Initial model upload

.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 +1,59 @@
+---
+base_model: Qwen/Qwen2.5-0.5B
+datasets: samhog/psychology-10k
+library_name: transformers
+model_name: Qwen2.5-0.5B-sft-lora-rel-therapy
+tags:
+- generated_from_trainer
+- sft
+- conversational
+licence: license
+---
+
+# Model Card for Qwen2.5-0.5B-sft-lora-rel-therapy
+
+This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B](https://huggingface.co/Qwen/Qwen2.5-0.5B) on the [samhog/psychology-10k](https://huggingface.co/datasets/samhog/psychology-10k) dataset.
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="None", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+ 
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 0.14.0
+- Transformers: 4.48.2
+- Pytorch: 2.5.1+cu124
+- Datasets: 3.2.0
+- Tokenizers: 0.21.0
+
+## Citations
+
+
+
+Cite TRL as:
+    
+```bibtex
+@misc{vonwerra2022trl,
+	title        = {{TRL: Transformer Reinforcement Learning}},
+	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
+	year         = 2020,
+	journal      = {GitHub repository},
+	publisher    = {GitHub},
+	howpublished = {\url{https://github.com/huggingface/trl}}
+}
+```

adapter_config.json

@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-0.5B",
+  "bias": "none",
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.1,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 4,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "v_proj",
+    "k_proj",
+    "q_proj",
+    "o_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9e9fb72d9c9e12493c84f61ca850d41f77ca9f8f1b63b54e5ac5622df4bf4fd
+size 2187608

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
+}

all_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 9.98120873539868,
+    "total_flos": 2.0218089447773184e+16,
+    "train_loss": 1.1380672057469685,
+    "train_runtime": 8182.6752,
+    "train_samples": 7876,
+    "train_samples_per_second": 9.625,
+    "train_steps_per_second": 0.601
+}

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": "<|endoftext|>",
+    "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 sha256:0bd8c258c419667ebe0fb2fdb04f0f374292349d5deba1cba179d29177e80fc0
+size 11421997

tokenizer_config.json

@@ -0,0 +1,208 @@
+{
+  "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 tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are a licensed assistant psychologist, please provide this patient with a helpful response to their concern.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are a licensed assistant psychologist, please provide this patient with a helpful response to their concern.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 9.98120873539868,
+    "total_flos": 2.0218089447773184e+16,
+    "train_loss": 1.1380672057469685,
+    "train_runtime": 8182.6752,
+    "train_samples": 7876,
+    "train_samples_per_second": 9.625,
+    "train_steps_per_second": 0.601
+}

trainer_state.json

@@ -0,0 +1,3566 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 9.98120873539868,
+  "eval_steps": 500,
+  "global_step": 4920,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.02031488065007618,
+      "grad_norm": 7.584718704223633,
+      "learning_rate": 1.999990002577244e-05,
+      "loss": 3.2828,
+      "step": 10
+    },
+    {
+      "epoch": 0.04062976130015236,
+      "grad_norm": 6.282063007354736,
+      "learning_rate": 1.999941036089858e-05,
+      "loss": 3.084,
+      "step": 20
+    },
+    {
+      "epoch": 0.06094464195022854,
+      "grad_norm": 4.432021617889404,
+      "learning_rate": 1.9998512662721364e-05,
+      "loss": 2.9074,
+      "step": 30
+    },
+    {
+      "epoch": 0.08125952260030472,
+      "grad_norm": 4.4949116706848145,
+      "learning_rate": 1.9997206967872025e-05,
+      "loss": 2.7432,
+      "step": 40
+    },
+    {
+      "epoch": 0.1015744032503809,
+      "grad_norm": 3.794994592666626,
+      "learning_rate": 1.99954933296304e-05,
+      "loss": 2.5308,
+      "step": 50
+    },
+    {
+      "epoch": 0.12188928390045708,
+      "grad_norm": 3.7512996196746826,
+      "learning_rate": 1.999337181792276e-05,
+      "loss": 2.3957,
+      "step": 60
+    },
+    {
+      "epoch": 0.14220416455053325,
+      "grad_norm": 3.3922431468963623,
+      "learning_rate": 1.999084251931896e-05,
+      "loss": 2.2018,
+      "step": 70
+    },
+    {
+      "epoch": 0.16251904520060945,
+      "grad_norm": 2.835357666015625,
+      "learning_rate": 1.998790553702888e-05,
+      "loss": 2.0505,
+      "step": 80
+    },
+    {
+      "epoch": 0.1828339258506856,
+      "grad_norm": 2.4077184200286865,
+      "learning_rate": 1.998456099089825e-05,
+      "loss": 1.8802,
+      "step": 90
+    },
+    {
+      "epoch": 0.2031488065007618,
+      "grad_norm": 2.9254958629608154,
+      "learning_rate": 1.998080901740374e-05,
+      "loss": 1.772,
+      "step": 100
+    },
+    {
+      "epoch": 0.22346368715083798,
+      "grad_norm": 2.102215051651001,
+      "learning_rate": 1.9976649769647392e-05,
+      "loss": 1.7375,
+      "step": 110
+    },
+    {
+      "epoch": 0.24377856780091417,
+      "grad_norm": 2.1416189670562744,
+      "learning_rate": 1.997208341735037e-05,
+      "loss": 1.6127,
+      "step": 120
+    },
+    {
+      "epoch": 0.26409344845099036,
+      "grad_norm": 1.9121668338775635,
+      "learning_rate": 1.9967110146846044e-05,
+      "loss": 1.5663,
+      "step": 130
+    },
+    {
+      "epoch": 0.2844083291010665,
+      "grad_norm": 1.7665735483169556,
+      "learning_rate": 1.9961730161072383e-05,
+      "loss": 1.5384,
+      "step": 140
+    },
+    {
+      "epoch": 0.3047232097511427,
+      "grad_norm": 2.695359468460083,
+      "learning_rate": 1.9955943679563655e-05,
+      "loss": 1.4752,
+      "step": 150
+    },
+    {
+      "epoch": 0.3250380904012189,
+      "grad_norm": 1.7461923360824585,
+      "learning_rate": 1.9949750938441517e-05,
+      "loss": 1.4581,
+      "step": 160
+    },
+    {
+      "epoch": 0.3453529710512951,
+      "grad_norm": 1.856467366218567,
+      "learning_rate": 1.994315219040532e-05,
+      "loss": 1.4243,
+      "step": 170
+    },
+    {
+      "epoch": 0.3656678517013712,
+      "grad_norm": 1.6817877292633057,
+      "learning_rate": 1.9936147704721836e-05,
+      "loss": 1.4437,
+      "step": 180
+    },
+    {
+      "epoch": 0.3859827323514474,
+      "grad_norm": 1.6875474452972412,
+      "learning_rate": 1.9928737767214267e-05,
+      "loss": 1.4146,
+      "step": 190
+    },
+    {
+      "epoch": 0.4062976130015236,
+      "grad_norm": 2.079127311706543,
+      "learning_rate": 1.992092268025057e-05,
+      "loss": 1.3857,
+      "step": 200
+    },
+    {
+      "epoch": 0.4266124936515998,
+      "grad_norm": 1.4948805570602417,
+      "learning_rate": 1.9912702762731118e-05,
+      "loss": 1.3748,
+      "step": 210
+    },
+    {
+      "epoch": 0.44692737430167595,
+      "grad_norm": 1.6643149852752686,
+      "learning_rate": 1.9904078350075703e-05,
+      "loss": 1.3904,
+      "step": 220
+    },
+    {
+      "epoch": 0.46724225495175215,
+      "grad_norm": 2.044997215270996,
+      "learning_rate": 1.9895049794209834e-05,
+      "loss": 1.3519,
+      "step": 230
+    },
+    {
+      "epoch": 0.48755713560182834,
+      "grad_norm": 1.9969228506088257,
+      "learning_rate": 1.9885617463550376e-05,
+      "loss": 1.3637,
+      "step": 240
+    },
+    {
+      "epoch": 0.5078720162519045,
+      "grad_norm": 2.5120513439178467,
+      "learning_rate": 1.987678345628239e-05,
+      "loss": 1.3516,
+      "step": 250
+    },
+    {
+      "epoch": 0.5281868969019807,
+      "grad_norm": 2.049866199493408,
+      "learning_rate": 1.9866585027500926e-05,
+      "loss": 1.3149,
+      "step": 260
+    },
+    {
+      "epoch": 0.5485017775520569,
+      "grad_norm": 2.5176775455474854,
+      "learning_rate": 1.9855983985451647e-05,
+      "loss": 1.2744,
+      "step": 270
+    },
+    {
+      "epoch": 0.568816658202133,
+      "grad_norm": 2.2515692710876465,
+      "learning_rate": 1.9844980762717877e-05,
+      "loss": 1.2986,
+      "step": 280
+    },
+    {
+      "epoch": 0.5891315388522093,
+      "grad_norm": 1.9790077209472656,
+      "learning_rate": 1.983357580829422e-05,
+      "loss": 1.2556,
+      "step": 290
+    },
+    {
+      "epoch": 0.6094464195022854,
+      "grad_norm": 2.0562939643859863,
+      "learning_rate": 1.982176958756825e-05,
+      "loss": 1.2565,
+      "step": 300
+    },
+    {
+      "epoch": 0.6297613001523616,
+      "grad_norm": 2.895200252532959,
+      "learning_rate": 1.9809562582301487e-05,
+      "loss": 1.2217,
+      "step": 310
+    },
+    {
+      "epoch": 0.6500761808024378,
+      "grad_norm": 2.379382371902466,
+      "learning_rate": 1.9796955290609775e-05,
+      "loss": 1.2186,
+      "step": 320
+    },
+    {
+      "epoch": 0.6703910614525139,
+      "grad_norm": 1.5008665323257446,
+      "learning_rate": 1.9783948226942936e-05,
+      "loss": 1.2093,
+      "step": 330
+    },
+    {
+      "epoch": 0.6907059421025902,
+      "grad_norm": 2.294449806213379,
+      "learning_rate": 1.9770541922063778e-05,
+      "loss": 1.228,
+      "step": 340
+    },
+    {
+      "epoch": 0.7110208227526663,
+      "grad_norm": 1.5949811935424805,
+      "learning_rate": 1.975673692302645e-05,
+      "loss": 1.211,
+      "step": 350
+    },
+    {
+      "epoch": 0.7313357034027425,
+      "grad_norm": 2.0285587310791016,
+      "learning_rate": 1.97425337931541e-05,
+      "loss": 1.2224,
+      "step": 360
+    },
+    {
+      "epoch": 0.7516505840528187,
+      "grad_norm": 1.7120383977890015,
+      "learning_rate": 1.97279331120159e-05,
+      "loss": 1.2181,
+      "step": 370
+    },
+    {
+      "epoch": 0.7719654647028948,
+      "grad_norm": 2.0460448265075684,
+      "learning_rate": 1.97129354754034e-05,
+      "loss": 1.206,
+      "step": 380
+    },
+    {
+      "epoch": 0.7922803453529711,
+      "grad_norm": 2.479123592376709,
+      "learning_rate": 1.96975414953062e-05,
+      "loss": 1.1987,
+      "step": 390
+    },
+    {
+      "epoch": 0.8125952260030472,
+      "grad_norm": 2.1566030979156494,
+      "learning_rate": 1.9681751799887e-05,
+      "loss": 1.2081,
+      "step": 400
+    },
+    {
+      "epoch": 0.8329101066531234,
+      "grad_norm": 1.7905540466308594,
+      "learning_rate": 1.9665567033455948e-05,
+      "loss": 1.205,
+      "step": 410
+    },
+    {
+      "epoch": 0.8532249873031996,
+      "grad_norm": 1.5104587078094482,
+      "learning_rate": 1.9648987856444356e-05,
+      "loss": 1.1808,
+      "step": 420
+    },
+    {
+      "epoch": 0.8735398679532758,
+      "grad_norm": 1.601035714149475,
+      "learning_rate": 1.9632014945377748e-05,
+      "loss": 1.2039,
+      "step": 430
+    },
+    {
+      "epoch": 0.8938547486033519,
+      "grad_norm": 1.4743677377700806,
+      "learning_rate": 1.9614648992848265e-05,
+      "loss": 1.1923,
+      "step": 440
+    },
+    {
+      "epoch": 0.9141696292534282,
+      "grad_norm": 2.2900314331054688,
+      "learning_rate": 1.959689070748638e-05,
+      "loss": 1.19,
+      "step": 450
+    },
+    {
+      "epoch": 0.9344845099035043,
+      "grad_norm": 2.500079393386841,
+      "learning_rate": 1.9578740813932017e-05,
+      "loss": 1.1826,
+      "step": 460
+    },
+    {
+      "epoch": 0.9547993905535805,
+      "grad_norm": 1.704733967781067,
+      "learning_rate": 1.9560200052804935e-05,
+      "loss": 1.1811,
+      "step": 470
+    },
+    {
+      "epoch": 0.9751142712036567,
+      "grad_norm": 2.3521337509155273,
+      "learning_rate": 1.954126918067455e-05,
+      "loss": 1.1868,
+      "step": 480
+    },
+    {
+      "epoch": 0.9954291518537328,
+      "grad_norm": 2.1910970211029053,
+      "learning_rate": 1.9521948970029045e-05,
+      "loss": 1.2032,
+      "step": 490
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 1.1905763149261475,
+      "eval_runtime": 57.9944,
+      "eval_samples_per_second": 33.969,
+      "eval_steps_per_second": 8.501,
+      "step": 493
+    },
+    {
+      "epoch": 1.0142204164550532,
+      "grad_norm": 1.6047745943069458,
+      "learning_rate": 1.9502240209243834e-05,
+      "loss": 1.1649,
+      "step": 500
+    },
+    {
+      "epoch": 1.0345352971051296,
+      "grad_norm": 2.1489083766937256,
+      "learning_rate": 1.948214370254942e-05,
+      "loss": 1.1811,
+      "step": 510
+    },
+    {
+      "epoch": 1.0548501777552057,
+      "grad_norm": 1.674709677696228,
+      "learning_rate": 1.9461660269998556e-05,
+      "loss": 1.1681,
+      "step": 520
+    },
+    {
+      "epoch": 1.0751650584052819,
+      "grad_norm": 1.839033603668213,
+      "learning_rate": 1.9440790747432797e-05,
+      "loss": 1.1683,
+      "step": 530
+    },
+    {
+      "epoch": 1.095479939055358,
+      "grad_norm": 2.0029451847076416,
+      "learning_rate": 1.941953598644837e-05,
+      "loss": 1.1471,
+      "step": 540
+    },
+    {
+      "epoch": 1.1157948197054341,
+      "grad_norm": 2.1408133506774902,
+      "learning_rate": 1.9397896854361454e-05,
+      "loss": 1.1865,
+      "step": 550
+    },
+    {
+      "epoch": 1.1361097003555103,
+      "grad_norm": 1.6449344158172607,
+      "learning_rate": 1.937587423417277e-05,
+      "loss": 1.1616,
+      "step": 560
+    },
+    {
+      "epoch": 1.1564245810055866,
+      "grad_norm": 2.530738592147827,
+      "learning_rate": 1.935346902453155e-05,
+      "loss": 1.1847,
+      "step": 570
+    },
+    {
+      "epoch": 1.1767394616556628,
+      "grad_norm": 1.7173138856887817,
+      "learning_rate": 1.9330682139698872e-05,
+      "loss": 1.1836,
+      "step": 580
+    },
+    {
+      "epoch": 1.197054342305739,
+      "grad_norm": 1.9412487745285034,
+      "learning_rate": 1.9307514509510358e-05,
+      "loss": 1.1745,
+      "step": 590
+    },
+    {
+      "epoch": 1.217369222955815,
+      "grad_norm": 1.9713300466537476,
+      "learning_rate": 1.928396707933821e-05,
+      "loss": 1.1498,
+      "step": 600
+    },
+    {
+      "epoch": 1.2376841036058912,
+      "grad_norm": 2.253971815109253,
+      "learning_rate": 1.926004081005266e-05,
+      "loss": 1.1554,
+      "step": 610
+    },
+    {
+      "epoch": 1.2579989842559676,
+      "grad_norm": 1.9793624877929688,
+      "learning_rate": 1.9235736677982747e-05,
+      "loss": 1.1445,
+      "step": 620
+    },
+    {
+      "epoch": 1.2783138649060437,
+      "grad_norm": 2.292801856994629,
+      "learning_rate": 1.9211055674876475e-05,
+      "loss": 1.1373,
+      "step": 630
+    },
+    {
+      "epoch": 1.2986287455561198,
+      "grad_norm": 1.8089720010757446,
+      "learning_rate": 1.9185998807860357e-05,
+      "loss": 1.1324,
+      "step": 640
+    },
+    {
+      "epoch": 1.318943626206196,
+      "grad_norm": 2.003289222717285,
+      "learning_rate": 1.9160567099398294e-05,
+      "loss": 1.173,
+      "step": 650
+    },
+    {
+      "epoch": 1.3392585068562721,
+      "grad_norm": 2.1809918880462646,
+      "learning_rate": 1.9134761587249886e-05,
+      "loss": 1.1408,
+      "step": 660
+    },
+    {
+      "epoch": 1.3595733875063485,
+      "grad_norm": 2.0070571899414062,
+      "learning_rate": 1.9108583324428054e-05,
+      "loss": 1.1811,
+      "step": 670
+    },
+    {
+      "epoch": 1.3798882681564246,
+      "grad_norm": 1.6615877151489258,
+      "learning_rate": 1.9082033379156096e-05,
+      "loss": 1.164,
+      "step": 680
+    },
+    {
+      "epoch": 1.4002031488065008,
+      "grad_norm": 2.217325210571289,
+      "learning_rate": 1.9055112834824075e-05,
+      "loss": 1.1534,
+      "step": 690
+    },
+    {
+      "epoch": 1.420518029456577,
+      "grad_norm": 1.8769993782043457,
+      "learning_rate": 1.9027822789944625e-05,
+      "loss": 1.1476,
+      "step": 700
+    },
+    {
+      "epoch": 1.440832910106653,
+      "grad_norm": 1.8755251169204712,
+      "learning_rate": 1.900016435810813e-05,
+      "loss": 1.1773,
+      "step": 710
+    },
+    {
+      "epoch": 1.4611477907567294,
+      "grad_norm": 2.059292793273926,
+      "learning_rate": 1.8972138667937268e-05,
+      "loss": 1.1512,
+      "step": 720
+    },
+    {
+      "epoch": 1.4814626714068055,
+      "grad_norm": 2.2438113689422607,
+      "learning_rate": 1.894374686304096e-05,
+      "loss": 1.1313,
+      "step": 730
+    },
+    {
+      "epoch": 1.5017775520568817,
+      "grad_norm": 1.6689549684524536,
+      "learning_rate": 1.8914990101967717e-05,
+      "loss": 1.1506,
+      "step": 740
+    },
+    {
+      "epoch": 1.5220924327069578,
+      "grad_norm": 1.9220091104507446,
+      "learning_rate": 1.8885869558158348e-05,
+      "loss": 1.1389,
+      "step": 750
+    },
+    {
+      "epoch": 1.542407313357034,
+      "grad_norm": 1.970219373703003,
+      "learning_rate": 1.885638641989809e-05,
+      "loss": 1.1535,
+      "step": 760
+    },
+    {
+      "epoch": 1.5627221940071103,
+      "grad_norm": 2.3392066955566406,
+      "learning_rate": 1.882654189026811e-05,
+      "loss": 1.1467,
+      "step": 770
+    },
+    {
+      "epoch": 1.5830370746571862,
+      "grad_norm": 1.9110764265060425,
+      "learning_rate": 1.8796337187096415e-05,
+      "loss": 1.1457,
+      "step": 780
+    },
+    {
+      "epoch": 1.6033519553072626,
+      "grad_norm": 3.4612278938293457,
+      "learning_rate": 1.8765773542908165e-05,
+      "loss": 1.1443,
+      "step": 790
+    },
+    {
+      "epoch": 1.6236668359573387,
+      "grad_norm": 1.8444197177886963,
+      "learning_rate": 1.8734852204875363e-05,
+      "loss": 1.1375,
+      "step": 800
+    },
+    {
+      "epoch": 1.6439817166074149,
+      "grad_norm": 1.7353184223175049,
+      "learning_rate": 1.8703574434765986e-05,
+      "loss": 1.1802,
+      "step": 810
+    },
+    {
+      "epoch": 1.6642965972574912,
+      "grad_norm": 2.283703327178955,
+      "learning_rate": 1.8671941508892467e-05,
+      "loss": 1.1457,
+      "step": 820
+    },
+    {
+      "epoch": 1.6846114779075672,
+      "grad_norm": 2.3982231616973877,
+      "learning_rate": 1.8639954718059647e-05,
+      "loss": 1.134,
+      "step": 830
+    },
+    {
+      "epoch": 1.7049263585576435,
+      "grad_norm": 2.6923394203186035,
+      "learning_rate": 1.8607615367512077e-05,
+      "loss": 1.1387,
+      "step": 840
+    },
+    {
+      "epoch": 1.7252412392077197,
+      "grad_norm": 1.870726466178894,
+      "learning_rate": 1.8574924776880777e-05,
+      "loss": 1.1479,
+      "step": 850
+    },
+    {
+      "epoch": 1.7455561198577958,
+      "grad_norm": 2.014890432357788,
+      "learning_rate": 1.8541884280129367e-05,
+      "loss": 1.1724,
+      "step": 860
+    },
+    {
+      "epoch": 1.7658710005078722,
+      "grad_norm": 1.8836543560028076,
+      "learning_rate": 1.8508495225499643e-05,
+      "loss": 1.1483,
+      "step": 870
+    },
+    {
+      "epoch": 1.786185881157948,
+      "grad_norm": 1.8277486562728882,
+      "learning_rate": 1.8474758975456576e-05,
+      "loss": 1.121,
+      "step": 880
+    },
+    {
+      "epoch": 1.8065007618080244,
+      "grad_norm": 1.8098642826080322,
+      "learning_rate": 1.8440676906632688e-05,
+      "loss": 1.1404,
+      "step": 890
+    },
+    {
+      "epoch": 1.8268156424581006,
+      "grad_norm": 2.0172536373138428,
+      "learning_rate": 1.8406250409771894e-05,
+      "loss": 1.1424,
+      "step": 900
+    },
+    {
+      "epoch": 1.8471305231081767,
+      "grad_norm": 2.2588918209075928,
+      "learning_rate": 1.8371480889672756e-05,
+      "loss": 1.1394,
+      "step": 910
+    },
+    {
+      "epoch": 1.867445403758253,
+      "grad_norm": 2.9289000034332275,
+      "learning_rate": 1.8336369765131137e-05,
+      "loss": 1.1273,
+      "step": 920
+    },
+    {
+      "epoch": 1.887760284408329,
+      "grad_norm": 1.9783216714859009,
+      "learning_rate": 1.8300918468882337e-05,
+      "loss": 1.1316,
+      "step": 930
+    },
+    {
+      "epoch": 1.9080751650584054,
+      "grad_norm": 2.6187262535095215,
+      "learning_rate": 1.8265128447542606e-05,
+      "loss": 1.1347,
+      "step": 940
+    },
+    {
+      "epoch": 1.9283900457084815,
+      "grad_norm": 1.7901383638381958,
+      "learning_rate": 1.822900116155012e-05,
+      "loss": 1.1419,
+      "step": 950
+    },
+    {
+      "epoch": 1.9487049263585576,
+      "grad_norm": 2.4337475299835205,
+      "learning_rate": 1.819253808510538e-05,
+      "loss": 1.1253,
+      "step": 960
+    },
+    {
+      "epoch": 1.9690198070086338,
+      "grad_norm": 1.88914155960083,
+      "learning_rate": 1.8155740706111086e-05,
+      "loss": 1.1278,
+      "step": 970
+    },
+    {
+      "epoch": 1.98933468765871,
+      "grad_norm": 2.087872266769409,
+      "learning_rate": 1.811861052611137e-05,
+      "loss": 1.1239,
+      "step": 980
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 1.1426180601119995,
+      "eval_runtime": 56.1438,
+      "eval_samples_per_second": 35.088,
+      "eval_steps_per_second": 8.781,
+      "step": 986
+    },
+    {
+      "epoch": 2.0081259522600305,
+      "grad_norm": 2.3936855792999268,
+      "learning_rate": 1.8081149060230584e-05,
+      "loss": 1.1307,
+      "step": 990
+    },
+    {
+      "epoch": 2.0284408329101065,
+      "grad_norm": 2.97658371925354,
+      "learning_rate": 1.804335783711142e-05,
+      "loss": 1.1115,
+      "step": 1000
+    },
+    {
+      "epoch": 2.048755713560183,
+      "grad_norm": 1.7273203134536743,
+      "learning_rate": 1.8005238398852567e-05,
+      "loss": 1.1145,
+      "step": 1010
+    },
+    {
+      "epoch": 2.069070594210259,
+      "grad_norm": 2.1559698581695557,
+      "learning_rate": 1.7966792300945782e-05,
+      "loss": 1.1333,
+      "step": 1020
+    },
+    {
+      "epoch": 2.089385474860335,
+      "grad_norm": 1.8381649255752563,
+      "learning_rate": 1.7928021112212397e-05,
+      "loss": 1.0982,
+      "step": 1030
+    },
+    {
+      "epoch": 2.1097003555104115,
+      "grad_norm": 2.330662727355957,
+      "learning_rate": 1.788892641473932e-05,
+      "loss": 1.1246,
+      "step": 1040
+    },
+    {
+      "epoch": 2.1300152361604874,
+      "grad_norm": 2.1321730613708496,
+      "learning_rate": 1.784950980381447e-05,
+      "loss": 1.1182,
+      "step": 1050
+    },
+    {
+      "epoch": 2.1503301168105637,
+      "grad_norm": 2.3504936695098877,
+      "learning_rate": 1.7809772887861684e-05,
+      "loss": 1.1449,
+      "step": 1060
+    },
+    {
+      "epoch": 2.17064499746064,
+      "grad_norm": 1.9942519664764404,
+      "learning_rate": 1.7769717288375078e-05,
+      "loss": 1.1355,
+      "step": 1070
+    },
+    {
+      "epoch": 2.190959878110716,
+      "grad_norm": 2.6634910106658936,
+      "learning_rate": 1.7729344639852886e-05,
+      "loss": 1.1317,
+      "step": 1080
+    },
+    {
+      "epoch": 2.2112747587607924,
+      "grad_norm": 2.0692458152770996,
+      "learning_rate": 1.768865658973076e-05,
+      "loss": 1.1349,
+      "step": 1090
+    },
+    {
+      "epoch": 2.2315896394108683,
+      "grad_norm": 2.0810651779174805,
+      "learning_rate": 1.7647654798314554e-05,
+      "loss": 1.1084,
+      "step": 1100
+    },
+    {
+      "epoch": 2.2519045200609447,
+      "grad_norm": 1.9821966886520386,
+      "learning_rate": 1.760634093871256e-05,
+      "loss": 1.1274,
+      "step": 1110
+    },
+    {
+      "epoch": 2.2722194007110206,
+      "grad_norm": 2.893826484680176,
+      "learning_rate": 1.756471669676723e-05,
+      "loss": 1.1262,
+      "step": 1120
+    },
+    {
+      "epoch": 2.292534281361097,
+      "grad_norm": 1.9358024597167969,
+      "learning_rate": 1.7522783770986425e-05,
+      "loss": 1.1138,
+      "step": 1130
+    },
+    {
+      "epoch": 2.3128491620111733,
+      "grad_norm": 1.746727466583252,
+      "learning_rate": 1.748054387247405e-05,
+      "loss": 1.1313,
+      "step": 1140
+    },
+    {
+      "epoch": 2.333164042661249,
+      "grad_norm": 1.878308892250061,
+      "learning_rate": 1.743799872486028e-05,
+      "loss": 1.1313,
+      "step": 1150
+    },
+    {
+      "epoch": 2.3534789233113256,
+      "grad_norm": 2.322291135787964,
+      "learning_rate": 1.739515006423118e-05,
+      "loss": 1.1173,
+      "step": 1160
+    },
+    {
+      "epoch": 2.3737938039614015,
+      "grad_norm": 1.9874950647354126,
+      "learning_rate": 1.7351999639057915e-05,
+      "loss": 1.1023,
+      "step": 1170
+    },
+    {
+      "epoch": 2.394108684611478,
+      "grad_norm": 2.070664882659912,
+      "learning_rate": 1.7308549210125346e-05,
+      "loss": 1.1208,
+      "step": 1180
+    },
+    {
+      "epoch": 2.414423565261554,
+      "grad_norm": 1.8656115531921387,
+      "learning_rate": 1.7264800550460235e-05,
+      "loss": 1.1324,
+      "step": 1190
+    },
+    {
+      "epoch": 2.43473844591163,
+      "grad_norm": 1.7181984186172485,
+      "learning_rate": 1.7220755445258847e-05,
+      "loss": 1.1299,
+      "step": 1200
+    },
+    {
+      "epoch": 2.4550533265617065,
+      "grad_norm": 2.0644876956939697,
+      "learning_rate": 1.717641569181414e-05,
+      "loss": 1.1057,
+      "step": 1210
+    },
+    {
+      "epoch": 2.4753682072117824,
+      "grad_norm": 2.54202938079834,
+      "learning_rate": 1.7131783099442408e-05,
+      "loss": 1.1371,
+      "step": 1220
+    },
+    {
+      "epoch": 2.4956830878618588,
+      "grad_norm": 1.8502358198165894,
+      "learning_rate": 1.708685948940944e-05,
+      "loss": 1.0938,
+      "step": 1230
+    },
+    {
+      "epoch": 2.515997968511935,
+      "grad_norm": 2.2226638793945312,
+      "learning_rate": 1.7041646694856237e-05,
+      "loss": 1.1306,
+      "step": 1240
+    },
+    {
+      "epoch": 2.536312849162011,
+      "grad_norm": 2.0658068656921387,
+      "learning_rate": 1.6996146560724163e-05,
+      "loss": 1.1465,
+      "step": 1250
+    },
+    {
+      "epoch": 2.5566277298120874,
+      "grad_norm": 2.187943458557129,
+      "learning_rate": 1.6950360943679708e-05,
+      "loss": 1.1071,
+      "step": 1260
+    },
+    {
+      "epoch": 2.5769426104621633,
+      "grad_norm": 2.104330062866211,
+      "learning_rate": 1.690429171203869e-05,
+      "loss": 1.1247,
+      "step": 1270
+    },
+    {
+      "epoch": 2.5972574911122397,
+      "grad_norm": 2.1676745414733887,
+      "learning_rate": 1.685794074569003e-05,
+      "loss": 1.112,
+      "step": 1280
+    },
+    {
+      "epoch": 2.617572371762316,
+      "grad_norm": 1.8994890451431274,
+      "learning_rate": 1.6811309936019034e-05,
+      "loss": 1.1129,
+      "step": 1290
+    },
+    {
+      "epoch": 2.637887252412392,
+      "grad_norm": 1.9885021448135376,
+      "learning_rate": 1.6764401185830248e-05,
+      "loss": 1.1074,
+      "step": 1300
+    },
+    {
+      "epoch": 2.6582021330624683,
+      "grad_norm": 1.8062092065811157,
+      "learning_rate": 1.6717216409269756e-05,
+      "loss": 1.1083,
+      "step": 1310
+    },
+    {
+      "epoch": 2.6785170137125442,
+      "grad_norm": 2.118276834487915,
+      "learning_rate": 1.666975753174711e-05,
+      "loss": 1.0967,
+      "step": 1320
+    },
+    {
+      "epoch": 2.6988318943626206,
+      "grad_norm": 2.219618082046509,
+      "learning_rate": 1.6622026489856767e-05,
+      "loss": 1.1294,
+      "step": 1330
+    },
+    {
+      "epoch": 2.719146775012697,
+      "grad_norm": 2.3651201725006104,
+      "learning_rate": 1.657402523129903e-05,
+      "loss": 1.1303,
+      "step": 1340
+    },
+    {
+      "epoch": 2.739461655662773,
+      "grad_norm": 2.132640838623047,
+      "learning_rate": 1.6525755714800605e-05,
+      "loss": 1.0982,
+      "step": 1350
+    },
+    {
+      "epoch": 2.7597765363128492,
+      "grad_norm": 2.0165741443634033,
+      "learning_rate": 1.6477219910034662e-05,
+      "loss": 1.1215,
+      "step": 1360
+    },
+    {
+      "epoch": 2.780091416962925,
+      "grad_norm": 2.050801992416382,
+      "learning_rate": 1.6428419797540455e-05,
+      "loss": 1.0938,
+      "step": 1370
+    },
+    {
+      "epoch": 2.8004062976130015,
+      "grad_norm": 2.0775163173675537,
+      "learning_rate": 1.637935736864251e-05,
+      "loss": 1.0976,
+      "step": 1380
+    },
+    {
+      "epoch": 2.820721178263078,
+      "grad_norm": 1.9527651071548462,
+      "learning_rate": 1.6330034625369366e-05,
+      "loss": 1.108,
+      "step": 1390
+    },
+    {
+      "epoch": 2.841036058913154,
+      "grad_norm": 2.5815951824188232,
+      "learning_rate": 1.6280453580371887e-05,
+      "loss": 1.1187,
+      "step": 1400
+    },
+    {
+      "epoch": 2.86135093956323,
+      "grad_norm": 2.0379745960235596,
+      "learning_rate": 1.6230616256841122e-05,
+      "loss": 1.1263,
+      "step": 1410
+    },
+    {
+      "epoch": 2.881665820213306,
+      "grad_norm": 2.2415878772735596,
+      "learning_rate": 1.618052468842576e-05,
+      "loss": 1.13,
+      "step": 1420
+    },
+    {
+      "epoch": 2.9019807008633824,
+      "grad_norm": 1.951768159866333,
+      "learning_rate": 1.6130180919149145e-05,
+      "loss": 1.1179,
+      "step": 1430
+    },
+    {
+      "epoch": 2.922295581513459,
+      "grad_norm": 2.3024680614471436,
+      "learning_rate": 1.607958700332584e-05,
+      "loss": 1.0993,
+      "step": 1440
+    },
+    {
+      "epoch": 2.9426104621635347,
+      "grad_norm": 1.86294686794281,
+      "learning_rate": 1.602874500547785e-05,
+      "loss": 1.1071,
+      "step": 1450
+    },
+    {
+      "epoch": 2.962925342813611,
+      "grad_norm": 2.225184202194214,
+      "learning_rate": 1.5977657000250332e-05,
+      "loss": 1.1189,
+      "step": 1460
+    },
+    {
+      "epoch": 2.983240223463687,
+      "grad_norm": 2.0135252475738525,
+      "learning_rate": 1.5926325072326963e-05,
+      "loss": 1.1019,
+      "step": 1470
+    },
+    {
+      "epoch": 3.0,
+      "eval_loss": 1.1209979057312012,
+      "eval_runtime": 55.7989,
+      "eval_samples_per_second": 35.305,
+      "eval_steps_per_second": 8.835,
+      "step": 1479
+    },
+    {
+      "epoch": 3.0020314880650076,
+      "grad_norm": 2.6400208473205566,
+      "learning_rate": 1.587475131634486e-05,
+      "loss": 1.1055,
+      "step": 1480
+    },
+    {
+      "epoch": 3.022346368715084,
+      "grad_norm": 1.8572973012924194,
+      "learning_rate": 1.582293783680912e-05,
+      "loss": 1.1214,
+      "step": 1490
+    },
+    {
+      "epoch": 3.04266124936516,
+      "grad_norm": 2.232374668121338,
+      "learning_rate": 1.5770886748006925e-05,
+      "loss": 1.1016,
+      "step": 1500
+    },
+    {
+      "epoch": 3.0629761300152363,
+      "grad_norm": 1.8693956136703491,
+      "learning_rate": 1.5718600173921287e-05,
+      "loss": 1.1011,
+      "step": 1510
+    },
+    {
+      "epoch": 3.083291010665312,
+      "grad_norm": 1.9302908182144165,
+      "learning_rate": 1.5666080248144364e-05,
+      "loss": 1.0913,
+      "step": 1520
+    },
+    {
+      "epoch": 3.1036058913153886,
+      "grad_norm": 2.3789591789245605,
+      "learning_rate": 1.5613329113790407e-05,
+      "loss": 1.1153,
+      "step": 1530
+    },
+    {
+      "epoch": 3.123920771965465,
+      "grad_norm": 2.4649009704589844,
+      "learning_rate": 1.5560348923408297e-05,
+      "loss": 1.1046,
+      "step": 1540
+    },
+    {
+      "epoch": 3.144235652615541,
+      "grad_norm": 1.7256394624710083,
+      "learning_rate": 1.5507141838893714e-05,
+      "loss": 1.1111,
+      "step": 1550
+    },
+    {
+      "epoch": 3.164550533265617,
+      "grad_norm": 2.8762755393981934,
+      "learning_rate": 1.545371003140093e-05,
+      "loss": 1.1196,
+      "step": 1560
+    },
+    {
+      "epoch": 3.184865413915693,
+      "grad_norm": 1.8774479627609253,
+      "learning_rate": 1.540005568125419e-05,
+      "loss": 1.0758,
+      "step": 1570
+    },
+    {
+      "epoch": 3.2051802945657695,
+      "grad_norm": 2.3003156185150146,
+      "learning_rate": 1.5346180977858766e-05,
+      "loss": 1.1024,
+      "step": 1580
+    },
+    {
+      "epoch": 3.225495175215846,
+      "grad_norm": 1.9327248334884644,
+      "learning_rate": 1.529208811961161e-05,
+      "loss": 1.1217,
+      "step": 1590
+    },
+    {
+      "epoch": 3.2458100558659218,
+      "grad_norm": 2.096296548843384,
+      "learning_rate": 1.5237779313811622e-05,
+      "loss": 1.0966,
+      "step": 1600
+    },
+    {
+      "epoch": 3.266124936515998,
+      "grad_norm": 1.896620750427246,
+      "learning_rate": 1.5183256776569625e-05,
+      "loss": 1.1019,
+      "step": 1610
+    },
+    {
+      "epoch": 3.286439817166074,
+      "grad_norm": 2.3992044925689697,
+      "learning_rate": 1.5128522732717907e-05,
+      "loss": 1.1048,
+      "step": 1620
+    },
+    {
+      "epoch": 3.3067546978161504,
+      "grad_norm": 2.170178174972534,
+      "learning_rate": 1.5073579415719422e-05,
+      "loss": 1.0886,
+      "step": 1630
+    },
+    {
+      "epoch": 3.3270695784662268,
+      "grad_norm": 2.299849510192871,
+      "learning_rate": 1.5018429067576691e-05,
+      "loss": 1.093,
+      "step": 1640
+    },
+    {
+      "epoch": 3.3473844591163027,
+      "grad_norm": 2.0816946029663086,
+      "learning_rate": 1.4963073938740276e-05,
+      "loss": 1.1009,
+      "step": 1650
+    },
+    {
+      "epoch": 3.367699339766379,
+      "grad_norm": 2.180161237716675,
+      "learning_rate": 1.4907516288016972e-05,
+      "loss": 1.1239,
+      "step": 1660
+    },
+    {
+      "epoch": 3.388014220416455,
+      "grad_norm": 2.5566606521606445,
+      "learning_rate": 1.4851758382477632e-05,
+      "loss": 1.1168,
+      "step": 1670
+    },
+    {
+      "epoch": 3.4083291010665313,
+      "grad_norm": 2.167325019836426,
+      "learning_rate": 1.4795802497364655e-05,
+      "loss": 1.0941,
+      "step": 1680
+    },
+    {
+      "epoch": 3.4286439817166072,
+      "grad_norm": 2.132948875427246,
+      "learning_rate": 1.4739650915999136e-05,
+      "loss": 1.0935,
+      "step": 1690
+    },
+    {
+      "epoch": 3.4489588623666836,
+      "grad_norm": 2.1108901500701904,
+      "learning_rate": 1.46833059296877e-05,
+      "loss": 1.0963,
+      "step": 1700
+    },
+    {
+      "epoch": 3.46927374301676,
+      "grad_norm": 2.1430904865264893,
+      "learning_rate": 1.4626769837629012e-05,
+      "loss": 1.1058,
+      "step": 1710
+    },
+    {
+      "epoch": 3.489588623666836,
+      "grad_norm": 2.2873406410217285,
+      "learning_rate": 1.4570044946819937e-05,
+      "loss": 1.0925,
+      "step": 1720
+    },
+    {
+      "epoch": 3.5099035043169122,
+      "grad_norm": 2.4914746284484863,
+      "learning_rate": 1.4513133571961415e-05,
+      "loss": 1.1094,
+      "step": 1730
+    },
+    {
+      "epoch": 3.5302183849669886,
+      "grad_norm": 2.074937582015991,
+      "learning_rate": 1.4456038035364014e-05,
+      "loss": 1.1123,
+      "step": 1740
+    },
+    {
+      "epoch": 3.5505332656170645,
+      "grad_norm": 2.091817855834961,
+      "learning_rate": 1.4398760666853143e-05,
+      "loss": 1.0965,
+      "step": 1750
+    },
+    {
+      "epoch": 3.5708481462671404,
+      "grad_norm": 1.8256065845489502,
+      "learning_rate": 1.434130380367401e-05,
+      "loss": 1.1179,
+      "step": 1760
+    },
+    {
+      "epoch": 3.591163026917217,
+      "grad_norm": 2.202068567276001,
+      "learning_rate": 1.428366979039622e-05,
+      "loss": 1.1102,
+      "step": 1770
+    },
+    {
+      "epoch": 3.611477907567293,
+      "grad_norm": 1.8671810626983643,
+      "learning_rate": 1.4225860978818134e-05,
+      "loss": 1.089,
+      "step": 1780
+    },
+    {
+      "epoch": 3.631792788217369,
+      "grad_norm": 2.394472599029541,
+      "learning_rate": 1.4167879727870873e-05,
+      "loss": 1.0968,
+      "step": 1790
+    },
+    {
+      "epoch": 3.6521076688674454,
+      "grad_norm": 1.918837308883667,
+      "learning_rate": 1.4109728403522088e-05,
+      "loss": 1.0971,
+      "step": 1800
+    },
+    {
+      "epoch": 3.6724225495175213,
+      "grad_norm": 2.170398235321045,
+      "learning_rate": 1.405140937867938e-05,
+      "loss": 1.0953,
+      "step": 1810
+    },
+    {
+      "epoch": 3.6927374301675977,
+      "grad_norm": 1.9695782661437988,
+      "learning_rate": 1.3992925033093514e-05,
+      "loss": 1.0954,
+      "step": 1820
+    },
+    {
+      "epoch": 3.713052310817674,
+      "grad_norm": 2.0776660442352295,
+      "learning_rate": 1.3934277753261274e-05,
+      "loss": 1.0989,
+      "step": 1830
+    },
+    {
+      "epoch": 3.73336719146775,
+      "grad_norm": 2.3341407775878906,
+      "learning_rate": 1.3875469932328104e-05,
+      "loss": 1.1108,
+      "step": 1840
+    },
+    {
+      "epoch": 3.7536820721178263,
+      "grad_norm": 2.07391095161438,
+      "learning_rate": 1.3816503969990444e-05,
+      "loss": 1.1194,
+      "step": 1850
+    },
+    {
+      "epoch": 3.7739969527679023,
+      "grad_norm": 1.9829392433166504,
+      "learning_rate": 1.3757382272397805e-05,
+      "loss": 1.0695,
+      "step": 1860
+    },
+    {
+      "epoch": 3.7943118334179786,
+      "grad_norm": 2.744811534881592,
+      "learning_rate": 1.3698107252054604e-05,
+      "loss": 1.0825,
+      "step": 1870
+    },
+    {
+      "epoch": 3.814626714068055,
+      "grad_norm": 2.293558120727539,
+      "learning_rate": 1.3638681327721686e-05,
+      "loss": 1.0942,
+      "step": 1880
+    },
+    {
+      "epoch": 3.834941594718131,
+      "grad_norm": 2.2765190601348877,
+      "learning_rate": 1.357910692431766e-05,
+      "loss": 1.0907,
+      "step": 1890
+    },
+    {
+      "epoch": 3.8552564753682073,
+      "grad_norm": 2.5928955078125,
+      "learning_rate": 1.3519386472819927e-05,
+      "loss": 1.0911,
+      "step": 1900
+    },
+    {
+      "epoch": 3.875571356018283,
+      "grad_norm": 2.018359661102295,
+      "learning_rate": 1.3459522410165488e-05,
+      "loss": 1.0965,
+      "step": 1910
+    },
+    {
+      "epoch": 3.8958862366683595,
+      "grad_norm": 3.01343035697937,
+      "learning_rate": 1.3399517179151504e-05,
+      "loss": 1.0946,
+      "step": 1920
+    },
+    {
+      "epoch": 3.916201117318436,
+      "grad_norm": 2.592369794845581,
+      "learning_rate": 1.3339373228335614e-05,
+      "loss": 1.0901,
+      "step": 1930
+    },
+    {
+      "epoch": 3.936515997968512,
+      "grad_norm": 2.4442763328552246,
+      "learning_rate": 1.3279093011936021e-05,
+      "loss": 1.0906,
+      "step": 1940
+    },
+    {
+      "epoch": 3.956830878618588,
+      "grad_norm": 2.2360129356384277,
+      "learning_rate": 1.3218678989731352e-05,
+      "loss": 1.0988,
+      "step": 1950
+    },
+    {
+      "epoch": 3.977145759268664,
+      "grad_norm": 2.1769418716430664,
+      "learning_rate": 1.3158133626960261e-05,
+      "loss": 1.0978,
+      "step": 1960
+    },
+    {
+      "epoch": 3.9974606399187405,
+      "grad_norm": 2.4375362396240234,
+      "learning_rate": 1.3097459394220867e-05,
+      "loss": 1.0874,
+      "step": 1970
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 1.1085847616195679,
+      "eval_runtime": 55.3659,
+      "eval_samples_per_second": 35.581,
+      "eval_steps_per_second": 8.904,
+      "step": 1972
+    },
+    {
+      "epoch": 4.016251904520061,
+      "grad_norm": 2.620096445083618,
+      "learning_rate": 1.3036658767369919e-05,
+      "loss": 1.1006,
+      "step": 1980
+    },
+    {
+      "epoch": 4.0365667851701375,
+      "grad_norm": 2.652712345123291,
+      "learning_rate": 1.2975734227421768e-05,
+      "loss": 1.0797,
+      "step": 1990
+    },
+    {
+      "epoch": 4.056881665820213,
+      "grad_norm": 2.4038774967193604,
+      "learning_rate": 1.291468826044713e-05,
+      "loss": 1.1085,
+      "step": 2000
+    },
+    {
+      "epoch": 4.077196546470289,
+      "grad_norm": 2.6778948307037354,
+      "learning_rate": 1.2853523357471644e-05,
+      "loss": 1.0818,
+      "step": 2010
+    },
+    {
+      "epoch": 4.097511427120366,
+      "grad_norm": 2.2972545623779297,
+      "learning_rate": 1.2792242014374216e-05,
+      "loss": 1.0921,
+      "step": 2020
+    },
+    {
+      "epoch": 4.117826307770442,
+      "grad_norm": 2.1259145736694336,
+      "learning_rate": 1.273084673178518e-05,
+      "loss": 1.0911,
+      "step": 2030
+    },
+    {
+      "epoch": 4.138141188420518,
+      "grad_norm": 2.6703901290893555,
+      "learning_rate": 1.2669340014984252e-05,
+      "loss": 1.0939,
+      "step": 2040
+    },
+    {
+      "epoch": 4.158456069070594,
+      "grad_norm": 2.7287003993988037,
+      "learning_rate": 1.260772437379831e-05,
+      "loss": 1.0896,
+      "step": 2050
+    },
+    {
+      "epoch": 4.17877094972067,
+      "grad_norm": 2.5945608615875244,
+      "learning_rate": 1.254600232249896e-05,
+      "loss": 1.0973,
+      "step": 2060
+    },
+    {
+      "epoch": 4.199085830370747,
+      "grad_norm": 2.0844831466674805,
+      "learning_rate": 1.248417637969996e-05,
+      "loss": 1.0883,
+      "step": 2070
+    },
+    {
+      "epoch": 4.219400711020823,
+      "grad_norm": 2.237929105758667,
+      "learning_rate": 1.2422249068254433e-05,
+      "loss": 1.0963,
+      "step": 2080
+    },
+    {
+      "epoch": 4.239715591670899,
+      "grad_norm": 2.147749662399292,
+      "learning_rate": 1.2360222915151929e-05,
+      "loss": 1.0939,
+      "step": 2090
+    },
+    {
+      "epoch": 4.260030472320975,
+      "grad_norm": 2.0512821674346924,
+      "learning_rate": 1.22981004514153e-05,
+      "loss": 1.0984,
+      "step": 2100
+    },
+    {
+      "epoch": 4.280345352971051,
+      "grad_norm": 2.3941662311553955,
+      "learning_rate": 1.223588421199743e-05,
+      "loss": 1.1019,
+      "step": 2110
+    },
+    {
+      "epoch": 4.3006602336211275,
+      "grad_norm": 2.5357017517089844,
+      "learning_rate": 1.2173576735677775e-05,
+      "loss": 1.1074,
+      "step": 2120
+    },
+    {
+      "epoch": 4.320975114271204,
+      "grad_norm": 2.3033499717712402,
+      "learning_rate": 1.2111180564958796e-05,
+      "loss": 1.0841,
+      "step": 2130
+    },
+    {
+      "epoch": 4.34128999492128,
+      "grad_norm": 2.249016046524048,
+      "learning_rate": 1.2048698245962184e-05,
+      "loss": 1.0746,
+      "step": 2140
+    },
+    {
+      "epoch": 4.361604875571356,
+      "grad_norm": 2.2466604709625244,
+      "learning_rate": 1.1986132328324986e-05,
+      "loss": 1.105,
+      "step": 2150
+    },
+    {
+      "epoch": 4.381919756221432,
+      "grad_norm": 2.7963321208953857,
+      "learning_rate": 1.192348536509554e-05,
+      "loss": 1.0876,
+      "step": 2160
+    },
+    {
+      "epoch": 4.402234636871508,
+      "grad_norm": 2.159446954727173,
+      "learning_rate": 1.1860759912629311e-05,
+      "loss": 1.1072,
+      "step": 2170
+    },
+    {
+      "epoch": 4.422549517521585,
+      "grad_norm": 2.521923542022705,
+      "learning_rate": 1.179795853048458e-05,
+      "loss": 1.1047,
+      "step": 2180
+    },
+    {
+      "epoch": 4.442864398171661,
+      "grad_norm": 2.2701151371002197,
+      "learning_rate": 1.1735083781317992e-05,
+      "loss": 1.0879,
+      "step": 2190
+    },
+    {
+      "epoch": 4.463179278821737,
+      "grad_norm": 2.706636667251587,
+      "learning_rate": 1.167213823077999e-05,
+      "loss": 1.0793,
+      "step": 2200
+    },
+    {
+      "epoch": 4.483494159471813,
+      "grad_norm": 2.425516128540039,
+      "learning_rate": 1.1609124447410114e-05,
+      "loss": 1.093,
+      "step": 2210
+    },
+    {
+      "epoch": 4.503809040121889,
+      "grad_norm": 2.4822216033935547,
+      "learning_rate": 1.1546045002532202e-05,
+      "loss": 1.0961,
+      "step": 2220
+    },
+    {
+      "epoch": 4.524123920771966,
+      "grad_norm": 2.3154852390289307,
+      "learning_rate": 1.148290247014945e-05,
+      "loss": 1.0861,
+      "step": 2230
+    },
+    {
+      "epoch": 4.544438801422041,
+      "grad_norm": 2.016235589981079,
+      "learning_rate": 1.1419699426839393e-05,
+      "loss": 1.0913,
+      "step": 2240
+    },
+    {
+      "epoch": 4.5647536820721175,
+      "grad_norm": 2.701549530029297,
+      "learning_rate": 1.1356438451648754e-05,
+      "loss": 1.0918,
+      "step": 2250
+    },
+    {
+      "epoch": 4.585068562722194,
+      "grad_norm": 2.02785587310791,
+      "learning_rate": 1.1293122125988211e-05,
+      "loss": 1.1,
+      "step": 2260
+    },
+    {
+      "epoch": 4.60538344337227,
+      "grad_norm": 2.196712017059326,
+      "learning_rate": 1.122975303352706e-05,
+      "loss": 1.0657,
+      "step": 2270
+    },
+    {
+      "epoch": 4.625698324022347,
+      "grad_norm": 2.8437767028808594,
+      "learning_rate": 1.1166333760087784e-05,
+      "loss": 1.1043,
+      "step": 2280
+    },
+    {
+      "epoch": 4.646013204672423,
+      "grad_norm": 2.5000803470611572,
+      "learning_rate": 1.1102866893540537e-05,
+      "loss": 1.0855,
+      "step": 2290
+    },
+    {
+      "epoch": 4.666328085322498,
+      "grad_norm": 2.488802194595337,
+      "learning_rate": 1.1039355023697547e-05,
+      "loss": 1.0792,
+      "step": 2300
+    },
+    {
+      "epoch": 4.686642965972575,
+      "grad_norm": 2.04819655418396,
+      "learning_rate": 1.097580074220743e-05,
+      "loss": 1.0746,
+      "step": 2310
+    },
+    {
+      "epoch": 4.706957846622651,
+      "grad_norm": 2.619755506515503,
+      "learning_rate": 1.091220664244945e-05,
+      "loss": 1.0807,
+      "step": 2320
+    },
+    {
+      "epoch": 4.7272727272727275,
+      "grad_norm": 2.009214401245117,
+      "learning_rate": 1.0848575319427673e-05,
+      "loss": 1.0881,
+      "step": 2330
+    },
+    {
+      "epoch": 4.747587607922803,
+      "grad_norm": 2.155071258544922,
+      "learning_rate": 1.0784909369665092e-05,
+      "loss": 1.0879,
+      "step": 2340
+    },
+    {
+      "epoch": 4.767902488572879,
+      "grad_norm": 2.4336414337158203,
+      "learning_rate": 1.0721211391097675e-05,
+      "loss": 1.0915,
+      "step": 2350
+    },
+    {
+      "epoch": 4.788217369222956,
+      "grad_norm": 2.942394733428955,
+      "learning_rate": 1.0657483982968341e-05,
+      "loss": 1.0753,
+      "step": 2360
+    },
+    {
+      "epoch": 4.808532249873032,
+      "grad_norm": 2.1195554733276367,
+      "learning_rate": 1.0593729745720913e-05,
+      "loss": 1.067,
+      "step": 2370
+    },
+    {
+      "epoch": 4.828847130523108,
+      "grad_norm": 2.387273073196411,
+      "learning_rate": 1.0529951280893986e-05,
+      "loss": 1.0828,
+      "step": 2380
+    },
+    {
+      "epoch": 4.849162011173185,
+      "grad_norm": 2.598560094833374,
+      "learning_rate": 1.0466151191014784e-05,
+      "loss": 1.0905,
+      "step": 2390
+    },
+    {
+      "epoch": 4.86947689182326,
+      "grad_norm": 2.338967800140381,
+      "learning_rate": 1.0402332079492959e-05,
+      "loss": 1.0888,
+      "step": 2400
+    },
+    {
+      "epoch": 4.889791772473337,
+      "grad_norm": 2.038918972015381,
+      "learning_rate": 1.033849655051435e-05,
+      "loss": 1.0713,
+      "step": 2410
+    },
+    {
+      "epoch": 4.910106653123413,
+      "grad_norm": 2.3864147663116455,
+      "learning_rate": 1.0274647208934732e-05,
+      "loss": 1.0997,
+      "step": 2420
+    },
+    {
+      "epoch": 4.930421533773489,
+      "grad_norm": 2.313906669616699,
+      "learning_rate": 1.02107866601735e-05,
+      "loss": 1.088,
+      "step": 2430
+    },
+    {
+      "epoch": 4.950736414423565,
+      "grad_norm": 1.9792237281799316,
+      "learning_rate": 1.0146917510107377e-05,
+      "loss": 1.0792,
+      "step": 2440
+    },
+    {
+      "epoch": 4.971051295073641,
+      "grad_norm": 2.2212326526641846,
+      "learning_rate": 1.0083042364964062e-05,
+      "loss": 1.0804,
+      "step": 2450
+    },
+    {
+      "epoch": 4.9913661757237175,
+      "grad_norm": 2.190819501876831,
+      "learning_rate": 1.0019163831215894e-05,
+      "loss": 1.0815,
+      "step": 2460
+    },
+    {
+      "epoch": 5.0,
+      "eval_loss": 1.1006184816360474,
+      "eval_runtime": 54.9569,
+      "eval_samples_per_second": 35.846,
+      "eval_steps_per_second": 8.971,
+      "step": 2465
+    },
+    {
+      "epoch": 5.010157440325038,
+      "grad_norm": 2.375826597213745,
+      "learning_rate": 9.955284515473478e-06,
+      "loss": 1.0816,
+      "step": 2470
+    },
+    {
+      "epoch": 5.0304723209751145,
+      "grad_norm": 2.3722262382507324,
+      "learning_rate": 9.89140702437934e-06,
+      "loss": 1.081,
+      "step": 2480
+    },
+    {
+      "epoch": 5.05078720162519,
+      "grad_norm": 2.717057943344116,
+      "learning_rate": 9.827533964501537e-06,
+      "loss": 1.1021,
+      "step": 2490
+    },
+    {
+      "epoch": 5.071102082275266,
+      "grad_norm": 2.2786624431610107,
+      "learning_rate": 9.763667942227323e-06,
+      "loss": 1.0983,
+      "step": 2500
+    },
+    {
+      "epoch": 5.091416962925343,
+      "grad_norm": 2.0201127529144287,
+      "learning_rate": 9.699811563656758e-06,
+      "loss": 1.0694,
+      "step": 2510
+    },
+    {
+      "epoch": 5.111731843575419,
+      "grad_norm": 2.4694550037384033,
+      "learning_rate": 9.635967434496396e-06,
+      "loss": 1.0899,
+      "step": 2520
+    },
+    {
+      "epoch": 5.1320467242254955,
+      "grad_norm": 2.8942248821258545,
+      "learning_rate": 9.572138159952945e-06,
+      "loss": 1.099,
+      "step": 2530
+    },
+    {
+      "epoch": 5.152361604875571,
+      "grad_norm": 2.079746723175049,
+      "learning_rate": 9.508326344626955e-06,
+      "loss": 1.0789,
+      "step": 2540
+    },
+    {
+      "epoch": 5.172676485525647,
+      "grad_norm": 2.178121566772461,
+      "learning_rate": 9.444534592406541e-06,
+      "loss": 1.0763,
+      "step": 2550
+    },
+    {
+      "epoch": 5.192991366175724,
+      "grad_norm": 2.4157040119171143,
+      "learning_rate": 9.380765506361132e-06,
+      "loss": 1.0735,
+      "step": 2560
+    },
+    {
+      "epoch": 5.2133062468258,
+      "grad_norm": 2.3019464015960693,
+      "learning_rate": 9.317021688635236e-06,
+      "loss": 1.0746,
+      "step": 2570
+    },
+    {
+      "epoch": 5.233621127475876,
+      "grad_norm": 2.6220922470092773,
+      "learning_rate": 9.253305740342287e-06,
+      "loss": 1.0663,
+      "step": 2580
+    },
+    {
+      "epoch": 5.253936008125952,
+      "grad_norm": 2.5131311416625977,
+      "learning_rate": 9.189620261458467e-06,
+      "loss": 1.0616,
+      "step": 2590
+    },
+    {
+      "epoch": 5.274250888776028,
+      "grad_norm": 2.5169596672058105,
+      "learning_rate": 9.125967850716648e-06,
+      "loss": 1.0709,
+      "step": 2600
+    },
+    {
+      "epoch": 5.294565769426105,
+      "grad_norm": 2.525052309036255,
+      "learning_rate": 9.062351105500316e-06,
+      "loss": 1.07,
+      "step": 2610
+    },
+    {
+      "epoch": 5.314880650076181,
+      "grad_norm": 2.4557714462280273,
+      "learning_rate": 8.998772621737609e-06,
+      "loss": 1.0802,
+      "step": 2620
+    },
+    {
+      "epoch": 5.335195530726257,
+      "grad_norm": 2.4241530895233154,
+      "learning_rate": 8.935234993795378e-06,
+      "loss": 1.0765,
+      "step": 2630
+    },
+    {
+      "epoch": 5.355510411376333,
+      "grad_norm": 2.6684749126434326,
+      "learning_rate": 8.87174081437332e-06,
+      "loss": 1.0793,
+      "step": 2640
+    },
+    {
+      "epoch": 5.375825292026409,
+      "grad_norm": 2.0009050369262695,
+      "learning_rate": 8.808292674398186e-06,
+      "loss": 1.0584,
+      "step": 2650
+    },
+    {
+      "epoch": 5.3961401726764855,
+      "grad_norm": 2.189467668533325,
+      "learning_rate": 8.744893162918045e-06,
+      "loss": 1.1002,
+      "step": 2660
+    },
+    {
+      "epoch": 5.416455053326562,
+      "grad_norm": 2.2791178226470947,
+      "learning_rate": 8.681544866996657e-06,
+      "loss": 1.0928,
+      "step": 2670
+    },
+    {
+      "epoch": 5.436769933976638,
+      "grad_norm": 2.332437753677368,
+      "learning_rate": 8.618250371607879e-06,
+      "loss": 1.0715,
+      "step": 2680
+    },
+    {
+      "epoch": 5.457084814626714,
+      "grad_norm": 2.518786907196045,
+      "learning_rate": 8.555012259530215e-06,
+      "loss": 1.074,
+      "step": 2690
+    },
+    {
+      "epoch": 5.47739969527679,
+      "grad_norm": 2.394573211669922,
+      "learning_rate": 8.491833111241392e-06,
+      "loss": 1.0846,
+      "step": 2700
+    },
+    {
+      "epoch": 5.497714575926866,
+      "grad_norm": 2.2103583812713623,
+      "learning_rate": 8.428715504813084e-06,
+      "loss": 1.0952,
+      "step": 2710
+    },
+    {
+      "epoch": 5.518029456576943,
+      "grad_norm": 2.51759934425354,
+      "learning_rate": 8.365662015805694e-06,
+      "loss": 1.0918,
+      "step": 2720
+    },
+    {
+      "epoch": 5.538344337227019,
+      "grad_norm": 2.085672616958618,
+      "learning_rate": 8.302675217163279e-06,
+      "loss": 1.0726,
+      "step": 2730
+    },
+    {
+      "epoch": 5.558659217877095,
+      "grad_norm": 2.341189384460449,
+      "learning_rate": 8.23975767910854e-06,
+      "loss": 1.0762,
+      "step": 2740
+    },
+    {
+      "epoch": 5.578974098527171,
+      "grad_norm": 2.3695602416992188,
+      "learning_rate": 8.176911969037951e-06,
+      "loss": 1.0794,
+      "step": 2750
+    },
+    {
+      "epoch": 5.599288979177247,
+      "grad_norm": 2.8153765201568604,
+      "learning_rate": 8.11414065141699e-06,
+      "loss": 1.0823,
+      "step": 2760
+    },
+    {
+      "epoch": 5.619603859827324,
+      "grad_norm": 2.3020527362823486,
+      "learning_rate": 8.051446287675499e-06,
+      "loss": 1.064,
+      "step": 2770
+    },
+    {
+      "epoch": 5.6399187404774,
+      "grad_norm": 2.436936378479004,
+      "learning_rate": 7.988831436103157e-06,
+      "loss": 1.0911,
+      "step": 2780
+    },
+    {
+      "epoch": 5.6602336211274755,
+      "grad_norm": 2.706698417663574,
+      "learning_rate": 7.926298651745095e-06,
+      "loss": 1.1105,
+      "step": 2790
+    },
+    {
+      "epoch": 5.680548501777552,
+      "grad_norm": 2.290712594985962,
+      "learning_rate": 7.863850486297622e-06,
+      "loss": 1.0668,
+      "step": 2800
+    },
+    {
+      "epoch": 5.700863382427628,
+      "grad_norm": 2.3723671436309814,
+      "learning_rate": 7.801489488004124e-06,
+      "loss": 1.0701,
+      "step": 2810
+    },
+    {
+      "epoch": 5.721178263077705,
+      "grad_norm": 2.1805179119110107,
+      "learning_rate": 7.739218201551054e-06,
+      "loss": 1.0718,
+      "step": 2820
+    },
+    {
+      "epoch": 5.741493143727781,
+      "grad_norm": 2.39038348197937,
+      "learning_rate": 7.677039167964114e-06,
+      "loss": 1.075,
+      "step": 2830
+    },
+    {
+      "epoch": 5.761808024377856,
+      "grad_norm": 2.164767265319824,
+      "learning_rate": 7.614954924504553e-06,
+      "loss": 1.0857,
+      "step": 2840
+    },
+    {
+      "epoch": 5.782122905027933,
+      "grad_norm": 2.3338029384613037,
+      "learning_rate": 7.552968004565653e-06,
+      "loss": 1.0692,
+      "step": 2850
+    },
+    {
+      "epoch": 5.802437785678009,
+      "grad_norm": 2.186400890350342,
+      "learning_rate": 7.491080937569321e-06,
+      "loss": 1.0759,
+      "step": 2860
+    },
+    {
+      "epoch": 5.8227526663280855,
+      "grad_norm": 2.548473834991455,
+      "learning_rate": 7.429296248862907e-06,
+      "loss": 1.0799,
+      "step": 2870
+    },
+    {
+      "epoch": 5.843067546978162,
+      "grad_norm": 2.4101195335388184,
+      "learning_rate": 7.367616459616127e-06,
+      "loss": 1.0772,
+      "step": 2880
+    },
+    {
+      "epoch": 5.863382427628237,
+      "grad_norm": 2.5707526206970215,
+      "learning_rate": 7.306044086718207e-06,
+      "loss": 1.0772,
+      "step": 2890
+    },
+    {
+      "epoch": 5.883697308278314,
+      "grad_norm": 2.447829246520996,
+      "learning_rate": 7.244581642675159e-06,
+      "loss": 1.0823,
+      "step": 2900
+    },
+    {
+      "epoch": 5.90401218892839,
+      "grad_norm": 2.3447105884552,
+      "learning_rate": 7.1832316355072785e-06,
+      "loss": 1.0927,
+      "step": 2910
+    },
+    {
+      "epoch": 5.924327069578466,
+      "grad_norm": 2.4636731147766113,
+      "learning_rate": 7.121996568646786e-06,
+      "loss": 1.0963,
+      "step": 2920
+    },
+    {
+      "epoch": 5.944641950228542,
+      "grad_norm": 2.2715587615966797,
+      "learning_rate": 7.060878940835674e-06,
+      "loss": 1.1112,
+      "step": 2930
+    },
+    {
+      "epoch": 5.964956830878618,
+      "grad_norm": 2.37288236618042,
+      "learning_rate": 6.999881246023754e-06,
+      "loss": 1.0959,
+      "step": 2940
+    },
+    {
+      "epoch": 5.985271711528695,
+      "grad_norm": 2.28794264793396,
+      "learning_rate": 6.9390059732668835e-06,
+      "loss": 1.0642,
+      "step": 2950
+    },
+    {
+      "epoch": 6.0,
+      "eval_loss": 1.0949872732162476,
+      "eval_runtime": 54.7572,
+      "eval_samples_per_second": 35.977,
+      "eval_steps_per_second": 9.003,
+      "step": 2958
+    },
+    {
+      "epoch": 6.004062976130015,
+      "grad_norm": 2.5767059326171875,
+      "learning_rate": 6.878255606625395e-06,
+      "loss": 1.0996,
+      "step": 2960
+    },
+    {
+      "epoch": 6.024377856780092,
+      "grad_norm": 2.55753755569458,
+      "learning_rate": 6.817632625062735e-06,
+      "loss": 1.0868,
+      "step": 2970
+    },
+    {
+      "epoch": 6.044692737430168,
+      "grad_norm": 2.5268876552581787,
+      "learning_rate": 6.757139502344312e-06,
+      "loss": 1.0957,
+      "step": 2980
+    },
+    {
+      "epoch": 6.0650076180802435,
+      "grad_norm": 2.596231698989868,
+      "learning_rate": 6.696778706936542e-06,
+      "loss": 1.078,
+      "step": 2990
+    },
+    {
+      "epoch": 6.08532249873032,
+      "grad_norm": 2.4920766353607178,
+      "learning_rate": 6.636552701906135e-06,
+      "loss": 1.0653,
+      "step": 3000
+    },
+    {
+      "epoch": 6.105637379380396,
+      "grad_norm": 2.75101900100708,
+      "learning_rate": 6.576463944819576e-06,
+      "loss": 1.1024,
+      "step": 3010
+    },
+    {
+      "epoch": 6.1259522600304726,
+      "grad_norm": 2.3427553176879883,
+      "learning_rate": 6.516514887642846e-06,
+      "loss": 1.0864,
+      "step": 3020
+    },
+    {
+      "epoch": 6.146267140680549,
+      "grad_norm": 2.1833336353302,
+      "learning_rate": 6.456707976641368e-06,
+      "loss": 1.066,
+      "step": 3030
+    },
+    {
+      "epoch": 6.166582021330624,
+      "grad_norm": 2.4288933277130127,
+      "learning_rate": 6.397045652280192e-06,
+      "loss": 1.0787,
+      "step": 3040
+    },
+    {
+      "epoch": 6.186896901980701,
+      "grad_norm": 2.3094112873077393,
+      "learning_rate": 6.3375303491243925e-06,
+      "loss": 1.0703,
+      "step": 3050
+    },
+    {
+      "epoch": 6.207211782630777,
+      "grad_norm": 2.3378853797912598,
+      "learning_rate": 6.278164495739749e-06,
+      "loss": 1.1107,
+      "step": 3060
+    },
+    {
+      "epoch": 6.2275266632808535,
+      "grad_norm": 2.429354667663574,
+      "learning_rate": 6.21895051459362e-06,
+      "loss": 1.0632,
+      "step": 3070
+    },
+    {
+      "epoch": 6.24784154393093,
+      "grad_norm": 2.5030083656311035,
+      "learning_rate": 6.159890821956118e-06,
+      "loss": 1.0794,
+      "step": 3080
+    },
+    {
+      "epoch": 6.268156424581005,
+      "grad_norm": 2.2454118728637695,
+      "learning_rate": 6.100987827801486e-06,
+      "loss": 1.0555,
+      "step": 3090
+    },
+    {
+      "epoch": 6.288471305231082,
+      "grad_norm": 2.3601973056793213,
+      "learning_rate": 6.042243935709782e-06,
+      "loss": 1.0758,
+      "step": 3100
+    },
+    {
+      "epoch": 6.308786185881158,
+      "grad_norm": 1.9250520467758179,
+      "learning_rate": 5.983661542768777e-06,
+      "loss": 1.066,
+      "step": 3110
+    },
+    {
+      "epoch": 6.329101066531234,
+      "grad_norm": 2.4363834857940674,
+      "learning_rate": 5.9252430394761625e-06,
+      "loss": 1.089,
+      "step": 3120
+    },
+    {
+      "epoch": 6.34941594718131,
+      "grad_norm": 2.1766016483306885,
+      "learning_rate": 5.866990809641978e-06,
+      "loss": 1.0695,
+      "step": 3130
+    },
+    {
+      "epoch": 6.369730827831386,
+      "grad_norm": 2.575039863586426,
+      "learning_rate": 5.808907230291359e-06,
+      "loss": 1.0874,
+      "step": 3140
+    },
+    {
+      "epoch": 6.390045708481463,
+      "grad_norm": 2.4749789237976074,
+      "learning_rate": 5.75099467156754e-06,
+      "loss": 1.1004,
+      "step": 3150
+    },
+    {
+      "epoch": 6.410360589131539,
+      "grad_norm": 2.669945240020752,
+      "learning_rate": 5.6932554966351185e-06,
+      "loss": 1.0756,
+      "step": 3160
+    },
+    {
+      "epoch": 6.430675469781615,
+      "grad_norm": 2.250049114227295,
+      "learning_rate": 5.635692061583651e-06,
+      "loss": 1.074,
+      "step": 3170
+    },
+    {
+      "epoch": 6.450990350431692,
+      "grad_norm": 2.271474599838257,
+      "learning_rate": 5.578306715331494e-06,
+      "loss": 1.0881,
+      "step": 3180
+    },
+    {
+      "epoch": 6.471305231081767,
+      "grad_norm": 2.2663955688476562,
+      "learning_rate": 5.521101799529965e-06,
+      "loss": 1.0601,
+      "step": 3190
+    },
+    {
+      "epoch": 6.4916201117318435,
+      "grad_norm": 2.209956407546997,
+      "learning_rate": 5.464079648467769e-06,
+      "loss": 1.0569,
+      "step": 3200
+    },
+    {
+      "epoch": 6.51193499238192,
+      "grad_norm": 2.395031213760376,
+      "learning_rate": 5.407242588975777e-06,
+      "loss": 1.0687,
+      "step": 3210
+    },
+    {
+      "epoch": 6.532249873031996,
+      "grad_norm": 2.570211410522461,
+      "learning_rate": 5.3505929403320575e-06,
+      "loss": 1.064,
+      "step": 3220
+    },
+    {
+      "epoch": 6.552564753682072,
+      "grad_norm": 2.569349765777588,
+      "learning_rate": 5.294133014167242e-06,
+      "loss": 1.0752,
+      "step": 3230
+    },
+    {
+      "epoch": 6.572879634332148,
+      "grad_norm": 2.717036008834839,
+      "learning_rate": 5.237865114370192e-06,
+      "loss": 1.06,
+      "step": 3240
+    },
+    {
+      "epoch": 6.593194514982224,
+      "grad_norm": 2.299722671508789,
+      "learning_rate": 5.18179153699399e-06,
+      "loss": 1.0828,
+      "step": 3250
+    },
+    {
+      "epoch": 6.613509395632301,
+      "grad_norm": 2.2542483806610107,
+      "learning_rate": 5.125914570162259e-06,
+      "loss": 1.0829,
+      "step": 3260
+    },
+    {
+      "epoch": 6.633824276282377,
+      "grad_norm": 2.26954984664917,
+      "learning_rate": 5.070236493975777e-06,
+      "loss": 1.0877,
+      "step": 3270
+    },
+    {
+      "epoch": 6.6541391569324535,
+      "grad_norm": 2.2126009464263916,
+      "learning_rate": 5.014759580419448e-06,
+      "loss": 1.0676,
+      "step": 3280
+    },
+    {
+      "epoch": 6.674454037582529,
+      "grad_norm": 2.658097505569458,
+      "learning_rate": 4.959486093269577e-06,
+      "loss": 1.0746,
+      "step": 3290
+    },
+    {
+      "epoch": 6.694768918232605,
+      "grad_norm": 2.2925167083740234,
+      "learning_rate": 4.904418288001511e-06,
+      "loss": 1.0689,
+      "step": 3300
+    },
+    {
+      "epoch": 6.715083798882682,
+      "grad_norm": 2.2561416625976562,
+      "learning_rate": 4.849558411697604e-06,
+      "loss": 1.0787,
+      "step": 3310
+    },
+    {
+      "epoch": 6.735398679532758,
+      "grad_norm": 2.590914726257324,
+      "learning_rate": 4.794908702955496e-06,
+      "loss": 1.0673,
+      "step": 3320
+    },
+    {
+      "epoch": 6.7557135601828335,
+      "grad_norm": 2.577042579650879,
+      "learning_rate": 4.740471391796807e-06,
+      "loss": 1.0841,
+      "step": 3330
+    },
+    {
+      "epoch": 6.77602844083291,
+      "grad_norm": 2.53352689743042,
+      "learning_rate": 4.686248699576098e-06,
+      "loss": 1.0603,
+      "step": 3340
+    },
+    {
+      "epoch": 6.796343321482986,
+      "grad_norm": 2.2674384117126465,
+      "learning_rate": 4.632242838890257e-06,
+      "loss": 1.0651,
+      "step": 3350
+    },
+    {
+      "epoch": 6.816658202133063,
+      "grad_norm": 2.618176221847534,
+      "learning_rate": 4.578456013488206e-06,
+      "loss": 1.0714,
+      "step": 3360
+    },
+    {
+      "epoch": 6.836973082783139,
+      "grad_norm": 2.610429286956787,
+      "learning_rate": 4.524890418180954e-06,
+      "loss": 1.1131,
+      "step": 3370
+    },
+    {
+      "epoch": 6.8572879634332145,
+      "grad_norm": 2.254110813140869,
+      "learning_rate": 4.471548238752068e-06,
+      "loss": 1.0697,
+      "step": 3380
+    },
+    {
+      "epoch": 6.877602844083291,
+      "grad_norm": 2.650230884552002,
+      "learning_rate": 4.4184316518684625e-06,
+      "loss": 1.0892,
+      "step": 3390
+    },
+    {
+      "epoch": 6.897917724733367,
+      "grad_norm": 2.45845890045166,
+      "learning_rate": 4.365542824991581e-06,
+      "loss": 1.0798,
+      "step": 3400
+    },
+    {
+      "epoch": 6.9182326053834435,
+      "grad_norm": 2.395923376083374,
+      "learning_rate": 4.312883916288947e-06,
+      "loss": 1.0596,
+      "step": 3410
+    },
+    {
+      "epoch": 6.93854748603352,
+      "grad_norm": 2.744338274002075,
+      "learning_rate": 4.26045707454611e-06,
+      "loss": 1.0624,
+      "step": 3420
+    },
+    {
+      "epoch": 6.958862366683595,
+      "grad_norm": 2.431194543838501,
+      "learning_rate": 4.2082644390789565e-06,
+      "loss": 1.0762,
+      "step": 3430
+    },
+    {
+      "epoch": 6.979177247333672,
+      "grad_norm": 2.344907522201538,
+      "learning_rate": 4.156308139646417e-06,
+      "loss": 1.0716,
+      "step": 3440
+    },
+    {
+      "epoch": 6.999492127983748,
+      "grad_norm": 2.289046049118042,
+      "learning_rate": 4.1045902963635355e-06,
+      "loss": 1.0749,
+      "step": 3450
+    },
+    {
+      "epoch": 7.0,
+      "eval_loss": 1.0920889377593994,
+      "eval_runtime": 54.5848,
+      "eval_samples_per_second": 36.091,
+      "eval_steps_per_second": 9.032,
+      "step": 3451
+    },
+    {
+      "epoch": 7.018283392585069,
+      "grad_norm": 2.9955337047576904,
+      "learning_rate": 4.053113019615001e-06,
+      "loss": 1.0685,
+      "step": 3460
+    },
+    {
+      "epoch": 7.038598273235145,
+      "grad_norm": 2.43476939201355,
+      "learning_rate": 4.001878409969e-06,
+      "loss": 1.0741,
+      "step": 3470
+    },
+    {
+      "epoch": 7.058913153885221,
+      "grad_norm": 2.4879932403564453,
+      "learning_rate": 3.95088855809151e-06,
+      "loss": 1.0906,
+      "step": 3480
+    },
+    {
+      "epoch": 7.079228034535297,
+      "grad_norm": 2.2968192100524902,
+      "learning_rate": 3.900145544660981e-06,
+      "loss": 1.0646,
+      "step": 3490
+    },
+    {
+      "epoch": 7.099542915185373,
+      "grad_norm": 2.6439547538757324,
+      "learning_rate": 3.849651440283443e-06,
+      "loss": 1.0802,
+      "step": 3500
+    },
+    {
+      "epoch": 7.11985779583545,
+      "grad_norm": 2.3785431385040283,
+      "learning_rate": 3.7994083054080123e-06,
+      "loss": 1.067,
+      "step": 3510
+    },
+    {
+      "epoch": 7.140172676485526,
+      "grad_norm": 2.7672765254974365,
+      "learning_rate": 3.7494181902428094e-06,
+      "loss": 1.082,
+      "step": 3520
+    },
+    {
+      "epoch": 7.1604875571356015,
+      "grad_norm": 2.634470224380493,
+      "learning_rate": 3.6996831346712892e-06,
+      "loss": 1.0749,
+      "step": 3530
+    },
+    {
+      "epoch": 7.180802437785678,
+      "grad_norm": 2.251033067703247,
+      "learning_rate": 3.6502051681690266e-06,
+      "loss": 1.0831,
+      "step": 3540
+    },
+    {
+      "epoch": 7.201117318435754,
+      "grad_norm": 2.855700969696045,
+      "learning_rate": 3.600986309720873e-06,
+      "loss": 1.0765,
+      "step": 3550
+    },
+    {
+      "epoch": 7.221432199085831,
+      "grad_norm": 2.3868894577026367,
+      "learning_rate": 3.552028567738596e-06,
+      "loss": 1.0614,
+      "step": 3560
+    },
+    {
+      "epoch": 7.241747079735907,
+      "grad_norm": 2.3770251274108887,
+      "learning_rate": 3.5033339399789125e-06,
+      "loss": 1.0764,
+      "step": 3570
+    },
+    {
+      "epoch": 7.262061960385982,
+      "grad_norm": 2.417048215866089,
+      "learning_rate": 3.4549044134619603e-06,
+      "loss": 1.0607,
+      "step": 3580
+    },
+    {
+      "epoch": 7.282376841036059,
+      "grad_norm": 2.5006279945373535,
+      "learning_rate": 3.4067419643902367e-06,
+      "loss": 1.0481,
+      "step": 3590
+    },
+    {
+      "epoch": 7.302691721686135,
+      "grad_norm": 2.4641923904418945,
+      "learning_rate": 3.3588485580679465e-06,
+      "loss": 1.0755,
+      "step": 3600
+    },
+    {
+      "epoch": 7.3230066023362115,
+      "grad_norm": 2.6034882068634033,
+      "learning_rate": 3.3112261488207962e-06,
+      "loss": 1.068,
+      "step": 3610
+    },
+    {
+      "epoch": 7.343321482986288,
+      "grad_norm": 2.483919858932495,
+      "learning_rate": 3.263876679916267e-06,
+      "loss": 1.0659,
+      "step": 3620
+    },
+    {
+      "epoch": 7.363636363636363,
+      "grad_norm": 2.473473072052002,
+      "learning_rate": 3.216802083484307e-06,
+      "loss": 1.0953,
+      "step": 3630
+    },
+    {
+      "epoch": 7.38395124428644,
+      "grad_norm": 2.57886004447937,
+      "learning_rate": 3.1700042804384924e-06,
+      "loss": 1.0537,
+      "step": 3640
+    },
+    {
+      "epoch": 7.404266124936516,
+      "grad_norm": 2.4494612216949463,
+      "learning_rate": 3.1234851803976352e-06,
+      "loss": 1.0834,
+      "step": 3650
+    },
+    {
+      "epoch": 7.424581005586592,
+      "grad_norm": 2.679572105407715,
+      "learning_rate": 3.077246681607865e-06,
+      "loss": 1.0538,
+      "step": 3660
+    },
+    {
+      "epoch": 7.444895886236669,
+      "grad_norm": 2.955970287322998,
+      "learning_rate": 3.03129067086518e-06,
+      "loss": 1.0855,
+      "step": 3670
+    },
+    {
+      "epoch": 7.465210766886744,
+      "grad_norm": 2.511077404022217,
+      "learning_rate": 2.98561902343844e-06,
+      "loss": 1.094,
+      "step": 3680
+    },
+    {
+      "epoch": 7.485525647536821,
+      "grad_norm": 3.1838951110839844,
+      "learning_rate": 2.9402336029928526e-06,
+      "loss": 1.0813,
+      "step": 3690
+    },
+    {
+      "epoch": 7.505840528186897,
+      "grad_norm": 2.397426128387451,
+      "learning_rate": 2.8951362615139167e-06,
+      "loss": 1.0666,
+      "step": 3700
+    },
+    {
+      "epoch": 7.526155408836973,
+      "grad_norm": 2.655226230621338,
+      "learning_rate": 2.8503288392318627e-06,
+      "loss": 1.0586,
+      "step": 3710
+    },
+    {
+      "epoch": 7.546470289487049,
+      "grad_norm": 2.5811400413513184,
+      "learning_rate": 2.8058131645465546e-06,
+      "loss": 1.0763,
+      "step": 3720
+    },
+    {
+      "epoch": 7.566785170137125,
+      "grad_norm": 2.3530173301696777,
+      "learning_rate": 2.7615910539528824e-06,
+      "loss": 1.069,
+      "step": 3730
+    },
+    {
+      "epoch": 7.5871000507872015,
+      "grad_norm": 2.2900733947753906,
+      "learning_rate": 2.717664311966627e-06,
+      "loss": 1.0621,
+      "step": 3740
+    },
+    {
+      "epoch": 7.607414931437278,
+      "grad_norm": 2.464092493057251,
+      "learning_rate": 2.6740347310508487e-06,
+      "loss": 1.0622,
+      "step": 3750
+    },
+    {
+      "epoch": 7.627729812087354,
+      "grad_norm": 2.5965607166290283,
+      "learning_rate": 2.6307040915427196e-06,
+      "loss": 1.0682,
+      "step": 3760
+    },
+    {
+      "epoch": 7.648044692737431,
+      "grad_norm": 2.3087034225463867,
+      "learning_rate": 2.5876741615809007e-06,
+      "loss": 1.0653,
+      "step": 3770
+    },
+    {
+      "epoch": 7.668359573387506,
+      "grad_norm": 2.2895405292510986,
+      "learning_rate": 2.5449466970333658e-06,
+      "loss": 1.0716,
+      "step": 3780
+    },
+    {
+      "epoch": 7.688674454037582,
+      "grad_norm": 2.888216972351074,
+      "learning_rate": 2.5025234414257717e-06,
+      "loss": 1.0812,
+      "step": 3790
+    },
+    {
+      "epoch": 7.708989334687659,
+      "grad_norm": 2.578557014465332,
+      "learning_rate": 2.460406125870305e-06,
+      "loss": 1.0694,
+      "step": 3800
+    },
+    {
+      "epoch": 7.729304215337735,
+      "grad_norm": 2.9586007595062256,
+      "learning_rate": 2.418596468995044e-06,
+      "loss": 1.0819,
+      "step": 3810
+    },
+    {
+      "epoch": 7.749619095987811,
+      "grad_norm": 2.469660520553589,
+      "learning_rate": 2.3770961768738222e-06,
+      "loss": 1.0815,
+      "step": 3820
+    },
+    {
+      "epoch": 7.769933976637887,
+      "grad_norm": 2.530141592025757,
+      "learning_rate": 2.3359069429566216e-06,
+      "loss": 1.0876,
+      "step": 3830
+    },
+    {
+      "epoch": 7.790248857287963,
+      "grad_norm": 2.477205514907837,
+      "learning_rate": 2.295030448000466e-06,
+      "loss": 1.0533,
+      "step": 3840
+    },
+    {
+      "epoch": 7.81056373793804,
+      "grad_norm": 2.493910789489746,
+      "learning_rate": 2.2544683600008345e-06,
+      "loss": 1.0663,
+      "step": 3850
+    },
+    {
+      "epoch": 7.830878618588116,
+      "grad_norm": 2.667234182357788,
+      "learning_rate": 2.214222334123596e-06,
+      "loss": 1.0802,
+      "step": 3860
+    },
+    {
+      "epoch": 7.851193499238192,
+      "grad_norm": 2.600020408630371,
+      "learning_rate": 2.17429401263747e-06,
+      "loss": 1.1002,
+      "step": 3870
+    },
+    {
+      "epoch": 7.871508379888268,
+      "grad_norm": 2.2591912746429443,
+      "learning_rate": 2.134685024847023e-06,
+      "loss": 1.0694,
+      "step": 3880
+    },
+    {
+      "epoch": 7.891823260538344,
+      "grad_norm": 2.7257189750671387,
+      "learning_rate": 2.0953969870261704e-06,
+      "loss": 1.0684,
+      "step": 3890
+    },
+    {
+      "epoch": 7.912138141188421,
+      "grad_norm": 2.2915289402008057,
+      "learning_rate": 2.0564315023522218e-06,
+      "loss": 1.0601,
+      "step": 3900
+    },
+    {
+      "epoch": 7.932453021838497,
+      "grad_norm": 2.529210090637207,
+      "learning_rate": 2.017790160840478e-06,
+      "loss": 1.0902,
+      "step": 3910
+    },
+    {
+      "epoch": 7.9527679024885725,
+      "grad_norm": 2.3367769718170166,
+      "learning_rate": 1.9794745392793344e-06,
+      "loss": 1.0629,
+      "step": 3920
+    },
+    {
+      "epoch": 7.973082783138649,
+      "grad_norm": 2.5318503379821777,
+      "learning_rate": 1.9414862011659462e-06,
+      "loss": 1.0851,
+      "step": 3930
+    },
+    {
+      "epoch": 7.993397663788725,
+      "grad_norm": 2.2982535362243652,
+      "learning_rate": 1.9038266966424213e-06,
+      "loss": 1.0725,
+      "step": 3940
+    },
+    {
+      "epoch": 8.0,
+      "eval_loss": 1.0901824235916138,
+      "eval_runtime": 55.0806,
+      "eval_samples_per_second": 35.766,
+      "eval_steps_per_second": 8.951,
+      "step": 3944
+    },
+    {
+      "epoch": 8.012188928390046,
+      "grad_norm": 2.5969467163085938,
+      "learning_rate": 1.8664975624325743e-06,
+      "loss": 1.0801,
+      "step": 3950
+    },
+    {
+      "epoch": 8.032503809040122,
+      "grad_norm": 2.4681787490844727,
+      "learning_rate": 1.8295003217792206e-06,
+      "loss": 1.0732,
+      "step": 3960
+    },
+    {
+      "epoch": 8.052818689690199,
+      "grad_norm": 2.256989002227783,
+      "learning_rate": 1.7928364843820034e-06,
+      "loss": 1.0615,
+      "step": 3970
+    },
+    {
+      "epoch": 8.073133570340275,
+      "grad_norm": 2.6549007892608643,
+      "learning_rate": 1.756507546335814e-06,
+      "loss": 1.0692,
+      "step": 3980
+    },
+    {
+      "epoch": 8.093448450990351,
+      "grad_norm": 2.311284065246582,
+      "learning_rate": 1.7205149900697138e-06,
+      "loss": 1.0669,
+      "step": 3990
+    },
+    {
+      "epoch": 8.113763331640426,
+      "grad_norm": 2.461761236190796,
+      "learning_rate": 1.6848602842864726e-06,
+      "loss": 1.0617,
+      "step": 4000
+    },
+    {
+      "epoch": 8.134078212290502,
+      "grad_norm": 2.347125291824341,
+      "learning_rate": 1.6495448839026128e-06,
+      "loss": 1.063,
+      "step": 4010
+    },
+    {
+      "epoch": 8.154393092940579,
+      "grad_norm": 2.2828891277313232,
+      "learning_rate": 1.6145702299890576e-06,
+      "loss": 1.1011,
+      "step": 4020
+    },
+    {
+      "epoch": 8.174707973590655,
+      "grad_norm": 2.536144256591797,
+      "learning_rate": 1.5799377497123104e-06,
+      "loss": 1.0696,
+      "step": 4030
+    },
+    {
+      "epoch": 8.195022854240731,
+      "grad_norm": 2.7368884086608887,
+      "learning_rate": 1.545648856276234e-06,
+      "loss": 1.0463,
+      "step": 4040
+    },
+    {
+      "epoch": 8.215337734890808,
+      "grad_norm": 2.709357976913452,
+      "learning_rate": 1.5117049488643787e-06,
+      "loss": 1.0894,
+      "step": 4050
+    },
+    {
+      "epoch": 8.235652615540884,
+      "grad_norm": 2.5076632499694824,
+      "learning_rate": 1.4781074125828799e-06,
+      "loss": 1.0951,
+      "step": 4060
+    },
+    {
+      "epoch": 8.25596749619096,
+      "grad_norm": 2.656967878341675,
+      "learning_rate": 1.4448576184039442e-06,
+      "loss": 1.0818,
+      "step": 4070
+    },
+    {
+      "epoch": 8.276282376841037,
+      "grad_norm": 2.170839786529541,
+      "learning_rate": 1.4119569231099107e-06,
+      "loss": 1.0555,
+      "step": 4080
+    },
+    {
+      "epoch": 8.296597257491113,
+      "grad_norm": 2.704313278198242,
+      "learning_rate": 1.3794066692378793e-06,
+      "loss": 1.0575,
+      "step": 4090
+    },
+    {
+      "epoch": 8.316912138141188,
+      "grad_norm": 2.512399911880493,
+      "learning_rate": 1.347208185024932e-06,
+      "loss": 1.067,
+      "step": 4100
+    },
+    {
+      "epoch": 8.337227018791264,
+      "grad_norm": 2.5209548473358154,
+      "learning_rate": 1.315362784353923e-06,
+      "loss": 1.0529,
+      "step": 4110
+    },
+    {
+      "epoch": 8.35754189944134,
+      "grad_norm": 2.6369845867156982,
+      "learning_rate": 1.2838717666998802e-06,
+      "loss": 1.0723,
+      "step": 4120
+    },
+    {
+      "epoch": 8.377856780091417,
+      "grad_norm": 3.272552013397217,
+      "learning_rate": 1.2527364170769686e-06,
+      "loss": 1.0733,
+      "step": 4130
+    },
+    {
+      "epoch": 8.398171660741493,
+      "grad_norm": 2.4088523387908936,
+      "learning_rate": 1.2219580059860602e-06,
+      "loss": 1.0554,
+      "step": 4140
+    },
+    {
+      "epoch": 8.41848654139157,
+      "grad_norm": 2.388200044631958,
+      "learning_rate": 1.191537789362881e-06,
+      "loss": 1.0691,
+      "step": 4150
+    },
+    {
+      "epoch": 8.438801422041646,
+      "grad_norm": 2.4707677364349365,
+      "learning_rate": 1.1614770085267724e-06,
+      "loss": 1.0583,
+      "step": 4160
+    },
+    {
+      "epoch": 8.459116302691722,
+      "grad_norm": 2.6969428062438965,
+      "learning_rate": 1.1317768901300375e-06,
+      "loss": 1.0633,
+      "step": 4170
+    },
+    {
+      "epoch": 8.479431183341799,
+      "grad_norm": 2.359264850616455,
+      "learning_rate": 1.1024386461078752e-06,
+      "loss": 1.0503,
+      "step": 4180
+    },
+    {
+      "epoch": 8.499746063991875,
+      "grad_norm": 2.645862579345703,
+      "learning_rate": 1.073463473628935e-06,
+      "loss": 1.0953,
+      "step": 4190
+    },
+    {
+      "epoch": 8.52006094464195,
+      "grad_norm": 2.5734922885894775,
+      "learning_rate": 1.0448525550464704e-06,
+      "loss": 1.0529,
+      "step": 4200
+    },
+    {
+      "epoch": 8.540375825292026,
+      "grad_norm": 2.237166404724121,
+      "learning_rate": 1.0166070578500842e-06,
+      "loss": 1.0684,
+      "step": 4210
+    },
+    {
+      "epoch": 8.560690705942102,
+      "grad_norm": 2.320554733276367,
+      "learning_rate": 9.887281346180888e-07,
+      "loss": 1.0654,
+      "step": 4220
+    },
+    {
+      "epoch": 8.581005586592179,
+      "grad_norm": 2.724879503250122,
+      "learning_rate": 9.612169229704748e-07,
+      "loss": 1.0667,
+      "step": 4230
+    },
+    {
+      "epoch": 8.601320467242255,
+      "grad_norm": 2.951643705368042,
+      "learning_rate": 9.340745455224943e-07,
+      "loss": 1.0892,
+      "step": 4240
+    },
+    {
+      "epoch": 8.621635347892331,
+      "grad_norm": 2.4628891944885254,
+      "learning_rate": 9.073021098388468e-07,
+      "loss": 1.0776,
+      "step": 4250
+    },
+    {
+      "epoch": 8.641950228542408,
+      "grad_norm": 2.5746138095855713,
+      "learning_rate": 8.809007083884868e-07,
+      "loss": 1.0644,
+      "step": 4260
+    },
+    {
+      "epoch": 8.662265109192484,
+      "grad_norm": 2.6229591369628906,
+      "learning_rate": 8.548714185000429e-07,
+      "loss": 1.0795,
+      "step": 4270
+    },
+    {
+      "epoch": 8.68257998984256,
+      "grad_norm": 2.9902350902557373,
+      "learning_rate": 8.292153023178517e-07,
+      "loss": 1.0791,
+      "step": 4280
+    },
+    {
+      "epoch": 8.702894870492635,
+      "grad_norm": 2.228269577026367,
+      "learning_rate": 8.039334067586325e-07,
+      "loss": 1.0862,
+      "step": 4290
+    },
+    {
+      "epoch": 8.723209751142711,
+      "grad_norm": 2.6434853076934814,
+      "learning_rate": 7.790267634687487e-07,
+      "loss": 1.0752,
+      "step": 4300
+    },
+    {
+      "epoch": 8.743524631792788,
+      "grad_norm": 2.551692008972168,
+      "learning_rate": 7.544963887821233e-07,
+      "loss": 1.0675,
+      "step": 4310
+    },
+    {
+      "epoch": 8.763839512442864,
+      "grad_norm": 2.4782984256744385,
+      "learning_rate": 7.303432836787572e-07,
+      "loss": 1.0768,
+      "step": 4320
+    },
+    {
+      "epoch": 8.78415439309294,
+      "grad_norm": 2.706620931625366,
+      "learning_rate": 7.065684337438916e-07,
+      "loss": 1.0881,
+      "step": 4330
+    },
+    {
+      "epoch": 8.804469273743017,
+      "grad_norm": 2.345564365386963,
+      "learning_rate": 6.831728091277867e-07,
+      "loss": 1.0771,
+      "step": 4340
+    },
+    {
+      "epoch": 8.824784154393093,
+      "grad_norm": 2.543198347091675,
+      "learning_rate": 6.601573645061321e-07,
+      "loss": 1.049,
+      "step": 4350
+    },
+    {
+      "epoch": 8.84509903504317,
+      "grad_norm": 2.7005691528320312,
+      "learning_rate": 6.375230390410947e-07,
+      "loss": 1.0871,
+      "step": 4360
+    },
+    {
+      "epoch": 8.865413915693246,
+      "grad_norm": 2.9565064907073975,
+      "learning_rate": 6.152707563429949e-07,
+      "loss": 1.0691,
+      "step": 4370
+    },
+    {
+      "epoch": 8.885728796343322,
+      "grad_norm": 2.2274577617645264,
+      "learning_rate": 5.934014244326114e-07,
+      "loss": 1.0697,
+      "step": 4380
+    },
+    {
+      "epoch": 8.906043676993399,
+      "grad_norm": 2.464477062225342,
+      "learning_rate": 5.719159357041415e-07,
+      "loss": 1.0937,
+      "step": 4390
+    },
+    {
+      "epoch": 8.926358557643473,
+      "grad_norm": 2.860454797744751,
+      "learning_rate": 5.508151668887696e-07,
+      "loss": 1.065,
+      "step": 4400
+    },
+    {
+      "epoch": 8.94667343829355,
+      "grad_norm": 2.2630953788757324,
+      "learning_rate": 5.300999790189088e-07,
+      "loss": 1.0709,
+      "step": 4410
+    },
+    {
+      "epoch": 8.966988318943626,
+      "grad_norm": 2.171909809112549,
+      "learning_rate": 5.11786680609313e-07,
+      "loss": 1.0732,
+      "step": 4420
+    },
+    {
+      "epoch": 8.987303199593702,
+      "grad_norm": 2.525136947631836,
+      "learning_rate": 4.918064123013932e-07,
+      "loss": 1.0703,
+      "step": 4430
+    },
+    {
+      "epoch": 9.0,
+      "eval_loss": 1.0895577669143677,
+      "eval_runtime": 54.6524,
+      "eval_samples_per_second": 36.046,
+      "eval_steps_per_second": 9.021,
+      "step": 4437
+    },
+    {
+      "epoch": 9.006094464195023,
+      "grad_norm": 2.3463656902313232,
+      "learning_rate": 4.722141328346708e-07,
+      "loss": 1.0563,
+      "step": 4440
+    },
+    {
+      "epoch": 9.0264093448451,
+      "grad_norm": 2.3120319843292236,
+      "learning_rate": 4.530106416865332e-07,
+      "loss": 1.0576,
+      "step": 4450
+    },
+    {
+      "epoch": 9.046724225495176,
+      "grad_norm": 2.431831121444702,
+      "learning_rate": 4.3419672246958313e-07,
+      "loss": 1.0689,
+      "step": 4460
+    },
+    {
+      "epoch": 9.067039106145252,
+      "grad_norm": 2.210456609725952,
+      "learning_rate": 4.1577314289964474e-07,
+      "loss": 1.0576,
+      "step": 4470
+    },
+    {
+      "epoch": 9.087353986795328,
+      "grad_norm": 2.559464931488037,
+      "learning_rate": 3.977406547644513e-07,
+      "loss": 1.0773,
+      "step": 4480
+    },
+    {
+      "epoch": 9.107668867445403,
+      "grad_norm": 2.405012845993042,
+      "learning_rate": 3.800999938929606e-07,
+      "loss": 1.0851,
+      "step": 4490
+    },
+    {
+      "epoch": 9.12798374809548,
+      "grad_norm": 2.8481688499450684,
+      "learning_rate": 3.628518801253311e-07,
+      "loss": 1.0486,
+      "step": 4500
+    },
+    {
+      "epoch": 9.148298628745556,
+      "grad_norm": 2.405825138092041,
+      "learning_rate": 3.4599701728354474e-07,
+      "loss": 1.0776,
+      "step": 4510
+    },
+    {
+      "epoch": 9.168613509395632,
+      "grad_norm": 2.5540664196014404,
+      "learning_rate": 3.2953609314269163e-07,
+      "loss": 1.0623,
+      "step": 4520
+    },
+    {
+      "epoch": 9.188928390045708,
+      "grad_norm": 2.9651260375976562,
+      "learning_rate": 3.1346977940290404e-07,
+      "loss": 1.0583,
+      "step": 4530
+    },
+    {
+      "epoch": 9.209243270695785,
+      "grad_norm": 2.4424266815185547,
+      "learning_rate": 2.9779873166194686e-07,
+      "loss": 1.0869,
+      "step": 4540
+    },
+    {
+      "epoch": 9.229558151345861,
+      "grad_norm": 2.3093559741973877,
+      "learning_rate": 2.8252358938846145e-07,
+      "loss": 1.0667,
+      "step": 4550
+    },
+    {
+      "epoch": 9.249873031995937,
+      "grad_norm": 2.28373646736145,
+      "learning_rate": 2.6764497589587544e-07,
+      "loss": 1.0634,
+      "step": 4560
+    },
+    {
+      "epoch": 9.270187912646014,
+      "grad_norm": 2.4830288887023926,
+      "learning_rate": 2.5316349831697175e-07,
+      "loss": 1.0964,
+      "step": 4570
+    },
+    {
+      "epoch": 9.29050279329609,
+      "grad_norm": 2.0313925743103027,
+      "learning_rate": 2.3907974757910735e-07,
+      "loss": 1.0625,
+      "step": 4580
+    },
+    {
+      "epoch": 9.310817673946165,
+      "grad_norm": 2.334379196166992,
+      "learning_rate": 2.253942983800994e-07,
+      "loss": 1.0654,
+      "step": 4590
+    },
+    {
+      "epoch": 9.331132554596241,
+      "grad_norm": 2.6533384323120117,
+      "learning_rate": 2.1210770916478384e-07,
+      "loss": 1.0769,
+      "step": 4600
+    },
+    {
+      "epoch": 9.351447435246317,
+      "grad_norm": 2.715524435043335,
+      "learning_rate": 1.9922052210221477e-07,
+      "loss": 1.0784,
+      "step": 4610
+    },
+    {
+      "epoch": 9.371762315896394,
+      "grad_norm": 2.238746404647827,
+      "learning_rate": 1.8673326306355432e-07,
+      "loss": 1.0694,
+      "step": 4620
+    },
+    {
+      "epoch": 9.39207719654647,
+      "grad_norm": 2.3140628337860107,
+      "learning_rate": 1.7464644160059885e-07,
+      "loss": 1.0733,
+      "step": 4630
+    },
+    {
+      "epoch": 9.412392077196547,
+      "grad_norm": 2.2057929039001465,
+      "learning_rate": 1.6296055092500207e-07,
+      "loss": 1.0837,
+      "step": 4640
+    },
+    {
+      "epoch": 9.432706957846623,
+      "grad_norm": 2.212195873260498,
+      "learning_rate": 1.5167606788813794e-07,
+      "loss": 1.0748,
+      "step": 4650
+    },
+    {
+      "epoch": 9.4530218384967,
+      "grad_norm": 2.6805107593536377,
+      "learning_rate": 1.4079345296164504e-07,
+      "loss": 1.0793,
+      "step": 4660
+    },
+    {
+      "epoch": 9.473336719146776,
+      "grad_norm": 2.494554042816162,
+      "learning_rate": 1.3031315021864054e-07,
+      "loss": 1.0564,
+      "step": 4670
+    },
+    {
+      "epoch": 9.493651599796852,
+      "grad_norm": 2.6305577754974365,
+      "learning_rate": 1.2023558731559582e-07,
+      "loss": 1.0756,
+      "step": 4680
+    },
+    {
+      "epoch": 9.513966480446927,
+      "grad_norm": 2.3631041049957275,
+      "learning_rate": 1.1056117547488588e-07,
+      "loss": 1.0828,
+      "step": 4690
+    },
+    {
+      "epoch": 9.534281361097003,
+      "grad_norm": 2.3476672172546387,
+      "learning_rate": 1.0129030946801066e-07,
+      "loss": 1.0885,
+      "step": 4700
+    },
+    {
+      "epoch": 9.55459624174708,
+      "grad_norm": 2.513430595397949,
+      "learning_rate": 9.242336759948457e-08,
+      "loss": 1.0698,
+      "step": 4710
+    },
+    {
+      "epoch": 9.574911122397156,
+      "grad_norm": 2.3962695598602295,
+      "learning_rate": 8.396071169139986e-08,
+      "loss": 1.082,
+      "step": 4720
+    },
+    {
+      "epoch": 9.595226003047232,
+      "grad_norm": 2.5314340591430664,
+      "learning_rate": 7.590268706866188e-08,
+      "loss": 1.075,
+      "step": 4730
+    },
+    {
+      "epoch": 9.615540883697308,
+      "grad_norm": 2.3048431873321533,
+      "learning_rate": 6.824962254490141e-08,
+      "loss": 1.0527,
+      "step": 4740
+    },
+    {
+      "epoch": 9.635855764347385,
+      "grad_norm": 2.2387807369232178,
+      "learning_rate": 6.100183040905095e-08,
+      "loss": 1.0573,
+      "step": 4750
+    },
+    {
+      "epoch": 9.656170644997461,
+      "grad_norm": 2.7502102851867676,
+      "learning_rate": 5.415960641260487e-08,
+      "loss": 1.0662,
+      "step": 4760
+    },
+    {
+      "epoch": 9.676485525647538,
+      "grad_norm": 2.980741024017334,
+      "learning_rate": 4.772322975755361e-08,
+      "loss": 1.099,
+      "step": 4770
+    },
+    {
+      "epoch": 9.696800406297612,
+      "grad_norm": 2.4962992668151855,
+      "learning_rate": 4.169296308498494e-08,
+      "loss": 1.065,
+      "step": 4780
+    },
+    {
+      "epoch": 9.717115286947688,
+      "grad_norm": 2.1819846630096436,
+      "learning_rate": 3.606905246437253e-08,
+      "loss": 1.0641,
+      "step": 4790
+    },
+    {
+      "epoch": 9.737430167597765,
+      "grad_norm": 2.336205244064331,
+      "learning_rate": 3.0851727383529596e-08,
+      "loss": 1.052,
+      "step": 4800
+    },
+    {
+      "epoch": 9.757745048247841,
+      "grad_norm": 2.4870941638946533,
+      "learning_rate": 2.604120073924965e-08,
+      "loss": 1.0729,
+      "step": 4810
+    },
+    {
+      "epoch": 9.778059928897918,
+      "grad_norm": 2.8149468898773193,
+      "learning_rate": 2.16376688286124e-08,
+      "loss": 1.0905,
+      "step": 4820
+    },
+    {
+      "epoch": 9.798374809547994,
+      "grad_norm": 2.3567795753479004,
+      "learning_rate": 1.7641311340982346e-08,
+      "loss": 1.0628,
+      "step": 4830
+    },
+    {
+      "epoch": 9.81868969019807,
+      "grad_norm": 2.4957797527313232,
+      "learning_rate": 1.40522913506691e-08,
+      "loss": 1.071,
+      "step": 4840
+    },
+    {
+      "epoch": 9.839004570848147,
+      "grad_norm": 2.4365956783294678,
+      "learning_rate": 1.0870755310276038e-08,
+      "loss": 1.0788,
+      "step": 4850
+    },
+    {
+      "epoch": 9.859319451498223,
+      "grad_norm": 2.4922902584075928,
+      "learning_rate": 8.096833044723973e-09,
+      "loss": 1.0908,
+      "step": 4860
+    },
+    {
+      "epoch": 9.8796343321483,
+      "grad_norm": 2.4978227615356445,
+      "learning_rate": 5.730637745954282e-09,
+      "loss": 1.0725,
+      "step": 4870
+    },
+    {
+      "epoch": 9.899949212798376,
+      "grad_norm": 2.3648996353149414,
+      "learning_rate": 3.772265968309263e-09,
+      "loss": 1.0646,
+      "step": 4880
+    },
+    {
+      "epoch": 9.92026409344845,
+      "grad_norm": 2.4374351501464844,
+      "learning_rate": 2.2217976245886286e-09,
+      "loss": 1.0571,
+      "step": 4890
+    },
+    {
+      "epoch": 9.940578974098527,
+      "grad_norm": 2.5737831592559814,
+      "learning_rate": 1.0792959827987671e-09,
+      "loss": 1.0577,
+      "step": 4900
+    },
+    {
+      "epoch": 9.960893854748603,
+      "grad_norm": 2.2821946144104004,
+      "learning_rate": 3.448076635570452e-10,
+      "loss": 1.0729,
+      "step": 4910
+    },
+    {
+      "epoch": 9.98120873539868,
+      "grad_norm": 2.4394965171813965,
+      "learning_rate": 1.8362638203317517e-11,
+      "loss": 1.068,
+      "step": 4920
+    },
+    {
+      "epoch": 9.98120873539868,
+      "eval_loss": 1.0894742012023926,
+      "eval_runtime": 55.0248,
+      "eval_samples_per_second": 35.802,
+      "eval_steps_per_second": 8.96,
+      "step": 4920
+    },
+    {
+      "epoch": 9.98120873539868,
+      "step": 4920,
+      "total_flos": 2.0218089447773184e+16,
+      "train_loss": 1.1380672057469685,
+      "train_runtime": 8182.6752,
+      "train_samples_per_second": 9.625,
+      "train_steps_per_second": 0.601
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 4920,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.0218089447773184e+16,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c208e4f8d60f42ed25b5bf8aec16cdad2d0c051236c96e2f5432c19844229f8e
+size 5816