Model save

.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 +1,58 @@
+---
+base_model: Qwen/Qwen2-VL-2B-Instruct
+library_name: transformers
+model_name: Qwen2-VL-2B-Instruct-SFT-imagenet_fewshot
+tags:
+- generated_from_trainer
+- trl
+- sft
+licence: license
+---
+
+# Model Card for Qwen2-VL-2B-Instruct-SFT-imagenet_fewshot
+
+This model is a fine-tuned version of [Qwen/Qwen2-VL-2B-Instruct](https://huggingface.co/Qwen/Qwen2-VL-2B-Instruct).
+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="afdsafas/Qwen2-VL-2B-Instruct-SFT-imagenet_fewshot", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/lm1640362161-university-of-tokyo/huggingface/runs/3jzluur4) 
+
+
+This model was trained with SFT.
+
+### Framework versions
+
+- TRL: 0.14.0
+- Transformers: 4.49.0.dev0
+- Pytorch: 2.5.1
+- Datasets: 3.3.0
+- 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}}
+}
+```

added_tokens.json

@@ -0,0 +1,16 @@
+{
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

all_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 10.0,
+    "total_flos": 2.880737320506491e+17,
+    "train_loss": 1.3902658111572266,
+    "train_runtime": 2102.4841,
+    "train_samples": 4000,
+    "train_samples_per_second": 19.025,
+    "train_steps_per_second": 1.189
+}

chat_template.json

@@ -0,0 +1,3 @@
+{
+  "chat_template": "{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n{% endif %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}<|im_end|>\n{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>\n{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}"
+}

config.json

@@ -0,0 +1,49 @@
+{
+  "_name_or_path": "Qwen/Qwen2-VL-2B-Instruct",
+  "architectures": [
+    "Qwen2VLForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "image_token_id": 151655,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "qwen2_vl",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": {
+    "mrope_section": [
+      16,
+      24,
+      24
+    ],
+    "rope_type": "default",
+    "type": "default"
+  },
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.49.0.dev0",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "video_token_id": 151656,
+  "vision_config": {
+    "hidden_size": 1536,
+    "in_chans": 3,
+    "model_type": "qwen2_vl",
+    "spatial_patch_size": 14,
+    "torch_dtype": "bfloat16"
+  },
+  "vision_end_token_id": 151653,
+  "vision_start_token_id": 151652,
+  "vision_token_id": 151654,
+  "vocab_size": 151936
+}

generation_config.json

@@ -0,0 +1,15 @@
+{
+  "attn_implementation": null,
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "temperature": 0.01,
+  "top_k": 1,
+  "top_p": 0.001,
+  "transformers_version": "4.49.0.dev0",
+  "use_cache": false
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:648347e5619c5aaa184a721ec36b9209790b1c2ead64a547d51046cefddd4427
+size 4884798456

preprocessor_config.json

@@ -0,0 +1,29 @@
+{
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "Qwen2VLImageProcessor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "max_pixels": 12845056,
+  "merge_size": 2,
+  "min_pixels": 3136,
+  "patch_size": 14,
+  "processor_class": "Qwen2VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "longest_edge": 12845056,
+    "shortest_edge": 3136
+  },
+  "temporal_patch_size": 2
+}

special_tokens_map.json

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

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88a3a6fcb80132f76da8aa40cdc3fccd7e5d8468ef15421f5b0c2715e85217d2
+size 11420538

tokenizer_config.json

@@ -0,0 +1,144 @@
+{
+  "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
+    }
+  },
+  "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": "{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n{% endif %}<|im_start|>{{ message['role'] }}\n{% if message['content'] is string %}{{ message['content'] }}<|im_end|>\n{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>\n{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant\n{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 10.0,
+    "total_flos": 2.880737320506491e+17,
+    "train_loss": 1.3902658111572266,
+    "train_runtime": 2102.4841,
+    "train_samples": 4000,
+    "train_samples_per_second": 19.025,
+    "train_steps_per_second": 1.189
+}

trainer_state.json

@@ -0,0 +1,3542 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 10.0,
+  "eval_steps": 500,
+  "global_step": 2500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.02,
+      "grad_norm": 30.21004867553711,
+      "learning_rate": 4.0000000000000003e-07,
+      "loss": 3.7961,
+      "step": 5
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 27.698633193969727,
+      "learning_rate": 8.000000000000001e-07,
+      "loss": 3.7529,
+      "step": 10
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 17.014102935791016,
+      "learning_rate": 1.2000000000000002e-06,
+      "loss": 3.6348,
+      "step": 15
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 12.189393997192383,
+      "learning_rate": 1.6000000000000001e-06,
+      "loss": 3.3817,
+      "step": 20
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 9.264037132263184,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 3.0922,
+      "step": 25
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 6.129642963409424,
+      "learning_rate": 2.4000000000000003e-06,
+      "loss": 2.8821,
+      "step": 30
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 4.392693996429443,
+      "learning_rate": 2.8000000000000003e-06,
+      "loss": 2.6029,
+      "step": 35
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 6.556579113006592,
+      "learning_rate": 3.2000000000000003e-06,
+      "loss": 2.4349,
+      "step": 40
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 3.495549440383911,
+      "learning_rate": 3.6000000000000003e-06,
+      "loss": 2.2218,
+      "step": 45
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 3.5002224445343018,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 2.0837,
+      "step": 50
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 3.8143324851989746,
+      "learning_rate": 4.4e-06,
+      "loss": 2.0036,
+      "step": 55
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 3.2101738452911377,
+      "learning_rate": 4.800000000000001e-06,
+      "loss": 1.9349,
+      "step": 60
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 2.8242392539978027,
+      "learning_rate": 5.2e-06,
+      "loss": 1.8964,
+      "step": 65
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 2.515568494796753,
+      "learning_rate": 5.600000000000001e-06,
+      "loss": 1.8728,
+      "step": 70
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 2.872358798980713,
+      "learning_rate": 6e-06,
+      "loss": 1.8441,
+      "step": 75
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 3.117579936981201,
+      "learning_rate": 6.4000000000000006e-06,
+      "loss": 1.8218,
+      "step": 80
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 2.6006624698638916,
+      "learning_rate": 6.800000000000001e-06,
+      "loss": 1.809,
+      "step": 85
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 2.7116575241088867,
+      "learning_rate": 7.2000000000000005e-06,
+      "loss": 1.7963,
+      "step": 90
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 2.297466278076172,
+      "learning_rate": 7.600000000000001e-06,
+      "loss": 1.7763,
+      "step": 95
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 2.5194053649902344,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 1.772,
+      "step": 100
+    },
+    {
+      "epoch": 0.42,
+      "grad_norm": 2.4360179901123047,
+      "learning_rate": 8.400000000000001e-06,
+      "loss": 1.7559,
+      "step": 105
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 2.224982261657715,
+      "learning_rate": 8.8e-06,
+      "loss": 1.7485,
+      "step": 110
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 2.297065258026123,
+      "learning_rate": 9.200000000000002e-06,
+      "loss": 1.7436,
+      "step": 115
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 2.390683889389038,
+      "learning_rate": 9.600000000000001e-06,
+      "loss": 1.7416,
+      "step": 120
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 2.2673444747924805,
+      "learning_rate": 1e-05,
+      "loss": 1.7449,
+      "step": 125
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 2.167109251022339,
+      "learning_rate": 1.04e-05,
+      "loss": 1.7413,
+      "step": 130
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 2.0062167644500732,
+      "learning_rate": 1.0800000000000002e-05,
+      "loss": 1.7111,
+      "step": 135
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 2.1711270809173584,
+      "learning_rate": 1.1200000000000001e-05,
+      "loss": 1.7177,
+      "step": 140
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 2.027642011642456,
+      "learning_rate": 1.16e-05,
+      "loss": 1.7261,
+      "step": 145
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 1.953282356262207,
+      "learning_rate": 1.2e-05,
+      "loss": 1.7124,
+      "step": 150
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 2.0645740032196045,
+      "learning_rate": 1.2400000000000002e-05,
+      "loss": 1.7249,
+      "step": 155
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 2.904872179031372,
+      "learning_rate": 1.2800000000000001e-05,
+      "loss": 1.7226,
+      "step": 160
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 2.2629189491271973,
+      "learning_rate": 1.3200000000000002e-05,
+      "loss": 1.7156,
+      "step": 165
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 2.0626604557037354,
+      "learning_rate": 1.3600000000000002e-05,
+      "loss": 1.7285,
+      "step": 170
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 2.1120095252990723,
+      "learning_rate": 1.4e-05,
+      "loss": 1.729,
+      "step": 175
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 2.1083803176879883,
+      "learning_rate": 1.4400000000000001e-05,
+      "loss": 1.7186,
+      "step": 180
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 2.4473488330841064,
+      "learning_rate": 1.48e-05,
+      "loss": 1.7169,
+      "step": 185
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 3.1553521156311035,
+      "learning_rate": 1.5200000000000002e-05,
+      "loss": 1.7141,
+      "step": 190
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 1.8050827980041504,
+      "learning_rate": 1.5600000000000003e-05,
+      "loss": 1.7211,
+      "step": 195
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 1.81538987159729,
+      "learning_rate": 1.6000000000000003e-05,
+      "loss": 1.718,
+      "step": 200
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 1.8642960786819458,
+      "learning_rate": 1.64e-05,
+      "loss": 1.7249,
+      "step": 205
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 1.9567952156066895,
+      "learning_rate": 1.6800000000000002e-05,
+      "loss": 1.7257,
+      "step": 210
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 2.035290002822876,
+      "learning_rate": 1.72e-05,
+      "loss": 1.7328,
+      "step": 215
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 1.8088799715042114,
+      "learning_rate": 1.76e-05,
+      "loss": 1.7182,
+      "step": 220
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 1.9366027116775513,
+      "learning_rate": 1.8e-05,
+      "loss": 1.7202,
+      "step": 225
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 1.8107857704162598,
+      "learning_rate": 1.8400000000000003e-05,
+      "loss": 1.7163,
+      "step": 230
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 1.8145129680633545,
+      "learning_rate": 1.88e-05,
+      "loss": 1.7214,
+      "step": 235
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 2.3850910663604736,
+      "learning_rate": 1.9200000000000003e-05,
+      "loss": 1.7193,
+      "step": 240
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 2.1443727016448975,
+      "learning_rate": 1.9600000000000002e-05,
+      "loss": 1.7152,
+      "step": 245
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 1.813616156578064,
+      "learning_rate": 2e-05,
+      "loss": 1.7176,
+      "step": 250
+    },
+    {
+      "epoch": 1.02,
+      "grad_norm": 1.8890033960342407,
+      "learning_rate": 1.9999756307053947e-05,
+      "loss": 1.7146,
+      "step": 255
+    },
+    {
+      "epoch": 1.04,
+      "grad_norm": 1.7228977680206299,
+      "learning_rate": 1.9999025240093045e-05,
+      "loss": 1.709,
+      "step": 260
+    },
+    {
+      "epoch": 1.06,
+      "grad_norm": 1.7775235176086426,
+      "learning_rate": 1.9997806834748455e-05,
+      "loss": 1.7132,
+      "step": 265
+    },
+    {
+      "epoch": 1.08,
+      "grad_norm": 1.6786800622940063,
+      "learning_rate": 1.9996101150403543e-05,
+      "loss": 1.7058,
+      "step": 270
+    },
+    {
+      "epoch": 1.1,
+      "grad_norm": 3.5919995307922363,
+      "learning_rate": 1.999390827019096e-05,
+      "loss": 1.7135,
+      "step": 275
+    },
+    {
+      "epoch": 1.12,
+      "grad_norm": 1.6665600538253784,
+      "learning_rate": 1.9991228300988586e-05,
+      "loss": 1.7138,
+      "step": 280
+    },
+    {
+      "epoch": 1.1400000000000001,
+      "grad_norm": 1.6498384475708008,
+      "learning_rate": 1.9988061373414342e-05,
+      "loss": 1.71,
+      "step": 285
+    },
+    {
+      "epoch": 1.16,
+      "grad_norm": 1.7032837867736816,
+      "learning_rate": 1.9984407641819812e-05,
+      "loss": 1.7103,
+      "step": 290
+    },
+    {
+      "epoch": 1.18,
+      "grad_norm": 1.7445223331451416,
+      "learning_rate": 1.9980267284282718e-05,
+      "loss": 1.7022,
+      "step": 295
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 1.8121885061264038,
+      "learning_rate": 1.9975640502598243e-05,
+      "loss": 1.7238,
+      "step": 300
+    },
+    {
+      "epoch": 1.22,
+      "grad_norm": 2.3573601245880127,
+      "learning_rate": 1.9970527522269204e-05,
+      "loss": 1.7069,
+      "step": 305
+    },
+    {
+      "epoch": 1.24,
+      "grad_norm": 1.790433406829834,
+      "learning_rate": 1.9964928592495046e-05,
+      "loss": 1.7151,
+      "step": 310
+    },
+    {
+      "epoch": 1.26,
+      "grad_norm": 1.7757713794708252,
+      "learning_rate": 1.9958843986159705e-05,
+      "loss": 1.7044,
+      "step": 315
+    },
+    {
+      "epoch": 1.28,
+      "grad_norm": 1.4678212404251099,
+      "learning_rate": 1.9952273999818312e-05,
+      "loss": 1.7023,
+      "step": 320
+    },
+    {
+      "epoch": 1.3,
+      "grad_norm": 1.7063349485397339,
+      "learning_rate": 1.9945218953682736e-05,
+      "loss": 1.7132,
+      "step": 325
+    },
+    {
+      "epoch": 1.32,
+      "grad_norm": 1.537674903869629,
+      "learning_rate": 1.9937679191605964e-05,
+      "loss": 1.7071,
+      "step": 330
+    },
+    {
+      "epoch": 1.34,
+      "grad_norm": 1.6040879487991333,
+      "learning_rate": 1.992965508106537e-05,
+      "loss": 1.7004,
+      "step": 335
+    },
+    {
+      "epoch": 1.3599999999999999,
+      "grad_norm": 1.7427316904067993,
+      "learning_rate": 1.9921147013144782e-05,
+      "loss": 1.7006,
+      "step": 340
+    },
+    {
+      "epoch": 1.38,
+      "grad_norm": 2.1476292610168457,
+      "learning_rate": 1.991215540251542e-05,
+      "loss": 1.7059,
+      "step": 345
+    },
+    {
+      "epoch": 1.4,
+      "grad_norm": 1.5894598960876465,
+      "learning_rate": 1.9902680687415704e-05,
+      "loss": 1.7066,
+      "step": 350
+    },
+    {
+      "epoch": 1.42,
+      "grad_norm": 1.5307775735855103,
+      "learning_rate": 1.9892723329629885e-05,
+      "loss": 1.7104,
+      "step": 355
+    },
+    {
+      "epoch": 1.44,
+      "grad_norm": 1.502901315689087,
+      "learning_rate": 1.988228381446553e-05,
+      "loss": 1.7045,
+      "step": 360
+    },
+    {
+      "epoch": 1.46,
+      "grad_norm": 1.6482329368591309,
+      "learning_rate": 1.987136265072988e-05,
+      "loss": 1.7235,
+      "step": 365
+    },
+    {
+      "epoch": 1.48,
+      "grad_norm": 1.5000663995742798,
+      "learning_rate": 1.985996037070505e-05,
+      "loss": 1.7026,
+      "step": 370
+    },
+    {
+      "epoch": 1.5,
+      "grad_norm": 1.5986205339431763,
+      "learning_rate": 1.9848077530122083e-05,
+      "loss": 1.7108,
+      "step": 375
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 2.0925049781799316,
+      "learning_rate": 1.983571470813386e-05,
+      "loss": 1.7049,
+      "step": 380
+    },
+    {
+      "epoch": 1.54,
+      "grad_norm": 1.4937970638275146,
+      "learning_rate": 1.982287250728689e-05,
+      "loss": 1.7075,
+      "step": 385
+    },
+    {
+      "epoch": 1.56,
+      "grad_norm": 1.7145628929138184,
+      "learning_rate": 1.9809551553491918e-05,
+      "loss": 1.704,
+      "step": 390
+    },
+    {
+      "epoch": 1.58,
+      "grad_norm": 1.4836574792861938,
+      "learning_rate": 1.979575249599344e-05,
+      "loss": 1.6981,
+      "step": 395
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 1.8676332235336304,
+      "learning_rate": 1.9781476007338058e-05,
+      "loss": 1.7138,
+      "step": 400
+    },
+    {
+      "epoch": 1.62,
+      "grad_norm": 1.549514651298523,
+      "learning_rate": 1.9766722783341682e-05,
+      "loss": 1.7077,
+      "step": 405
+    },
+    {
+      "epoch": 1.6400000000000001,
+      "grad_norm": 1.6212232112884521,
+      "learning_rate": 1.9751493543055634e-05,
+      "loss": 1.7103,
+      "step": 410
+    },
+    {
+      "epoch": 1.6600000000000001,
+      "grad_norm": 1.6188278198242188,
+      "learning_rate": 1.9735789028731603e-05,
+      "loss": 1.7064,
+      "step": 415
+    },
+    {
+      "epoch": 1.6800000000000002,
+      "grad_norm": 1.6380282640457153,
+      "learning_rate": 1.9719610005785466e-05,
+      "loss": 1.6988,
+      "step": 420
+    },
+    {
+      "epoch": 1.7,
+      "grad_norm": 1.4315824508666992,
+      "learning_rate": 1.9702957262759964e-05,
+      "loss": 1.7076,
+      "step": 425
+    },
+    {
+      "epoch": 1.72,
+      "grad_norm": 1.497166395187378,
+      "learning_rate": 1.9685831611286312e-05,
+      "loss": 1.6965,
+      "step": 430
+    },
+    {
+      "epoch": 1.74,
+      "grad_norm": 1.5363743305206299,
+      "learning_rate": 1.9668233886044597e-05,
+      "loss": 1.7073,
+      "step": 435
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 1.4399864673614502,
+      "learning_rate": 1.9650164944723116e-05,
+      "loss": 1.7028,
+      "step": 440
+    },
+    {
+      "epoch": 1.78,
+      "grad_norm": 1.5128633975982666,
+      "learning_rate": 1.9631625667976584e-05,
+      "loss": 1.7078,
+      "step": 445
+    },
+    {
+      "epoch": 1.8,
+      "grad_norm": 1.4235734939575195,
+      "learning_rate": 1.961261695938319e-05,
+      "loss": 1.7093,
+      "step": 450
+    },
+    {
+      "epoch": 1.8199999999999998,
+      "grad_norm": 1.4514182806015015,
+      "learning_rate": 1.9593139745400575e-05,
+      "loss": 1.7059,
+      "step": 455
+    },
+    {
+      "epoch": 1.8399999999999999,
+      "grad_norm": 1.360684871673584,
+      "learning_rate": 1.9573194975320672e-05,
+      "loss": 1.7017,
+      "step": 460
+    },
+    {
+      "epoch": 1.8599999999999999,
+      "grad_norm": 1.4784973859786987,
+      "learning_rate": 1.9552783621223437e-05,
+      "loss": 1.7054,
+      "step": 465
+    },
+    {
+      "epoch": 1.88,
+      "grad_norm": 1.4710580110549927,
+      "learning_rate": 1.9531906677929472e-05,
+      "loss": 1.707,
+      "step": 470
+    },
+    {
+      "epoch": 1.9,
+      "grad_norm": 1.3699547052383423,
+      "learning_rate": 1.9510565162951538e-05,
+      "loss": 1.7047,
+      "step": 475
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 1.395521879196167,
+      "learning_rate": 1.9488760116444966e-05,
+      "loss": 1.7121,
+      "step": 480
+    },
+    {
+      "epoch": 1.94,
+      "grad_norm": 1.368019938468933,
+      "learning_rate": 1.9466492601156964e-05,
+      "loss": 1.7013,
+      "step": 485
+    },
+    {
+      "epoch": 1.96,
+      "grad_norm": 1.7290316820144653,
+      "learning_rate": 1.944376370237481e-05,
+      "loss": 1.7042,
+      "step": 490
+    },
+    {
+      "epoch": 1.98,
+      "grad_norm": 1.4156850576400757,
+      "learning_rate": 1.942057452787297e-05,
+      "loss": 1.7041,
+      "step": 495
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.374906063079834,
+      "learning_rate": 1.9396926207859085e-05,
+      "loss": 1.6999,
+      "step": 500
+    },
+    {
+      "epoch": 2.02,
+      "grad_norm": 1.431320309638977,
+      "learning_rate": 1.937281989491892e-05,
+      "loss": 1.7057,
+      "step": 505
+    },
+    {
+      "epoch": 2.04,
+      "grad_norm": 1.4097741842269897,
+      "learning_rate": 1.9348256763960146e-05,
+      "loss": 1.6996,
+      "step": 510
+    },
+    {
+      "epoch": 2.06,
+      "grad_norm": 1.4540244340896606,
+      "learning_rate": 1.9323238012155125e-05,
+      "loss": 1.7065,
+      "step": 515
+    },
+    {
+      "epoch": 2.08,
+      "grad_norm": 1.3888367414474487,
+      "learning_rate": 1.9297764858882516e-05,
+      "loss": 1.6991,
+      "step": 520
+    },
+    {
+      "epoch": 2.1,
+      "grad_norm": 1.3523085117340088,
+      "learning_rate": 1.9271838545667876e-05,
+      "loss": 1.7039,
+      "step": 525
+    },
+    {
+      "epoch": 2.12,
+      "grad_norm": 1.4312589168548584,
+      "learning_rate": 1.9245460336123136e-05,
+      "loss": 1.6876,
+      "step": 530
+    },
+    {
+      "epoch": 2.14,
+      "grad_norm": 1.4729927778244019,
+      "learning_rate": 1.9218631515885007e-05,
+      "loss": 1.7074,
+      "step": 535
+    },
+    {
+      "epoch": 2.16,
+      "grad_norm": 1.3183554410934448,
+      "learning_rate": 1.9191353392552346e-05,
+      "loss": 1.7018,
+      "step": 540
+    },
+    {
+      "epoch": 2.18,
+      "grad_norm": 1.4427063465118408,
+      "learning_rate": 1.9163627295622397e-05,
+      "loss": 1.6896,
+      "step": 545
+    },
+    {
+      "epoch": 2.2,
+      "grad_norm": 1.4542627334594727,
+      "learning_rate": 1.913545457642601e-05,
+      "loss": 1.6937,
+      "step": 550
+    },
+    {
+      "epoch": 2.22,
+      "grad_norm": 1.3560303449630737,
+      "learning_rate": 1.910683660806177e-05,
+      "loss": 1.7066,
+      "step": 555
+    },
+    {
+      "epoch": 2.24,
+      "grad_norm": 1.6534428596496582,
+      "learning_rate": 1.907777478532909e-05,
+      "loss": 1.6913,
+      "step": 560
+    },
+    {
+      "epoch": 2.26,
+      "grad_norm": 1.334754467010498,
+      "learning_rate": 1.9048270524660197e-05,
+      "loss": 1.7045,
+      "step": 565
+    },
+    {
+      "epoch": 2.2800000000000002,
+      "grad_norm": 1.367875576019287,
+      "learning_rate": 1.901832526405114e-05,
+      "loss": 1.6936,
+      "step": 570
+    },
+    {
+      "epoch": 2.3,
+      "grad_norm": 1.3227001428604126,
+      "learning_rate": 1.8987940462991673e-05,
+      "loss": 1.6991,
+      "step": 575
+    },
+    {
+      "epoch": 2.32,
+      "grad_norm": 1.2901933193206787,
+      "learning_rate": 1.895711760239413e-05,
+      "loss": 1.696,
+      "step": 580
+    },
+    {
+      "epoch": 2.34,
+      "grad_norm": 1.33935546875,
+      "learning_rate": 1.892585818452126e-05,
+      "loss": 1.7008,
+      "step": 585
+    },
+    {
+      "epoch": 2.36,
+      "grad_norm": 1.487889051437378,
+      "learning_rate": 1.889416373291298e-05,
+      "loss": 1.6883,
+      "step": 590
+    },
+    {
+      "epoch": 2.38,
+      "grad_norm": 1.3563237190246582,
+      "learning_rate": 1.8862035792312148e-05,
+      "loss": 1.6967,
+      "step": 595
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 1.2794183492660522,
+      "learning_rate": 1.8829475928589272e-05,
+      "loss": 1.6922,
+      "step": 600
+    },
+    {
+      "epoch": 2.42,
+      "grad_norm": 1.4755550622940063,
+      "learning_rate": 1.879648572866617e-05,
+      "loss": 1.7054,
+      "step": 605
+    },
+    {
+      "epoch": 2.44,
+      "grad_norm": 1.307241678237915,
+      "learning_rate": 1.8763066800438638e-05,
+      "loss": 1.6897,
+      "step": 610
+    },
+    {
+      "epoch": 2.46,
+      "grad_norm": 1.4721298217773438,
+      "learning_rate": 1.8729220772698096e-05,
+      "loss": 1.6969,
+      "step": 615
+    },
+    {
+      "epoch": 2.48,
+      "grad_norm": 1.8397995233535767,
+      "learning_rate": 1.869494929505219e-05,
+      "loss": 1.697,
+      "step": 620
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 1.2730300426483154,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 1.6994,
+      "step": 625
+    },
+    {
+      "epoch": 2.52,
+      "grad_norm": 1.252538800239563,
+      "learning_rate": 1.8625136692072577e-05,
+      "loss": 1.7046,
+      "step": 630
+    },
+    {
+      "epoch": 2.54,
+      "grad_norm": 1.2064297199249268,
+      "learning_rate": 1.8589598969306646e-05,
+      "loss": 1.6935,
+      "step": 635
+    },
+    {
+      "epoch": 2.56,
+      "grad_norm": 1.2591032981872559,
+      "learning_rate": 1.855364260160507e-05,
+      "loss": 1.7,
+      "step": 640
+    },
+    {
+      "epoch": 2.58,
+      "grad_norm": 1.380873203277588,
+      "learning_rate": 1.851726934143048e-05,
+      "loss": 1.6952,
+      "step": 645
+    },
+    {
+      "epoch": 2.6,
+      "grad_norm": 1.301558256149292,
+      "learning_rate": 1.848048096156426e-05,
+      "loss": 1.7029,
+      "step": 650
+    },
+    {
+      "epoch": 2.62,
+      "grad_norm": 1.3414067029953003,
+      "learning_rate": 1.8443279255020153e-05,
+      "loss": 1.6973,
+      "step": 655
+    },
+    {
+      "epoch": 2.64,
+      "grad_norm": 1.2605617046356201,
+      "learning_rate": 1.8405666034956842e-05,
+      "loss": 1.6883,
+      "step": 660
+    },
+    {
+      "epoch": 2.66,
+      "grad_norm": 1.2734073400497437,
+      "learning_rate": 1.836764313458962e-05,
+      "loss": 1.6925,
+      "step": 665
+    },
+    {
+      "epoch": 2.68,
+      "grad_norm": 1.3956694602966309,
+      "learning_rate": 1.8329212407100996e-05,
+      "loss": 1.7057,
+      "step": 670
+    },
+    {
+      "epoch": 2.7,
+      "grad_norm": 1.463300108909607,
+      "learning_rate": 1.8290375725550417e-05,
+      "loss": 1.695,
+      "step": 675
+    },
+    {
+      "epoch": 2.7199999999999998,
+      "grad_norm": 1.276495099067688,
+      "learning_rate": 1.8251134982782952e-05,
+      "loss": 1.7108,
+      "step": 680
+    },
+    {
+      "epoch": 2.74,
+      "grad_norm": 1.205746054649353,
+      "learning_rate": 1.821149209133704e-05,
+      "loss": 1.6987,
+      "step": 685
+    },
+    {
+      "epoch": 2.76,
+      "grad_norm": 1.5678800344467163,
+      "learning_rate": 1.8171448983351284e-05,
+      "loss": 1.6933,
+      "step": 690
+    },
+    {
+      "epoch": 2.7800000000000002,
+      "grad_norm": 1.2740401029586792,
+      "learning_rate": 1.8131007610470278e-05,
+      "loss": 1.7092,
+      "step": 695
+    },
+    {
+      "epoch": 2.8,
+      "grad_norm": 1.2071486711502075,
+      "learning_rate": 1.8090169943749477e-05,
+      "loss": 1.6895,
+      "step": 700
+    },
+    {
+      "epoch": 2.82,
+      "grad_norm": 1.2892779111862183,
+      "learning_rate": 1.804893797355914e-05,
+      "loss": 1.7124,
+      "step": 705
+    },
+    {
+      "epoch": 2.84,
+      "grad_norm": 1.339904546737671,
+      "learning_rate": 1.8007313709487334e-05,
+      "loss": 1.7018,
+      "step": 710
+    },
+    {
+      "epoch": 2.86,
+      "grad_norm": 1.2110613584518433,
+      "learning_rate": 1.7965299180241963e-05,
+      "loss": 1.6975,
+      "step": 715
+    },
+    {
+      "epoch": 2.88,
+      "grad_norm": 1.2822394371032715,
+      "learning_rate": 1.792289643355191e-05,
+      "loss": 1.7022,
+      "step": 720
+    },
+    {
+      "epoch": 2.9,
+      "grad_norm": 1.3015908002853394,
+      "learning_rate": 1.788010753606722e-05,
+      "loss": 1.6998,
+      "step": 725
+    },
+    {
+      "epoch": 2.92,
+      "grad_norm": 1.2471498250961304,
+      "learning_rate": 1.78369345732584e-05,
+      "loss": 1.6939,
+      "step": 730
+    },
+    {
+      "epoch": 2.94,
+      "grad_norm": 1.5627294778823853,
+      "learning_rate": 1.7793379649314743e-05,
+      "loss": 1.6913,
+      "step": 735
+    },
+    {
+      "epoch": 2.96,
+      "grad_norm": 1.384990930557251,
+      "learning_rate": 1.7749444887041797e-05,
+      "loss": 1.6877,
+      "step": 740
+    },
+    {
+      "epoch": 2.98,
+      "grad_norm": 1.2845375537872314,
+      "learning_rate": 1.7705132427757895e-05,
+      "loss": 1.6985,
+      "step": 745
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.3291269540786743,
+      "learning_rate": 1.766044443118978e-05,
+      "loss": 1.6925,
+      "step": 750
+    },
+    {
+      "epoch": 3.02,
+      "grad_norm": 1.144539475440979,
+      "learning_rate": 1.761538307536737e-05,
+      "loss": 1.6943,
+      "step": 755
+    },
+    {
+      "epoch": 3.04,
+      "grad_norm": 1.256880521774292,
+      "learning_rate": 1.7569950556517566e-05,
+      "loss": 1.6949,
+      "step": 760
+    },
+    {
+      "epoch": 3.06,
+      "grad_norm": 1.3574849367141724,
+      "learning_rate": 1.7524149088957244e-05,
+      "loss": 1.6912,
+      "step": 765
+    },
+    {
+      "epoch": 3.08,
+      "grad_norm": 1.5494495630264282,
+      "learning_rate": 1.747798090498532e-05,
+      "loss": 1.6831,
+      "step": 770
+    },
+    {
+      "epoch": 3.1,
+      "grad_norm": 1.711756944656372,
+      "learning_rate": 1.7431448254773943e-05,
+      "loss": 1.689,
+      "step": 775
+    },
+    {
+      "epoch": 3.12,
+      "grad_norm": 1.3662241697311401,
+      "learning_rate": 1.7384553406258842e-05,
+      "loss": 1.6989,
+      "step": 780
+    },
+    {
+      "epoch": 3.14,
+      "grad_norm": 1.2295734882354736,
+      "learning_rate": 1.7337298645028764e-05,
+      "loss": 1.6812,
+      "step": 785
+    },
+    {
+      "epoch": 3.16,
+      "grad_norm": 1.1872354745864868,
+      "learning_rate": 1.7289686274214116e-05,
+      "loss": 1.6978,
+      "step": 790
+    },
+    {
+      "epoch": 3.18,
+      "grad_norm": 1.2291423082351685,
+      "learning_rate": 1.7241718614374678e-05,
+      "loss": 1.6937,
+      "step": 795
+    },
+    {
+      "epoch": 3.2,
+      "grad_norm": 1.1743723154067993,
+      "learning_rate": 1.7193398003386514e-05,
+      "loss": 1.6958,
+      "step": 800
+    },
+    {
+      "epoch": 3.22,
+      "grad_norm": 1.2758196592330933,
+      "learning_rate": 1.7144726796328034e-05,
+      "loss": 1.6886,
+      "step": 805
+    },
+    {
+      "epoch": 3.24,
+      "grad_norm": 1.2271584272384644,
+      "learning_rate": 1.709570736536521e-05,
+      "loss": 1.6896,
+      "step": 810
+    },
+    {
+      "epoch": 3.26,
+      "grad_norm": 1.4545894861221313,
+      "learning_rate": 1.7046342099635948e-05,
+      "loss": 1.6934,
+      "step": 815
+    },
+    {
+      "epoch": 3.2800000000000002,
+      "grad_norm": 1.218001365661621,
+      "learning_rate": 1.6996633405133656e-05,
+      "loss": 1.6984,
+      "step": 820
+    },
+    {
+      "epoch": 3.3,
+      "grad_norm": 1.7679849863052368,
+      "learning_rate": 1.6946583704589973e-05,
+      "loss": 1.6965,
+      "step": 825
+    },
+    {
+      "epoch": 3.32,
+      "grad_norm": 1.437711477279663,
+      "learning_rate": 1.68961954373567e-05,
+      "loss": 1.7012,
+      "step": 830
+    },
+    {
+      "epoch": 3.34,
+      "grad_norm": 1.165284514427185,
+      "learning_rate": 1.684547105928689e-05,
+      "loss": 1.6817,
+      "step": 835
+    },
+    {
+      "epoch": 3.36,
+      "grad_norm": 1.228750467300415,
+      "learning_rate": 1.6794413042615168e-05,
+      "loss": 1.6972,
+      "step": 840
+    },
+    {
+      "epoch": 3.38,
+      "grad_norm": 1.3721333742141724,
+      "learning_rate": 1.6743023875837233e-05,
+      "loss": 1.6953,
+      "step": 845
+    },
+    {
+      "epoch": 3.4,
+      "grad_norm": 1.221617341041565,
+      "learning_rate": 1.6691306063588583e-05,
+      "loss": 1.684,
+      "step": 850
+    },
+    {
+      "epoch": 3.42,
+      "grad_norm": 1.4893969297409058,
+      "learning_rate": 1.6639262126522417e-05,
+      "loss": 1.6976,
+      "step": 855
+    },
+    {
+      "epoch": 3.44,
+      "grad_norm": 1.3793864250183105,
+      "learning_rate": 1.6586894601186804e-05,
+      "loss": 1.6917,
+      "step": 860
+    },
+    {
+      "epoch": 3.46,
+      "grad_norm": 1.2841719388961792,
+      "learning_rate": 1.6534206039901057e-05,
+      "loss": 1.688,
+      "step": 865
+    },
+    {
+      "epoch": 3.48,
+      "grad_norm": 1.1975661516189575,
+      "learning_rate": 1.6481199010631312e-05,
+      "loss": 1.701,
+      "step": 870
+    },
+    {
+      "epoch": 3.5,
+      "grad_norm": 1.4768658876419067,
+      "learning_rate": 1.6427876096865394e-05,
+      "loss": 1.6828,
+      "step": 875
+    },
+    {
+      "epoch": 3.52,
+      "grad_norm": 1.1539041996002197,
+      "learning_rate": 1.63742398974869e-05,
+      "loss": 1.6807,
+      "step": 880
+    },
+    {
+      "epoch": 3.54,
+      "grad_norm": 1.301548719406128,
+      "learning_rate": 1.632029302664851e-05,
+      "loss": 1.6826,
+      "step": 885
+    },
+    {
+      "epoch": 3.56,
+      "grad_norm": 1.2049092054367065,
+      "learning_rate": 1.6266038113644605e-05,
+      "loss": 1.7009,
+      "step": 890
+    },
+    {
+      "epoch": 3.58,
+      "grad_norm": 1.151050090789795,
+      "learning_rate": 1.6211477802783105e-05,
+      "loss": 1.6855,
+      "step": 895
+    },
+    {
+      "epoch": 3.6,
+      "grad_norm": 1.1691709756851196,
+      "learning_rate": 1.6156614753256583e-05,
+      "loss": 1.692,
+      "step": 900
+    },
+    {
+      "epoch": 3.62,
+      "grad_norm": 1.2301398515701294,
+      "learning_rate": 1.610145163901268e-05,
+      "loss": 1.6893,
+      "step": 905
+    },
+    {
+      "epoch": 3.64,
+      "grad_norm": 1.1491998434066772,
+      "learning_rate": 1.6045991148623752e-05,
+      "loss": 1.6737,
+      "step": 910
+    },
+    {
+      "epoch": 3.66,
+      "grad_norm": 1.6054513454437256,
+      "learning_rate": 1.599023598515586e-05,
+      "loss": 1.6843,
+      "step": 915
+    },
+    {
+      "epoch": 3.68,
+      "grad_norm": 1.224255919456482,
+      "learning_rate": 1.5934188866037017e-05,
+      "loss": 1.6828,
+      "step": 920
+    },
+    {
+      "epoch": 3.7,
+      "grad_norm": 1.2744394540786743,
+      "learning_rate": 1.5877852522924733e-05,
+      "loss": 1.6954,
+      "step": 925
+    },
+    {
+      "epoch": 3.7199999999999998,
+      "grad_norm": 1.6310608386993408,
+      "learning_rate": 1.5821229701572897e-05,
+      "loss": 1.6992,
+      "step": 930
+    },
+    {
+      "epoch": 3.74,
+      "grad_norm": 1.1329830884933472,
+      "learning_rate": 1.5764323161697933e-05,
+      "loss": 1.6858,
+      "step": 935
+    },
+    {
+      "epoch": 3.76,
+      "grad_norm": 1.1588170528411865,
+      "learning_rate": 1.570713567684432e-05,
+      "loss": 1.6868,
+      "step": 940
+    },
+    {
+      "epoch": 3.7800000000000002,
+      "grad_norm": 1.1529875993728638,
+      "learning_rate": 1.564967003424938e-05,
+      "loss": 1.688,
+      "step": 945
+    },
+    {
+      "epoch": 3.8,
+      "grad_norm": 1.172899842262268,
+      "learning_rate": 1.5591929034707468e-05,
+      "loss": 1.6989,
+      "step": 950
+    },
+    {
+      "epoch": 3.82,
+      "grad_norm": 1.1531150341033936,
+      "learning_rate": 1.553391549243344e-05,
+      "loss": 1.6896,
+      "step": 955
+    },
+    {
+      "epoch": 3.84,
+      "grad_norm": 1.2427817583084106,
+      "learning_rate": 1.5475632234925505e-05,
+      "loss": 1.6944,
+      "step": 960
+    },
+    {
+      "epoch": 3.86,
+      "grad_norm": 1.168960452079773,
+      "learning_rate": 1.54170821028274e-05,
+      "loss": 1.6789,
+      "step": 965
+    },
+    {
+      "epoch": 3.88,
+      "grad_norm": 1.1754002571105957,
+      "learning_rate": 1.5358267949789968e-05,
+      "loss": 1.6926,
+      "step": 970
+    },
+    {
+      "epoch": 3.9,
+      "grad_norm": 1.373303771018982,
+      "learning_rate": 1.529919264233205e-05,
+      "loss": 1.7003,
+      "step": 975
+    },
+    {
+      "epoch": 3.92,
+      "grad_norm": 2.1144957542419434,
+      "learning_rate": 1.5239859059700794e-05,
+      "loss": 1.6954,
+      "step": 980
+    },
+    {
+      "epoch": 3.94,
+      "grad_norm": 1.2079601287841797,
+      "learning_rate": 1.5180270093731305e-05,
+      "loss": 1.692,
+      "step": 985
+    },
+    {
+      "epoch": 3.96,
+      "grad_norm": 1.2189019918441772,
+      "learning_rate": 1.5120428648705716e-05,
+      "loss": 1.6848,
+      "step": 990
+    },
+    {
+      "epoch": 3.98,
+      "grad_norm": 1.1501588821411133,
+      "learning_rate": 1.5060337641211637e-05,
+      "loss": 1.7002,
+      "step": 995
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 1.1473348140716553,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 1.6857,
+      "step": 1000
+    },
+    {
+      "epoch": 4.02,
+      "grad_norm": 1.1277931928634644,
+      "learning_rate": 1.493941866584231e-05,
+      "loss": 1.6764,
+      "step": 1005
+    },
+    {
+      "epoch": 4.04,
+      "grad_norm": 1.2237178087234497,
+      "learning_rate": 1.4878596591387329e-05,
+      "loss": 1.6793,
+      "step": 1010
+    },
+    {
+      "epoch": 4.06,
+      "grad_norm": 1.3738579750061035,
+      "learning_rate": 1.4817536741017153e-05,
+      "loss": 1.6807,
+      "step": 1015
+    },
+    {
+      "epoch": 4.08,
+      "grad_norm": 1.2409491539001465,
+      "learning_rate": 1.4756242090702756e-05,
+      "loss": 1.6794,
+      "step": 1020
+    },
+    {
+      "epoch": 4.1,
+      "grad_norm": 1.2782703638076782,
+      "learning_rate": 1.469471562785891e-05,
+      "loss": 1.6749,
+      "step": 1025
+    },
+    {
+      "epoch": 4.12,
+      "grad_norm": 1.2754226922988892,
+      "learning_rate": 1.463296035119862e-05,
+      "loss": 1.6734,
+      "step": 1030
+    },
+    {
+      "epoch": 4.14,
+      "grad_norm": 1.2375459671020508,
+      "learning_rate": 1.4570979270586944e-05,
+      "loss": 1.6694,
+      "step": 1035
+    },
+    {
+      "epoch": 4.16,
+      "grad_norm": 1.3407933712005615,
+      "learning_rate": 1.4508775406894308e-05,
+      "loss": 1.6766,
+      "step": 1040
+    },
+    {
+      "epoch": 4.18,
+      "grad_norm": 1.4103169441223145,
+      "learning_rate": 1.4446351791849276e-05,
+      "loss": 1.6756,
+      "step": 1045
+    },
+    {
+      "epoch": 4.2,
+      "grad_norm": 1.2496464252471924,
+      "learning_rate": 1.4383711467890776e-05,
+      "loss": 1.6804,
+      "step": 1050
+    },
+    {
+      "epoch": 4.22,
+      "grad_norm": 1.2907923460006714,
+      "learning_rate": 1.4320857488019826e-05,
+      "loss": 1.6841,
+      "step": 1055
+    },
+    {
+      "epoch": 4.24,
+      "grad_norm": 1.3250927925109863,
+      "learning_rate": 1.4257792915650728e-05,
+      "loss": 1.6815,
+      "step": 1060
+    },
+    {
+      "epoch": 4.26,
+      "grad_norm": 1.2333563566207886,
+      "learning_rate": 1.4194520824461773e-05,
+      "loss": 1.6641,
+      "step": 1065
+    },
+    {
+      "epoch": 4.28,
+      "grad_norm": 1.340520977973938,
+      "learning_rate": 1.413104429824542e-05,
+      "loss": 1.6791,
+      "step": 1070
+    },
+    {
+      "epoch": 4.3,
+      "grad_norm": 1.379290223121643,
+      "learning_rate": 1.4067366430758004e-05,
+      "loss": 1.6995,
+      "step": 1075
+    },
+    {
+      "epoch": 4.32,
+      "grad_norm": 1.1723048686981201,
+      "learning_rate": 1.4003490325568953e-05,
+      "loss": 1.6702,
+      "step": 1080
+    },
+    {
+      "epoch": 4.34,
+      "grad_norm": 1.2853553295135498,
+      "learning_rate": 1.3939419095909513e-05,
+      "loss": 1.676,
+      "step": 1085
+    },
+    {
+      "epoch": 4.36,
+      "grad_norm": 1.2285099029541016,
+      "learning_rate": 1.3875155864521031e-05,
+      "loss": 1.68,
+      "step": 1090
+    },
+    {
+      "epoch": 4.38,
+      "grad_norm": 1.2212533950805664,
+      "learning_rate": 1.3810703763502744e-05,
+      "loss": 1.6758,
+      "step": 1095
+    },
+    {
+      "epoch": 4.4,
+      "grad_norm": 1.2971107959747314,
+      "learning_rate": 1.3746065934159123e-05,
+      "loss": 1.6779,
+      "step": 1100
+    },
+    {
+      "epoch": 4.42,
+      "grad_norm": 1.2474068403244019,
+      "learning_rate": 1.3681245526846782e-05,
+      "loss": 1.6857,
+      "step": 1105
+    },
+    {
+      "epoch": 4.44,
+      "grad_norm": 1.330390453338623,
+      "learning_rate": 1.3616245700820922e-05,
+      "loss": 1.6797,
+      "step": 1110
+    },
+    {
+      "epoch": 4.46,
+      "grad_norm": 1.258988857269287,
+      "learning_rate": 1.3551069624081372e-05,
+      "loss": 1.6794,
+      "step": 1115
+    },
+    {
+      "epoch": 4.48,
+      "grad_norm": 1.2740941047668457,
+      "learning_rate": 1.3485720473218153e-05,
+      "loss": 1.6711,
+      "step": 1120
+    },
+    {
+      "epoch": 4.5,
+      "grad_norm": 1.2095681428909302,
+      "learning_rate": 1.342020143325669e-05,
+      "loss": 1.6831,
+      "step": 1125
+    },
+    {
+      "epoch": 4.52,
+      "grad_norm": 1.3167965412139893,
+      "learning_rate": 1.3354515697502552e-05,
+      "loss": 1.6789,
+      "step": 1130
+    },
+    {
+      "epoch": 4.54,
+      "grad_norm": 1.2791593074798584,
+      "learning_rate": 1.3288666467385834e-05,
+      "loss": 1.6762,
+      "step": 1135
+    },
+    {
+      "epoch": 4.5600000000000005,
+      "grad_norm": 1.2301644086837769,
+      "learning_rate": 1.3222656952305113e-05,
+      "loss": 1.6773,
+      "step": 1140
+    },
+    {
+      "epoch": 4.58,
+      "grad_norm": 1.253196120262146,
+      "learning_rate": 1.3156490369471026e-05,
+      "loss": 1.6793,
+      "step": 1145
+    },
+    {
+      "epoch": 4.6,
+      "grad_norm": 1.2421847581863403,
+      "learning_rate": 1.3090169943749475e-05,
+      "loss": 1.6898,
+      "step": 1150
+    },
+    {
+      "epoch": 4.62,
+      "grad_norm": 1.237874984741211,
+      "learning_rate": 1.3023698907504447e-05,
+      "loss": 1.6703,
+      "step": 1155
+    },
+    {
+      "epoch": 4.64,
+      "grad_norm": 1.2837252616882324,
+      "learning_rate": 1.2957080500440469e-05,
+      "loss": 1.6747,
+      "step": 1160
+    },
+    {
+      "epoch": 4.66,
+      "grad_norm": 1.2899264097213745,
+      "learning_rate": 1.2890317969444716e-05,
+      "loss": 1.6826,
+      "step": 1165
+    },
+    {
+      "epoch": 4.68,
+      "grad_norm": 1.166990041732788,
+      "learning_rate": 1.2823414568428767e-05,
+      "loss": 1.6687,
+      "step": 1170
+    },
+    {
+      "epoch": 4.7,
+      "grad_norm": 1.1460736989974976,
+      "learning_rate": 1.2756373558169992e-05,
+      "loss": 1.6774,
+      "step": 1175
+    },
+    {
+      "epoch": 4.72,
+      "grad_norm": 1.2305958271026611,
+      "learning_rate": 1.2689198206152657e-05,
+      "loss": 1.6846,
+      "step": 1180
+    },
+    {
+      "epoch": 4.74,
+      "grad_norm": 1.1499146223068237,
+      "learning_rate": 1.2621891786408648e-05,
+      "loss": 1.6738,
+      "step": 1185
+    },
+    {
+      "epoch": 4.76,
+      "grad_norm": 1.3568623065948486,
+      "learning_rate": 1.2554457579357906e-05,
+      "loss": 1.6868,
+      "step": 1190
+    },
+    {
+      "epoch": 4.78,
+      "grad_norm": 1.7656750679016113,
+      "learning_rate": 1.2486898871648552e-05,
+      "loss": 1.6829,
+      "step": 1195
+    },
+    {
+      "epoch": 4.8,
+      "grad_norm": 1.4720224142074585,
+      "learning_rate": 1.2419218955996677e-05,
+      "loss": 1.6739,
+      "step": 1200
+    },
+    {
+      "epoch": 4.82,
+      "grad_norm": 1.2218985557556152,
+      "learning_rate": 1.23514211310259e-05,
+      "loss": 1.6699,
+      "step": 1205
+    },
+    {
+      "epoch": 4.84,
+      "grad_norm": 1.2355974912643433,
+      "learning_rate": 1.2283508701106559e-05,
+      "loss": 1.669,
+      "step": 1210
+    },
+    {
+      "epoch": 4.86,
+      "grad_norm": 1.2267754077911377,
+      "learning_rate": 1.2215484976194675e-05,
+      "loss": 1.6785,
+      "step": 1215
+    },
+    {
+      "epoch": 4.88,
+      "grad_norm": 1.3336670398712158,
+      "learning_rate": 1.2147353271670634e-05,
+      "loss": 1.6811,
+      "step": 1220
+    },
+    {
+      "epoch": 4.9,
+      "grad_norm": 1.1803137063980103,
+      "learning_rate": 1.2079116908177592e-05,
+      "loss": 1.6802,
+      "step": 1225
+    },
+    {
+      "epoch": 4.92,
+      "grad_norm": 2.2649755477905273,
+      "learning_rate": 1.2010779211459649e-05,
+      "loss": 1.6807,
+      "step": 1230
+    },
+    {
+      "epoch": 4.9399999999999995,
+      "grad_norm": 1.1885944604873657,
+      "learning_rate": 1.194234351219972e-05,
+      "loss": 1.6784,
+      "step": 1235
+    },
+    {
+      "epoch": 4.96,
+      "grad_norm": 1.2544772624969482,
+      "learning_rate": 1.187381314585725e-05,
+      "loss": 1.6766,
+      "step": 1240
+    },
+    {
+      "epoch": 4.98,
+      "grad_norm": 1.3752778768539429,
+      "learning_rate": 1.1805191452505602e-05,
+      "loss": 1.6865,
+      "step": 1245
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 1.2073532342910767,
+      "learning_rate": 1.1736481776669307e-05,
+      "loss": 1.6776,
+      "step": 1250
+    },
+    {
+      "epoch": 5.02,
+      "grad_norm": 1.3473936319351196,
+      "learning_rate": 1.1667687467161025e-05,
+      "loss": 1.6418,
+      "step": 1255
+    },
+    {
+      "epoch": 5.04,
+      "grad_norm": 1.5514543056488037,
+      "learning_rate": 1.159881187691835e-05,
+      "loss": 1.6321,
+      "step": 1260
+    },
+    {
+      "epoch": 5.06,
+      "grad_norm": 1.534508228302002,
+      "learning_rate": 1.1529858362840383e-05,
+      "loss": 1.6177,
+      "step": 1265
+    },
+    {
+      "epoch": 5.08,
+      "grad_norm": 1.8792202472686768,
+      "learning_rate": 1.1460830285624119e-05,
+      "loss": 1.6193,
+      "step": 1270
+    },
+    {
+      "epoch": 5.1,
+      "grad_norm": 1.6183016300201416,
+      "learning_rate": 1.1391731009600655e-05,
+      "loss": 1.6298,
+      "step": 1275
+    },
+    {
+      "epoch": 5.12,
+      "grad_norm": 1.7961289882659912,
+      "learning_rate": 1.1322563902571227e-05,
+      "loss": 1.6209,
+      "step": 1280
+    },
+    {
+      "epoch": 5.14,
+      "grad_norm": 1.5732005834579468,
+      "learning_rate": 1.1253332335643043e-05,
+      "loss": 1.6191,
+      "step": 1285
+    },
+    {
+      "epoch": 5.16,
+      "grad_norm": 1.7134405374526978,
+      "learning_rate": 1.1184039683065014e-05,
+      "loss": 1.6236,
+      "step": 1290
+    },
+    {
+      "epoch": 5.18,
+      "grad_norm": 1.8280987739562988,
+      "learning_rate": 1.1114689322063255e-05,
+      "loss": 1.6228,
+      "step": 1295
+    },
+    {
+      "epoch": 5.2,
+      "grad_norm": 1.6293045282363892,
+      "learning_rate": 1.1045284632676535e-05,
+      "loss": 1.6207,
+      "step": 1300
+    },
+    {
+      "epoch": 5.22,
+      "grad_norm": 1.6155800819396973,
+      "learning_rate": 1.0975828997591496e-05,
+      "loss": 1.6233,
+      "step": 1305
+    },
+    {
+      "epoch": 5.24,
+      "grad_norm": 1.6458238363265991,
+      "learning_rate": 1.0906325801977804e-05,
+      "loss": 1.6252,
+      "step": 1310
+    },
+    {
+      "epoch": 5.26,
+      "grad_norm": 1.6302796602249146,
+      "learning_rate": 1.083677843332316e-05,
+      "loss": 1.629,
+      "step": 1315
+    },
+    {
+      "epoch": 5.28,
+      "grad_norm": 1.6061460971832275,
+      "learning_rate": 1.0767190281268187e-05,
+      "loss": 1.6288,
+      "step": 1320
+    },
+    {
+      "epoch": 5.3,
+      "grad_norm": 1.6623554229736328,
+      "learning_rate": 1.0697564737441254e-05,
+      "loss": 1.6336,
+      "step": 1325
+    },
+    {
+      "epoch": 5.32,
+      "grad_norm": 1.5957063436508179,
+      "learning_rate": 1.0627905195293135e-05,
+      "loss": 1.6236,
+      "step": 1330
+    },
+    {
+      "epoch": 5.34,
+      "grad_norm": 1.6672245264053345,
+      "learning_rate": 1.055821504993164e-05,
+      "loss": 1.6304,
+      "step": 1335
+    },
+    {
+      "epoch": 5.36,
+      "grad_norm": 1.7344614267349243,
+      "learning_rate": 1.0488497697956134e-05,
+      "loss": 1.6276,
+      "step": 1340
+    },
+    {
+      "epoch": 5.38,
+      "grad_norm": 1.7076168060302734,
+      "learning_rate": 1.0418756537291996e-05,
+      "loss": 1.637,
+      "step": 1345
+    },
+    {
+      "epoch": 5.4,
+      "grad_norm": 1.7165381908416748,
+      "learning_rate": 1.0348994967025012e-05,
+      "loss": 1.6308,
+      "step": 1350
+    },
+    {
+      "epoch": 5.42,
+      "grad_norm": 1.69627046585083,
+      "learning_rate": 1.0279216387235691e-05,
+      "loss": 1.6368,
+      "step": 1355
+    },
+    {
+      "epoch": 5.44,
+      "grad_norm": 1.8979768753051758,
+      "learning_rate": 1.0209424198833571e-05,
+      "loss": 1.6414,
+      "step": 1360
+    },
+    {
+      "epoch": 5.46,
+      "grad_norm": 1.769880771636963,
+      "learning_rate": 1.0139621803391454e-05,
+      "loss": 1.6368,
+      "step": 1365
+    },
+    {
+      "epoch": 5.48,
+      "grad_norm": 1.628905177116394,
+      "learning_rate": 1.0069812602979617e-05,
+      "loss": 1.6459,
+      "step": 1370
+    },
+    {
+      "epoch": 5.5,
+      "grad_norm": 1.581150770187378,
+      "learning_rate": 1e-05,
+      "loss": 1.6233,
+      "step": 1375
+    },
+    {
+      "epoch": 5.52,
+      "grad_norm": 1.5890735387802124,
+      "learning_rate": 9.930187397020385e-06,
+      "loss": 1.6264,
+      "step": 1380
+    },
+    {
+      "epoch": 5.54,
+      "grad_norm": 1.6208202838897705,
+      "learning_rate": 9.860378196608549e-06,
+      "loss": 1.6264,
+      "step": 1385
+    },
+    {
+      "epoch": 5.5600000000000005,
+      "grad_norm": 1.627545952796936,
+      "learning_rate": 9.790575801166432e-06,
+      "loss": 1.6296,
+      "step": 1390
+    },
+    {
+      "epoch": 5.58,
+      "grad_norm": 1.5746124982833862,
+      "learning_rate": 9.720783612764314e-06,
+      "loss": 1.6331,
+      "step": 1395
+    },
+    {
+      "epoch": 5.6,
+      "grad_norm": 1.6535131931304932,
+      "learning_rate": 9.651005032974994e-06,
+      "loss": 1.629,
+      "step": 1400
+    },
+    {
+      "epoch": 5.62,
+      "grad_norm": 1.559142827987671,
+      "learning_rate": 9.581243462708007e-06,
+      "loss": 1.6187,
+      "step": 1405
+    },
+    {
+      "epoch": 5.64,
+      "grad_norm": 1.5918751955032349,
+      "learning_rate": 9.511502302043867e-06,
+      "loss": 1.635,
+      "step": 1410
+    },
+    {
+      "epoch": 5.66,
+      "grad_norm": 1.6395398378372192,
+      "learning_rate": 9.441784950068362e-06,
+      "loss": 1.6323,
+      "step": 1415
+    },
+    {
+      "epoch": 5.68,
+      "grad_norm": 1.6209341287612915,
+      "learning_rate": 9.372094804706867e-06,
+      "loss": 1.6342,
+      "step": 1420
+    },
+    {
+      "epoch": 5.7,
+      "grad_norm": 1.7134851217269897,
+      "learning_rate": 9.302435262558748e-06,
+      "loss": 1.6312,
+      "step": 1425
+    },
+    {
+      "epoch": 5.72,
+      "grad_norm": 2.302839994430542,
+      "learning_rate": 9.232809718731815e-06,
+      "loss": 1.6326,
+      "step": 1430
+    },
+    {
+      "epoch": 5.74,
+      "grad_norm": 1.6598634719848633,
+      "learning_rate": 9.163221566676847e-06,
+      "loss": 1.6416,
+      "step": 1435
+    },
+    {
+      "epoch": 5.76,
+      "grad_norm": 1.692478060722351,
+      "learning_rate": 9.093674198022201e-06,
+      "loss": 1.6366,
+      "step": 1440
+    },
+    {
+      "epoch": 5.78,
+      "grad_norm": 1.6801451444625854,
+      "learning_rate": 9.024171002408507e-06,
+      "loss": 1.6338,
+      "step": 1445
+    },
+    {
+      "epoch": 5.8,
+      "grad_norm": 1.7073163986206055,
+      "learning_rate": 8.954715367323468e-06,
+      "loss": 1.6325,
+      "step": 1450
+    },
+    {
+      "epoch": 5.82,
+      "grad_norm": 1.6709482669830322,
+      "learning_rate": 8.885310677936746e-06,
+      "loss": 1.6296,
+      "step": 1455
+    },
+    {
+      "epoch": 5.84,
+      "grad_norm": 1.6847496032714844,
+      "learning_rate": 8.815960316934991e-06,
+      "loss": 1.63,
+      "step": 1460
+    },
+    {
+      "epoch": 5.86,
+      "grad_norm": 1.6837451457977295,
+      "learning_rate": 8.746667664356957e-06,
+      "loss": 1.6413,
+      "step": 1465
+    },
+    {
+      "epoch": 5.88,
+      "grad_norm": 1.6769607067108154,
+      "learning_rate": 8.677436097428775e-06,
+      "loss": 1.6393,
+      "step": 1470
+    },
+    {
+      "epoch": 5.9,
+      "grad_norm": 1.7436274290084839,
+      "learning_rate": 8.60826899039935e-06,
+      "loss": 1.632,
+      "step": 1475
+    },
+    {
+      "epoch": 5.92,
+      "grad_norm": 1.6370117664337158,
+      "learning_rate": 8.539169714375885e-06,
+      "loss": 1.6294,
+      "step": 1480
+    },
+    {
+      "epoch": 5.9399999999999995,
+      "grad_norm": 1.633933186531067,
+      "learning_rate": 8.47014163715962e-06,
+      "loss": 1.6318,
+      "step": 1485
+    },
+    {
+      "epoch": 5.96,
+      "grad_norm": 1.7138166427612305,
+      "learning_rate": 8.401188123081653e-06,
+      "loss": 1.6333,
+      "step": 1490
+    },
+    {
+      "epoch": 5.98,
+      "grad_norm": 1.6793943643569946,
+      "learning_rate": 8.332312532838978e-06,
+      "loss": 1.625,
+      "step": 1495
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 1.6901153326034546,
+      "learning_rate": 8.263518223330698e-06,
+      "loss": 1.6276,
+      "step": 1500
+    },
+    {
+      "epoch": 6.02,
+      "grad_norm": 2.6299827098846436,
+      "learning_rate": 8.194808547494401e-06,
+      "loss": 1.4932,
+      "step": 1505
+    },
+    {
+      "epoch": 6.04,
+      "grad_norm": 3.6387939453125,
+      "learning_rate": 8.126186854142752e-06,
+      "loss": 1.4582,
+      "step": 1510
+    },
+    {
+      "epoch": 6.06,
+      "grad_norm": 2.7163808345794678,
+      "learning_rate": 8.057656487800283e-06,
+      "loss": 1.4205,
+      "step": 1515
+    },
+    {
+      "epoch": 6.08,
+      "grad_norm": 3.661618947982788,
+      "learning_rate": 7.989220788540356e-06,
+      "loss": 1.4164,
+      "step": 1520
+    },
+    {
+      "epoch": 6.1,
+      "grad_norm": 3.3539645671844482,
+      "learning_rate": 7.92088309182241e-06,
+      "loss": 1.4098,
+      "step": 1525
+    },
+    {
+      "epoch": 6.12,
+      "grad_norm": 3.1834163665771484,
+      "learning_rate": 7.852646728329368e-06,
+      "loss": 1.4189,
+      "step": 1530
+    },
+    {
+      "epoch": 6.14,
+      "grad_norm": 3.607142448425293,
+      "learning_rate": 7.784515023805328e-06,
+      "loss": 1.4078,
+      "step": 1535
+    },
+    {
+      "epoch": 6.16,
+      "grad_norm": 3.505796194076538,
+      "learning_rate": 7.716491298893443e-06,
+      "loss": 1.4103,
+      "step": 1540
+    },
+    {
+      "epoch": 6.18,
+      "grad_norm": 3.1681082248687744,
+      "learning_rate": 7.6485788689741e-06,
+      "loss": 1.4122,
+      "step": 1545
+    },
+    {
+      "epoch": 6.2,
+      "grad_norm": 3.3361103534698486,
+      "learning_rate": 7.580781044003324e-06,
+      "loss": 1.4108,
+      "step": 1550
+    },
+    {
+      "epoch": 6.22,
+      "grad_norm": 3.2785024642944336,
+      "learning_rate": 7.513101128351454e-06,
+      "loss": 1.4101,
+      "step": 1555
+    },
+    {
+      "epoch": 6.24,
+      "grad_norm": 3.350015640258789,
+      "learning_rate": 7.445542420642097e-06,
+      "loss": 1.419,
+      "step": 1560
+    },
+    {
+      "epoch": 6.26,
+      "grad_norm": 3.3478310108184814,
+      "learning_rate": 7.378108213591355e-06,
+      "loss": 1.4295,
+      "step": 1565
+    },
+    {
+      "epoch": 6.28,
+      "grad_norm": 3.2591845989227295,
+      "learning_rate": 7.310801793847344e-06,
+      "loss": 1.4049,
+      "step": 1570
+    },
+    {
+      "epoch": 6.3,
+      "grad_norm": 3.5392940044403076,
+      "learning_rate": 7.243626441830009e-06,
+      "loss": 1.4217,
+      "step": 1575
+    },
+    {
+      "epoch": 6.32,
+      "grad_norm": 3.395463228225708,
+      "learning_rate": 7.176585431571235e-06,
+      "loss": 1.4155,
+      "step": 1580
+    },
+    {
+      "epoch": 6.34,
+      "grad_norm": 3.304579496383667,
+      "learning_rate": 7.109682030555283e-06,
+      "loss": 1.4195,
+      "step": 1585
+    },
+    {
+      "epoch": 6.36,
+      "grad_norm": 3.369657039642334,
+      "learning_rate": 7.042919499559538e-06,
+      "loss": 1.4129,
+      "step": 1590
+    },
+    {
+      "epoch": 6.38,
+      "grad_norm": 3.4654033184051514,
+      "learning_rate": 6.976301092495556e-06,
+      "loss": 1.4068,
+      "step": 1595
+    },
+    {
+      "epoch": 6.4,
+      "grad_norm": 3.3557608127593994,
+      "learning_rate": 6.909830056250527e-06,
+      "loss": 1.4095,
+      "step": 1600
+    },
+    {
+      "epoch": 6.42,
+      "grad_norm": 3.554288625717163,
+      "learning_rate": 6.843509630528977e-06,
+      "loss": 1.4188,
+      "step": 1605
+    },
+    {
+      "epoch": 6.44,
+      "grad_norm": 3.3676187992095947,
+      "learning_rate": 6.777343047694891e-06,
+      "loss": 1.4055,
+      "step": 1610
+    },
+    {
+      "epoch": 6.46,
+      "grad_norm": 3.2888877391815186,
+      "learning_rate": 6.711333532614168e-06,
+      "loss": 1.418,
+      "step": 1615
+    },
+    {
+      "epoch": 6.48,
+      "grad_norm": 3.4561023712158203,
+      "learning_rate": 6.645484302497452e-06,
+      "loss": 1.4232,
+      "step": 1620
+    },
+    {
+      "epoch": 6.5,
+      "grad_norm": 3.327456474304199,
+      "learning_rate": 6.579798566743314e-06,
+      "loss": 1.4278,
+      "step": 1625
+    },
+    {
+      "epoch": 6.52,
+      "grad_norm": 3.2338459491729736,
+      "learning_rate": 6.5142795267818505e-06,
+      "loss": 1.4156,
+      "step": 1630
+    },
+    {
+      "epoch": 6.54,
+      "grad_norm": 3.3685622215270996,
+      "learning_rate": 6.448930375918632e-06,
+      "loss": 1.4114,
+      "step": 1635
+    },
+    {
+      "epoch": 6.5600000000000005,
+      "grad_norm": 3.458646774291992,
+      "learning_rate": 6.383754299179079e-06,
+      "loss": 1.4156,
+      "step": 1640
+    },
+    {
+      "epoch": 6.58,
+      "grad_norm": 3.4090073108673096,
+      "learning_rate": 6.318754473153221e-06,
+      "loss": 1.4046,
+      "step": 1645
+    },
+    {
+      "epoch": 6.6,
+      "grad_norm": 3.409212112426758,
+      "learning_rate": 6.25393406584088e-06,
+      "loss": 1.4137,
+      "step": 1650
+    },
+    {
+      "epoch": 6.62,
+      "grad_norm": 3.3928050994873047,
+      "learning_rate": 6.18929623649726e-06,
+      "loss": 1.4198,
+      "step": 1655
+    },
+    {
+      "epoch": 6.64,
+      "grad_norm": 3.4656596183776855,
+      "learning_rate": 6.124844135478971e-06,
+      "loss": 1.4149,
+      "step": 1660
+    },
+    {
+      "epoch": 6.66,
+      "grad_norm": 3.529662847518921,
+      "learning_rate": 6.06058090409049e-06,
+      "loss": 1.4195,
+      "step": 1665
+    },
+    {
+      "epoch": 6.68,
+      "grad_norm": 3.4222183227539062,
+      "learning_rate": 5.996509674431053e-06,
+      "loss": 1.4084,
+      "step": 1670
+    },
+    {
+      "epoch": 6.7,
+      "grad_norm": 3.577754020690918,
+      "learning_rate": 5.932633569242e-06,
+      "loss": 1.4155,
+      "step": 1675
+    },
+    {
+      "epoch": 6.72,
+      "grad_norm": 3.3030552864074707,
+      "learning_rate": 5.868955701754584e-06,
+      "loss": 1.4193,
+      "step": 1680
+    },
+    {
+      "epoch": 6.74,
+      "grad_norm": 3.3563826084136963,
+      "learning_rate": 5.8054791755382286e-06,
+      "loss": 1.4242,
+      "step": 1685
+    },
+    {
+      "epoch": 6.76,
+      "grad_norm": 3.5344948768615723,
+      "learning_rate": 5.742207084349274e-06,
+      "loss": 1.4201,
+      "step": 1690
+    },
+    {
+      "epoch": 6.78,
+      "grad_norm": 3.388164520263672,
+      "learning_rate": 5.679142511980176e-06,
+      "loss": 1.4106,
+      "step": 1695
+    },
+    {
+      "epoch": 6.8,
+      "grad_norm": 3.513644218444824,
+      "learning_rate": 5.616288532109225e-06,
+      "loss": 1.4088,
+      "step": 1700
+    },
+    {
+      "epoch": 6.82,
+      "grad_norm": 3.5908257961273193,
+      "learning_rate": 5.553648208150728e-06,
+      "loss": 1.414,
+      "step": 1705
+    },
+    {
+      "epoch": 6.84,
+      "grad_norm": 3.547783374786377,
+      "learning_rate": 5.491224593105695e-06,
+      "loss": 1.4183,
+      "step": 1710
+    },
+    {
+      "epoch": 6.86,
+      "grad_norm": 3.5997567176818848,
+      "learning_rate": 5.429020729413062e-06,
+      "loss": 1.4082,
+      "step": 1715
+    },
+    {
+      "epoch": 6.88,
+      "grad_norm": 3.3622734546661377,
+      "learning_rate": 5.367039648801386e-06,
+      "loss": 1.4178,
+      "step": 1720
+    },
+    {
+      "epoch": 6.9,
+      "grad_norm": 3.4612014293670654,
+      "learning_rate": 5.305284372141095e-06,
+      "loss": 1.3978,
+      "step": 1725
+    },
+    {
+      "epoch": 6.92,
+      "grad_norm": 3.395056962966919,
+      "learning_rate": 5.243757909297247e-06,
+      "loss": 1.4018,
+      "step": 1730
+    },
+    {
+      "epoch": 6.9399999999999995,
+      "grad_norm": 3.3911194801330566,
+      "learning_rate": 5.1824632589828465e-06,
+      "loss": 1.4298,
+      "step": 1735
+    },
+    {
+      "epoch": 6.96,
+      "grad_norm": 3.307224750518799,
+      "learning_rate": 5.121403408612672e-06,
+      "loss": 1.4006,
+      "step": 1740
+    },
+    {
+      "epoch": 6.98,
+      "grad_norm": 3.5052733421325684,
+      "learning_rate": 5.060581334157693e-06,
+      "loss": 1.4161,
+      "step": 1745
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 3.5450494289398193,
+      "learning_rate": 5.000000000000003e-06,
+      "loss": 1.3935,
+      "step": 1750
+    },
+    {
+      "epoch": 7.02,
+      "grad_norm": 4.5564093589782715,
+      "learning_rate": 4.939662358788364e-06,
+      "loss": 1.074,
+      "step": 1755
+    },
+    {
+      "epoch": 7.04,
+      "grad_norm": 7.566091060638428,
+      "learning_rate": 4.879571351294287e-06,
+      "loss": 1.0397,
+      "step": 1760
+    },
+    {
+      "epoch": 7.06,
+      "grad_norm": 5.21124792098999,
+      "learning_rate": 4.8197299062687e-06,
+      "loss": 1.0191,
+      "step": 1765
+    },
+    {
+      "epoch": 7.08,
+      "grad_norm": 5.251906394958496,
+      "learning_rate": 4.76014094029921e-06,
+      "loss": 1.0124,
+      "step": 1770
+    },
+    {
+      "epoch": 7.1,
+      "grad_norm": 5.793355464935303,
+      "learning_rate": 4.700807357667953e-06,
+      "loss": 0.9893,
+      "step": 1775
+    },
+    {
+      "epoch": 7.12,
+      "grad_norm": 5.138261318206787,
+      "learning_rate": 4.641732050210032e-06,
+      "loss": 0.9938,
+      "step": 1780
+    },
+    {
+      "epoch": 7.14,
+      "grad_norm": 5.370635032653809,
+      "learning_rate": 4.582917897172603e-06,
+      "loss": 0.9988,
+      "step": 1785
+    },
+    {
+      "epoch": 7.16,
+      "grad_norm": 5.199043273925781,
+      "learning_rate": 4.524367765074499e-06,
+      "loss": 0.9776,
+      "step": 1790
+    },
+    {
+      "epoch": 7.18,
+      "grad_norm": 5.165049076080322,
+      "learning_rate": 4.46608450756656e-06,
+      "loss": 0.9923,
+      "step": 1795
+    },
+    {
+      "epoch": 7.2,
+      "grad_norm": 5.3443779945373535,
+      "learning_rate": 4.408070965292534e-06,
+      "loss": 0.9809,
+      "step": 1800
+    },
+    {
+      "epoch": 7.22,
+      "grad_norm": 5.287759780883789,
+      "learning_rate": 4.350329965750622e-06,
+      "loss": 0.986,
+      "step": 1805
+    },
+    {
+      "epoch": 7.24,
+      "grad_norm": 5.187274932861328,
+      "learning_rate": 4.292864323155684e-06,
+      "loss": 1.002,
+      "step": 1810
+    },
+    {
+      "epoch": 7.26,
+      "grad_norm": 5.28734016418457,
+      "learning_rate": 4.235676838302069e-06,
+      "loss": 0.9989,
+      "step": 1815
+    },
+    {
+      "epoch": 7.28,
+      "grad_norm": 5.318136215209961,
+      "learning_rate": 4.178770298427107e-06,
+      "loss": 0.981,
+      "step": 1820
+    },
+    {
+      "epoch": 7.3,
+      "grad_norm": 5.267811298370361,
+      "learning_rate": 4.12214747707527e-06,
+      "loss": 0.9907,
+      "step": 1825
+    },
+    {
+      "epoch": 7.32,
+      "grad_norm": 5.451784610748291,
+      "learning_rate": 4.065811133962987e-06,
+      "loss": 0.9826,
+      "step": 1830
+    },
+    {
+      "epoch": 7.34,
+      "grad_norm": 5.516955375671387,
+      "learning_rate": 4.009764014844143e-06,
+      "loss": 0.9877,
+      "step": 1835
+    },
+    {
+      "epoch": 7.36,
+      "grad_norm": 5.202047348022461,
+      "learning_rate": 3.954008851376252e-06,
+      "loss": 0.9866,
+      "step": 1840
+    },
+    {
+      "epoch": 7.38,
+      "grad_norm": 5.0173540115356445,
+      "learning_rate": 3.898548360987325e-06,
+      "loss": 0.9887,
+      "step": 1845
+    },
+    {
+      "epoch": 7.4,
+      "grad_norm": 5.356690883636475,
+      "learning_rate": 3.8433852467434175e-06,
+      "loss": 0.9897,
+      "step": 1850
+    },
+    {
+      "epoch": 7.42,
+      "grad_norm": 5.419712543487549,
+      "learning_rate": 3.7885221972168974e-06,
+      "loss": 0.988,
+      "step": 1855
+    },
+    {
+      "epoch": 7.44,
+      "grad_norm": 5.474433898925781,
+      "learning_rate": 3.7339618863553983e-06,
+      "loss": 1.001,
+      "step": 1860
+    },
+    {
+      "epoch": 7.46,
+      "grad_norm": 5.467788219451904,
+      "learning_rate": 3.679706973351491e-06,
+      "loss": 0.978,
+      "step": 1865
+    },
+    {
+      "epoch": 7.48,
+      "grad_norm": 5.380366802215576,
+      "learning_rate": 3.625760102513103e-06,
+      "loss": 0.9791,
+      "step": 1870
+    },
+    {
+      "epoch": 7.5,
+      "grad_norm": 5.486819744110107,
+      "learning_rate": 3.5721239031346067e-06,
+      "loss": 0.9987,
+      "step": 1875
+    },
+    {
+      "epoch": 7.52,
+      "grad_norm": 5.298658847808838,
+      "learning_rate": 3.5188009893686916e-06,
+      "loss": 0.9926,
+      "step": 1880
+    },
+    {
+      "epoch": 7.54,
+      "grad_norm": 5.251545429229736,
+      "learning_rate": 3.4657939600989453e-06,
+      "loss": 0.9863,
+      "step": 1885
+    },
+    {
+      "epoch": 7.5600000000000005,
+      "grad_norm": 5.376091480255127,
+      "learning_rate": 3.4131053988131947e-06,
+      "loss": 0.9786,
+      "step": 1890
+    },
+    {
+      "epoch": 7.58,
+      "grad_norm": 5.347581386566162,
+      "learning_rate": 3.360737873477584e-06,
+      "loss": 0.9853,
+      "step": 1895
+    },
+    {
+      "epoch": 7.6,
+      "grad_norm": 5.291326522827148,
+      "learning_rate": 3.308693936411421e-06,
+      "loss": 0.9867,
+      "step": 1900
+    },
+    {
+      "epoch": 7.62,
+      "grad_norm": 5.290416240692139,
+      "learning_rate": 3.2569761241627694e-06,
+      "loss": 0.9824,
+      "step": 1905
+    },
+    {
+      "epoch": 7.64,
+      "grad_norm": 5.326883792877197,
+      "learning_rate": 3.2055869573848374e-06,
+      "loss": 0.9879,
+      "step": 1910
+    },
+    {
+      "epoch": 7.66,
+      "grad_norm": 5.346325397491455,
+      "learning_rate": 3.1545289407131128e-06,
+      "loss": 0.9931,
+      "step": 1915
+    },
+    {
+      "epoch": 7.68,
+      "grad_norm": 5.48749303817749,
+      "learning_rate": 3.103804562643302e-06,
+      "loss": 0.9901,
+      "step": 1920
+    },
+    {
+      "epoch": 7.7,
+      "grad_norm": 5.999955654144287,
+      "learning_rate": 3.0534162954100264e-06,
+      "loss": 1.0058,
+      "step": 1925
+    },
+    {
+      "epoch": 7.72,
+      "grad_norm": 5.255881309509277,
+      "learning_rate": 3.003366594866345e-06,
+      "loss": 0.9939,
+      "step": 1930
+    },
+    {
+      "epoch": 7.74,
+      "grad_norm": 5.439182758331299,
+      "learning_rate": 2.953657900364053e-06,
+      "loss": 0.9915,
+      "step": 1935
+    },
+    {
+      "epoch": 7.76,
+      "grad_norm": 5.336669445037842,
+      "learning_rate": 2.9042926346347932e-06,
+      "loss": 0.991,
+      "step": 1940
+    },
+    {
+      "epoch": 7.78,
+      "grad_norm": 5.281376361846924,
+      "learning_rate": 2.855273203671969e-06,
+      "loss": 0.9847,
+      "step": 1945
+    },
+    {
+      "epoch": 7.8,
+      "grad_norm": 5.2418413162231445,
+      "learning_rate": 2.8066019966134907e-06,
+      "loss": 0.9808,
+      "step": 1950
+    },
+    {
+      "epoch": 7.82,
+      "grad_norm": 5.544259071350098,
+      "learning_rate": 2.7582813856253276e-06,
+      "loss": 0.9833,
+      "step": 1955
+    },
+    {
+      "epoch": 7.84,
+      "grad_norm": 5.195487976074219,
+      "learning_rate": 2.7103137257858867e-06,
+      "loss": 0.9857,
+      "step": 1960
+    },
+    {
+      "epoch": 7.86,
+      "grad_norm": 5.333943843841553,
+      "learning_rate": 2.6627013549712355e-06,
+      "loss": 0.9994,
+      "step": 1965
+    },
+    {
+      "epoch": 7.88,
+      "grad_norm": 5.322220325469971,
+      "learning_rate": 2.615446593741161e-06,
+      "loss": 0.9742,
+      "step": 1970
+    },
+    {
+      "epoch": 7.9,
+      "grad_norm": 5.36502742767334,
+      "learning_rate": 2.5685517452260566e-06,
+      "loss": 0.9775,
+      "step": 1975
+    },
+    {
+      "epoch": 7.92,
+      "grad_norm": 5.417782783508301,
+      "learning_rate": 2.522019095014683e-06,
+      "loss": 0.9779,
+      "step": 1980
+    },
+    {
+      "epoch": 7.9399999999999995,
+      "grad_norm": 5.225217819213867,
+      "learning_rate": 2.4758509110427576e-06,
+      "loss": 0.9829,
+      "step": 1985
+    },
+    {
+      "epoch": 7.96,
+      "grad_norm": 5.371156215667725,
+      "learning_rate": 2.4300494434824373e-06,
+      "loss": 0.9835,
+      "step": 1990
+    },
+    {
+      "epoch": 7.98,
+      "grad_norm": 5.273430824279785,
+      "learning_rate": 2.3846169246326345e-06,
+      "loss": 0.9869,
+      "step": 1995
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 5.407411575317383,
+      "learning_rate": 2.339555568810221e-06,
+      "loss": 0.983,
+      "step": 2000
+    },
+    {
+      "epoch": 8.02,
+      "grad_norm": 5.448741436004639,
+      "learning_rate": 2.2948675722421086e-06,
+      "loss": 0.6566,
+      "step": 2005
+    },
+    {
+      "epoch": 8.04,
+      "grad_norm": 6.663678169250488,
+      "learning_rate": 2.2505551129582047e-06,
+      "loss": 0.6327,
+      "step": 2010
+    },
+    {
+      "epoch": 8.06,
+      "grad_norm": 6.39850378036499,
+      "learning_rate": 2.206620350685257e-06,
+      "loss": 0.6286,
+      "step": 2015
+    },
+    {
+      "epoch": 8.08,
+      "grad_norm": 5.739835262298584,
+      "learning_rate": 2.163065426741603e-06,
+      "loss": 0.6437,
+      "step": 2020
+    },
+    {
+      "epoch": 8.1,
+      "grad_norm": 5.893033981323242,
+      "learning_rate": 2.119892463932781e-06,
+      "loss": 0.6179,
+      "step": 2025
+    },
+    {
+      "epoch": 8.12,
+      "grad_norm": 6.056233882904053,
+      "learning_rate": 2.0771035664480944e-06,
+      "loss": 0.6167,
+      "step": 2030
+    },
+    {
+      "epoch": 8.14,
+      "grad_norm": 6.03655481338501,
+      "learning_rate": 2.0347008197580376e-06,
+      "loss": 0.6305,
+      "step": 2035
+    },
+    {
+      "epoch": 8.16,
+      "grad_norm": 6.043553352355957,
+      "learning_rate": 1.9926862905126663e-06,
+      "loss": 0.6165,
+      "step": 2040
+    },
+    {
+      "epoch": 8.18,
+      "grad_norm": 5.853250503540039,
+      "learning_rate": 1.95106202644086e-06,
+      "loss": 0.6307,
+      "step": 2045
+    },
+    {
+      "epoch": 8.2,
+      "grad_norm": 5.7116780281066895,
+      "learning_rate": 1.9098300562505266e-06,
+      "loss": 0.6103,
+      "step": 2050
+    },
+    {
+      "epoch": 8.22,
+      "grad_norm": 5.799511432647705,
+      "learning_rate": 1.8689923895297247e-06,
+      "loss": 0.6174,
+      "step": 2055
+    },
+    {
+      "epoch": 8.24,
+      "grad_norm": 5.831727504730225,
+      "learning_rate": 1.8285510166487154e-06,
+      "loss": 0.6292,
+      "step": 2060
+    },
+    {
+      "epoch": 8.26,
+      "grad_norm": 6.006520748138428,
+      "learning_rate": 1.7885079086629598e-06,
+      "loss": 0.6243,
+      "step": 2065
+    },
+    {
+      "epoch": 8.28,
+      "grad_norm": 6.150590419769287,
+      "learning_rate": 1.7488650172170496e-06,
+      "loss": 0.616,
+      "step": 2070
+    },
+    {
+      "epoch": 8.3,
+      "grad_norm": 6.151932239532471,
+      "learning_rate": 1.709624274449584e-06,
+      "loss": 0.632,
+      "step": 2075
+    },
+    {
+      "epoch": 8.32,
+      "grad_norm": 5.84307336807251,
+      "learning_rate": 1.6707875928990059e-06,
+      "loss": 0.6216,
+      "step": 2080
+    },
+    {
+      "epoch": 8.34,
+      "grad_norm": 6.241573810577393,
+      "learning_rate": 1.6323568654103838e-06,
+      "loss": 0.6106,
+      "step": 2085
+    },
+    {
+      "epoch": 8.36,
+      "grad_norm": 6.045923233032227,
+      "learning_rate": 1.5943339650431578e-06,
+      "loss": 0.6349,
+      "step": 2090
+    },
+    {
+      "epoch": 8.38,
+      "grad_norm": 6.1905598640441895,
+      "learning_rate": 1.5567207449798517e-06,
+      "loss": 0.6338,
+      "step": 2095
+    },
+    {
+      "epoch": 8.4,
+      "grad_norm": 5.788094520568848,
+      "learning_rate": 1.5195190384357405e-06,
+      "loss": 0.6191,
+      "step": 2100
+    },
+    {
+      "epoch": 8.42,
+      "grad_norm": 5.833618640899658,
+      "learning_rate": 1.4827306585695234e-06,
+      "loss": 0.6309,
+      "step": 2105
+    },
+    {
+      "epoch": 8.44,
+      "grad_norm": 6.230561256408691,
+      "learning_rate": 1.446357398394934e-06,
+      "loss": 0.6558,
+      "step": 2110
+    },
+    {
+      "epoch": 8.46,
+      "grad_norm": 6.213106632232666,
+      "learning_rate": 1.4104010306933558e-06,
+      "loss": 0.6292,
+      "step": 2115
+    },
+    {
+      "epoch": 8.48,
+      "grad_norm": 5.9851975440979,
+      "learning_rate": 1.3748633079274254e-06,
+      "loss": 0.6263,
+      "step": 2120
+    },
+    {
+      "epoch": 8.5,
+      "grad_norm": 5.79006290435791,
+      "learning_rate": 1.339745962155613e-06,
+      "loss": 0.6192,
+      "step": 2125
+    },
+    {
+      "epoch": 8.52,
+      "grad_norm": 5.929935932159424,
+      "learning_rate": 1.30505070494781e-06,
+      "loss": 0.6353,
+      "step": 2130
+    },
+    {
+      "epoch": 8.54,
+      "grad_norm": 6.356067180633545,
+      "learning_rate": 1.2707792273019049e-06,
+      "loss": 0.624,
+      "step": 2135
+    },
+    {
+      "epoch": 8.56,
+      "grad_norm": 6.018552303314209,
+      "learning_rate": 1.2369331995613664e-06,
+      "loss": 0.6158,
+      "step": 2140
+    },
+    {
+      "epoch": 8.58,
+      "grad_norm": 6.076599597930908,
+      "learning_rate": 1.2035142713338366e-06,
+      "loss": 0.6242,
+      "step": 2145
+    },
+    {
+      "epoch": 8.6,
+      "grad_norm": 5.8761701583862305,
+      "learning_rate": 1.1705240714107301e-06,
+      "loss": 0.616,
+      "step": 2150
+    },
+    {
+      "epoch": 8.62,
+      "grad_norm": 5.857326984405518,
+      "learning_rate": 1.1379642076878528e-06,
+      "loss": 0.6187,
+      "step": 2155
+    },
+    {
+      "epoch": 8.64,
+      "grad_norm": 6.0746636390686035,
+      "learning_rate": 1.1058362670870248e-06,
+      "loss": 0.6359,
+      "step": 2160
+    },
+    {
+      "epoch": 8.66,
+      "grad_norm": 6.17335844039917,
+      "learning_rate": 1.0741418154787443e-06,
+      "loss": 0.6234,
+      "step": 2165
+    },
+    {
+      "epoch": 8.68,
+      "grad_norm": 6.193073749542236,
+      "learning_rate": 1.042882397605871e-06,
+      "loss": 0.6144,
+      "step": 2170
+    },
+    {
+      "epoch": 8.7,
+      "grad_norm": 5.796996593475342,
+      "learning_rate": 1.012059537008332e-06,
+      "loss": 0.6187,
+      "step": 2175
+    },
+    {
+      "epoch": 8.72,
+      "grad_norm": 6.0771164894104,
+      "learning_rate": 9.816747359488632e-07,
+      "loss": 0.6352,
+      "step": 2180
+    },
+    {
+      "epoch": 8.74,
+      "grad_norm": 5.912444114685059,
+      "learning_rate": 9.517294753398066e-07,
+      "loss": 0.6201,
+      "step": 2185
+    },
+    {
+      "epoch": 8.76,
+      "grad_norm": 5.874046325683594,
+      "learning_rate": 9.222252146709143e-07,
+      "loss": 0.6308,
+      "step": 2190
+    },
+    {
+      "epoch": 8.78,
+      "grad_norm": 6.162435054779053,
+      "learning_rate": 8.931633919382299e-07,
+      "loss": 0.6282,
+      "step": 2195
+    },
+    {
+      "epoch": 8.8,
+      "grad_norm": 6.2155914306640625,
+      "learning_rate": 8.645454235739903e-07,
+      "loss": 0.6188,
+      "step": 2200
+    },
+    {
+      "epoch": 8.82,
+      "grad_norm": 6.127189636230469,
+      "learning_rate": 8.363727043776037e-07,
+      "loss": 0.6267,
+      "step": 2205
+    },
+    {
+      "epoch": 8.84,
+      "grad_norm": 6.407212734222412,
+      "learning_rate": 8.086466074476562e-07,
+      "loss": 0.6268,
+      "step": 2210
+    },
+    {
+      "epoch": 8.86,
+      "grad_norm": 6.001618385314941,
+      "learning_rate": 7.81368484114996e-07,
+      "loss": 0.6274,
+      "step": 2215
+    },
+    {
+      "epoch": 8.88,
+      "grad_norm": 6.074345111846924,
+      "learning_rate": 7.545396638768698e-07,
+      "loss": 0.6183,
+      "step": 2220
+    },
+    {
+      "epoch": 8.9,
+      "grad_norm": 6.457765579223633,
+      "learning_rate": 7.281614543321269e-07,
+      "loss": 0.6112,
+      "step": 2225
+    },
+    {
+      "epoch": 8.92,
+      "grad_norm": 5.76880407333374,
+      "learning_rate": 7.022351411174866e-07,
+      "loss": 0.6086,
+      "step": 2230
+    },
+    {
+      "epoch": 8.94,
+      "grad_norm": 6.315970420837402,
+      "learning_rate": 6.767619878448783e-07,
+      "loss": 0.6381,
+      "step": 2235
+    },
+    {
+      "epoch": 8.96,
+      "grad_norm": 5.965588569641113,
+      "learning_rate": 6.517432360398556e-07,
+      "loss": 0.6116,
+      "step": 2240
+    },
+    {
+      "epoch": 8.98,
+      "grad_norm": 5.967585563659668,
+      "learning_rate": 6.271801050810856e-07,
+      "loss": 0.643,
+      "step": 2245
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 6.2677226066589355,
+      "learning_rate": 6.030737921409169e-07,
+      "loss": 0.62,
+      "step": 2250
+    },
+    {
+      "epoch": 9.02,
+      "grad_norm": 5.486161231994629,
+      "learning_rate": 5.794254721270331e-07,
+      "loss": 0.464,
+      "step": 2255
+    },
+    {
+      "epoch": 9.04,
+      "grad_norm": 5.577027797698975,
+      "learning_rate": 5.562362976251901e-07,
+      "loss": 0.4631,
+      "step": 2260
+    },
+    {
+      "epoch": 9.06,
+      "grad_norm": 5.71274995803833,
+      "learning_rate": 5.335073988430373e-07,
+      "loss": 0.4612,
+      "step": 2265
+    },
+    {
+      "epoch": 9.08,
+      "grad_norm": 5.8267083168029785,
+      "learning_rate": 5.112398835550348e-07,
+      "loss": 0.4524,
+      "step": 2270
+    },
+    {
+      "epoch": 9.1,
+      "grad_norm": 5.931846618652344,
+      "learning_rate": 4.894348370484648e-07,
+      "loss": 0.4674,
+      "step": 2275
+    },
+    {
+      "epoch": 9.12,
+      "grad_norm": 5.749095916748047,
+      "learning_rate": 4.6809332207053083e-07,
+      "loss": 0.4544,
+      "step": 2280
+    },
+    {
+      "epoch": 9.14,
+      "grad_norm": 6.270796298980713,
+      "learning_rate": 4.4721637877656377e-07,
+      "loss": 0.4581,
+      "step": 2285
+    },
+    {
+      "epoch": 9.16,
+      "grad_norm": 5.706393718719482,
+      "learning_rate": 4.268050246793276e-07,
+      "loss": 0.4534,
+      "step": 2290
+    },
+    {
+      "epoch": 9.18,
+      "grad_norm": 5.894087791442871,
+      "learning_rate": 4.068602545994249e-07,
+      "loss": 0.456,
+      "step": 2295
+    },
+    {
+      "epoch": 9.2,
+      "grad_norm": 5.531440734863281,
+      "learning_rate": 3.8738304061681107e-07,
+      "loss": 0.4561,
+      "step": 2300
+    },
+    {
+      "epoch": 9.22,
+      "grad_norm": 5.7872314453125,
+      "learning_rate": 3.68374332023419e-07,
+      "loss": 0.4683,
+      "step": 2305
+    },
+    {
+      "epoch": 9.24,
+      "grad_norm": 5.619744777679443,
+      "learning_rate": 3.498350552768859e-07,
+      "loss": 0.4468,
+      "step": 2310
+    },
+    {
+      "epoch": 9.26,
+      "grad_norm": 5.65980863571167,
+      "learning_rate": 3.3176611395540625e-07,
+      "loss": 0.4558,
+      "step": 2315
+    },
+    {
+      "epoch": 9.28,
+      "grad_norm": 5.706433296203613,
+      "learning_rate": 3.1416838871368925e-07,
+      "loss": 0.451,
+      "step": 2320
+    },
+    {
+      "epoch": 9.3,
+      "grad_norm": 5.564632892608643,
+      "learning_rate": 2.970427372400353e-07,
+      "loss": 0.4597,
+      "step": 2325
+    },
+    {
+      "epoch": 9.32,
+      "grad_norm": 5.819154262542725,
+      "learning_rate": 2.8038999421453827e-07,
+      "loss": 0.4589,
+      "step": 2330
+    },
+    {
+      "epoch": 9.34,
+      "grad_norm": 5.930497169494629,
+      "learning_rate": 2.6421097126839714e-07,
+      "loss": 0.4668,
+      "step": 2335
+    },
+    {
+      "epoch": 9.36,
+      "grad_norm": 5.8165059089660645,
+      "learning_rate": 2.4850645694436736e-07,
+      "loss": 0.4481,
+      "step": 2340
+    },
+    {
+      "epoch": 9.38,
+      "grad_norm": 5.936621189117432,
+      "learning_rate": 2.332772166583208e-07,
+      "loss": 0.4527,
+      "step": 2345
+    },
+    {
+      "epoch": 9.4,
+      "grad_norm": 5.787903308868408,
+      "learning_rate": 2.1852399266194312e-07,
+      "loss": 0.4645,
+      "step": 2350
+    },
+    {
+      "epoch": 9.42,
+      "grad_norm": 6.234328269958496,
+      "learning_rate": 2.0424750400655947e-07,
+      "loss": 0.4722,
+      "step": 2355
+    },
+    {
+      "epoch": 9.44,
+      "grad_norm": 5.86660623550415,
+      "learning_rate": 1.9044844650808468e-07,
+      "loss": 0.4692,
+      "step": 2360
+    },
+    {
+      "epoch": 9.46,
+      "grad_norm": 6.366551876068115,
+      "learning_rate": 1.7712749271311392e-07,
+      "loss": 0.4828,
+      "step": 2365
+    },
+    {
+      "epoch": 9.48,
+      "grad_norm": 5.825527191162109,
+      "learning_rate": 1.6428529186614195e-07,
+      "loss": 0.4536,
+      "step": 2370
+    },
+    {
+      "epoch": 9.5,
+      "grad_norm": 5.744025230407715,
+      "learning_rate": 1.519224698779198e-07,
+      "loss": 0.4635,
+      "step": 2375
+    },
+    {
+      "epoch": 9.52,
+      "grad_norm": 6.037289142608643,
+      "learning_rate": 1.400396292949513e-07,
+      "loss": 0.4679,
+      "step": 2380
+    },
+    {
+      "epoch": 9.54,
+      "grad_norm": 5.628565311431885,
+      "learning_rate": 1.2863734927012094e-07,
+      "loss": 0.48,
+      "step": 2385
+    },
+    {
+      "epoch": 9.56,
+      "grad_norm": 6.118548393249512,
+      "learning_rate": 1.1771618553447217e-07,
+      "loss": 0.4501,
+      "step": 2390
+    },
+    {
+      "epoch": 9.58,
+      "grad_norm": 5.726400375366211,
+      "learning_rate": 1.0727667037011668e-07,
+      "loss": 0.4647,
+      "step": 2395
+    },
+    {
+      "epoch": 9.6,
+      "grad_norm": 5.941714286804199,
+      "learning_rate": 9.731931258429638e-08,
+      "loss": 0.4756,
+      "step": 2400
+    },
+    {
+      "epoch": 9.62,
+      "grad_norm": 5.818594455718994,
+      "learning_rate": 8.784459748458318e-08,
+      "loss": 0.4471,
+      "step": 2405
+    },
+    {
+      "epoch": 9.64,
+      "grad_norm": 5.669724464416504,
+      "learning_rate": 7.885298685522235e-08,
+      "loss": 0.4629,
+      "step": 2410
+    },
+    {
+      "epoch": 9.66,
+      "grad_norm": 6.057743549346924,
+      "learning_rate": 7.034491893463059e-08,
+      "loss": 0.4573,
+      "step": 2415
+    },
+    {
+      "epoch": 9.68,
+      "grad_norm": 5.836614608764648,
+      "learning_rate": 6.232080839403631e-08,
+      "loss": 0.4519,
+      "step": 2420
+    },
+    {
+      "epoch": 9.7,
+      "grad_norm": 5.849831581115723,
+      "learning_rate": 5.4781046317267103e-08,
+      "loss": 0.4679,
+      "step": 2425
+    },
+    {
+      "epoch": 9.72,
+      "grad_norm": 5.624325275421143,
+      "learning_rate": 4.772600018168816e-08,
+      "loss": 0.4458,
+      "step": 2430
+    },
+    {
+      "epoch": 9.74,
+      "grad_norm": 5.884892463684082,
+      "learning_rate": 4.115601384029666e-08,
+      "loss": 0.4679,
+      "step": 2435
+    },
+    {
+      "epoch": 9.76,
+      "grad_norm": 5.915401935577393,
+      "learning_rate": 3.50714075049563e-08,
+      "loss": 0.4603,
+      "step": 2440
+    },
+    {
+      "epoch": 9.78,
+      "grad_norm": 5.965377330780029,
+      "learning_rate": 2.947247773079753e-08,
+      "loss": 0.4726,
+      "step": 2445
+    },
+    {
+      "epoch": 9.8,
+      "grad_norm": 6.21063232421875,
+      "learning_rate": 2.4359497401758026e-08,
+      "loss": 0.4775,
+      "step": 2450
+    },
+    {
+      "epoch": 9.82,
+      "grad_norm": 5.652218341827393,
+      "learning_rate": 1.973271571728441e-08,
+      "loss": 0.435,
+      "step": 2455
+    },
+    {
+      "epoch": 9.84,
+      "grad_norm": 5.780144214630127,
+      "learning_rate": 1.5592358180189782e-08,
+      "loss": 0.4834,
+      "step": 2460
+    },
+    {
+      "epoch": 9.86,
+      "grad_norm": 5.7750563621521,
+      "learning_rate": 1.1938626585660252e-08,
+      "loss": 0.463,
+      "step": 2465
+    },
+    {
+      "epoch": 9.88,
+      "grad_norm": 5.517821788787842,
+      "learning_rate": 8.771699011416169e-09,
+      "loss": 0.4486,
+      "step": 2470
+    },
+    {
+      "epoch": 9.9,
+      "grad_norm": 5.808730602264404,
+      "learning_rate": 6.091729809042379e-09,
+      "loss": 0.468,
+      "step": 2475
+    },
+    {
+      "epoch": 9.92,
+      "grad_norm": 5.867649555206299,
+      "learning_rate": 3.898849596456477e-09,
+      "loss": 0.4439,
+      "step": 2480
+    },
+    {
+      "epoch": 9.94,
+      "grad_norm": 6.076911926269531,
+      "learning_rate": 2.193165251545004e-09,
+      "loss": 0.455,
+      "step": 2485
+    },
+    {
+      "epoch": 9.96,
+      "grad_norm": 5.9915971755981445,
+      "learning_rate": 9.74759906957612e-10,
+      "loss": 0.4597,
+      "step": 2490
+    },
+    {
+      "epoch": 9.98,
+      "grad_norm": 5.625442028045654,
+      "learning_rate": 2.436929460525317e-10,
+      "loss": 0.4575,
+      "step": 2495
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 5.879223346710205,
+      "learning_rate": 0.0,
+      "loss": 0.4807,
+      "step": 2500
+    },
+    {
+      "epoch": 10.0,
+      "step": 2500,
+      "total_flos": 2.880737320506491e+17,
+      "train_loss": 1.3902658111572266,
+      "train_runtime": 2102.4841,
+      "train_samples_per_second": 19.025,
+      "train_steps_per_second": 1.189
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 2500,
+  "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": 2.880737320506491e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e1ca13df9a03f9838bfc5697e73e66727f7e14098bf738d87ca59bd50f2ad51
+size 7160