Upload PEFT adapter for router (Gemma 3 27B)

.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

README.md

@@ -0,0 +1,208 @@
+---
+base_model: google/gemma-3-27b-it
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- axolotl
+- base_model:adapter:/tmp/base_model_dir/vertex-model-garden-restricted-us/gemma3/gemma-3-27b-it
+- lora
+- transformers
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.17.1

adapter_config.json

@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/tmp/base_model_dir/vertex-model-garden-restricted-us/gemma3/gemma-3-27b-it",
+  "bias": "none",
+  "corda_config": null,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": null,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 32,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": "model.language_model.layers.[\\d]+.(mlp|cross_attn|self_attn).(up|down|gate|q|k|v|o)_proj",
+  "target_parameters": [],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e21d15860fe7d9cb2819679fb3177b25c6239c8a641fb3f15eead5ec091801a5
+size 454196424

added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<image_soft_token>": 262144
+}

chat_template.jinja

@@ -0,0 +1,47 @@
+{{ bos_token }}
+{%- if messages[0]['role'] == 'system' -%}
+    {%- if messages[0]['content'] is string -%}
+        {%- set first_user_prefix = messages[0]['content'] + '
+
+' -%}
+    {%- else -%}
+        {%- set first_user_prefix = messages[0]['content'][0]['text'] + '
+
+' -%}
+    {%- endif -%}
+    {%- set loop_messages = messages[1:] -%}
+{%- else -%}
+    {%- set first_user_prefix = "" -%}
+    {%- set loop_messages = messages -%}
+{%- endif -%}
+{%- for message in loop_messages -%}
+    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
+        {{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
+    {%- endif -%}
+    {%- if (message['role'] == 'assistant') -%}
+        {%- set role = "model" -%}
+    {%- else -%}
+        {%- set role = message['role'] -%}
+    {%- endif -%}
+    {{ '<start_of_turn>' + role + '
+' + (first_user_prefix if loop.first else "") }}
+    {%- if message['content'] is string -%}
+        {{ message['content'] | trim }}
+    {%- elif message['content'] is iterable -%}
+        {%- for item in message['content'] -%}
+            {%- if item['type'] == 'image' -%}
+                {{ '<start_of_image>' }}
+            {%- elif item['type'] == 'text' -%}
+                {{ item['text'] | trim }}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- else -%}
+        {{ raise_exception("Invalid content type") }}
+    {%- endif -%}
+    {{ '<end_of_turn>
+' }}
+{%- endfor -%}
+{%- if add_generation_prompt -%}
+    {{'<start_of_turn>model
+'}}
+{%- endif -%}

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7b85a0f877b2a027130ef33bb5e3bd0e4a33796590fefe0470b1b20ae56c989
+size 231600373

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c02c96188c753217eaed2cce8bd49c10cdd24c8733c073701acc6be9aa22ea3
+size 16389

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e98ecb6cf1b8c208b6b576b0e1841164cc32019b7ec4a27ca3d1ef90567d72bd
+size 16389

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2978b2afaaf3e65982e51202789a50a0a71b75a263114181333ad97c50d69ecb
+size 16389

rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef2689baced8212074ad1c520d7304b3a8e4dedffee069f90dae7f650310dea1
+size 16389

rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ced30595f225935935a8e1883dcf3a0d246703ec673f27a6ab99c891dde86d10
+size 16389

rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0033ed388ef2b38db06e24aca4f6e632a2588efcf526eeb2bc6833513a772c16
+size 16389

rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:326782d010f7eb214ef9e4541f77c72433847562dd692fd9f97fd5f6eb21eb7c
+size 16389

rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d26cc39928423d7d5b6f37f018dc249af97bf5b444da2d8fc97cf681400a6f1e
+size 16389

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5179285edeec2f930b98e9072708f794e4a735a124f02b95a7c981c9c975ca0
+size 1465

special_tokens_map.json

@@ -0,0 +1,33 @@
+{
+  "boi_token": "<start_of_image>",
+  "bos_token": {
+    "content": "<bos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eoi_token": "<end_of_image>",
+  "eos_token": {
+    "content": "<eos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "image_token": "<image_soft_token>",
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4667f2089529e8e7657cfb6d1c19910ae71ff5f28aa7ab2ff2763330affad795
+size 33384568

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1299c11d7cf632ef3b4e11937501358ada021bbdf7c47638d13c0ee982f2e79c
+size 4689074

trainer_state.json

@@ -0,0 +1,1908 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 61,
+  "global_step": 183,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0,
+      "eval_loss": 1.987602710723877,
+      "eval_runtime": 21.27,
+      "eval_samples_per_second": 38.458,
+      "eval_steps_per_second": 2.445,
+      "memory/device_reserved (GiB)": 65.8,
+      "memory/max_active (GiB)": 48.55,
+      "memory/max_allocated (GiB)": 48.55,
+      "step": 0
+    },
+    {
+      "epoch": 0.01639344262295082,
+      "grad_norm": 2.1048970222473145,
+      "learning_rate": 0.0,
+      "loss": 1.9829,
+      "memory/device_reserved (GiB)": 33.32,
+      "memory/max_active (GiB)": 32.48,
+      "memory/max_allocated (GiB)": 32.48,
+      "step": 1
+    },
+    {
+      "epoch": 0.03278688524590164,
+      "grad_norm": 2.132734537124634,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 1.9692,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 2
+    },
+    {
+      "epoch": 0.04918032786885246,
+      "grad_norm": 2.0775797367095947,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 1.9615,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 3
+    },
+    {
+      "epoch": 0.06557377049180328,
+      "grad_norm": 1.9622066020965576,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 1.9353,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 4
+    },
+    {
+      "epoch": 0.08196721311475409,
+      "grad_norm": 1.344926118850708,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 1.83,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 5
+    },
+    {
+      "epoch": 0.09836065573770492,
+      "grad_norm": 0.977587878704071,
+      "learning_rate": 5.555555555555556e-05,
+      "loss": 1.7399,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 6
+    },
+    {
+      "epoch": 0.11475409836065574,
+      "grad_norm": 0.8041461706161499,
+      "learning_rate": 6.666666666666667e-05,
+      "loss": 1.6341,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 7
+    },
+    {
+      "epoch": 0.13114754098360656,
+      "grad_norm": 0.6574768424034119,
+      "learning_rate": 7.777777777777778e-05,
+      "loss": 1.5328,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 8
+    },
+    {
+      "epoch": 0.14754098360655737,
+      "grad_norm": 0.5925188660621643,
+      "learning_rate": 8.888888888888889e-05,
+      "loss": 1.4402,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 9
+    },
+    {
+      "epoch": 0.16393442622950818,
+      "grad_norm": 0.6970644593238831,
+      "learning_rate": 0.0001,
+      "loss": 1.4063,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 10
+    },
+    {
+      "epoch": 0.18032786885245902,
+      "grad_norm": 0.8010526895523071,
+      "learning_rate": 0.00011111111111111112,
+      "loss": 1.3407,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 11
+    },
+    {
+      "epoch": 0.19672131147540983,
+      "grad_norm": 0.759565532207489,
+      "learning_rate": 0.00012222222222222224,
+      "loss": 1.2844,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 12
+    },
+    {
+      "epoch": 0.21311475409836064,
+      "grad_norm": 0.6551400423049927,
+      "learning_rate": 0.00013333333333333334,
+      "loss": 1.223,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 13
+    },
+    {
+      "epoch": 0.22950819672131148,
+      "grad_norm": 0.557424008846283,
+      "learning_rate": 0.00014444444444444444,
+      "loss": 1.146,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 14
+    },
+    {
+      "epoch": 0.2459016393442623,
+      "grad_norm": 0.4815196990966797,
+      "learning_rate": 0.00015555555555555556,
+      "loss": 1.087,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 15
+    },
+    {
+      "epoch": 0.26229508196721313,
+      "grad_norm": 0.4362974762916565,
+      "learning_rate": 0.0001666666666666667,
+      "loss": 1.0415,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 16
+    },
+    {
+      "epoch": 0.2786885245901639,
+      "grad_norm": 0.44927817583084106,
+      "learning_rate": 0.00017777777777777779,
+      "loss": 1.0049,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 17
+    },
+    {
+      "epoch": 0.29508196721311475,
+      "grad_norm": 0.49147677421569824,
+      "learning_rate": 0.00018888888888888888,
+      "loss": 0.9776,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 18
+    },
+    {
+      "epoch": 0.3114754098360656,
+      "grad_norm": 0.39037176966667175,
+      "learning_rate": 0.0002,
+      "loss": 0.9379,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 19
+    },
+    {
+      "epoch": 0.32786885245901637,
+      "grad_norm": 0.28441116213798523,
+      "learning_rate": 0.0001999818745523526,
+      "loss": 0.9064,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 20
+    },
+    {
+      "epoch": 0.3442622950819672,
+      "grad_norm": 0.3153410851955414,
+      "learning_rate": 0.00019992750478004738,
+      "loss": 0.8899,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 21
+    },
+    {
+      "epoch": 0.36065573770491804,
+      "grad_norm": 0.28423598408699036,
+      "learning_rate": 0.00019983691039261357,
+      "loss": 0.8671,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 22
+    },
+    {
+      "epoch": 0.3770491803278688,
+      "grad_norm": 0.2221442312002182,
+      "learning_rate": 0.00019971012423132775,
+      "loss": 0.8383,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 23
+    },
+    {
+      "epoch": 0.39344262295081966,
+      "grad_norm": 0.20366857945919037,
+      "learning_rate": 0.00019954719225730847,
+      "loss": 0.8386,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 24
+    },
+    {
+      "epoch": 0.4098360655737705,
+      "grad_norm": 0.22895343601703644,
+      "learning_rate": 0.00019934817353485501,
+      "loss": 0.8219,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 25
+    },
+    {
+      "epoch": 0.4262295081967213,
+      "grad_norm": 0.2062416672706604,
+      "learning_rate": 0.00019911314021003613,
+      "loss": 0.8057,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 26
+    },
+    {
+      "epoch": 0.4426229508196721,
+      "grad_norm": 0.18311655521392822,
+      "learning_rate": 0.00019884217748453623,
+      "loss": 0.804,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 27
+    },
+    {
+      "epoch": 0.45901639344262296,
+      "grad_norm": 0.19581766426563263,
+      "learning_rate": 0.00019853538358476932,
+      "loss": 0.7831,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 28
+    },
+    {
+      "epoch": 0.47540983606557374,
+      "grad_norm": 0.19657522439956665,
+      "learning_rate": 0.00019819286972627066,
+      "loss": 0.7815,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 29
+    },
+    {
+      "epoch": 0.4918032786885246,
+      "grad_norm": 0.1872270107269287,
+      "learning_rate": 0.00019781476007338058,
+      "loss": 0.7798,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 30
+    },
+    {
+      "epoch": 0.5081967213114754,
+      "grad_norm": 0.17483267188072205,
+      "learning_rate": 0.00019740119169423337,
+      "loss": 0.7738,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 31
+    },
+    {
+      "epoch": 0.5245901639344263,
+      "grad_norm": 0.18843160569667816,
+      "learning_rate": 0.00019695231451106912,
+      "loss": 0.7684,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 32
+    },
+    {
+      "epoch": 0.5409836065573771,
+      "grad_norm": 0.1772245317697525,
+      "learning_rate": 0.0001964682912458856,
+      "loss": 0.7661,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 33
+    },
+    {
+      "epoch": 0.5573770491803278,
+      "grad_norm": 0.1664779931306839,
+      "learning_rate": 0.00019594929736144976,
+      "loss": 0.7521,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 34
+    },
+    {
+      "epoch": 0.5737704918032787,
+      "grad_norm": 0.17712299525737762,
+      "learning_rate": 0.00019539552099769126,
+      "loss": 0.7545,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 35
+    },
+    {
+      "epoch": 0.5901639344262295,
+      "grad_norm": 0.1669715940952301,
+      "learning_rate": 0.00019480716290349995,
+      "loss": 0.7549,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 36
+    },
+    {
+      "epoch": 0.6065573770491803,
+      "grad_norm": 0.17512178421020508,
+      "learning_rate": 0.00019418443636395248,
+      "loss": 0.7451,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 37
+    },
+    {
+      "epoch": 0.6229508196721312,
+      "grad_norm": 0.1587410718202591,
+      "learning_rate": 0.00019352756712299468,
+      "loss": 0.7317,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 38
+    },
+    {
+      "epoch": 0.639344262295082,
+      "grad_norm": 0.1550321877002716,
+      "learning_rate": 0.00019283679330160726,
+      "loss": 0.7294,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 39
+    },
+    {
+      "epoch": 0.6557377049180327,
+      "grad_norm": 0.19032655656337738,
+      "learning_rate": 0.000192112365311485,
+      "loss": 0.7241,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 40
+    },
+    {
+      "epoch": 0.6721311475409836,
+      "grad_norm": 0.15234725177288055,
+      "learning_rate": 0.0001913545457642601,
+      "loss": 0.7165,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 41
+    },
+    {
+      "epoch": 0.6885245901639344,
+      "grad_norm": 0.1558808833360672,
+      "learning_rate": 0.0001905636093763031,
+      "loss": 0.7212,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 42
+    },
+    {
+      "epoch": 0.7049180327868853,
+      "grad_norm": 0.15694938600063324,
+      "learning_rate": 0.00018973984286913584,
+      "loss": 0.7055,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 43
+    },
+    {
+      "epoch": 0.7213114754098361,
+      "grad_norm": 0.14805985987186432,
+      "learning_rate": 0.00018888354486549237,
+      "loss": 0.7137,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 44
+    },
+    {
+      "epoch": 0.7377049180327869,
+      "grad_norm": 0.14736340939998627,
+      "learning_rate": 0.00018799502578106534,
+      "loss": 0.6931,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 45
+    },
+    {
+      "epoch": 0.7540983606557377,
+      "grad_norm": 0.22682452201843262,
+      "learning_rate": 0.00018707460771197774,
+      "loss": 0.6979,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 46
+    },
+    {
+      "epoch": 0.7704918032786885,
+      "grad_norm": 0.14981690049171448,
+      "learning_rate": 0.00018612262431802007,
+      "loss": 0.7045,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 47
+    },
+    {
+      "epoch": 0.7868852459016393,
+      "grad_norm": 0.14373798668384552,
+      "learning_rate": 0.0001851394207016957,
+      "loss": 0.7089,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 48
+    },
+    {
+      "epoch": 0.8032786885245902,
+      "grad_norm": 0.1548561304807663,
+      "learning_rate": 0.00018412535328311814,
+      "loss": 0.6927,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 49
+    },
+    {
+      "epoch": 0.819672131147541,
+      "grad_norm": 0.1519075483083725,
+      "learning_rate": 0.00018308078967080546,
+      "loss": 0.6908,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 50
+    },
+    {
+      "epoch": 0.8360655737704918,
+      "grad_norm": 0.15933729708194733,
+      "learning_rate": 0.00018200610852841913,
+      "loss": 0.6732,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 51
+    },
+    {
+      "epoch": 0.8524590163934426,
+      "grad_norm": 0.14739681780338287,
+      "learning_rate": 0.00018090169943749476,
+      "loss": 0.6868,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 52
+    },
+    {
+      "epoch": 0.8688524590163934,
+      "grad_norm": 0.20240117609500885,
+      "learning_rate": 0.00017976796275621555,
+      "loss": 0.6789,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 53
+    },
+    {
+      "epoch": 0.8852459016393442,
+      "grad_norm": 0.15849758684635162,
+      "learning_rate": 0.00017860530947427875,
+      "loss": 0.6948,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 54
+    },
+    {
+      "epoch": 0.9016393442622951,
+      "grad_norm": 0.19192887842655182,
+      "learning_rate": 0.00017741416106390826,
+      "loss": 0.6705,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 55
+    },
+    {
+      "epoch": 0.9180327868852459,
+      "grad_norm": 0.14879536628723145,
+      "learning_rate": 0.0001761949493270671,
+      "loss": 0.6765,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 56
+    },
+    {
+      "epoch": 0.9344262295081968,
+      "grad_norm": 0.15834036469459534,
+      "learning_rate": 0.0001749481162389254,
+      "loss": 0.6639,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 57
+    },
+    {
+      "epoch": 0.9508196721311475,
+      "grad_norm": 0.1480400264263153,
+      "learning_rate": 0.0001736741137876405,
+      "loss": 0.6801,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 58
+    },
+    {
+      "epoch": 0.9672131147540983,
+      "grad_norm": 0.15552930533885956,
+      "learning_rate": 0.00017237340381050703,
+      "loss": 0.6721,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 59
+    },
+    {
+      "epoch": 0.9836065573770492,
+      "grad_norm": 0.15891212224960327,
+      "learning_rate": 0.0001710464578265369,
+      "loss": 0.669,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 60
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.154049351811409,
+      "learning_rate": 0.00016969375686552937,
+      "loss": 0.6585,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 61
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 0.6697277426719666,
+      "eval_runtime": 15.4593,
+      "eval_samples_per_second": 52.913,
+      "eval_steps_per_second": 3.364,
+      "memory/device_reserved (GiB)": 73.37,
+      "memory/max_active (GiB)": 48.8,
+      "memory/max_allocated (GiB)": 48.8,
+      "step": 61
+    },
+    {
+      "epoch": 1.0163934426229508,
+      "grad_norm": 0.15614724159240723,
+      "learning_rate": 0.00016831579129369346,
+      "loss": 0.6731,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 62
+    },
+    {
+      "epoch": 1.0327868852459017,
+      "grad_norm": 0.16120535135269165,
+      "learning_rate": 0.00016691306063588583,
+      "loss": 0.654,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 63
+    },
+    {
+      "epoch": 1.0491803278688525,
+      "grad_norm": 0.1574525684118271,
+      "learning_rate": 0.00016548607339452853,
+      "loss": 0.6625,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 64
+    },
+    {
+      "epoch": 1.0655737704918034,
+      "grad_norm": 0.1553254872560501,
+      "learning_rate": 0.00016403534686527225,
+      "loss": 0.6608,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 65
+    },
+    {
+      "epoch": 1.0819672131147542,
+      "grad_norm": 0.1639672964811325,
+      "learning_rate": 0.00016256140694947217,
+      "loss": 0.6601,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 66
+    },
+    {
+      "epoch": 1.098360655737705,
+      "grad_norm": 0.1632416546344757,
+      "learning_rate": 0.00016106478796354382,
+      "loss": 0.6588,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 67
+    },
+    {
+      "epoch": 1.1147540983606556,
+      "grad_norm": 0.15430399775505066,
+      "learning_rate": 0.0001595460324452688,
+      "loss": 0.6559,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 68
+    },
+    {
+      "epoch": 1.1311475409836065,
+      "grad_norm": 0.14951854944229126,
+      "learning_rate": 0.00015800569095711982,
+      "loss": 0.6511,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 69
+    },
+    {
+      "epoch": 1.1475409836065573,
+      "grad_norm": 0.15766219794750214,
+      "learning_rate": 0.00015644432188667695,
+      "loss": 0.6476,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 70
+    },
+    {
+      "epoch": 1.1639344262295082,
+      "grad_norm": 0.15970833599567413,
+      "learning_rate": 0.000154862491244207,
+      "loss": 0.6491,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 71
+    },
+    {
+      "epoch": 1.180327868852459,
+      "grad_norm": 0.1682189404964447,
+      "learning_rate": 0.00015326077245747999,
+      "loss": 0.6448,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 72
+    },
+    {
+      "epoch": 1.1967213114754098,
+      "grad_norm": 0.1654193103313446,
+      "learning_rate": 0.0001516397461638962,
+      "loss": 0.647,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 73
+    },
+    {
+      "epoch": 1.2131147540983607,
+      "grad_norm": 0.1643504798412323,
+      "learning_rate": 0.00015000000000000001,
+      "loss": 0.6411,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 74
+    },
+    {
+      "epoch": 1.2295081967213115,
+      "grad_norm": 0.15997686982154846,
+      "learning_rate": 0.00014834212838845637,
+      "loss": 0.634,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 75
+    },
+    {
+      "epoch": 1.2459016393442623,
+      "grad_norm": 0.16089341044425964,
+      "learning_rate": 0.00014666673232256738,
+      "loss": 0.6323,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 76
+    },
+    {
+      "epoch": 1.2622950819672132,
+      "grad_norm": 0.1618468165397644,
+      "learning_rate": 0.0001449744191484066,
+      "loss": 0.6363,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 77
+    },
+    {
+      "epoch": 1.278688524590164,
+      "grad_norm": 0.16787968575954437,
+      "learning_rate": 0.00014326580234465085,
+      "loss": 0.6393,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 78
+    },
+    {
+      "epoch": 1.2950819672131146,
+      "grad_norm": 0.16223575174808502,
+      "learning_rate": 0.00014154150130018866,
+      "loss": 0.6343,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 79
+    },
+    {
+      "epoch": 1.3114754098360657,
+      "grad_norm": 0.1583593338727951,
+      "learning_rate": 0.00013980214108958624,
+      "loss": 0.6279,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 80
+    },
+    {
+      "epoch": 1.3278688524590163,
+      "grad_norm": 0.17309962213039398,
+      "learning_rate": 0.0001380483522464923,
+      "loss": 0.6309,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 81
+    },
+    {
+      "epoch": 1.3442622950819672,
+      "grad_norm": 0.1649285852909088,
+      "learning_rate": 0.0001362807705350641,
+      "loss": 0.6278,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 82
+    },
+    {
+      "epoch": 1.360655737704918,
+      "grad_norm": 0.1694308966398239,
+      "learning_rate": 0.00013450003671949706,
+      "loss": 0.6268,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 83
+    },
+    {
+      "epoch": 1.3770491803278688,
+      "grad_norm": 0.15790660679340363,
+      "learning_rate": 0.00013270679633174218,
+      "loss": 0.6124,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 84
+    },
+    {
+      "epoch": 1.3934426229508197,
+      "grad_norm": 0.17729204893112183,
+      "learning_rate": 0.00013090169943749476,
+      "loss": 0.625,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 85
+    },
+    {
+      "epoch": 1.4098360655737705,
+      "grad_norm": 0.1664123386144638,
+      "learning_rate": 0.0001290854004005399,
+      "loss": 0.6103,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 86
+    },
+    {
+      "epoch": 1.4262295081967213,
+      "grad_norm": 0.17104221880435944,
+      "learning_rate": 0.0001272585576455398,
+      "loss": 0.6105,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 87
+    },
+    {
+      "epoch": 1.4426229508196722,
+      "grad_norm": 0.18972143530845642,
+      "learning_rate": 0.00012542183341934872,
+      "loss": 0.6135,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 88
+    },
+    {
+      "epoch": 1.459016393442623,
+      "grad_norm": 0.17770206928253174,
+      "learning_rate": 0.00012357589355094275,
+      "loss": 0.6057,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 89
+    },
+    {
+      "epoch": 1.4754098360655736,
+      "grad_norm": 0.1827758103609085,
+      "learning_rate": 0.00012172140721005079,
+      "loss": 0.6115,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 90
+    },
+    {
+      "epoch": 1.4918032786885247,
+      "grad_norm": 0.17028623819351196,
+      "learning_rate": 0.00011985904666457455,
+      "loss": 0.6143,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 91
+    },
+    {
+      "epoch": 1.5081967213114753,
+      "grad_norm": 0.17973949015140533,
+      "learning_rate": 0.00011798948703688539,
+      "loss": 0.6155,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 92
+    },
+    {
+      "epoch": 1.5245901639344264,
+      "grad_norm": 0.17090703547000885,
+      "learning_rate": 0.00011611340605908642,
+      "loss": 0.6121,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 93
+    },
+    {
+      "epoch": 1.540983606557377,
+      "grad_norm": 0.16992104053497314,
+      "learning_rate": 0.00011423148382732853,
+      "loss": 0.6176,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 94
+    },
+    {
+      "epoch": 1.5573770491803278,
+      "grad_norm": 0.17985564470291138,
+      "learning_rate": 0.00011234440255526948,
+      "loss": 0.6104,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 95
+    },
+    {
+      "epoch": 1.5737704918032787,
+      "grad_norm": 0.1746932864189148,
+      "learning_rate": 0.00011045284632676536,
+      "loss": 0.6123,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 96
+    },
+    {
+      "epoch": 1.5901639344262295,
+      "grad_norm": 0.17701321840286255,
+      "learning_rate": 0.00010855750084788398,
+      "loss": 0.6174,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 97
+    },
+    {
+      "epoch": 1.6065573770491803,
+      "grad_norm": 0.17487658560276031,
+      "learning_rate": 0.00010665905319833041,
+      "loss": 0.6065,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 98
+    },
+    {
+      "epoch": 1.6229508196721312,
+      "grad_norm": 0.1731928288936615,
+      "learning_rate": 0.00010475819158237425,
+      "loss": 0.6054,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 99
+    },
+    {
+      "epoch": 1.639344262295082,
+      "grad_norm": 0.17205819487571716,
+      "learning_rate": 0.00010285560507936961,
+      "loss": 0.6033,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 100
+    },
+    {
+      "epoch": 1.6557377049180326,
+      "grad_norm": 0.17849475145339966,
+      "learning_rate": 0.00010095198339395769,
+      "loss": 0.6027,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 101
+    },
+    {
+      "epoch": 1.6721311475409837,
+      "grad_norm": 0.17146815359592438,
+      "learning_rate": 9.904801660604234e-05,
+      "loss": 0.5987,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 102
+    },
+    {
+      "epoch": 1.6885245901639343,
+      "grad_norm": 0.47297370433807373,
+      "learning_rate": 9.71443949206304e-05,
+      "loss": 0.6055,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 103
+    },
+    {
+      "epoch": 1.7049180327868854,
+      "grad_norm": 0.1885872781276703,
+      "learning_rate": 9.524180841762577e-05,
+      "loss": 0.5936,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 104
+    },
+    {
+      "epoch": 1.721311475409836,
+      "grad_norm": 0.1687517911195755,
+      "learning_rate": 9.334094680166962e-05,
+      "loss": 0.6003,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 105
+    },
+    {
+      "epoch": 1.737704918032787,
+      "grad_norm": 0.16893097758293152,
+      "learning_rate": 9.144249915211605e-05,
+      "loss": 0.5874,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 106
+    },
+    {
+      "epoch": 1.7540983606557377,
+      "grad_norm": 0.18470335006713867,
+      "learning_rate": 8.954715367323468e-05,
+      "loss": 0.5953,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 107
+    },
+    {
+      "epoch": 1.7704918032786885,
+      "grad_norm": 0.18364070355892181,
+      "learning_rate": 8.765559744473053e-05,
+      "loss": 0.6019,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 108
+    },
+    {
+      "epoch": 1.7868852459016393,
+      "grad_norm": 0.17885719239711761,
+      "learning_rate": 8.57685161726715e-05,
+      "loss": 0.6081,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 109
+    },
+    {
+      "epoch": 1.8032786885245902,
+      "grad_norm": 0.17162950336933136,
+      "learning_rate": 8.38865939409136e-05,
+      "loss": 0.5941,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 110
+    },
+    {
+      "epoch": 1.819672131147541,
+      "grad_norm": 0.1798235923051834,
+      "learning_rate": 8.201051296311462e-05,
+      "loss": 0.5948,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 111
+    },
+    {
+      "epoch": 1.8360655737704918,
+      "grad_norm": 0.18231722712516785,
+      "learning_rate": 8.014095333542548e-05,
+      "loss": 0.5784,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 112
+    },
+    {
+      "epoch": 1.8524590163934427,
+      "grad_norm": 0.17540279030799866,
+      "learning_rate": 7.827859278994925e-05,
+      "loss": 0.593,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 113
+    },
+    {
+      "epoch": 1.8688524590163933,
+      "grad_norm": 0.17537765204906464,
+      "learning_rate": 7.642410644905726e-05,
+      "loss": 0.5902,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 114
+    },
+    {
+      "epoch": 1.8852459016393444,
+      "grad_norm": 0.17418642342090607,
+      "learning_rate": 7.457816658065134e-05,
+      "loss": 0.6039,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 115
+    },
+    {
+      "epoch": 1.901639344262295,
+      "grad_norm": 0.17102915048599243,
+      "learning_rate": 7.274144235446023e-05,
+      "loss": 0.5853,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 116
+    },
+    {
+      "epoch": 1.918032786885246,
+      "grad_norm": 0.1854507327079773,
+      "learning_rate": 7.09145995994601e-05,
+      "loss": 0.592,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 117
+    },
+    {
+      "epoch": 1.9344262295081966,
+      "grad_norm": 0.17946381866931915,
+      "learning_rate": 6.909830056250527e-05,
+      "loss": 0.5818,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 118
+    },
+    {
+      "epoch": 1.9508196721311475,
+      "grad_norm": 0.17315614223480225,
+      "learning_rate": 6.729320366825784e-05,
+      "loss": 0.5939,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 119
+    },
+    {
+      "epoch": 1.9672131147540983,
+      "grad_norm": 0.17754870653152466,
+      "learning_rate": 6.549996328050296e-05,
+      "loss": 0.5875,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 120
+    },
+    {
+      "epoch": 1.9836065573770492,
+      "grad_norm": 0.17100785672664642,
+      "learning_rate": 6.371922946493591e-05,
+      "loss": 0.5864,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 121
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.1712140440940857,
+      "learning_rate": 6.19516477535077e-05,
+      "loss": 0.5808,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 122
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 0.6171819567680359,
+      "eval_runtime": 15.4867,
+      "eval_samples_per_second": 52.82,
+      "eval_steps_per_second": 3.358,
+      "memory/device_reserved (GiB)": 73.37,
+      "memory/max_active (GiB)": 48.8,
+      "memory/max_allocated (GiB)": 48.8,
+      "step": 122
+    },
+    {
+      "epoch": 2.0163934426229506,
+      "grad_norm": 0.17188939452171326,
+      "learning_rate": 6.019785891041381e-05,
+      "loss": 0.5943,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 123
+    },
+    {
+      "epoch": 2.0327868852459017,
+      "grad_norm": 0.17433030903339386,
+      "learning_rate": 5.845849869981137e-05,
+      "loss": 0.5771,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 124
+    },
+    {
+      "epoch": 2.0491803278688523,
+      "grad_norm": 0.17646503448486328,
+      "learning_rate": 5.6734197655349156e-05,
+      "loss": 0.5889,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 125
+    },
+    {
+      "epoch": 2.0655737704918034,
+      "grad_norm": 0.18202351033687592,
+      "learning_rate": 5.5025580851593436e-05,
+      "loss": 0.5864,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 126
+    },
+    {
+      "epoch": 2.081967213114754,
+      "grad_norm": 0.17786148190498352,
+      "learning_rate": 5.333326767743263e-05,
+      "loss": 0.5869,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 127
+    },
+    {
+      "epoch": 2.098360655737705,
+      "grad_norm": 0.18139372766017914,
+      "learning_rate": 5.1657871611543605e-05,
+      "loss": 0.5852,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 128
+    },
+    {
+      "epoch": 2.1147540983606556,
+      "grad_norm": 0.17242832481861115,
+      "learning_rate": 5.000000000000002e-05,
+      "loss": 0.5842,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 129
+    },
+    {
+      "epoch": 2.1311475409836067,
+      "grad_norm": 0.1741384118795395,
+      "learning_rate": 4.836025383610382e-05,
+      "loss": 0.5798,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 130
+    },
+    {
+      "epoch": 2.1475409836065573,
+      "grad_norm": 0.18157395720481873,
+      "learning_rate": 4.673922754252002e-05,
+      "loss": 0.5794,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 131
+    },
+    {
+      "epoch": 2.1639344262295084,
+      "grad_norm": 0.18070365488529205,
+      "learning_rate": 4.513750875579303e-05,
+      "loss": 0.5807,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 132
+    },
+    {
+      "epoch": 2.180327868852459,
+      "grad_norm": 0.17728549242019653,
+      "learning_rate": 4.355567811332311e-05,
+      "loss": 0.5794,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 133
+    },
+    {
+      "epoch": 2.19672131147541,
+      "grad_norm": 0.17409607768058777,
+      "learning_rate": 4.19943090428802e-05,
+      "loss": 0.5805,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 134
+    },
+    {
+      "epoch": 2.2131147540983607,
+      "grad_norm": 0.17317527532577515,
+      "learning_rate": 4.045396755473121e-05,
+      "loss": 0.5761,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 135
+    },
+    {
+      "epoch": 2.2295081967213113,
+      "grad_norm": 0.18648019433021545,
+      "learning_rate": 3.893521203645618e-05,
+      "loss": 0.5706,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 136
+    },
+    {
+      "epoch": 2.2459016393442623,
+      "grad_norm": 0.1756281554698944,
+      "learning_rate": 3.7438593050527845e-05,
+      "loss": 0.5698,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 137
+    },
+    {
+      "epoch": 2.262295081967213,
+      "grad_norm": 0.17808155715465546,
+      "learning_rate": 3.5964653134727776e-05,
+      "loss": 0.5727,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 138
+    },
+    {
+      "epoch": 2.278688524590164,
+      "grad_norm": 0.2911638021469116,
+      "learning_rate": 3.45139266054715e-05,
+      "loss": 0.5782,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 139
+    },
+    {
+      "epoch": 2.2950819672131146,
+      "grad_norm": 0.18621869385242462,
+      "learning_rate": 3.308693936411421e-05,
+      "loss": 0.5746,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 140
+    },
+    {
+      "epoch": 2.3114754098360657,
+      "grad_norm": 0.18035250902175903,
+      "learning_rate": 3.1684208706306574e-05,
+      "loss": 0.5685,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 141
+    },
+    {
+      "epoch": 2.3278688524590163,
+      "grad_norm": 0.17170463502407074,
+      "learning_rate": 3.030624313447067e-05,
+      "loss": 0.5708,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 142
+    },
+    {
+      "epoch": 2.3442622950819674,
+      "grad_norm": 0.1746087372303009,
+      "learning_rate": 2.8953542173463133e-05,
+      "loss": 0.5706,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 143
+    },
+    {
+      "epoch": 2.360655737704918,
+      "grad_norm": 0.18402113020420074,
+      "learning_rate": 2.7626596189492983e-05,
+      "loss": 0.5712,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 144
+    },
+    {
+      "epoch": 2.3770491803278686,
+      "grad_norm": 0.1822223961353302,
+      "learning_rate": 2.6325886212359498e-05,
+      "loss": 0.5579,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 145
+    },
+    {
+      "epoch": 2.3934426229508197,
+      "grad_norm": 0.18002824485301971,
+      "learning_rate": 2.5051883761074614e-05,
+      "loss": 0.5693,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 146
+    },
+    {
+      "epoch": 2.4098360655737707,
+      "grad_norm": 0.17535454034805298,
+      "learning_rate": 2.3805050672932928e-05,
+      "loss": 0.5554,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 147
+    },
+    {
+      "epoch": 2.4262295081967213,
+      "grad_norm": 0.17632795870304108,
+      "learning_rate": 2.2585838936091754e-05,
+      "loss": 0.5584,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 148
+    },
+    {
+      "epoch": 2.442622950819672,
+      "grad_norm": 0.1735614687204361,
+      "learning_rate": 2.139469052572127e-05,
+      "loss": 0.5603,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 149
+    },
+    {
+      "epoch": 2.459016393442623,
+      "grad_norm": 0.17682839930057526,
+      "learning_rate": 2.0232037243784475e-05,
+      "loss": 0.5545,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 150
+    },
+    {
+      "epoch": 2.4754098360655736,
+      "grad_norm": 0.18232908844947815,
+      "learning_rate": 1.9098300562505266e-05,
+      "loss": 0.5619,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 151
+    },
+    {
+      "epoch": 2.4918032786885247,
+      "grad_norm": 0.1740136593580246,
+      "learning_rate": 1.7993891471580893e-05,
+      "loss": 0.5652,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 152
+    },
+    {
+      "epoch": 2.5081967213114753,
+      "grad_norm": 0.17907975614070892,
+      "learning_rate": 1.6919210329194533e-05,
+      "loss": 0.5675,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 153
+    },
+    {
+      "epoch": 2.5245901639344264,
+      "grad_norm": 0.17249107360839844,
+      "learning_rate": 1.587464671688187e-05,
+      "loss": 0.5646,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 154
+    },
+    {
+      "epoch": 2.540983606557377,
+      "grad_norm": 0.17240655422210693,
+      "learning_rate": 1.4860579298304312e-05,
+      "loss": 0.5706,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 155
+    },
+    {
+      "epoch": 2.557377049180328,
+      "grad_norm": 0.17784570157527924,
+      "learning_rate": 1.3877375681979943e-05,
+      "loss": 0.5641,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 156
+    },
+    {
+      "epoch": 2.5737704918032787,
+      "grad_norm": 0.17441287636756897,
+      "learning_rate": 1.2925392288022298e-05,
+      "loss": 0.5652,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 157
+    },
+    {
+      "epoch": 2.5901639344262293,
+      "grad_norm": 0.17398421466350555,
+      "learning_rate": 1.2004974218934695e-05,
+      "loss": 0.5731,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 158
+    },
+    {
+      "epoch": 2.6065573770491803,
+      "grad_norm": 0.1747160106897354,
+      "learning_rate": 1.1116455134507664e-05,
+      "loss": 0.5609,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 159
+    },
+    {
+      "epoch": 2.6229508196721314,
+      "grad_norm": 0.17699241638183594,
+      "learning_rate": 1.026015713086418e-05,
+      "loss": 0.5622,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 160
+    },
+    {
+      "epoch": 2.639344262295082,
+      "grad_norm": 0.17194418609142303,
+      "learning_rate": 9.436390623696911e-06,
+      "loss": 0.5603,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 161
+    },
+    {
+      "epoch": 2.6557377049180326,
+      "grad_norm": 0.1708948016166687,
+      "learning_rate": 8.645454235739903e-06,
+      "loss": 0.5606,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 162
+    },
+    {
+      "epoch": 2.6721311475409837,
+      "grad_norm": 0.17234554886817932,
+      "learning_rate": 7.887634688515e-06,
+      "loss": 0.5577,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 163
+    },
+    {
+      "epoch": 2.6885245901639343,
+      "grad_norm": 0.17192111909389496,
+      "learning_rate": 7.163206698392744e-06,
+      "loss": 0.5649,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 164
+    },
+    {
+      "epoch": 2.7049180327868854,
+      "grad_norm": 0.17286840081214905,
+      "learning_rate": 6.472432877005341e-06,
+      "loss": 0.5534,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 165
+    },
+    {
+      "epoch": 2.721311475409836,
+      "grad_norm": 0.1688247174024582,
+      "learning_rate": 5.8155636360475385e-06,
+      "loss": 0.5613,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 166
+    },
+    {
+      "epoch": 2.737704918032787,
+      "grad_norm": 0.168919175863266,
+      "learning_rate": 5.192837096500058e-06,
+      "loss": 0.5501,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 167
+    },
+    {
+      "epoch": 2.7540983606557377,
+      "grad_norm": 0.17062221467494965,
+      "learning_rate": 4.604479002308737e-06,
+      "loss": 0.5569,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 168
+    },
+    {
+      "epoch": 2.7704918032786887,
+      "grad_norm": 0.1695357859134674,
+      "learning_rate": 4.050702638550275e-06,
+      "loss": 0.563,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 169
+    },
+    {
+      "epoch": 2.7868852459016393,
+      "grad_norm": 0.1687244176864624,
+      "learning_rate": 3.5317087541144377e-06,
+      "loss": 0.5703,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 170
+    },
+    {
+      "epoch": 2.80327868852459,
+      "grad_norm": 0.1708020716905594,
+      "learning_rate": 3.047685488930874e-06,
+      "loss": 0.558,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 171
+    },
+    {
+      "epoch": 2.819672131147541,
+      "grad_norm": 0.17512136697769165,
+      "learning_rate": 2.5988083057666533e-06,
+      "loss": 0.5588,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 172
+    },
+    {
+      "epoch": 2.836065573770492,
+      "grad_norm": 0.16872957348823547,
+      "learning_rate": 2.1852399266194314e-06,
+      "loss": 0.544,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 173
+    },
+    {
+      "epoch": 2.8524590163934427,
+      "grad_norm": 0.16911788284778595,
+      "learning_rate": 1.8071302737293295e-06,
+      "loss": 0.5596,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 174
+    },
+    {
+      "epoch": 2.8688524590163933,
+      "grad_norm": 0.17218004167079926,
+      "learning_rate": 1.4646164152307018e-06,
+      "loss": 0.5569,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 175
+    },
+    {
+      "epoch": 2.8852459016393444,
+      "grad_norm": 0.17117170989513397,
+      "learning_rate": 1.157822515463758e-06,
+      "loss": 0.5707,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 176
+    },
+    {
+      "epoch": 2.901639344262295,
+      "grad_norm": 0.1660652756690979,
+      "learning_rate": 8.868597899638898e-07,
+      "loss": 0.5546,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 177
+    },
+    {
+      "epoch": 2.918032786885246,
+      "grad_norm": 0.5194974541664124,
+      "learning_rate": 6.518264651449779e-07,
+      "loss": 0.5608,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 178
+    },
+    {
+      "epoch": 2.9344262295081966,
+      "grad_norm": 0.1674128770828247,
+      "learning_rate": 4.5280774269154115e-07,
+      "loss": 0.5513,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 179
+    },
+    {
+      "epoch": 2.9508196721311473,
+      "grad_norm": 0.1717360019683838,
+      "learning_rate": 2.898757686722542e-07,
+      "loss": 0.5644,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 180
+    },
+    {
+      "epoch": 2.9672131147540983,
+      "grad_norm": 0.1789117455482483,
+      "learning_rate": 1.630896073864352e-07,
+      "loss": 0.5587,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 181
+    },
+    {
+      "epoch": 2.9836065573770494,
+      "grad_norm": 0.16960012912750244,
+      "learning_rate": 7.249521995263964e-08,
+      "loss": 0.5582,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 182
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.16996316611766815,
+      "learning_rate": 1.81254476474213e-08,
+      "loss": 0.5541,
+      "memory/device_reserved (GiB)": 33.36,
+      "memory/max_active (GiB)": 32.7,
+      "memory/max_allocated (GiB)": 32.7,
+      "step": 183
+    },
+    {
+      "epoch": 3.0,
+      "eval_loss": 0.6080317497253418,
+      "eval_runtime": 15.4278,
+      "eval_samples_per_second": 53.021,
+      "eval_steps_per_second": 3.371,
+      "memory/device_reserved (GiB)": 73.37,
+      "memory/max_active (GiB)": 48.8,
+      "memory/max_allocated (GiB)": 48.8,
+      "step": 183
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 183,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 61,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.760779115062362e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a13265aba8ad676a02c620fd19cdfeefd887469b26900fa7bc597ad8654742e5
+size 9041