_rate": 6.5765484596429905e-06,
+      "loss": 0.1297,
+      "step": 347
+    },
+    {
+      "epoch": 0.3511604439959637,
+      "grad_norm": 6.541982650756836,
+      "learning_rate": 6.565744426654449e-06,
+      "loss": 0.2545,
+      "step": 348
+    },
+    {
+      "epoch": 0.3521695257315843,
+      "grad_norm": 6.773824214935303,
+      "learning_rate": 6.554908496119232e-06,
+      "loss": 0.2685,
+      "step": 349
+    },
+    {
+      "epoch": 0.35317860746720486,
+      "grad_norm": 5.670194625854492,
+      "learning_rate": 6.544040802750526e-06,
+      "loss": 0.3245,
+      "step": 350
+    },
+    {
+      "epoch": 0.35418768920282545,
+      "grad_norm": 5.502919673919678,
+      "learning_rate": 6.5331414816563914e-06,
+      "loss": 0.282,
+      "step": 351
+    },
+    {
+      "epoch": 0.35519677093844604,
+      "grad_norm": 6.195550918579102,
+      "learning_rate": 6.52221066833809e-06,
+      "loss": 0.3791,
+      "step": 352
+    },
+    {
+      "epoch": 0.3562058526740666,
+      "grad_norm": 6.497583866119385,
+      "learning_rate": 6.511248498688395e-06,
+      "loss": 0.3853,
+      "step": 353
+    },
+    {
+      "epoch": 0.35721493440968716,
+      "grad_norm": 4.045083045959473,
+      "learning_rate": 6.500255108989904e-06,
+      "loss": 0.1618,
+      "step": 354
+    },
+    {
+      "epoch": 0.35822401614530774,
+      "grad_norm": 4.328484058380127,
+      "learning_rate": 6.489230635913346e-06,
+      "loss": 0.2393,
+      "step": 355
+    },
+    {
+      "epoch": 0.35923309788092833,
+      "grad_norm": 7.9082183837890625,
+      "learning_rate": 6.478175216515884e-06,
+      "loss": 0.2272,
+      "step": 356
+    },
+    {
+      "epoch": 0.3602421796165489,
+      "grad_norm": 3.21086049079895,
+      "learning_rate": 6.467088988239402e-06,
+      "loss": 0.0689,
+      "step": 357
+    },
+    {
+      "epoch": 0.3612512613521695,
+      "grad_norm": 5.211169242858887,
+      "learning_rate": 6.455972088908807e-06,
+      "loss": 0.1841,
+      "step": 358
+    },
+    {
+      "epoch": 0.3622603430877901,
+      "grad_norm": 5.808556079864502,
+      "learning_rate": 6.444824656730311e-06,
+      "loss": 0.1076,
+      "step": 359
+    },
+    {
+      "epoch": 0.3632694248234107,
+      "grad_norm": 16.821155548095703,
+      "learning_rate": 6.43364683028971e-06,
+      "loss": 0.4606,
+      "step": 360
+    },
+    {
+      "epoch": 0.3632694248234107,
+      "eval_accuracy": 0.8484676503972758,
+      "eval_loss": 0.3792904019355774,
+      "eval_runtime": 62.8768,
+      "eval_samples_per_second": 28.023,
+      "eval_steps_per_second": 3.515,
+      "step": 360
+    },
+    {
+      "epoch": 0.36427850655903127,
+      "grad_norm": 12.411088943481445,
+      "learning_rate": 6.422438748550666e-06,
+      "loss": 0.2406,
+      "step": 361
+    },
+    {
+      "epoch": 0.36528758829465185,
+      "grad_norm": 13.036988258361816,
+      "learning_rate": 6.411200550852978e-06,
+      "loss": 0.3551,
+      "step": 362
+    },
+    {
+      "epoch": 0.36629667003027244,
+      "grad_norm": 13.768930435180664,
+      "learning_rate": 6.3999323769108485e-06,
+      "loss": 0.3561,
+      "step": 363
+    },
+    {
+      "epoch": 0.36730575176589303,
+      "grad_norm": 4.591919422149658,
+      "learning_rate": 6.388634366811145e-06,
+      "loss": 0.1434,
+      "step": 364
+    },
+    {
+      "epoch": 0.3683148335015136,
+      "grad_norm": 9.278558731079102,
+      "learning_rate": 6.377306661011664e-06,
+      "loss": 0.445,
+      "step": 365
+    },
+    {
+      "epoch": 0.3693239152371342,
+      "grad_norm": 12.307679176330566,
+      "learning_rate": 6.365949400339378e-06,
+      "loss": 0.3966,
+      "step": 366
+    },
+    {
+      "epoch": 0.3703329969727548,
+      "grad_norm": 5.971735000610352,
+      "learning_rate": 6.354562725988691e-06,
+      "loss": 0.2508,
+      "step": 367
+    },
+    {
+      "epoch": 0.3713420787083754,
+      "grad_norm": 5.919165134429932,
+      "learning_rate": 6.343146779519681e-06,
+      "loss": 0.287,
+      "step": 368
+    },
+    {
+      "epoch": 0.37235116044399597,
+      "grad_norm": 4.112932205200195,
+      "learning_rate": 6.331701702856335e-06,
+      "loss": 0.1366,
+      "step": 369
+    },
+    {
+      "epoch": 0.37336024217961655,
+      "grad_norm": 9.273783683776855,
+      "learning_rate": 6.3202276382847925e-06,
+      "loss": 0.4608,
+      "step": 370
+    },
+    {
+      "epoch": 0.37436932391523714,
+      "grad_norm": 6.013513088226318,
+      "learning_rate": 6.308724728451572e-06,
+      "loss": 0.1759,
+      "step": 371
+    },
+    {
+      "epoch": 0.3753784056508577,
+      "grad_norm": 4.636185169219971,
+      "learning_rate": 6.2971931163618e-06,
+      "loss": 0.1464,
+      "step": 372
+    },
+    {
+      "epoch": 0.3763874873864783,
+      "grad_norm": 9.652426719665527,
+      "learning_rate": 6.285632945377429e-06,
+      "loss": 0.3786,
+      "step": 373
+    },
+    {
+      "epoch": 0.3773965691220989,
+      "grad_norm": 4.286848545074463,
+      "learning_rate": 6.274044359215461e-06,
+      "loss": 0.1569,
+      "step": 374
+    },
+    {
+      "epoch": 0.3784056508577195,
+      "grad_norm": 6.9270853996276855,
+      "learning_rate": 6.2624275019461545e-06,
+      "loss": 0.3606,
+      "step": 375
+    },
+    {
+      "epoch": 0.3794147325933401,
+      "grad_norm": 4.69766092300415,
+      "learning_rate": 6.250782517991241e-06,
+      "loss": 0.1682,
+      "step": 376
+    },
+    {
+      "epoch": 0.38042381432896066,
+      "grad_norm": 4.445621967315674,
+      "learning_rate": 6.239109552122122e-06,
+      "loss": 0.1959,
+      "step": 377
+    },
+    {
+      "epoch": 0.38143289606458125,
+      "grad_norm": 8.952390670776367,
+      "learning_rate": 6.227408749458073e-06,
+      "loss": 0.3696,
+      "step": 378
+    },
+    {
+      "epoch": 0.38244197780020184,
+      "grad_norm": 8.195367813110352,
+      "learning_rate": 6.215680255464441e-06,
+      "loss": 0.3198,
+      "step": 379
+    },
+    {
+      "epoch": 0.3834510595358224,
+      "grad_norm": 14.417852401733398,
+      "learning_rate": 6.203924215950831e-06,
+      "loss": 0.6757,
+      "step": 380
+    },
+    {
+      "epoch": 0.3834510595358224,
+      "eval_accuracy": 0.8513053348467651,
+      "eval_loss": 0.29195961356163025,
+      "eval_runtime": 62.6263,
+      "eval_samples_per_second": 28.135,
+      "eval_steps_per_second": 3.529,
+      "step": 380
+    },
+    {
+      "epoch": 0.384460141271443,
+      "grad_norm": 10.577422142028809,
+      "learning_rate": 6.192140777069298e-06,
+      "loss": 0.3435,
+      "step": 381
+    },
+    {
+      "epoch": 0.3854692230070636,
+      "grad_norm": 5.178398132324219,
+      "learning_rate": 6.180330085312526e-06,
+      "loss": 0.2577,
+      "step": 382
+    },
+    {
+      "epoch": 0.38647830474268413,
+      "grad_norm": 5.634678363800049,
+      "learning_rate": 6.168492287512014e-06,
+      "loss": 0.2692,
+      "step": 383
+    },
+    {
+      "epoch": 0.3874873864783047,
+      "grad_norm": 7.128296852111816,
+      "learning_rate": 6.156627530836239e-06,
+      "loss": 0.2777,
+      "step": 384
+    },
+    {
+      "epoch": 0.3884964682139253,
+      "grad_norm": 6.315018653869629,
+      "learning_rate": 6.144735962788837e-06,
+      "loss": 0.3489,
+      "step": 385
+    },
+    {
+      "epoch": 0.3895055499495459,
+      "grad_norm": 3.5851023197174072,
+      "learning_rate": 6.132817731206765e-06,
+      "loss": 0.1926,
+      "step": 386
+    },
+    {
+      "epoch": 0.3905146316851665,
+      "grad_norm": 10.211600303649902,
+      "learning_rate": 6.120872984258462e-06,
+      "loss": 0.4683,
+      "step": 387
+    },
+    {
+      "epoch": 0.39152371342078707,
+      "grad_norm": 12.310287475585938,
+      "learning_rate": 6.108901870442009e-06,
+      "loss": 0.3685,
+      "step": 388
+    },
+    {
+      "epoch": 0.39253279515640765,
+      "grad_norm": 6.385642051696777,
+      "learning_rate": 6.096904538583283e-06,
+      "loss": 0.2559,
+      "step": 389
+    },
+    {
+      "epoch": 0.39354187689202824,
+      "grad_norm": 7.468835830688477,
+      "learning_rate": 6.084881137834103e-06,
+      "loss": 0.4374,
+      "step": 390
+    },
+    {
+      "epoch": 0.39455095862764883,
+      "grad_norm": 2.5986368656158447,
+      "learning_rate": 6.072831817670382e-06,
+      "loss": 0.0992,
+      "step": 391
+    },
+    {
+      "epoch": 0.3955600403632694,
+      "grad_norm": 2.776925802230835,
+      "learning_rate": 6.060756727890262e-06,
+      "loss": 0.1379,
+      "step": 392
+    },
+    {
+      "epoch": 0.39656912209889,
+      "grad_norm": 5.025798797607422,
+      "learning_rate": 6.04865601861226e-06,
+      "loss": 0.37,
+      "step": 393
+    },
+    {
+      "epoch": 0.3975782038345106,
+      "grad_norm": 4.2647857666015625,
+      "learning_rate": 6.036529840273388e-06,
+      "loss": 0.2485,
+      "step": 394
+    },
+    {
+      "epoch": 0.3985872855701312,
+      "grad_norm": 3.0298309326171875,
+      "learning_rate": 6.0243783436273e-06,
+      "loss": 0.1594,
+      "step": 395
+    },
+    {
+      "epoch": 0.39959636730575177,
+      "grad_norm": 5.6464338302612305,
+      "learning_rate": 6.012201679742408e-06,
+      "loss": 0.3974,
+      "step": 396
+    },
+    {
+      "epoch": 0.40060544904137235,
+      "grad_norm": 4.060618877410889,
+      "learning_rate": 6e-06,
+      "loss": 0.2286,
+      "step": 397
+    },
+    {
+      "epoch": 0.40161453077699294,
+      "grad_norm": 4.622409820556641,
+      "learning_rate": 5.987773456092368e-06,
+      "loss": 0.315,
+      "step": 398
+    },
+    {
+      "epoch": 0.4026236125126135,
+      "grad_norm": 5.471047401428223,
+      "learning_rate": 5.9755222000209165e-06,
+      "loss": 0.4354,
+      "step": 399
+    },
+    {
+      "epoch": 0.4036326942482341,
+      "grad_norm": 3.929957628250122,
+      "learning_rate": 5.963246384094273e-06,
+      "loss": 0.2409,
+      "step": 400
+    },
+    {
+      "epoch": 0.4036326942482341,
+      "eval_accuracy": 0.8376844494892168,
+      "eval_loss": 0.28469741344451904,
+      "eval_runtime": 62.4393,
+      "eval_samples_per_second": 28.219,
+      "eval_steps_per_second": 3.539,
+      "step": 400
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 991,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 20,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9ce9b17901cea44c9a33ee1c2531e6752c15621e9455c217e5c6708f0a52225
+size 7096

value_head.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b6b7478af4de4f8e44c5b31fd204e116cb8a68b710bdbf4a4caa214a7ecd9f5
+size 10442

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)