Upload folder using huggingface_hub

.gitattributes

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

added_tokens.json

@@ -0,0 +1,28 @@
+{
+  "</think>": 151668,
+  "</tool_call>": 151658,
+  "</tool_response>": 151666,
+  "<think>": 151667,
+  "<tool_call>": 151657,
+  "<tool_response>": 151665,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

chat_template.jinja

@@ -0,0 +1,61 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if message.content is string %}
+        {%- set content = message.content %}
+    {%- else %}
+        {%- set content = '' %}
+    {%- endif %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

config.json

@@ -0,0 +1,39 @@
+{
+  "architectures": [
+    "Qwen3MoeForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 151644,
+  "decoder_sparse_step": 1,
+  "eos_token_id": 151645,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "max_position_embeddings": 262144,
+  "max_window_layers": 48,
+  "mlp_only_layers": [],
+  "model_type": "qwen3_moe",
+  "moe_intermediate_size": 768,
+  "norm_topk_prob": true,
+  "num_attention_heads": 32,
+  "num_experts": 128,
+  "num_experts_per_tok": 8,
+  "num_hidden_layers": 48,
+  "num_key_value_heads": 4,
+  "output_router_logits": false,
+  "pad_token_id": 151645,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 10000000,
+  "router_aux_loss_coef": 0.001,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.52.4",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}

generation_config.json

@@ -0,0 +1,10 @@
+{
+  "bos_token_id": 151644,
+  "do_sample": true,
+  "eos_token_id": 151645,
+  "pad_token_id": 151643,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.52.4"
+}

model-00001-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a2db94b3783dec0ade10d45e010bd1c41af6916d5a78b0bcac5ad71212fb81a4
+size 4997184968

model-00002-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a395363d19334a1cd94ad65a7899e4360122899b451211d8ebe392eae7427f03
+size 4997741608

model-00003-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d872b5d2178c702a9ef7ac611ac1ec26d4b86f6733f9cd57212cd0166996c6f
+size 4997742208

model-00004-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6665c751e86254e3af2aed212af6da5f6a7f4adde08e8f1840605b8da08617b1
+size 4997743184

model-00005-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa711dcca894ae074c0543b3579f713209a1a35eb5c1844cd9a39dc52b534d84
+size 4997743184

model-00006-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82bc78dc66aef8279ac1f7ba851156fe754caa2bcca5336e6af7ee3fedb3241f
+size 4997743184

model-00007-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e7255a4dea8d7e8be084d919f7ff5eb3170946b2c46c40a4492abb4ce0635b9
+size 4997743184

model-00008-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b8403cda24cabb037e4a7d85748bc88ec90361c60fae088fcca501943c04bfc
+size 4997743184

model-00009-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb62f93ee34830905e5541c1cc025ea963687ec184b380201ba144eaf73d6f5c
+size 4997743184

model-00010-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6cbbe1035d00031373c413d3cec8cb21a607f080178042d30c9bffc3657b21be
+size 4997743184

model-00011-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:160946733cabe8c57c74ec5e8a836e335345a92b3502c241a4e458e9065a78c5
+size 4997743184

model-00012-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75cdc9b259cb1aec199044983f93c91811dae1e4a1b10ddf35258bdb761a7789
+size 4997743184

model-00013-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ec5fddbff519abc8886a12224fbe5623fbd71aad95a1d302c669b96c57d82a4
+size 1094220288

special_tokens_map.json

@@ -0,0 +1,20 @@
+{
+  "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": "<|im_start|>",
+  "eos_token": "<|im_end|>",
+  "pad_token": "<|im_end|>"
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
+size 11422654

tokenizer_config.json

@@ -0,0 +1,239 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": "<|im_start|>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 1010000,
+  "pad_token": "<|im_end|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

trainer_state.json

@@ -0,0 +1,3739 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 528,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.005681818181818182,
+      "grad_norm": 4.13636899582678,
+      "learning_rate": 0.0,
+      "loss": 0.427,
+      "step": 1
+    },
+    {
+      "epoch": 0.011363636363636364,
+      "grad_norm": 5.690904269421442,
+      "learning_rate": 2.2727272727272729e-07,
+      "loss": 0.4263,
+      "step": 2
+    },
+    {
+      "epoch": 0.017045454545454544,
+      "grad_norm": 3.8643937954330965,
+      "learning_rate": 4.5454545454545457e-07,
+      "loss": 0.3878,
+      "step": 3
+    },
+    {
+      "epoch": 0.022727272727272728,
+      "grad_norm": 3.869278700723408,
+      "learning_rate": 6.818181818181818e-07,
+      "loss": 0.3965,
+      "step": 4
+    },
+    {
+      "epoch": 0.028409090909090908,
+      "grad_norm": 4.9499247122631065,
+      "learning_rate": 9.090909090909091e-07,
+      "loss": 0.5151,
+      "step": 5
+    },
+    {
+      "epoch": 0.03409090909090909,
+      "grad_norm": 4.008175065559179,
+      "learning_rate": 1.1363636363636364e-06,
+      "loss": 0.3988,
+      "step": 6
+    },
+    {
+      "epoch": 0.03977272727272727,
+      "grad_norm": 3.802187440592589,
+      "learning_rate": 1.3636363636363636e-06,
+      "loss": 0.3982,
+      "step": 7
+    },
+    {
+      "epoch": 0.045454545454545456,
+      "grad_norm": 3.5944086963537805,
+      "learning_rate": 1.590909090909091e-06,
+      "loss": 0.4356,
+      "step": 8
+    },
+    {
+      "epoch": 0.05113636363636364,
+      "grad_norm": 3.300357884289983,
+      "learning_rate": 1.8181818181818183e-06,
+      "loss": 0.3881,
+      "step": 9
+    },
+    {
+      "epoch": 0.056818181818181816,
+      "grad_norm": 2.419752440718465,
+      "learning_rate": 2.0454545454545457e-06,
+      "loss": 0.3891,
+      "step": 10
+    },
+    {
+      "epoch": 0.0625,
+      "grad_norm": 2.376164291933713,
+      "learning_rate": 2.2727272727272728e-06,
+      "loss": 0.4075,
+      "step": 11
+    },
+    {
+      "epoch": 0.06818181818181818,
+      "grad_norm": 1.8451276554587914,
+      "learning_rate": 2.5e-06,
+      "loss": 0.3605,
+      "step": 12
+    },
+    {
+      "epoch": 0.07386363636363637,
+      "grad_norm": 1.6725738451562537,
+      "learning_rate": 2.7272727272727272e-06,
+      "loss": 0.3423,
+      "step": 13
+    },
+    {
+      "epoch": 0.07954545454545454,
+      "grad_norm": 0.7493280852305411,
+      "learning_rate": 2.954545454545455e-06,
+      "loss": 0.3163,
+      "step": 14
+    },
+    {
+      "epoch": 0.08522727272727272,
+      "grad_norm": 0.6405813170837762,
+      "learning_rate": 3.181818181818182e-06,
+      "loss": 0.3241,
+      "step": 15
+    },
+    {
+      "epoch": 0.09090909090909091,
+      "grad_norm": 0.6010914848016558,
+      "learning_rate": 3.409090909090909e-06,
+      "loss": 0.351,
+      "step": 16
+    },
+    {
+      "epoch": 0.09659090909090909,
+      "grad_norm": 0.49883812733122435,
+      "learning_rate": 3.6363636363636366e-06,
+      "loss": 0.3153,
+      "step": 17
+    },
+    {
+      "epoch": 0.10227272727272728,
+      "grad_norm": 0.3797160566248851,
+      "learning_rate": 3.863636363636364e-06,
+      "loss": 0.3291,
+      "step": 18
+    },
+    {
+      "epoch": 0.10795454545454546,
+      "grad_norm": 0.6583932591823167,
+      "learning_rate": 4.0909090909090915e-06,
+      "loss": 0.3324,
+      "step": 19
+    },
+    {
+      "epoch": 0.11363636363636363,
+      "grad_norm": 0.7109283198888129,
+      "learning_rate": 4.3181818181818185e-06,
+      "loss": 0.2966,
+      "step": 20
+    },
+    {
+      "epoch": 0.11931818181818182,
+      "grad_norm": 0.7599768039513521,
+      "learning_rate": 4.5454545454545455e-06,
+      "loss": 0.3554,
+      "step": 21
+    },
+    {
+      "epoch": 0.125,
+      "grad_norm": 0.6406620909580979,
+      "learning_rate": 4.772727272727273e-06,
+      "loss": 0.3214,
+      "step": 22
+    },
+    {
+      "epoch": 0.13068181818181818,
+      "grad_norm": 0.6002184638663769,
+      "learning_rate": 5e-06,
+      "loss": 0.3522,
+      "step": 23
+    },
+    {
+      "epoch": 0.13636363636363635,
+      "grad_norm": 0.5064858699473549,
+      "learning_rate": 4.999951815503011e-06,
+      "loss": 0.315,
+      "step": 24
+    },
+    {
+      "epoch": 0.14204545454545456,
+      "grad_norm": 0.4253893803928414,
+      "learning_rate": 4.999807263869441e-06,
+      "loss": 0.3258,
+      "step": 25
+    },
+    {
+      "epoch": 0.14772727272727273,
+      "grad_norm": 0.3692592226306194,
+      "learning_rate": 4.999566350671405e-06,
+      "loss": 0.3332,
+      "step": 26
+    },
+    {
+      "epoch": 0.1534090909090909,
+      "grad_norm": 0.3351959335334293,
+      "learning_rate": 4.999229085195532e-06,
+      "loss": 0.3226,
+      "step": 27
+    },
+    {
+      "epoch": 0.1590909090909091,
+      "grad_norm": 0.3078710542803697,
+      "learning_rate": 4.998795480442595e-06,
+      "loss": 0.2877,
+      "step": 28
+    },
+    {
+      "epoch": 0.16477272727272727,
+      "grad_norm": 0.3717545484104801,
+      "learning_rate": 4.998265553127013e-06,
+      "loss": 0.2903,
+      "step": 29
+    },
+    {
+      "epoch": 0.17045454545454544,
+      "grad_norm": 0.3050884065792019,
+      "learning_rate": 4.997639323676214e-06,
+      "loss": 0.2902,
+      "step": 30
+    },
+    {
+      "epoch": 0.17613636363636365,
+      "grad_norm": 0.3009538798031033,
+      "learning_rate": 4.996916816229838e-06,
+      "loss": 0.2978,
+      "step": 31
+    },
+    {
+      "epoch": 0.18181818181818182,
+      "grad_norm": 0.34861390917413415,
+      "learning_rate": 4.99609805863881e-06,
+      "loss": 0.2778,
+      "step": 32
+    },
+    {
+      "epoch": 0.1875,
+      "grad_norm": 0.2884008619731284,
+      "learning_rate": 4.995183082464269e-06,
+      "loss": 0.2942,
+      "step": 33
+    },
+    {
+      "epoch": 0.19318181818181818,
+      "grad_norm": 0.2901730865490431,
+      "learning_rate": 4.994171922976349e-06,
+      "loss": 0.2987,
+      "step": 34
+    },
+    {
+      "epoch": 0.19886363636363635,
+      "grad_norm": 0.29290478785731666,
+      "learning_rate": 4.993064619152818e-06,
+      "loss": 0.3209,
+      "step": 35
+    },
+    {
+      "epoch": 0.20454545454545456,
+      "grad_norm": 0.35688808962883994,
+      "learning_rate": 4.991861213677578e-06,
+      "loss": 0.2962,
+      "step": 36
+    },
+    {
+      "epoch": 0.21022727272727273,
+      "grad_norm": 0.24675751243710786,
+      "learning_rate": 4.99056175293902e-06,
+      "loss": 0.2766,
+      "step": 37
+    },
+    {
+      "epoch": 0.2159090909090909,
+      "grad_norm": 0.3267376051155682,
+      "learning_rate": 4.989166287028234e-06,
+      "loss": 0.308,
+      "step": 38
+    },
+    {
+      "epoch": 0.2215909090909091,
+      "grad_norm": 0.22493166793665828,
+      "learning_rate": 4.987674869737078e-06,
+      "loss": 0.2742,
+      "step": 39
+    },
+    {
+      "epoch": 0.22727272727272727,
+      "grad_norm": 0.5317875361368901,
+      "learning_rate": 4.986087558556104e-06,
+      "loss": 0.2888,
+      "step": 40
+    },
+    {
+      "epoch": 0.23295454545454544,
+      "grad_norm": 0.29801026545030884,
+      "learning_rate": 4.984404414672346e-06,
+      "loss": 0.2924,
+      "step": 41
+    },
+    {
+      "epoch": 0.23863636363636365,
+      "grad_norm": 0.25799269953369336,
+      "learning_rate": 4.9826255029669575e-06,
+      "loss": 0.2816,
+      "step": 42
+    },
+    {
+      "epoch": 0.24431818181818182,
+      "grad_norm": 0.22240409365014543,
+      "learning_rate": 4.980750892012711e-06,
+      "loss": 0.2758,
+      "step": 43
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.2799907974670139,
+      "learning_rate": 4.978780654071355e-06,
+      "loss": 0.3098,
+      "step": 44
+    },
+    {
+      "epoch": 0.2556818181818182,
+      "grad_norm": 0.28876080918518277,
+      "learning_rate": 4.976714865090827e-06,
+      "loss": 0.2802,
+      "step": 45
+    },
+    {
+      "epoch": 0.26136363636363635,
+      "grad_norm": 0.24583878016647032,
+      "learning_rate": 4.974553604702332e-06,
+      "loss": 0.3026,
+      "step": 46
+    },
+    {
+      "epoch": 0.26704545454545453,
+      "grad_norm": 0.2656598065177574,
+      "learning_rate": 4.972296956217265e-06,
+      "loss": 0.2961,
+      "step": 47
+    },
+    {
+      "epoch": 0.2727272727272727,
+      "grad_norm": 0.21012707248222598,
+      "learning_rate": 4.969945006624003e-06,
+      "loss": 0.2861,
+      "step": 48
+    },
+    {
+      "epoch": 0.2784090909090909,
+      "grad_norm": 0.2943742503501893,
+      "learning_rate": 4.967497846584552e-06,
+      "loss": 0.2948,
+      "step": 49
+    },
+    {
+      "epoch": 0.2840909090909091,
+      "grad_norm": 0.20799525994994228,
+      "learning_rate": 4.9649555704310545e-06,
+      "loss": 0.3063,
+      "step": 50
+    },
+    {
+      "epoch": 0.2897727272727273,
+      "grad_norm": 0.3249940372924672,
+      "learning_rate": 4.962318276162148e-06,
+      "loss": 0.2503,
+      "step": 51
+    },
+    {
+      "epoch": 0.29545454545454547,
+      "grad_norm": 0.3426841066756318,
+      "learning_rate": 4.959586065439189e-06,
+      "loss": 0.2914,
+      "step": 52
+    },
+    {
+      "epoch": 0.30113636363636365,
+      "grad_norm": 0.22855678787775643,
+      "learning_rate": 4.956759043582339e-06,
+      "loss": 0.2823,
+      "step": 53
+    },
+    {
+      "epoch": 0.3068181818181818,
+      "grad_norm": 0.247964442806115,
+      "learning_rate": 4.953837319566498e-06,
+      "loss": 0.2796,
+      "step": 54
+    },
+    {
+      "epoch": 0.3125,
+      "grad_norm": 0.2332518969431367,
+      "learning_rate": 4.950821006017107e-06,
+      "loss": 0.2611,
+      "step": 55
+    },
+    {
+      "epoch": 0.3181818181818182,
+      "grad_norm": 0.21011622253491574,
+      "learning_rate": 4.947710219205808e-06,
+      "loss": 0.2738,
+      "step": 56
+    },
+    {
+      "epoch": 0.32386363636363635,
+      "grad_norm": 0.3365637065974888,
+      "learning_rate": 4.9445050790459585e-06,
+      "loss": 0.2984,
+      "step": 57
+    },
+    {
+      "epoch": 0.32954545454545453,
+      "grad_norm": 0.2539746858405654,
+      "learning_rate": 4.9412057090880115e-06,
+      "loss": 0.2725,
+      "step": 58
+    },
+    {
+      "epoch": 0.3352272727272727,
+      "grad_norm": 0.3967914747485111,
+      "learning_rate": 4.937812236514754e-06,
+      "loss": 0.2752,
+      "step": 59
+    },
+    {
+      "epoch": 0.3409090909090909,
+      "grad_norm": 0.2392453733971367,
+      "learning_rate": 4.9343247921364e-06,
+      "loss": 0.2608,
+      "step": 60
+    },
+    {
+      "epoch": 0.3465909090909091,
+      "grad_norm": 0.24816417648181116,
+      "learning_rate": 4.930743510385551e-06,
+      "loss": 0.2968,
+      "step": 61
+    },
+    {
+      "epoch": 0.3522727272727273,
+      "grad_norm": 0.19303548298632028,
+      "learning_rate": 4.927068529312017e-06,
+      "loss": 0.2766,
+      "step": 62
+    },
+    {
+      "epoch": 0.35795454545454547,
+      "grad_norm": 0.31205417275979963,
+      "learning_rate": 4.923299990577488e-06,
+      "loss": 0.2646,
+      "step": 63
+    },
+    {
+      "epoch": 0.36363636363636365,
+      "grad_norm": 0.27131751442758495,
+      "learning_rate": 4.919438039450079e-06,
+      "loss": 0.294,
+      "step": 64
+    },
+    {
+      "epoch": 0.3693181818181818,
+      "grad_norm": 0.22222852606773144,
+      "learning_rate": 4.915482824798728e-06,
+      "loss": 0.2781,
+      "step": 65
+    },
+    {
+      "epoch": 0.375,
+      "grad_norm": 0.2118792282893684,
+      "learning_rate": 4.911434499087457e-06,
+      "loss": 0.2741,
+      "step": 66
+    },
+    {
+      "epoch": 0.3806818181818182,
+      "grad_norm": 0.5758672380246893,
+      "learning_rate": 4.907293218369499e-06,
+      "loss": 0.2896,
+      "step": 67
+    },
+    {
+      "epoch": 0.38636363636363635,
+      "grad_norm": 0.20085333233060845,
+      "learning_rate": 4.903059142281273e-06,
+      "loss": 0.2884,
+      "step": 68
+    },
+    {
+      "epoch": 0.39204545454545453,
+      "grad_norm": 0.2654667792604945,
+      "learning_rate": 4.8987324340362445e-06,
+      "loss": 0.2972,
+      "step": 69
+    },
+    {
+      "epoch": 0.3977272727272727,
+      "grad_norm": 0.19331704344701295,
+      "learning_rate": 4.894313260418617e-06,
+      "loss": 0.2667,
+      "step": 70
+    },
+    {
+      "epoch": 0.4034090909090909,
+      "grad_norm": 0.2171929345597126,
+      "learning_rate": 4.889801791776921e-06,
+      "loss": 0.2781,
+      "step": 71
+    },
+    {
+      "epoch": 0.4090909090909091,
+      "grad_norm": 0.2684319221291861,
+      "learning_rate": 4.885198202017431e-06,
+      "loss": 0.282,
+      "step": 72
+    },
+    {
+      "epoch": 0.4147727272727273,
+      "grad_norm": 0.23776136068594306,
+      "learning_rate": 4.880502668597475e-06,
+      "loss": 0.2863,
+      "step": 73
+    },
+    {
+      "epoch": 0.42045454545454547,
+      "grad_norm": 0.2198068186762879,
+      "learning_rate": 4.875715372518585e-06,
+      "loss": 0.2791,
+      "step": 74
+    },
+    {
+      "epoch": 0.42613636363636365,
+      "grad_norm": 0.21520734900564198,
+      "learning_rate": 4.870836498319523e-06,
+      "loss": 0.2857,
+      "step": 75
+    },
+    {
+      "epoch": 0.4318181818181818,
+      "grad_norm": 0.43054319607393365,
+      "learning_rate": 4.865866234069169e-06,
+      "loss": 0.2907,
+      "step": 76
+    },
+    {
+      "epoch": 0.4375,
+      "grad_norm": 0.26218468258954264,
+      "learning_rate": 4.86080477135927e-06,
+      "loss": 0.2798,
+      "step": 77
+    },
+    {
+      "epoch": 0.4431818181818182,
+      "grad_norm": 0.24308410366408476,
+      "learning_rate": 4.855652305297052e-06,
+      "loss": 0.2886,
+      "step": 78
+    },
+    {
+      "epoch": 0.44886363636363635,
+      "grad_norm": 0.17664891091831125,
+      "learning_rate": 4.8504090344977036e-06,
+      "loss": 0.2511,
+      "step": 79
+    },
+    {
+      "epoch": 0.45454545454545453,
+      "grad_norm": 0.197319451098218,
+      "learning_rate": 4.84507516107672e-06,
+      "loss": 0.2749,
+      "step": 80
+    },
+    {
+      "epoch": 0.4602272727272727,
+      "grad_norm": 0.6302897592568023,
+      "learning_rate": 4.839650890642104e-06,
+      "loss": 0.2869,
+      "step": 81
+    },
+    {
+      "epoch": 0.4659090909090909,
+      "grad_norm": 0.23377209749002553,
+      "learning_rate": 4.834136432286452e-06,
+      "loss": 0.2939,
+      "step": 82
+    },
+    {
+      "epoch": 0.4715909090909091,
+      "grad_norm": 0.25817574948385447,
+      "learning_rate": 4.828531998578885e-06,
+      "loss": 0.2802,
+      "step": 83
+    },
+    {
+      "epoch": 0.4772727272727273,
+      "grad_norm": 0.27536793241671365,
+      "learning_rate": 4.822837805556858e-06,
+      "loss": 0.289,
+      "step": 84
+    },
+    {
+      "epoch": 0.48295454545454547,
+      "grad_norm": 0.25733930454992526,
+      "learning_rate": 4.817054072717833e-06,
+      "loss": 0.2574,
+      "step": 85
+    },
+    {
+      "epoch": 0.48863636363636365,
+      "grad_norm": 0.19536546286609527,
+      "learning_rate": 4.811181023010815e-06,
+      "loss": 0.2692,
+      "step": 86
+    },
+    {
+      "epoch": 0.4943181818181818,
+      "grad_norm": 0.227177625973573,
+      "learning_rate": 4.805218882827761e-06,
+      "loss": 0.254,
+      "step": 87
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.22612797471836377,
+      "learning_rate": 4.799167881994852e-06,
+      "loss": 0.2716,
+      "step": 88
+    },
+    {
+      "epoch": 0.5056818181818182,
+      "grad_norm": 0.3023768741873149,
+      "learning_rate": 4.793028253763633e-06,
+      "loss": 0.2911,
+      "step": 89
+    },
+    {
+      "epoch": 0.5113636363636364,
+      "grad_norm": 0.22087623835911116,
+      "learning_rate": 4.786800234802022e-06,
+      "loss": 0.2821,
+      "step": 90
+    },
+    {
+      "epoch": 0.5170454545454546,
+      "grad_norm": 0.21731422655095442,
+      "learning_rate": 4.780484065185188e-06,
+      "loss": 0.2695,
+      "step": 91
+    },
+    {
+      "epoch": 0.5227272727272727,
+      "grad_norm": 0.2979687908352143,
+      "learning_rate": 4.7740799883862966e-06,
+      "loss": 0.2956,
+      "step": 92
+    },
+    {
+      "epoch": 0.5284090909090909,
+      "grad_norm": 0.24392237794403746,
+      "learning_rate": 4.767588251267121e-06,
+      "loss": 0.2448,
+      "step": 93
+    },
+    {
+      "epoch": 0.5340909090909091,
+      "grad_norm": 0.49788496215188893,
+      "learning_rate": 4.761009104068533e-06,
+      "loss": 0.2829,
+      "step": 94
+    },
+    {
+      "epoch": 0.5397727272727273,
+      "grad_norm": 0.18919038139111577,
+      "learning_rate": 4.754342800400852e-06,
+      "loss": 0.2617,
+      "step": 95
+    },
+    {
+      "epoch": 0.5454545454545454,
+      "grad_norm": 0.2422601005721465,
+      "learning_rate": 4.747589597234068e-06,
+      "loss": 0.2511,
+      "step": 96
+    },
+    {
+      "epoch": 0.5511363636363636,
+      "grad_norm": 0.22463058582826498,
+      "learning_rate": 4.740749754887939e-06,
+      "loss": 0.2551,
+      "step": 97
+    },
+    {
+      "epoch": 0.5568181818181818,
+      "grad_norm": 0.24946343736182136,
+      "learning_rate": 4.7338235370219556e-06,
+      "loss": 0.2568,
+      "step": 98
+    },
+    {
+      "epoch": 0.5625,
+      "grad_norm": 0.22126538984664773,
+      "learning_rate": 4.726811210625176e-06,
+      "loss": 0.245,
+      "step": 99
+    },
+    {
+      "epoch": 0.5681818181818182,
+      "grad_norm": 0.20738868709242012,
+      "learning_rate": 4.7197130460059385e-06,
+      "loss": 0.2622,
+      "step": 100
+    },
+    {
+      "epoch": 0.5738636363636364,
+      "grad_norm": 0.2683170690250489,
+      "learning_rate": 4.712529316781435e-06,
+      "loss": 0.2514,
+      "step": 101
+    },
+    {
+      "epoch": 0.5795454545454546,
+      "grad_norm": 0.23542588925837882,
+      "learning_rate": 4.705260299867169e-06,
+      "loss": 0.2845,
+      "step": 102
+    },
+    {
+      "epoch": 0.5852272727272727,
+      "grad_norm": 0.24502311943242538,
+      "learning_rate": 4.697906275466279e-06,
+      "loss": 0.2776,
+      "step": 103
+    },
+    {
+      "epoch": 0.5909090909090909,
+      "grad_norm": 0.20760065932738048,
+      "learning_rate": 4.69046752705874e-06,
+      "loss": 0.2466,
+      "step": 104
+    },
+    {
+      "epoch": 0.5965909090909091,
+      "grad_norm": 0.20102638793251298,
+      "learning_rate": 4.682944341390431e-06,
+      "loss": 0.2525,
+      "step": 105
+    },
+    {
+      "epoch": 0.6022727272727273,
+      "grad_norm": 0.2065096502849592,
+      "learning_rate": 4.675337008462085e-06,
+      "loss": 0.2868,
+      "step": 106
+    },
+    {
+      "epoch": 0.6079545454545454,
+      "grad_norm": 0.2508016307640096,
+      "learning_rate": 4.667645821518111e-06,
+      "loss": 0.3024,
+      "step": 107
+    },
+    {
+      "epoch": 0.6136363636363636,
+      "grad_norm": 0.21666808347183772,
+      "learning_rate": 4.659871077035289e-06,
+      "loss": 0.2734,
+      "step": 108
+    },
+    {
+      "epoch": 0.6193181818181818,
+      "grad_norm": 0.23182976627931445,
+      "learning_rate": 4.65201307471134e-06,
+      "loss": 0.2647,
+      "step": 109
+    },
+    {
+      "epoch": 0.625,
+      "grad_norm": 0.20981496328273563,
+      "learning_rate": 4.644072117453377e-06,
+      "loss": 0.2754,
+      "step": 110
+    },
+    {
+      "epoch": 0.6306818181818182,
+      "grad_norm": 0.26326725184511524,
+      "learning_rate": 4.636048511366222e-06,
+      "loss": 0.2919,
+      "step": 111
+    },
+    {
+      "epoch": 0.6363636363636364,
+      "grad_norm": 0.262563807536366,
+      "learning_rate": 4.6279425657406154e-06,
+      "loss": 0.2926,
+      "step": 112
+    },
+    {
+      "epoch": 0.6420454545454546,
+      "grad_norm": 0.19306379242380328,
+      "learning_rate": 4.619754593041287e-06,
+      "loss": 0.2881,
+      "step": 113
+    },
+    {
+      "epoch": 0.6477272727272727,
+      "grad_norm": 0.26239388814091164,
+      "learning_rate": 4.6114849088949146e-06,
+      "loss": 0.2784,
+      "step": 114
+    },
+    {
+      "epoch": 0.6534090909090909,
+      "grad_norm": 0.2843429012137256,
+      "learning_rate": 4.603133832077953e-06,
+      "loss": 0.2596,
+      "step": 115
+    },
+    {
+      "epoch": 0.6590909090909091,
+      "grad_norm": 0.21245751576381866,
+      "learning_rate": 4.594701684504352e-06,
+      "loss": 0.2398,
+      "step": 116
+    },
+    {
+      "epoch": 0.6647727272727273,
+      "grad_norm": 0.22787695260596558,
+      "learning_rate": 4.586188791213143e-06,
+      "loss": 0.2711,
+      "step": 117
+    },
+    {
+      "epoch": 0.6704545454545454,
+      "grad_norm": 0.3818816711935563,
+      "learning_rate": 4.577595480355911e-06,
+      "loss": 0.2661,
+      "step": 118
+    },
+    {
+      "epoch": 0.6761363636363636,
+      "grad_norm": 0.23355384107293414,
+      "learning_rate": 4.568922083184144e-06,
+      "loss": 0.234,
+      "step": 119
+    },
+    {
+      "epoch": 0.6818181818181818,
+      "grad_norm": 0.18288553831071802,
+      "learning_rate": 4.560168934036467e-06,
+      "loss": 0.2552,
+      "step": 120
+    },
+    {
+      "epoch": 0.6875,
+      "grad_norm": 0.24924773346049447,
+      "learning_rate": 4.55133637032575e-06,
+      "loss": 0.2653,
+      "step": 121
+    },
+    {
+      "epoch": 0.6931818181818182,
+      "grad_norm": 0.20852872060779476,
+      "learning_rate": 4.542424732526105e-06,
+      "loss": 0.2505,
+      "step": 122
+    },
+    {
+      "epoch": 0.6988636363636364,
+      "grad_norm": 0.24118571168870664,
+      "learning_rate": 4.533434364159761e-06,
+      "loss": 0.2399,
+      "step": 123
+    },
+    {
+      "epoch": 0.7045454545454546,
+      "grad_norm": 0.20144416490181125,
+      "learning_rate": 4.524365611783818e-06,
+      "loss": 0.2417,
+      "step": 124
+    },
+    {
+      "epoch": 0.7102272727272727,
+      "grad_norm": 0.2262954076430847,
+      "learning_rate": 4.515218824976895e-06,
+      "loss": 0.2672,
+      "step": 125
+    },
+    {
+      "epoch": 0.7159090909090909,
+      "grad_norm": 0.20857659711513643,
+      "learning_rate": 4.505994356325648e-06,
+      "loss": 0.2413,
+      "step": 126
+    },
+    {
+      "epoch": 0.7215909090909091,
+      "grad_norm": 0.1840485932451498,
+      "learning_rate": 4.496692561411182e-06,
+      "loss": 0.2431,
+      "step": 127
+    },
+    {
+      "epoch": 0.7272727272727273,
+      "grad_norm": 0.21559446578225816,
+      "learning_rate": 4.487313798795347e-06,
+      "loss": 0.2266,
+      "step": 128
+    },
+    {
+      "epoch": 0.7329545454545454,
+      "grad_norm": 0.31855427106346423,
+      "learning_rate": 4.477858430006906e-06,
+      "loss": 0.2578,
+      "step": 129
+    },
+    {
+      "epoch": 0.7386363636363636,
+      "grad_norm": 0.2327594094385503,
+      "learning_rate": 4.468326819527613e-06,
+      "loss": 0.2493,
+      "step": 130
+    },
+    {
+      "epoch": 0.7443181818181818,
+      "grad_norm": 0.27069782646168544,
+      "learning_rate": 4.458719334778153e-06,
+      "loss": 0.2853,
+      "step": 131
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.20290888911590216,
+      "learning_rate": 4.449036346103982e-06,
+      "loss": 0.2525,
+      "step": 132
+    },
+    {
+      "epoch": 0.7556818181818182,
+      "grad_norm": 0.38497870978848986,
+      "learning_rate": 4.43927822676105e-06,
+      "loss": 0.2573,
+      "step": 133
+    },
+    {
+      "epoch": 0.7613636363636364,
+      "grad_norm": 0.2134895810331637,
+      "learning_rate": 4.429445352901415e-06,
+      "loss": 0.2681,
+      "step": 134
+    },
+    {
+      "epoch": 0.7670454545454546,
+      "grad_norm": 0.18517860586217338,
+      "learning_rate": 4.419538103558742e-06,
+      "loss": 0.2199,
+      "step": 135
+    },
+    {
+      "epoch": 0.7727272727272727,
+      "grad_norm": 0.30958029717950675,
+      "learning_rate": 4.409556860633692e-06,
+      "loss": 0.2784,
+      "step": 136
+    },
+    {
+      "epoch": 0.7784090909090909,
+      "grad_norm": 0.21154899257882034,
+      "learning_rate": 4.3995020088792e-06,
+      "loss": 0.2684,
+      "step": 137
+    },
+    {
+      "epoch": 0.7840909090909091,
+      "grad_norm": 0.23099439084231488,
+      "learning_rate": 4.3893739358856465e-06,
+      "loss": 0.3017,
+      "step": 138
+    },
+    {
+      "epoch": 0.7897727272727273,
+      "grad_norm": 0.2319823375879864,
+      "learning_rate": 4.379173032065912e-06,
+      "loss": 0.2368,
+      "step": 139
+    },
+    {
+      "epoch": 0.7954545454545454,
+      "grad_norm": 0.36801732527860587,
+      "learning_rate": 4.368899690640333e-06,
+      "loss": 0.2699,
+      "step": 140
+    },
+    {
+      "epoch": 0.8011363636363636,
+      "grad_norm": 0.23529268911329576,
+      "learning_rate": 4.3585543076215405e-06,
+      "loss": 0.2698,
+      "step": 141
+    },
+    {
+      "epoch": 0.8068181818181818,
+      "grad_norm": 0.2184490414418197,
+      "learning_rate": 4.3481372817991976e-06,
+      "loss": 0.2631,
+      "step": 142
+    },
+    {
+      "epoch": 0.8125,
+      "grad_norm": 0.439487873206488,
+      "learning_rate": 4.3376490147246205e-06,
+      "loss": 0.2673,
+      "step": 143
+    },
+    {
+      "epoch": 0.8181818181818182,
+      "grad_norm": 0.26855989609963915,
+      "learning_rate": 4.32708991069531e-06,
+      "loss": 0.2698,
+      "step": 144
+    },
+    {
+      "epoch": 0.8238636363636364,
+      "grad_norm": 0.3543130370224819,
+      "learning_rate": 4.31646037673936e-06,
+      "loss": 0.2627,
+      "step": 145
+    },
+    {
+      "epoch": 0.8295454545454546,
+      "grad_norm": 0.20174717662069117,
+      "learning_rate": 4.305760822599766e-06,
+      "loss": 0.2806,
+      "step": 146
+    },
+    {
+      "epoch": 0.8352272727272727,
+      "grad_norm": 0.21956421754738986,
+      "learning_rate": 4.294991660718636e-06,
+      "loss": 0.2592,
+      "step": 147
+    },
+    {
+      "epoch": 0.8409090909090909,
+      "grad_norm": 0.2101546633291597,
+      "learning_rate": 4.284153306221289e-06,
+      "loss": 0.272,
+      "step": 148
+    },
+    {
+      "epoch": 0.8465909090909091,
+      "grad_norm": 0.20128651667264633,
+      "learning_rate": 4.273246176900252e-06,
+      "loss": 0.236,
+      "step": 149
+    },
+    {
+      "epoch": 0.8522727272727273,
+      "grad_norm": 0.24737021101483003,
+      "learning_rate": 4.262270693199159e-06,
+      "loss": 0.2634,
+      "step": 150
+    },
+    {
+      "epoch": 0.8579545454545454,
+      "grad_norm": 0.22448605023421997,
+      "learning_rate": 4.2512272781965355e-06,
+      "loss": 0.2438,
+      "step": 151
+    },
+    {
+      "epoch": 0.8636363636363636,
+      "grad_norm": 0.34900609419270606,
+      "learning_rate": 4.240116357589502e-06,
+      "loss": 0.2514,
+      "step": 152
+    },
+    {
+      "epoch": 0.8693181818181818,
+      "grad_norm": 0.1935622280843455,
+      "learning_rate": 4.228938359677354e-06,
+      "loss": 0.2388,
+      "step": 153
+    },
+    {
+      "epoch": 0.875,
+      "grad_norm": 0.25773072503962624,
+      "learning_rate": 4.217693715345057e-06,
+      "loss": 0.2531,
+      "step": 154
+    },
+    {
+      "epoch": 0.8806818181818182,
+      "grad_norm": 0.22041420459104752,
+      "learning_rate": 4.206382858046636e-06,
+      "loss": 0.2636,
+      "step": 155
+    },
+    {
+      "epoch": 0.8863636363636364,
+      "grad_norm": 0.22347272836568208,
+      "learning_rate": 4.195006223788466e-06,
+      "loss": 0.3136,
+      "step": 156
+    },
+    {
+      "epoch": 0.8920454545454546,
+      "grad_norm": 0.23516522146462165,
+      "learning_rate": 4.183564251112466e-06,
+      "loss": 0.2627,
+      "step": 157
+    },
+    {
+      "epoch": 0.8977272727272727,
+      "grad_norm": 0.2245388610780939,
+      "learning_rate": 4.172057381079196e-06,
+      "loss": 0.2652,
+      "step": 158
+    },
+    {
+      "epoch": 0.9034090909090909,
+      "grad_norm": 0.26988373559964796,
+      "learning_rate": 4.160486057250849e-06,
+      "loss": 0.2557,
+      "step": 159
+    },
+    {
+      "epoch": 0.9090909090909091,
+      "grad_norm": 0.6872792618362236,
+      "learning_rate": 4.148850725674162e-06,
+      "loss": 0.2839,
+      "step": 160
+    },
+    {
+      "epoch": 0.9147727272727273,
+      "grad_norm": 0.20758056768557093,
+      "learning_rate": 4.137151834863213e-06,
+      "loss": 0.2677,
+      "step": 161
+    },
+    {
+      "epoch": 0.9204545454545454,
+      "grad_norm": 0.19698361252403598,
+      "learning_rate": 4.125389835782138e-06,
+      "loss": 0.2611,
+      "step": 162
+    },
+    {
+      "epoch": 0.9261363636363636,
+      "grad_norm": 0.2634401952443913,
+      "learning_rate": 4.113565181827745e-06,
+      "loss": 0.2541,
+      "step": 163
+    },
+    {
+      "epoch": 0.9318181818181818,
+      "grad_norm": 0.22955498966434157,
+      "learning_rate": 4.101678328812035e-06,
+      "loss": 0.2519,
+      "step": 164
+    },
+    {
+      "epoch": 0.9375,
+      "grad_norm": 0.21505253408914526,
+      "learning_rate": 4.0897297349446345e-06,
+      "loss": 0.2366,
+      "step": 165
+    },
+    {
+      "epoch": 0.9431818181818182,
+      "grad_norm": 0.22541600048527471,
+      "learning_rate": 4.077719860815132e-06,
+      "loss": 0.2787,
+      "step": 166
+    },
+    {
+      "epoch": 0.9488636363636364,
+      "grad_norm": 0.2270297479035574,
+      "learning_rate": 4.065649169375324e-06,
+      "loss": 0.2745,
+      "step": 167
+    },
+    {
+      "epoch": 0.9545454545454546,
+      "grad_norm": 0.21939653618157418,
+      "learning_rate": 4.053518125921365e-06,
+      "loss": 0.2381,
+      "step": 168
+    },
+    {
+      "epoch": 0.9602272727272727,
+      "grad_norm": 0.23630408488271606,
+      "learning_rate": 4.041327198075838e-06,
+      "loss": 0.2524,
+      "step": 169
+    },
+    {
+      "epoch": 0.9659090909090909,
+      "grad_norm": 1.0985204504913177,
+      "learning_rate": 4.029076855769722e-06,
+      "loss": 0.2398,
+      "step": 170
+    },
+    {
+      "epoch": 0.9715909090909091,
+      "grad_norm": 0.20401818951431577,
+      "learning_rate": 4.016767571224285e-06,
+      "loss": 0.2618,
+      "step": 171
+    },
+    {
+      "epoch": 0.9772727272727273,
+      "grad_norm": 0.27265229854495676,
+      "learning_rate": 4.0043998189328705e-06,
+      "loss": 0.2384,
+      "step": 172
+    },
+    {
+      "epoch": 0.9829545454545454,
+      "grad_norm": 0.18755154669027838,
+      "learning_rate": 3.991974075642621e-06,
+      "loss": 0.2105,
+      "step": 173
+    },
+    {
+      "epoch": 0.9886363636363636,
+      "grad_norm": 0.23111041332184715,
+      "learning_rate": 3.9794908203360865e-06,
+      "loss": 0.2796,
+      "step": 174
+    },
+    {
+      "epoch": 0.9943181818181818,
+      "grad_norm": 0.22836211042533255,
+      "learning_rate": 3.966950534212769e-06,
+      "loss": 0.2476,
+      "step": 175
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.25532820133220246,
+      "learning_rate": 3.954353700670573e-06,
+      "loss": 0.2407,
+      "step": 176
+    },
+    {
+      "epoch": 1.0056818181818181,
+      "grad_norm": 0.3070527658349344,
+      "learning_rate": 3.941700805287169e-06,
+      "loss": 0.2739,
+      "step": 177
+    },
+    {
+      "epoch": 1.0113636363636365,
+      "grad_norm": 0.27024017776926557,
+      "learning_rate": 3.9289923358012735e-06,
+      "loss": 0.2906,
+      "step": 178
+    },
+    {
+      "epoch": 1.0170454545454546,
+      "grad_norm": 0.3601678496407945,
+      "learning_rate": 3.9162287820938575e-06,
+      "loss": 0.2576,
+      "step": 179
+    },
+    {
+      "epoch": 1.0227272727272727,
+      "grad_norm": 0.234846494439447,
+      "learning_rate": 3.903410636169252e-06,
+      "loss": 0.2721,
+      "step": 180
+    },
+    {
+      "epoch": 1.0284090909090908,
+      "grad_norm": 0.21894893703936477,
+      "learning_rate": 3.890538392136188e-06,
+      "loss": 0.2204,
+      "step": 181
+    },
+    {
+      "epoch": 1.0340909090909092,
+      "grad_norm": 0.23511239991354554,
+      "learning_rate": 3.877612546188749e-06,
+      "loss": 0.2821,
+      "step": 182
+    },
+    {
+      "epoch": 1.0397727272727273,
+      "grad_norm": 0.26371937696669423,
+      "learning_rate": 3.864633596587242e-06,
+      "loss": 0.2992,
+      "step": 183
+    },
+    {
+      "epoch": 1.0454545454545454,
+      "grad_norm": 0.21822331732453468,
+      "learning_rate": 3.8516020436389945e-06,
+      "loss": 0.2366,
+      "step": 184
+    },
+    {
+      "epoch": 1.0511363636363635,
+      "grad_norm": 0.19304217881994865,
+      "learning_rate": 3.838518389679065e-06,
+      "loss": 0.2236,
+      "step": 185
+    },
+    {
+      "epoch": 1.0568181818181819,
+      "grad_norm": 0.46074983555399596,
+      "learning_rate": 3.825383139050881e-06,
+      "loss": 0.2579,
+      "step": 186
+    },
+    {
+      "epoch": 1.0625,
+      "grad_norm": 0.3864313032500556,
+      "learning_rate": 3.812196798086799e-06,
+      "loss": 0.2604,
+      "step": 187
+    },
+    {
+      "epoch": 1.0681818181818181,
+      "grad_norm": 0.22907818535523758,
+      "learning_rate": 3.798959875088584e-06,
+      "loss": 0.2429,
+      "step": 188
+    },
+    {
+      "epoch": 1.0738636363636365,
+      "grad_norm": 0.4564861968720899,
+      "learning_rate": 3.7856728803078168e-06,
+      "loss": 0.288,
+      "step": 189
+    },
+    {
+      "epoch": 1.0795454545454546,
+      "grad_norm": 0.21128235148404068,
+      "learning_rate": 3.7723363259262253e-06,
+      "loss": 0.2596,
+      "step": 190
+    },
+    {
+      "epoch": 1.0852272727272727,
+      "grad_norm": 0.24544408764038442,
+      "learning_rate": 3.75895072603594e-06,
+      "loss": 0.2966,
+      "step": 191
+    },
+    {
+      "epoch": 1.0909090909090908,
+      "grad_norm": 0.4425637441419803,
+      "learning_rate": 3.7455165966196817e-06,
+      "loss": 0.2813,
+      "step": 192
+    },
+    {
+      "epoch": 1.0965909090909092,
+      "grad_norm": 0.23114254677192722,
+      "learning_rate": 3.732034455530863e-06,
+      "loss": 0.2416,
+      "step": 193
+    },
+    {
+      "epoch": 1.1022727272727273,
+      "grad_norm": 0.5892208013523996,
+      "learning_rate": 3.718504822473634e-06,
+      "loss": 0.2575,
+      "step": 194
+    },
+    {
+      "epoch": 1.1079545454545454,
+      "grad_norm": 0.27120988659913997,
+      "learning_rate": 3.704928218982845e-06,
+      "loss": 0.2483,
+      "step": 195
+    },
+    {
+      "epoch": 1.1136363636363635,
+      "grad_norm": 0.27030517090763195,
+      "learning_rate": 3.6913051684039435e-06,
+      "loss": 0.2869,
+      "step": 196
+    },
+    {
+      "epoch": 1.1193181818181819,
+      "grad_norm": 0.24721081702866537,
+      "learning_rate": 3.6776361958728025e-06,
+      "loss": 0.2536,
+      "step": 197
+    },
+    {
+      "epoch": 1.125,
+      "grad_norm": 0.2257562078807439,
+      "learning_rate": 3.663921828295474e-06,
+      "loss": 0.2498,
+      "step": 198
+    },
+    {
+      "epoch": 1.1306818181818181,
+      "grad_norm": 0.2599699973969289,
+      "learning_rate": 3.650162594327881e-06,
+      "loss": 0.2632,
+      "step": 199
+    },
+    {
+      "epoch": 1.1363636363636362,
+      "grad_norm": 0.29494278610069274,
+      "learning_rate": 3.6363590243554362e-06,
+      "loss": 0.2758,
+      "step": 200
+    },
+    {
+      "epoch": 1.1420454545454546,
+      "grad_norm": 0.250508527226537,
+      "learning_rate": 3.6225116504726014e-06,
+      "loss": 0.2579,
+      "step": 201
+    },
+    {
+      "epoch": 1.1477272727272727,
+      "grad_norm": 0.24509958653162817,
+      "learning_rate": 3.6086210064623735e-06,
+      "loss": 0.2694,
+      "step": 202
+    },
+    {
+      "epoch": 1.1534090909090908,
+      "grad_norm": 0.22830303655343187,
+      "learning_rate": 3.5946876277757066e-06,
+      "loss": 0.2417,
+      "step": 203
+    },
+    {
+      "epoch": 1.1590909090909092,
+      "grad_norm": 0.2886944456973541,
+      "learning_rate": 3.580712051510876e-06,
+      "loss": 0.2581,
+      "step": 204
+    },
+    {
+      "epoch": 1.1647727272727273,
+      "grad_norm": 0.27254932607772603,
+      "learning_rate": 3.5666948163927716e-06,
+      "loss": 0.2663,
+      "step": 205
+    },
+    {
+      "epoch": 1.1704545454545454,
+      "grad_norm": 0.30632533777086607,
+      "learning_rate": 3.5526364627521322e-06,
+      "loss": 0.302,
+      "step": 206
+    },
+    {
+      "epoch": 1.1761363636363638,
+      "grad_norm": 0.22063246630024258,
+      "learning_rate": 3.5385375325047167e-06,
+      "loss": 0.2541,
+      "step": 207
+    },
+    {
+      "epoch": 1.1818181818181819,
+      "grad_norm": 0.2199009929510123,
+      "learning_rate": 3.5243985691304146e-06,
+      "loss": 0.2785,
+      "step": 208
+    },
+    {
+      "epoch": 1.1875,
+      "grad_norm": 0.22508931676812532,
+      "learning_rate": 3.5102201176522966e-06,
+      "loss": 0.254,
+      "step": 209
+    },
+    {
+      "epoch": 1.1931818181818181,
+      "grad_norm": 0.2655168492080692,
+      "learning_rate": 3.4960027246156043e-06,
+      "loss": 0.2734,
+      "step": 210
+    },
+    {
+      "epoch": 1.1988636363636362,
+      "grad_norm": 0.2397043633507901,
+      "learning_rate": 3.4817469380666834e-06,
+      "loss": 0.2434,
+      "step": 211
+    },
+    {
+      "epoch": 1.2045454545454546,
+      "grad_norm": 0.2219952128067738,
+      "learning_rate": 3.467453307531858e-06,
+      "loss": 0.2268,
+      "step": 212
+    },
+    {
+      "epoch": 1.2102272727272727,
+      "grad_norm": 0.27311261074430904,
+      "learning_rate": 3.453122383996245e-06,
+      "loss": 0.2644,
+      "step": 213
+    },
+    {
+      "epoch": 1.2159090909090908,
+      "grad_norm": 0.3734046673169832,
+      "learning_rate": 3.4387547198825187e-06,
+      "loss": 0.2354,
+      "step": 214
+    },
+    {
+      "epoch": 1.2215909090909092,
+      "grad_norm": 0.21182889889501536,
+      "learning_rate": 3.4243508690296133e-06,
+      "loss": 0.2515,
+      "step": 215
+    },
+    {
+      "epoch": 1.2272727272727273,
+      "grad_norm": 0.21938148808558747,
+      "learning_rate": 3.409911386671375e-06,
+      "loss": 0.2414,
+      "step": 216
+    },
+    {
+      "epoch": 1.2329545454545454,
+      "grad_norm": 0.22588281820458495,
+      "learning_rate": 3.3954368294151603e-06,
+      "loss": 0.251,
+      "step": 217
+    },
+    {
+      "epoch": 1.2386363636363638,
+      "grad_norm": 0.29774703377407774,
+      "learning_rate": 3.380927755220376e-06,
+      "loss": 0.234,
+      "step": 218
+    },
+    {
+      "epoch": 1.2443181818181819,
+      "grad_norm": 0.21627968204006795,
+      "learning_rate": 3.366384723376977e-06,
+      "loss": 0.2522,
+      "step": 219
+    },
+    {
+      "epoch": 1.25,
+      "grad_norm": 0.3187385148259123,
+      "learning_rate": 3.351808294483902e-06,
+      "loss": 0.2542,
+      "step": 220
+    },
+    {
+      "epoch": 1.2556818181818181,
+      "grad_norm": 0.24935352708926728,
+      "learning_rate": 3.3371990304274654e-06,
+      "loss": 0.2394,
+      "step": 221
+    },
+    {
+      "epoch": 1.2613636363636362,
+      "grad_norm": 0.23693314587613898,
+      "learning_rate": 3.3225574943597005e-06,
+      "loss": 0.2469,
+      "step": 222
+    },
+    {
+      "epoch": 1.2670454545454546,
+      "grad_norm": 0.28110255475362683,
+      "learning_rate": 3.3078842506766484e-06,
+      "loss": 0.2603,
+      "step": 223
+    },
+    {
+      "epoch": 1.2727272727272727,
+      "grad_norm": 0.2625043471276174,
+      "learning_rate": 3.2931798649966e-06,
+      "loss": 0.2593,
+      "step": 224
+    },
+    {
+      "epoch": 1.2784090909090908,
+      "grad_norm": 0.2540347477642736,
+      "learning_rate": 3.2784449041382973e-06,
+      "loss": 0.2752,
+      "step": 225
+    },
+    {
+      "epoch": 1.2840909090909092,
+      "grad_norm": 0.22673580657535364,
+      "learning_rate": 3.263679936099083e-06,
+      "loss": 0.2496,
+      "step": 226
+    },
+    {
+      "epoch": 1.2897727272727273,
+      "grad_norm": 0.2504539013854373,
+      "learning_rate": 3.248885530033004e-06,
+      "loss": 0.2578,
+      "step": 227
+    },
+    {
+      "epoch": 1.2954545454545454,
+      "grad_norm": 0.2181778867441864,
+      "learning_rate": 3.2340622562288717e-06,
+      "loss": 0.2394,
+      "step": 228
+    },
+    {
+      "epoch": 1.3011363636363638,
+      "grad_norm": 0.33020816509435613,
+      "learning_rate": 3.2192106860882782e-06,
+      "loss": 0.2733,
+      "step": 229
+    },
+    {
+      "epoch": 1.3068181818181819,
+      "grad_norm": 0.42372937786391984,
+      "learning_rate": 3.2043313921035747e-06,
+      "loss": 0.2602,
+      "step": 230
+    },
+    {
+      "epoch": 1.3125,
+      "grad_norm": 0.25468090847913644,
+      "learning_rate": 3.189424947835797e-06,
+      "loss": 0.2716,
+      "step": 231
+    },
+    {
+      "epoch": 1.3181818181818181,
+      "grad_norm": 0.2215655194050218,
+      "learning_rate": 3.174491927892561e-06,
+      "loss": 0.2718,
+      "step": 232
+    },
+    {
+      "epoch": 1.3238636363636362,
+      "grad_norm": 0.2754635828663956,
+      "learning_rate": 3.1595329079059102e-06,
+      "loss": 0.272,
+      "step": 233
+    },
+    {
+      "epoch": 1.3295454545454546,
+      "grad_norm": 0.2024807906225442,
+      "learning_rate": 3.144548464510127e-06,
+      "loss": 0.2466,
+      "step": 234
+    },
+    {
+      "epoch": 1.3352272727272727,
+      "grad_norm": 0.20548663623518323,
+      "learning_rate": 3.129539175319505e-06,
+      "loss": 0.2582,
+      "step": 235
+    },
+    {
+      "epoch": 1.3409090909090908,
+      "grad_norm": 0.23352172661093246,
+      "learning_rate": 3.114505618906086e-06,
+      "loss": 0.2434,
+      "step": 236
+    },
+    {
+      "epoch": 1.3465909090909092,
+      "grad_norm": 0.2916666966380751,
+      "learning_rate": 3.0994483747773508e-06,
+      "loss": 0.2602,
+      "step": 237
+    },
+    {
+      "epoch": 1.3522727272727273,
+      "grad_norm": 0.23761083831599722,
+      "learning_rate": 3.0843680233538885e-06,
+      "loss": 0.2702,
+      "step": 238
+    },
+    {
+      "epoch": 1.3579545454545454,
+      "grad_norm": 0.219481280373593,
+      "learning_rate": 3.0692651459470164e-06,
+      "loss": 0.2532,
+      "step": 239
+    },
+    {
+      "epoch": 1.3636363636363638,
+      "grad_norm": 0.2285839114976421,
+      "learning_rate": 3.0541403247363756e-06,
+      "loss": 0.2387,
+      "step": 240
+    },
+    {
+      "epoch": 1.3693181818181819,
+      "grad_norm": 0.24095505608449277,
+      "learning_rate": 3.0389941427474873e-06,
+      "loss": 0.2509,
+      "step": 241
+    },
+    {
+      "epoch": 1.375,
+      "grad_norm": 0.4493662508268078,
+      "learning_rate": 3.0238271838292815e-06,
+      "loss": 0.2595,
+      "step": 242
+    },
+    {
+      "epoch": 1.3806818181818181,
+      "grad_norm": 0.23328787857171382,
+      "learning_rate": 3.0086400326315853e-06,
+      "loss": 0.231,
+      "step": 243
+    },
+    {
+      "epoch": 1.3863636363636362,
+      "grad_norm": 0.5159195368293701,
+      "learning_rate": 2.9934332745825924e-06,
+      "loss": 0.2559,
+      "step": 244
+    },
+    {
+      "epoch": 1.3920454545454546,
+      "grad_norm": 0.4703634817361865,
+      "learning_rate": 2.9782074958662923e-06,
+      "loss": 0.2704,
+      "step": 245
+    },
+    {
+      "epoch": 1.3977272727272727,
+      "grad_norm": 0.2062134175125288,
+      "learning_rate": 2.962963283399877e-06,
+      "loss": 0.2478,
+      "step": 246
+    },
+    {
+      "epoch": 1.4034090909090908,
+      "grad_norm": 0.21282792221823255,
+      "learning_rate": 2.947701224811113e-06,
+      "loss": 0.2507,
+      "step": 247
+    },
+    {
+      "epoch": 1.4090909090909092,
+      "grad_norm": 0.29276392029658793,
+      "learning_rate": 2.932421908415694e-06,
+      "loss": 0.2277,
+      "step": 248
+    },
+    {
+      "epoch": 1.4147727272727273,
+      "grad_norm": 0.24683118700261866,
+      "learning_rate": 2.9171259231945598e-06,
+      "loss": 0.2607,
+      "step": 249
+    },
+    {
+      "epoch": 1.4204545454545454,
+      "grad_norm": 0.22971842427970574,
+      "learning_rate": 2.901813858771193e-06,
+      "loss": 0.2782,
+      "step": 250
+    },
+    {
+      "epoch": 1.4261363636363638,
+      "grad_norm": 0.22116314296102615,
+      "learning_rate": 2.8864863053888927e-06,
+      "loss": 0.2441,
+      "step": 251
+    },
+    {
+      "epoch": 1.4318181818181819,
+      "grad_norm": 0.19906723333631912,
+      "learning_rate": 2.871143853888017e-06,
+      "loss": 0.2461,
+      "step": 252
+    },
+    {
+      "epoch": 1.4375,
+      "grad_norm": 0.23452835492911767,
+      "learning_rate": 2.8557870956832135e-06,
+      "loss": 0.2511,
+      "step": 253
+    },
+    {
+      "epoch": 1.4431818181818181,
+      "grad_norm": 0.2146671991697931,
+      "learning_rate": 2.840416622740617e-06,
+      "loss": 0.2408,
+      "step": 254
+    },
+    {
+      "epoch": 1.4488636363636362,
+      "grad_norm": 0.262172959878327,
+      "learning_rate": 2.8250330275550337e-06,
+      "loss": 0.2803,
+      "step": 255
+    },
+    {
+      "epoch": 1.4545454545454546,
+      "grad_norm": 0.21360936224318358,
+      "learning_rate": 2.8096369031271e-06,
+      "loss": 0.2175,
+      "step": 256
+    },
+    {
+      "epoch": 1.4602272727272727,
+      "grad_norm": 0.2120651082005393,
+      "learning_rate": 2.7942288429404256e-06,
+      "loss": 0.2451,
+      "step": 257
+    },
+    {
+      "epoch": 1.4659090909090908,
+      "grad_norm": 0.21958260295487422,
+      "learning_rate": 2.778809440938714e-06,
+      "loss": 0.2781,
+      "step": 258
+    },
+    {
+      "epoch": 1.4715909090909092,
+      "grad_norm": 0.22983903008037593,
+      "learning_rate": 2.763379291502868e-06,
+      "loss": 0.2665,
+      "step": 259
+    },
+    {
+      "epoch": 1.4772727272727273,
+      "grad_norm": 0.19467024370331582,
+      "learning_rate": 2.7479389894280793e-06,
+      "loss": 0.2375,
+      "step": 260
+    },
+    {
+      "epoch": 1.4829545454545454,
+      "grad_norm": 0.21667848301184672,
+      "learning_rate": 2.7324891299008987e-06,
+      "loss": 0.2493,
+      "step": 261
+    },
+    {
+      "epoch": 1.4886363636363638,
+      "grad_norm": 0.24658186742801239,
+      "learning_rate": 2.7170303084762958e-06,
+      "loss": 0.2238,
+      "step": 262
+    },
+    {
+      "epoch": 1.4943181818181819,
+      "grad_norm": 0.48492645337740303,
+      "learning_rate": 2.701563121054695e-06,
+      "loss": 0.2307,
+      "step": 263
+    },
+    {
+      "epoch": 1.5,
+      "grad_norm": 0.23128616464818613,
+      "learning_rate": 2.6860881638590146e-06,
+      "loss": 0.2505,
+      "step": 264
+    },
+    {
+      "epoch": 1.5056818181818183,
+      "grad_norm": 0.2546561917759946,
+      "learning_rate": 2.670606033411678e-06,
+      "loss": 0.2528,
+      "step": 265
+    },
+    {
+      "epoch": 1.5113636363636362,
+      "grad_norm": 0.24741495860424792,
+      "learning_rate": 2.6551173265116182e-06,
+      "loss": 0.2508,
+      "step": 266
+    },
+    {
+      "epoch": 1.5170454545454546,
+      "grad_norm": 0.2692088472000367,
+      "learning_rate": 2.6396226402112768e-06,
+      "loss": 0.2601,
+      "step": 267
+    },
+    {
+      "epoch": 1.5227272727272727,
+      "grad_norm": 0.20089589499919636,
+      "learning_rate": 2.624122571793586e-06,
+      "loss": 0.2254,
+      "step": 268
+    },
+    {
+      "epoch": 1.5284090909090908,
+      "grad_norm": 0.2857287556843859,
+      "learning_rate": 2.6086177187489453e-06,
+      "loss": 0.2487,
+      "step": 269
+    },
+    {
+      "epoch": 1.5340909090909092,
+      "grad_norm": 0.2432557344935323,
+      "learning_rate": 2.593108678752191e-06,
+      "loss": 0.2688,
+      "step": 270
+    },
+    {
+      "epoch": 1.5397727272727273,
+      "grad_norm": 0.24063084990361086,
+      "learning_rate": 2.5775960496395565e-06,
+      "loss": 0.2277,
+      "step": 271
+    },
+    {
+      "epoch": 1.5454545454545454,
+      "grad_norm": 0.216657762192942,
+      "learning_rate": 2.562080429385626e-06,
+      "loss": 0.257,
+      "step": 272
+    },
+    {
+      "epoch": 1.5511363636363638,
+      "grad_norm": 0.35847342618755507,
+      "learning_rate": 2.5465624160802847e-06,
+      "loss": 0.238,
+      "step": 273
+    },
+    {
+      "epoch": 1.5568181818181817,
+      "grad_norm": 0.369555743490675,
+      "learning_rate": 2.531042607905665e-06,
+      "loss": 0.2354,
+      "step": 274
+    },
+    {
+      "epoch": 1.5625,
+      "grad_norm": 0.24361226061790756,
+      "learning_rate": 2.5155216031130882e-06,
+      "loss": 0.2816,
+      "step": 275
+    },
+    {
+      "epoch": 1.5681818181818183,
+      "grad_norm": 0.1922285560902828,
+      "learning_rate": 2.5e-06,
+      "loss": 0.2157,
+      "step": 276
+    },
+    {
+      "epoch": 1.5738636363636362,
+      "grad_norm": 0.21414334547048944,
+      "learning_rate": 2.4844783968869126e-06,
+      "loss": 0.2267,
+      "step": 277
+    },
+    {
+      "epoch": 1.5795454545454546,
+      "grad_norm": 1.2184972998385,
+      "learning_rate": 2.4689573920943358e-06,
+      "loss": 0.2626,
+      "step": 278
+    },
+    {
+      "epoch": 1.5852272727272727,
+      "grad_norm": 0.2756105324536501,
+      "learning_rate": 2.4534375839197166e-06,
+      "loss": 0.2925,
+      "step": 279
+    },
+    {
+      "epoch": 1.5909090909090908,
+      "grad_norm": 0.2494305680424624,
+      "learning_rate": 2.4379195706143755e-06,
+      "loss": 0.284,
+      "step": 280
+    },
+    {
+      "epoch": 1.5965909090909092,
+      "grad_norm": 0.4449397002609527,
+      "learning_rate": 2.422403950360444e-06,
+      "loss": 0.2277,
+      "step": 281
+    },
+    {
+      "epoch": 1.6022727272727273,
+      "grad_norm": 0.262193720551146,
+      "learning_rate": 2.406891321247809e-06,
+      "loss": 0.2748,
+      "step": 282
+    },
+    {
+      "epoch": 1.6079545454545454,
+      "grad_norm": 0.1902602150146509,
+      "learning_rate": 2.391382281251055e-06,
+      "loss": 0.2275,
+      "step": 283
+    },
+    {
+      "epoch": 1.6136363636363638,
+      "grad_norm": 0.22553747837476507,
+      "learning_rate": 2.375877428206415e-06,
+      "loss": 0.2385,
+      "step": 284
+    },
+    {
+      "epoch": 1.6193181818181817,
+      "grad_norm": 0.19774421293388952,
+      "learning_rate": 2.360377359788724e-06,
+      "loss": 0.2294,
+      "step": 285
+    },
+    {
+      "epoch": 1.625,
+      "grad_norm": 0.246379182223565,
+      "learning_rate": 2.3448826734883826e-06,
+      "loss": 0.2153,
+      "step": 286
+    },
+    {
+      "epoch": 1.6306818181818183,
+      "grad_norm": 0.25298356695474494,
+      "learning_rate": 2.3293939665883233e-06,
+      "loss": 0.2776,
+      "step": 287
+    },
+    {
+      "epoch": 1.6363636363636362,
+      "grad_norm": 0.2701672778631383,
+      "learning_rate": 2.313911836140986e-06,
+      "loss": 0.2717,
+      "step": 288
+    },
+    {
+      "epoch": 1.6420454545454546,
+      "grad_norm": 0.20559053992614199,
+      "learning_rate": 2.298436878945306e-06,
+      "loss": 0.2386,
+      "step": 289
+    },
+    {
+      "epoch": 1.6477272727272727,
+      "grad_norm": 0.3382944574936427,
+      "learning_rate": 2.2829696915237055e-06,
+      "loss": 0.2499,
+      "step": 290
+    },
+    {
+      "epoch": 1.6534090909090908,
+      "grad_norm": 0.3486190206377146,
+      "learning_rate": 2.267510870099101e-06,
+      "loss": 0.2363,
+      "step": 291
+    },
+    {
+      "epoch": 1.6590909090909092,
+      "grad_norm": 0.20281777636510792,
+      "learning_rate": 2.252061010571921e-06,
+      "loss": 0.2402,
+      "step": 292
+    },
+    {
+      "epoch": 1.6647727272727273,
+      "grad_norm": 0.20620257894673166,
+      "learning_rate": 2.2366207084971326e-06,
+      "loss": 0.2383,
+      "step": 293
+    },
+    {
+      "epoch": 1.6704545454545454,
+      "grad_norm": 0.43286460952304634,
+      "learning_rate": 2.2211905590612864e-06,
+      "loss": 0.2494,
+      "step": 294
+    },
+    {
+      "epoch": 1.6761363636363638,
+      "grad_norm": 0.3012071629038352,
+      "learning_rate": 2.205771157059575e-06,
+      "loss": 0.2076,
+      "step": 295
+    },
+    {
+      "epoch": 1.6818181818181817,
+      "grad_norm": 0.27532102490196564,
+      "learning_rate": 2.1903630968729003e-06,
+      "loss": 0.2554,
+      "step": 296
+    },
+    {
+      "epoch": 1.6875,
+      "grad_norm": 0.21603766103964822,
+      "learning_rate": 2.174966972444967e-06,
+      "loss": 0.2313,
+      "step": 297
+    },
+    {
+      "epoch": 1.6931818181818183,
+      "grad_norm": 0.2267972545695054,
+      "learning_rate": 2.159583377259384e-06,
+      "loss": 0.2647,
+      "step": 298
+    },
+    {
+      "epoch": 1.6988636363636362,
+      "grad_norm": 0.23145153019790213,
+      "learning_rate": 2.1442129043167877e-06,
+      "loss": 0.2488,
+      "step": 299
+    },
+    {
+      "epoch": 1.7045454545454546,
+      "grad_norm": 0.30814691128071964,
+      "learning_rate": 2.1288561461119837e-06,
+      "loss": 0.2278,
+      "step": 300
+    },
+    {
+      "epoch": 1.7102272727272727,
+      "grad_norm": 0.2934543144230605,
+      "learning_rate": 2.1135136946111077e-06,
+      "loss": 0.2753,
+      "step": 301
+    },
+    {
+      "epoch": 1.7159090909090908,
+      "grad_norm": 0.20303068805720303,
+      "learning_rate": 2.098186141228807e-06,
+      "loss": 0.2416,
+      "step": 302
+    },
+    {
+      "epoch": 1.7215909090909092,
+      "grad_norm": 0.20739352247286236,
+      "learning_rate": 2.0828740768054406e-06,
+      "loss": 0.2639,
+      "step": 303
+    },
+    {
+      "epoch": 1.7272727272727273,
+      "grad_norm": 0.2217270011433331,
+      "learning_rate": 2.0675780915843068e-06,
+      "loss": 0.2441,
+      "step": 304
+    },
+    {
+      "epoch": 1.7329545454545454,
+      "grad_norm": 0.19345781924887342,
+      "learning_rate": 2.052298775188888e-06,
+      "loss": 0.2378,
+      "step": 305
+    },
+    {
+      "epoch": 1.7386363636363638,
+      "grad_norm": 0.25373894853099455,
+      "learning_rate": 2.0370367166001245e-06,
+      "loss": 0.2592,
+      "step": 306
+    },
+    {
+      "epoch": 1.7443181818181817,
+      "grad_norm": 0.22910464762227148,
+      "learning_rate": 2.021792504133709e-06,
+      "loss": 0.2264,
+      "step": 307
+    },
+    {
+      "epoch": 1.75,
+      "grad_norm": 0.2303297961637058,
+      "learning_rate": 2.006566725417409e-06,
+      "loss": 0.2554,
+      "step": 308
+    },
+    {
+      "epoch": 1.7556818181818183,
+      "grad_norm": 0.21711583814497334,
+      "learning_rate": 1.991359967368416e-06,
+      "loss": 0.2456,
+      "step": 309
+    },
+    {
+      "epoch": 1.7613636363636362,
+      "grad_norm": 0.26241126899754347,
+      "learning_rate": 1.97617281617072e-06,
+      "loss": 0.2753,
+      "step": 310
+    },
+    {
+      "epoch": 1.7670454545454546,
+      "grad_norm": 0.22750312954422816,
+      "learning_rate": 1.9610058572525127e-06,
+      "loss": 0.2871,
+      "step": 311
+    },
+    {
+      "epoch": 1.7727272727272727,
+      "grad_norm": 0.18431040551647146,
+      "learning_rate": 1.945859675263625e-06,
+      "loss": 0.2486,
+      "step": 312
+    },
+    {
+      "epoch": 1.7784090909090908,
+      "grad_norm": 0.27552518413839977,
+      "learning_rate": 1.9307348540529845e-06,
+      "loss": 0.2422,
+      "step": 313
+    },
+    {
+      "epoch": 1.7840909090909092,
+      "grad_norm": 0.24464895994518496,
+      "learning_rate": 1.9156319766461124e-06,
+      "loss": 0.2463,
+      "step": 314
+    },
+    {
+      "epoch": 1.7897727272727273,
+      "grad_norm": 0.2382418996025972,
+      "learning_rate": 1.90055162522265e-06,
+      "loss": 0.2789,
+      "step": 315
+    },
+    {
+      "epoch": 1.7954545454545454,
+      "grad_norm": 0.28239156175807245,
+      "learning_rate": 1.8854943810939152e-06,
+      "loss": 0.2531,
+      "step": 316
+    },
+    {
+      "epoch": 1.8011363636363638,
+      "grad_norm": 0.2895263471358066,
+      "learning_rate": 1.8704608246804956e-06,
+      "loss": 0.2405,
+      "step": 317
+    },
+    {
+      "epoch": 1.8068181818181817,
+      "grad_norm": 0.20758099301934713,
+      "learning_rate": 1.8554515354898744e-06,
+      "loss": 0.2248,
+      "step": 318
+    },
+    {
+      "epoch": 1.8125,
+      "grad_norm": 0.22705536402699977,
+      "learning_rate": 1.840467092094091e-06,
+      "loss": 0.2684,
+      "step": 319
+    },
+    {
+      "epoch": 1.8181818181818183,
+      "grad_norm": 0.1928306955966746,
+      "learning_rate": 1.8255080721074391e-06,
+      "loss": 0.2298,
+      "step": 320
+    },
+    {
+      "epoch": 1.8238636363636362,
+      "grad_norm": 0.20499241539562923,
+      "learning_rate": 1.8105750521642035e-06,
+      "loss": 0.2207,
+      "step": 321
+    },
+    {
+      "epoch": 1.8295454545454546,
+      "grad_norm": 0.20797732732157784,
+      "learning_rate": 1.7956686078964257e-06,
+      "loss": 0.2474,
+      "step": 322
+    },
+    {
+      "epoch": 1.8352272727272727,
+      "grad_norm": 0.21542074815945023,
+      "learning_rate": 1.7807893139117222e-06,
+      "loss": 0.2331,
+      "step": 323
+    },
+    {
+      "epoch": 1.8409090909090908,
+      "grad_norm": 0.2118090262141367,
+      "learning_rate": 1.7659377437711294e-06,
+      "loss": 0.2546,
+      "step": 324
+    },
+    {
+      "epoch": 1.8465909090909092,
+      "grad_norm": 0.2841859057779013,
+      "learning_rate": 1.7511144699669967e-06,
+      "loss": 0.2749,
+      "step": 325
+    },
+    {
+      "epoch": 1.8522727272727273,
+      "grad_norm": 0.2987554078016277,
+      "learning_rate": 1.7363200639009176e-06,
+      "loss": 0.2288,
+      "step": 326
+    },
+    {
+      "epoch": 1.8579545454545454,
+      "grad_norm": 0.18732882489325686,
+      "learning_rate": 1.7215550958617036e-06,
+      "loss": 0.2289,
+      "step": 327
+    },
+    {
+      "epoch": 1.8636363636363638,
+      "grad_norm": 0.2608504460685447,
+      "learning_rate": 1.7068201350034017e-06,
+      "loss": 0.2631,
+      "step": 328
+    },
+    {
+      "epoch": 1.8693181818181817,
+      "grad_norm": 0.2721253524262622,
+      "learning_rate": 1.692115749323353e-06,
+      "loss": 0.2271,
+      "step": 329
+    },
+    {
+      "epoch": 1.875,
+      "grad_norm": 0.2379430708667795,
+      "learning_rate": 1.6774425056402993e-06,
+      "loss": 0.2412,
+      "step": 330
+    },
+    {
+      "epoch": 1.8806818181818183,
+      "grad_norm": 0.24257409750516146,
+      "learning_rate": 1.6628009695725348e-06,
+      "loss": 0.2683,
+      "step": 331
+    },
+    {
+      "epoch": 1.8863636363636362,
+      "grad_norm": 0.19301492232724798,
+      "learning_rate": 1.648191705516099e-06,
+      "loss": 0.2026,
+      "step": 332
+    },
+    {
+      "epoch": 1.8920454545454546,
+      "grad_norm": 0.25069740638984933,
+      "learning_rate": 1.6336152766230235e-06,
+      "loss": 0.2714,
+      "step": 333
+    },
+    {
+      "epoch": 1.8977272727272727,
+      "grad_norm": 0.24153255545005992,
+      "learning_rate": 1.6190722447796242e-06,
+      "loss": 0.2445,
+      "step": 334
+    },
+    {
+      "epoch": 1.9034090909090908,
+      "grad_norm": 0.21336850539372973,
+      "learning_rate": 1.6045631705848405e-06,
+      "loss": 0.2662,
+      "step": 335
+    },
+    {
+      "epoch": 1.9090909090909092,
+      "grad_norm": 0.2435556265678761,
+      "learning_rate": 1.5900886133286254e-06,
+      "loss": 0.2626,
+      "step": 336
+    },
+    {
+      "epoch": 1.9147727272727273,
+      "grad_norm": 0.19661831324809856,
+      "learning_rate": 1.5756491309703875e-06,
+      "loss": 0.228,
+      "step": 337
+    },
+    {
+      "epoch": 1.9204545454545454,
+      "grad_norm": 0.20715327562589703,
+      "learning_rate": 1.561245280117482e-06,
+      "loss": 0.2377,
+      "step": 338
+    },
+    {
+      "epoch": 1.9261363636363638,
+      "grad_norm": 0.2546762373725108,
+      "learning_rate": 1.5468776160037558e-06,
+      "loss": 0.2378,
+      "step": 339
+    },
+    {
+      "epoch": 1.9318181818181817,
+      "grad_norm": 0.21893927747495043,
+      "learning_rate": 1.5325466924681425e-06,
+      "loss": 0.2542,
+      "step": 340
+    },
+    {
+      "epoch": 1.9375,
+      "grad_norm": 0.29705395872406126,
+      "learning_rate": 1.5182530619333168e-06,
+      "loss": 0.2536,
+      "step": 341
+    },
+    {
+      "epoch": 1.9431818181818183,
+      "grad_norm": 0.2627707493027739,
+      "learning_rate": 1.5039972753843966e-06,
+      "loss": 0.2462,
+      "step": 342
+    },
+    {
+      "epoch": 1.9488636363636362,
+      "grad_norm": 0.196727357906668,
+      "learning_rate": 1.4897798823477045e-06,
+      "loss": 0.2417,
+      "step": 343
+    },
+    {
+      "epoch": 1.9545454545454546,
+      "grad_norm": 0.18765721167603922,
+      "learning_rate": 1.4756014308695865e-06,
+      "loss": 0.2371,
+      "step": 344
+    },
+    {
+      "epoch": 1.9602272727272727,
+      "grad_norm": 0.2169021146401285,
+      "learning_rate": 1.4614624674952843e-06,
+      "loss": 0.2417,
+      "step": 345
+    },
+    {
+      "epoch": 1.9659090909090908,
+      "grad_norm": 0.1724333522092845,
+      "learning_rate": 1.4473635372478692e-06,
+      "loss": 0.2252,
+      "step": 346
+    },
+    {
+      "epoch": 1.9715909090909092,
+      "grad_norm": 0.27729471470089767,
+      "learning_rate": 1.4333051836072298e-06,
+      "loss": 0.2072,
+      "step": 347
+    },
+    {
+      "epoch": 1.9772727272727273,
+      "grad_norm": 0.21309523464495775,
+      "learning_rate": 1.4192879484891253e-06,
+      "loss": 0.2423,
+      "step": 348
+    },
+    {
+      "epoch": 1.9829545454545454,
+      "grad_norm": 0.2579270534818131,
+      "learning_rate": 1.405312372224294e-06,
+      "loss": 0.2463,
+      "step": 349
+    },
+    {
+      "epoch": 1.9886363636363638,
+      "grad_norm": 0.20458929831058947,
+      "learning_rate": 1.3913789935376271e-06,
+      "loss": 0.2382,
+      "step": 350
+    },
+    {
+      "epoch": 1.9943181818181817,
+      "grad_norm": 0.264219593485995,
+      "learning_rate": 1.3774883495273986e-06,
+      "loss": 0.2306,
+      "step": 351
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.20532725547286926,
+      "learning_rate": 1.363640975644564e-06,
+      "loss": 0.2116,
+      "step": 352
+    },
+    {
+      "epoch": 2.0056818181818183,
+      "grad_norm": 0.2992343797336132,
+      "learning_rate": 1.3498374056721198e-06,
+      "loss": 0.2604,
+      "step": 353
+    },
+    {
+      "epoch": 2.0113636363636362,
+      "grad_norm": 0.2029411119629725,
+      "learning_rate": 1.3360781717045266e-06,
+      "loss": 0.2295,
+      "step": 354
+    },
+    {
+      "epoch": 2.0170454545454546,
+      "grad_norm": 0.20429497877305577,
+      "learning_rate": 1.322363804127198e-06,
+      "loss": 0.2384,
+      "step": 355
+    },
+    {
+      "epoch": 2.022727272727273,
+      "grad_norm": 0.26030944241332393,
+      "learning_rate": 1.3086948315960567e-06,
+      "loss": 0.2289,
+      "step": 356
+    },
+    {
+      "epoch": 2.028409090909091,
+      "grad_norm": 0.23879201027045563,
+      "learning_rate": 1.295071781017156e-06,
+      "loss": 0.2489,
+      "step": 357
+    },
+    {
+      "epoch": 2.034090909090909,
+      "grad_norm": 0.2065949866931435,
+      "learning_rate": 1.2814951775263671e-06,
+      "loss": 0.2402,
+      "step": 358
+    },
+    {
+      "epoch": 2.039772727272727,
+      "grad_norm": 0.3304715872732868,
+      "learning_rate": 1.267965544469137e-06,
+      "loss": 0.2621,
+      "step": 359
+    },
+    {
+      "epoch": 2.0454545454545454,
+      "grad_norm": 0.19051837959097934,
+      "learning_rate": 1.2544834033803183e-06,
+      "loss": 0.2213,
+      "step": 360
+    },
+    {
+      "epoch": 2.0511363636363638,
+      "grad_norm": 0.23880675922567549,
+      "learning_rate": 1.2410492739640592e-06,
+      "loss": 0.2269,
+      "step": 361
+    },
+    {
+      "epoch": 2.0568181818181817,
+      "grad_norm": 0.2018598929268875,
+      "learning_rate": 1.227663674073775e-06,
+      "loss": 0.2267,
+      "step": 362
+    },
+    {
+      "epoch": 2.0625,
+      "grad_norm": 0.25110946606632356,
+      "learning_rate": 1.2143271196921832e-06,
+      "loss": 0.2415,
+      "step": 363
+    },
+    {
+      "epoch": 2.0681818181818183,
+      "grad_norm": 0.22102547627853814,
+      "learning_rate": 1.2010401249114166e-06,
+      "loss": 0.2262,
+      "step": 364
+    },
+    {
+      "epoch": 2.0738636363636362,
+      "grad_norm": 0.21927465830616452,
+      "learning_rate": 1.1878032019132016e-06,
+      "loss": 0.2304,
+      "step": 365
+    },
+    {
+      "epoch": 2.0795454545454546,
+      "grad_norm": 0.23137588236465315,
+      "learning_rate": 1.1746168609491198e-06,
+      "loss": 0.2409,
+      "step": 366
+    },
+    {
+      "epoch": 2.085227272727273,
+      "grad_norm": 0.26540697815717607,
+      "learning_rate": 1.1614816103209363e-06,
+      "loss": 0.2491,
+      "step": 367
+    },
+    {
+      "epoch": 2.090909090909091,
+      "grad_norm": 0.2240465647803614,
+      "learning_rate": 1.148397956361007e-06,
+      "loss": 0.2676,
+      "step": 368
+    },
+    {
+      "epoch": 2.096590909090909,
+      "grad_norm": 0.21344405314643286,
+      "learning_rate": 1.1353664034127585e-06,
+      "loss": 0.2315,
+      "step": 369
+    },
+    {
+      "epoch": 2.102272727272727,
+      "grad_norm": 0.3017527199834084,
+      "learning_rate": 1.122387453811252e-06,
+      "loss": 0.2246,
+      "step": 370
+    },
+    {
+      "epoch": 2.1079545454545454,
+      "grad_norm": 0.22871251640001977,
+      "learning_rate": 1.1094616078638123e-06,
+      "loss": 0.2263,
+      "step": 371
+    },
+    {
+      "epoch": 2.1136363636363638,
+      "grad_norm": 0.22922429353524632,
+      "learning_rate": 1.0965893638307484e-06,
+      "loss": 0.2479,
+      "step": 372
+    },
+    {
+      "epoch": 2.1193181818181817,
+      "grad_norm": 0.23625823763756973,
+      "learning_rate": 1.083771217906143e-06,
+      "loss": 0.2654,
+      "step": 373
+    },
+    {
+      "epoch": 2.125,
+      "grad_norm": 0.27625305972483133,
+      "learning_rate": 1.071007664198727e-06,
+      "loss": 0.2365,
+      "step": 374
+    },
+    {
+      "epoch": 2.1306818181818183,
+      "grad_norm": 0.22438286954207515,
+      "learning_rate": 1.0582991947128324e-06,
+      "loss": 0.2549,
+      "step": 375
+    },
+    {
+      "epoch": 2.1363636363636362,
+      "grad_norm": 0.3702343135252277,
+      "learning_rate": 1.0456462993294273e-06,
+      "loss": 0.2296,
+      "step": 376
+    },
+    {
+      "epoch": 2.1420454545454546,
+      "grad_norm": 0.22114715704005006,
+      "learning_rate": 1.0330494657872312e-06,
+      "loss": 0.2502,
+      "step": 377
+    },
+    {
+      "epoch": 2.147727272727273,
+      "grad_norm": 0.30333787634431497,
+      "learning_rate": 1.0205091796639143e-06,
+      "loss": 0.2736,
+      "step": 378
+    },
+    {
+      "epoch": 2.153409090909091,
+      "grad_norm": 0.20447974690727722,
+      "learning_rate": 1.008025924357379e-06,
+      "loss": 0.2405,
+      "step": 379
+    },
+    {
+      "epoch": 2.159090909090909,
+      "grad_norm": 0.20209966437613497,
+      "learning_rate": 9.95600181067129e-07,
+      "loss": 0.2413,
+      "step": 380
+    },
+    {
+      "epoch": 2.164772727272727,
+      "grad_norm": 0.232405336227515,
+      "learning_rate": 9.832324287757158e-07,
+      "loss": 0.2573,
+      "step": 381
+    },
+    {
+      "epoch": 2.1704545454545454,
+      "grad_norm": 0.27892846481633526,
+      "learning_rate": 9.709231442302777e-07,
+      "loss": 0.2314,
+      "step": 382
+    },
+    {
+      "epoch": 2.1761363636363638,
+      "grad_norm": 0.23501245795670683,
+      "learning_rate": 9.586728019241622e-07,
+      "loss": 0.2375,
+      "step": 383
+    },
+    {
+      "epoch": 2.1818181818181817,
+      "grad_norm": 0.23046728304284175,
+      "learning_rate": 9.464818740786357e-07,
+      "loss": 0.2514,
+      "step": 384
+    },
+    {
+      "epoch": 2.1875,
+      "grad_norm": 0.23024328925839851,
+      "learning_rate": 9.343508306246771e-07,
+      "loss": 0.2562,
+      "step": 385
+    },
+    {
+      "epoch": 2.1931818181818183,
+      "grad_norm": 0.21907595115075354,
+      "learning_rate": 9.222801391848688e-07,
+      "loss": 0.2364,
+      "step": 386
+    },
+    {
+      "epoch": 2.1988636363636362,
+      "grad_norm": 0.2577759642925678,
+      "learning_rate": 9.102702650553672e-07,
+      "loss": 0.2452,
+      "step": 387
+    },
+    {
+      "epoch": 2.2045454545454546,
+      "grad_norm": 0.20572031186271958,
+      "learning_rate": 8.983216711879663e-07,
+      "loss": 0.2366,
+      "step": 388
+    },
+    {
+      "epoch": 2.210227272727273,
+      "grad_norm": 0.21329319069858652,
+      "learning_rate": 8.86434818172256e-07,
+      "loss": 0.2332,
+      "step": 389
+    },
+    {
+      "epoch": 2.215909090909091,
+      "grad_norm": 0.26280014382816824,
+      "learning_rate": 8.746101642178623e-07,
+      "loss": 0.2731,
+      "step": 390
+    },
+    {
+      "epoch": 2.221590909090909,
+      "grad_norm": 0.23251275990425788,
+      "learning_rate": 8.628481651367876e-07,
+      "loss": 0.2285,
+      "step": 391
+    },
+    {
+      "epoch": 2.227272727272727,
+      "grad_norm": 0.213052782052081,
+      "learning_rate": 8.51149274325839e-07,
+      "loss": 0.269,
+      "step": 392
+    },
+    {
+      "epoch": 2.2329545454545454,
+      "grad_norm": 0.2390787111598495,
+      "learning_rate": 8.395139427491517e-07,
+      "loss": 0.228,
+      "step": 393
+    },
+    {
+      "epoch": 2.2386363636363638,
+      "grad_norm": 0.2801854339324972,
+      "learning_rate": 8.279426189208057e-07,
+      "loss": 0.2496,
+      "step": 394
+    },
+    {
+      "epoch": 2.2443181818181817,
+      "grad_norm": 0.21989720758556128,
+      "learning_rate": 8.164357488875349e-07,
+      "loss": 0.2335,
+      "step": 395
+    },
+    {
+      "epoch": 2.25,
+      "grad_norm": 0.22453398641005057,
+      "learning_rate": 8.049937762115354e-07,
+      "loss": 0.2299,
+      "step": 396
+    },
+    {
+      "epoch": 2.2556818181818183,
+      "grad_norm": 0.22778711590056566,
+      "learning_rate": 7.936171419533653e-07,
+      "loss": 0.2274,
+      "step": 397
+    },
+    {
+      "epoch": 2.2613636363636362,
+      "grad_norm": 0.22100874802253603,
+      "learning_rate": 7.823062846549432e-07,
+      "loss": 0.2501,
+      "step": 398
+    },
+    {
+      "epoch": 2.2670454545454546,
+      "grad_norm": 0.2234622648942781,
+      "learning_rate": 7.71061640322646e-07,
+      "loss": 0.2592,
+      "step": 399
+    },
+    {
+      "epoch": 2.2727272727272725,
+      "grad_norm": 0.19533941288191214,
+      "learning_rate": 7.59883642410498e-07,
+      "loss": 0.2433,
+      "step": 400
+    },
+    {
+      "epoch": 2.278409090909091,
+      "grad_norm": 0.2611615221654739,
+      "learning_rate": 7.487727218034646e-07,
+      "loss": 0.2463,
+      "step": 401
+    },
+    {
+      "epoch": 2.284090909090909,
+      "grad_norm": 0.227647391322542,
+      "learning_rate": 7.377293068008421e-07,
+      "loss": 0.2293,
+      "step": 402
+    },
+    {
+      "epoch": 2.2897727272727275,
+      "grad_norm": 0.2765238173548462,
+      "learning_rate": 7.267538230997487e-07,
+      "loss": 0.2458,
+      "step": 403
+    },
+    {
+      "epoch": 2.2954545454545454,
+      "grad_norm": 0.2736767495641602,
+      "learning_rate": 7.15846693778712e-07,
+      "loss": 0.2244,
+      "step": 404
+    },
+    {
+      "epoch": 2.3011363636363638,
+      "grad_norm": 0.29322810329004506,
+      "learning_rate": 7.050083392813651e-07,
+      "loss": 0.2429,
+      "step": 405
+    },
+    {
+      "epoch": 2.3068181818181817,
+      "grad_norm": 0.23672260332490674,
+      "learning_rate": 6.942391774002352e-07,
+      "loss": 0.2224,
+      "step": 406
+    },
+    {
+      "epoch": 2.3125,
+      "grad_norm": 0.20842228485707248,
+      "learning_rate": 6.835396232606414e-07,
+      "loss": 0.2505,
+      "step": 407
+    },
+    {
+      "epoch": 2.3181818181818183,
+      "grad_norm": 0.22128940362648533,
+      "learning_rate": 6.729100893046897e-07,
+      "loss": 0.2355,
+      "step": 408
+    },
+    {
+      "epoch": 2.3238636363636362,
+      "grad_norm": 0.22752723544755182,
+      "learning_rate": 6.623509852753798e-07,
+      "loss": 0.2143,
+      "step": 409
+    },
+    {
+      "epoch": 2.3295454545454546,
+      "grad_norm": 0.22237504317670226,
+      "learning_rate": 6.518627182008034e-07,
+      "loss": 0.2296,
+      "step": 410
+    },
+    {
+      "epoch": 2.3352272727272725,
+      "grad_norm": 0.21204727758454492,
+      "learning_rate": 6.414456923784593e-07,
+      "loss": 0.209,
+      "step": 411
+    },
+    {
+      "epoch": 2.340909090909091,
+      "grad_norm": 0.2390107432904483,
+      "learning_rate": 6.311003093596674e-07,
+      "loss": 0.2532,
+      "step": 412
+    },
+    {
+      "epoch": 2.346590909090909,
+      "grad_norm": 0.19149924277202832,
+      "learning_rate": 6.208269679340886e-07,
+      "loss": 0.2268,
+      "step": 413
+    },
+    {
+      "epoch": 2.3522727272727275,
+      "grad_norm": 0.29290190158043866,
+      "learning_rate": 6.106260641143547e-07,
+      "loss": 0.2413,
+      "step": 414
+    },
+    {
+      "epoch": 2.3579545454545454,
+      "grad_norm": 0.2640131071106126,
+      "learning_rate": 6.004979911208006e-07,
+      "loss": 0.2203,
+      "step": 415
+    },
+    {
+      "epoch": 2.3636363636363638,
+      "grad_norm": 0.26467170247291516,
+      "learning_rate": 5.904431393663088e-07,
+      "loss": 0.2671,
+      "step": 416
+    },
+    {
+      "epoch": 2.3693181818181817,
+      "grad_norm": 0.2047663364954759,
+      "learning_rate": 5.804618964412587e-07,
+      "loss": 0.2826,
+      "step": 417
+    },
+    {
+      "epoch": 2.375,
+      "grad_norm": 0.2157657875191716,
+      "learning_rate": 5.705546470985851e-07,
+      "loss": 0.2181,
+      "step": 418
+    },
+    {
+      "epoch": 2.3806818181818183,
+      "grad_norm": 0.21019411350145403,
+      "learning_rate": 5.607217732389503e-07,
+      "loss": 0.23,
+      "step": 419
+    },
+    {
+      "epoch": 2.3863636363636362,
+      "grad_norm": 0.26327220650995004,
+      "learning_rate": 5.509636538960183e-07,
+      "loss": 0.2461,
+      "step": 420
+    },
+    {
+      "epoch": 2.3920454545454546,
+      "grad_norm": 0.22603288999069882,
+      "learning_rate": 5.412806652218469e-07,
+      "loss": 0.2277,
+      "step": 421
+    },
+    {
+      "epoch": 2.3977272727272725,
+      "grad_norm": 0.24543228404047565,
+      "learning_rate": 5.316731804723877e-07,
+      "loss": 0.2654,
+      "step": 422
+    },
+    {
+      "epoch": 2.403409090909091,
+      "grad_norm": 0.2511069445289598,
+      "learning_rate": 5.221415699930952e-07,
+      "loss": 0.25,
+      "step": 423
+    },
+    {
+      "epoch": 2.409090909090909,
+      "grad_norm": 0.27112636787166394,
+      "learning_rate": 5.126862012046551e-07,
+      "loss": 0.2354,
+      "step": 424
+    },
+    {
+      "epoch": 2.4147727272727275,
+      "grad_norm": 0.24057911667311155,
+      "learning_rate": 5.03307438588819e-07,
+      "loss": 0.2555,
+      "step": 425
+    },
+    {
+      "epoch": 2.4204545454545454,
+      "grad_norm": 0.20526839995277346,
+      "learning_rate": 4.940056436743534e-07,
+      "loss": 0.2023,
+      "step": 426
+    },
+    {
+      "epoch": 2.4261363636363638,
+      "grad_norm": 0.3463263547587994,
+      "learning_rate": 4.847811750231057e-07,
+      "loss": 0.2423,
+      "step": 427
+    },
+    {
+      "epoch": 2.4318181818181817,
+      "grad_norm": 0.3230515716711578,
+      "learning_rate": 4.7563438821618236e-07,
+      "loss": 0.2364,
+      "step": 428
+    },
+    {
+      "epoch": 2.4375,
+      "grad_norm": 0.19216365713273992,
+      "learning_rate": 4.6656563584023955e-07,
+      "loss": 0.2592,
+      "step": 429
+    },
+    {
+      "epoch": 2.4431818181818183,
+      "grad_norm": 0.213493956757016,
+      "learning_rate": 4.5757526747389506e-07,
+      "loss": 0.2348,
+      "step": 430
+    },
+    {
+      "epoch": 2.4488636363636362,
+      "grad_norm": 0.22855349656144805,
+      "learning_rate": 4.4866362967425054e-07,
+      "loss": 0.211,
+      "step": 431
+    },
+    {
+      "epoch": 2.4545454545454546,
+      "grad_norm": 0.2010595906652801,
+      "learning_rate": 4.398310659635338e-07,
+      "loss": 0.2203,
+      "step": 432
+    },
+    {
+      "epoch": 2.4602272727272725,
+      "grad_norm": 0.28214757374784766,
+      "learning_rate": 4.310779168158566e-07,
+      "loss": 0.2636,
+      "step": 433
+    },
+    {
+      "epoch": 2.465909090909091,
+      "grad_norm": 0.2512424715594609,
+      "learning_rate": 4.2240451964408984e-07,
+      "loss": 0.244,
+      "step": 434
+    },
+    {
+      "epoch": 2.471590909090909,
+      "grad_norm": 0.5451128021916027,
+      "learning_rate": 4.138112087868576e-07,
+      "loss": 0.2562,
+      "step": 435
+    },
+    {
+      "epoch": 2.4772727272727275,
+      "grad_norm": 0.2828376249470446,
+      "learning_rate": 4.052983154956483e-07,
+      "loss": 0.2384,
+      "step": 436
+    },
+    {
+      "epoch": 2.4829545454545454,
+      "grad_norm": 0.2557661881508072,
+      "learning_rate": 3.9686616792204677e-07,
+      "loss": 0.2509,
+      "step": 437
+    },
+    {
+      "epoch": 2.4886363636363638,
+      "grad_norm": 0.25795762829235164,
+      "learning_rate": 3.885150911050856e-07,
+      "loss": 0.2218,
+      "step": 438
+    },
+    {
+      "epoch": 2.4943181818181817,
+      "grad_norm": 0.25563496442518113,
+      "learning_rate": 3.8024540695871275e-07,
+      "loss": 0.2344,
+      "step": 439
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 0.28154467860900795,
+      "learning_rate": 3.720574342593847e-07,
+      "loss": 0.2444,
+      "step": 440
+    },
+    {
+      "epoch": 2.5056818181818183,
+      "grad_norm": 0.22139203111003553,
+      "learning_rate": 3.639514886337786e-07,
+      "loss": 0.285,
+      "step": 441
+    },
+    {
+      "epoch": 2.5113636363636362,
+      "grad_norm": 0.21532391608850526,
+      "learning_rate": 3.559278825466245e-07,
+      "loss": 0.2397,
+      "step": 442
+    },
+    {
+      "epoch": 2.5170454545454546,
+      "grad_norm": 0.20959198276551452,
+      "learning_rate": 3.4798692528866057e-07,
+      "loss": 0.2444,
+      "step": 443
+    },
+    {
+      "epoch": 2.5227272727272725,
+      "grad_norm": 0.2587316337980766,
+      "learning_rate": 3.4012892296471173e-07,
+      "loss": 0.2394,
+      "step": 444
+    },
+    {
+      "epoch": 2.528409090909091,
+      "grad_norm": 0.21690366841626674,
+      "learning_rate": 3.3235417848188985e-07,
+      "loss": 0.2428,
+      "step": 445
+    },
+    {
+      "epoch": 2.534090909090909,
+      "grad_norm": 0.26971941485731793,
+      "learning_rate": 3.2466299153791626e-07,
+      "loss": 0.2611,
+      "step": 446
+    },
+    {
+      "epoch": 2.5397727272727275,
+      "grad_norm": 0.22784776958627076,
+      "learning_rate": 3.1705565860956994e-07,
+      "loss": 0.2672,
+      "step": 447
+    },
+    {
+      "epoch": 2.5454545454545454,
+      "grad_norm": 0.22730153530504624,
+      "learning_rate": 3.095324729412602e-07,
+      "loss": 0.2316,
+      "step": 448
+    },
+    {
+      "epoch": 2.5511363636363638,
+      "grad_norm": 0.23472487686066498,
+      "learning_rate": 3.020937245337208e-07,
+      "loss": 0.2557,
+      "step": 449
+    },
+    {
+      "epoch": 2.5568181818181817,
+      "grad_norm": 0.19943308794934309,
+      "learning_rate": 2.947397001328314e-07,
+      "loss": 0.2359,
+      "step": 450
+    },
+    {
+      "epoch": 2.5625,
+      "grad_norm": 0.2370879570780672,
+      "learning_rate": 2.874706832185656e-07,
+      "loss": 0.24,
+      "step": 451
+    },
+    {
+      "epoch": 2.5681818181818183,
+      "grad_norm": 0.28430114512901383,
+      "learning_rate": 2.80286953994062e-07,
+      "loss": 0.2509,
+      "step": 452
+    },
+    {
+      "epoch": 2.5738636363636362,
+      "grad_norm": 1.2939227200205659,
+      "learning_rate": 2.731887893748242e-07,
+      "loss": 0.2452,
+      "step": 453
+    },
+    {
+      "epoch": 2.5795454545454546,
+      "grad_norm": 0.6171012295566409,
+      "learning_rate": 2.6617646297804554e-07,
+      "loss": 0.2586,
+      "step": 454
+    },
+    {
+      "epoch": 2.5852272727272725,
+      "grad_norm": 0.22954149192739645,
+      "learning_rate": 2.5925024511206207e-07,
+      "loss": 0.238,
+      "step": 455
+    },
+    {
+      "epoch": 2.590909090909091,
+      "grad_norm": 0.2237123859341468,
+      "learning_rate": 2.52410402765933e-07,
+      "loss": 0.2073,
+      "step": 456
+    },
+    {
+      "epoch": 2.596590909090909,
+      "grad_norm": 0.24492186112063838,
+      "learning_rate": 2.45657199599148e-07,
+      "loss": 0.2319,
+      "step": 457
+    },
+    {
+      "epoch": 2.6022727272727275,
+      "grad_norm": 0.2699569181048303,
+      "learning_rate": 2.389908959314663e-07,
+      "loss": 0.2275,
+      "step": 458
+    },
+    {
+      "epoch": 2.6079545454545454,
+      "grad_norm": 0.3385798703479409,
+      "learning_rate": 2.3241174873287892e-07,
+      "loss": 0.2353,
+      "step": 459
+    },
+    {
+      "epoch": 2.6136363636363638,
+      "grad_norm": 0.24587689957508765,
+      "learning_rate": 2.2592001161370392e-07,
+      "loss": 0.2269,
+      "step": 460
+    },
+    {
+      "epoch": 2.6193181818181817,
+      "grad_norm": 0.2546095534311607,
+      "learning_rate": 2.1951593481481236e-07,
+      "loss": 0.2435,
+      "step": 461
+    },
+    {
+      "epoch": 2.625,
+      "grad_norm": 0.3425334424188858,
+      "learning_rate": 2.1319976519797862e-07,
+      "loss": 0.2277,
+      "step": 462
+    },
+    {
+      "epoch": 2.6306818181818183,
+      "grad_norm": 0.2266424757647852,
+      "learning_rate": 2.0697174623636795e-07,
+      "loss": 0.2515,
+      "step": 463
+    },
+    {
+      "epoch": 2.6363636363636362,
+      "grad_norm": 0.3478031919431262,
+      "learning_rate": 2.0083211800514868e-07,
+      "loss": 0.2473,
+      "step": 464
+    },
+    {
+      "epoch": 2.6420454545454546,
+      "grad_norm": 0.25923655545004287,
+      "learning_rate": 1.9478111717223968e-07,
+      "loss": 0.2284,
+      "step": 465
+    },
+    {
+      "epoch": 2.6477272727272725,
+      "grad_norm": 0.20656881520592626,
+      "learning_rate": 1.8881897698918544e-07,
+      "loss": 0.2117,
+      "step": 466
+    },
+    {
+      "epoch": 2.653409090909091,
+      "grad_norm": 0.23728611207119654,
+      "learning_rate": 1.8294592728216764e-07,
+      "loss": 0.2404,
+      "step": 467
+    },
+    {
+      "epoch": 2.659090909090909,
+      "grad_norm": 0.2220771487213648,
+      "learning_rate": 1.7716219444314204e-07,
+      "loss": 0.2696,
+      "step": 468
+    },
+    {
+      "epoch": 2.6647727272727275,
+      "grad_norm": 0.20381048055659795,
+      "learning_rate": 1.7146800142111536e-07,
+      "loss": 0.2339,
+      "step": 469
+    },
+    {
+      "epoch": 2.6704545454545454,
+      "grad_norm": 0.2626669618263925,
+      "learning_rate": 1.658635677135484e-07,
+      "loss": 0.2531,
+      "step": 470
+    },
+    {
+      "epoch": 2.6761363636363638,
+      "grad_norm": 0.28962842443201436,
+      "learning_rate": 1.6034910935789628e-07,
+      "loss": 0.2364,
+      "step": 471
+    },
+    {
+      "epoch": 2.6818181818181817,
+      "grad_norm": 0.22258233394626142,
+      "learning_rate": 1.5492483892328104e-07,
+      "loss": 0.2313,
+      "step": 472
+    },
+    {
+      "epoch": 2.6875,
+      "grad_norm": 0.22898287693164499,
+      "learning_rate": 1.4959096550229645e-07,
+      "loss": 0.2304,
+      "step": 473
+    },
+    {
+      "epoch": 2.6931818181818183,
+      "grad_norm": 0.21073886345157697,
+      "learning_rate": 1.44347694702949e-07,
+      "loss": 0.2416,
+      "step": 474
+    },
+    {
+      "epoch": 2.6988636363636362,
+      "grad_norm": 1.0729633168708006,
+      "learning_rate": 1.391952286407311e-07,
+      "loss": 0.2563,
+      "step": 475
+    },
+    {
+      "epoch": 2.7045454545454546,
+      "grad_norm": 0.3533455916134876,
+      "learning_rate": 1.341337659308309e-07,
+      "loss": 0.2493,
+      "step": 476
+    },
+    {
+      "epoch": 2.7102272727272725,
+      "grad_norm": 0.2577397312681227,
+      "learning_rate": 1.291635016804768e-07,
+      "loss": 0.2366,
+      "step": 477
+    },
+    {
+      "epoch": 2.715909090909091,
+      "grad_norm": 0.2306912910968671,
+      "learning_rate": 1.2428462748141523e-07,
+      "loss": 0.2292,
+      "step": 478
+    },
+    {
+      "epoch": 2.721590909090909,
+      "grad_norm": 0.22698482056742056,
+      "learning_rate": 1.1949733140252468e-07,
+      "loss": 0.2414,
+      "step": 479
+    },
+    {
+      "epoch": 2.7272727272727275,
+      "grad_norm": 0.18931106378770035,
+      "learning_rate": 1.1480179798256857e-07,
+      "loss": 0.2211,
+      "step": 480
+    },
+    {
+      "epoch": 2.7329545454545454,
+      "grad_norm": 0.21904198372370234,
+      "learning_rate": 1.1019820822307986e-07,
+      "loss": 0.2386,
+      "step": 481
+    },
+    {
+      "epoch": 2.7386363636363638,
+      "grad_norm": 0.2033101000039358,
+      "learning_rate": 1.056867395813832e-07,
+      "loss": 0.251,
+      "step": 482
+    },
+    {
+      "epoch": 2.7443181818181817,
+      "grad_norm": 0.19353522082992652,
+      "learning_rate": 1.0126756596375687e-07,
+      "loss": 0.2344,
+      "step": 483
+    },
+    {
+      "epoch": 2.75,
+      "grad_norm": 0.30934234312647046,
+      "learning_rate": 9.694085771872697e-08,
+      "loss": 0.2314,
+      "step": 484
+    },
+    {
+      "epoch": 2.7556818181818183,
+      "grad_norm": 0.19788186109536335,
+      "learning_rate": 9.270678163050218e-08,
+      "loss": 0.227,
+      "step": 485
+    },
+    {
+      "epoch": 2.7613636363636362,
+      "grad_norm": 0.22823739130011358,
+      "learning_rate": 8.856550091254302e-08,
+      "loss": 0.252,
+      "step": 486
+    },
+    {
+      "epoch": 2.7670454545454546,
+      "grad_norm": 0.2740360143378895,
+      "learning_rate": 8.451717520127272e-08,
+      "loss": 0.21,
+      "step": 487
+    },
+    {
+      "epoch": 2.7727272727272725,
+      "grad_norm": 0.24604734634179243,
+      "learning_rate": 8.056196054992193e-08,
+      "loss": 0.2724,
+      "step": 488
+    },
+    {
+      "epoch": 2.778409090909091,
+      "grad_norm": 0.6552142496694562,
+      "learning_rate": 7.670000942251288e-08,
+      "loss": 0.2484,
+      "step": 489
+    },
+    {
+      "epoch": 2.784090909090909,
+      "grad_norm": 0.2192780573268507,
+      "learning_rate": 7.293147068798384e-08,
+      "loss": 0.2367,
+      "step": 490
+    },
+    {
+      "epoch": 2.7897727272727275,
+      "grad_norm": 0.42226315361363354,
+      "learning_rate": 6.92564896144493e-08,
+      "loss": 0.2283,
+      "step": 491
+    },
+    {
+      "epoch": 2.7954545454545454,
+      "grad_norm": 0.19585545271426144,
+      "learning_rate": 6.56752078636011e-08,
+      "loss": 0.2049,
+      "step": 492
+    },
+    {
+      "epoch": 2.8011363636363638,
+      "grad_norm": 0.28626860845495716,
+      "learning_rate": 6.218776348524663e-08,
+      "loss": 0.237,
+      "step": 493
+    },
+    {
+      "epoch": 2.8068181818181817,
+      "grad_norm": 0.2317197207710347,
+      "learning_rate": 5.879429091198846e-08,
+      "loss": 0.2348,
+      "step": 494
+    },
+    {
+      "epoch": 2.8125,
+      "grad_norm": 0.22290822975196342,
+      "learning_rate": 5.549492095404202e-08,
+      "loss": 0.229,
+      "step": 495
+    },
+    {
+      "epoch": 2.8181818181818183,
+      "grad_norm": 0.2189315388332226,
+      "learning_rate": 5.2289780794192726e-08,
+      "loss": 0.2148,
+      "step": 496
+    },
+    {
+      "epoch": 2.8238636363636362,
+      "grad_norm": 0.2992947334378841,
+      "learning_rate": 4.917899398289378e-08,
+      "loss": 0.2462,
+      "step": 497
+    },
+    {
+      "epoch": 2.8295454545454546,
+      "grad_norm": 0.22994410451461197,
+      "learning_rate": 4.6162680433503024e-08,
+      "loss": 0.2499,
+      "step": 498
+    },
+    {
+      "epoch": 2.8352272727272725,
+      "grad_norm": 0.20864426108924167,
+      "learning_rate": 4.3240956417661685e-08,
+      "loss": 0.2396,
+      "step": 499
+    },
+    {
+      "epoch": 2.840909090909091,
+      "grad_norm": 0.30251108346981104,
+      "learning_rate": 4.0413934560811216e-08,
+      "loss": 0.2644,
+      "step": 500
+    },
+    {
+      "epoch": 2.846590909090909,
+      "grad_norm": 0.18457127510671542,
+      "learning_rate": 3.768172383785268e-08,
+      "loss": 0.2195,
+      "step": 501
+    },
+    {
+      "epoch": 2.8522727272727275,
+      "grad_norm": 0.23763669865593898,
+      "learning_rate": 3.504442956894533e-08,
+      "loss": 0.2429,
+      "step": 502
+    },
+    {
+      "epoch": 2.8579545454545454,
+      "grad_norm": 0.22394648398208236,
+      "learning_rate": 3.250215341544766e-08,
+      "loss": 0.2227,
+      "step": 503
+    },
+    {
+      "epoch": 2.8636363636363638,
+      "grad_norm": 0.28772938681487004,
+      "learning_rate": 3.005499337599777e-08,
+      "loss": 0.2512,
+      "step": 504
+    },
+    {
+      "epoch": 2.8693181818181817,
+      "grad_norm": 0.20045670964039816,
+      "learning_rate": 2.7703043782735527e-08,
+      "loss": 0.2207,
+      "step": 505
+    },
+    {
+      "epoch": 2.875,
+      "grad_norm": 0.2988401612109738,
+      "learning_rate": 2.544639529766829e-08,
+      "loss": 0.2518,
+      "step": 506
+    },
+    {
+      "epoch": 2.8806818181818183,
+      "grad_norm": 0.23304873762081754,
+      "learning_rate": 2.3285134909173113e-08,
+      "loss": 0.2713,
+      "step": 507
+    },
+    {
+      "epoch": 2.8863636363636362,
+      "grad_norm": 0.20304650842985458,
+      "learning_rate": 2.1219345928646107e-08,
+      "loss": 0.2342,
+      "step": 508
+    },
+    {
+      "epoch": 2.8920454545454546,
+      "grad_norm": 0.22037485443764462,
+      "learning_rate": 1.924910798728946e-08,
+      "loss": 0.2711,
+      "step": 509
+    },
+    {
+      "epoch": 2.8977272727272725,
+      "grad_norm": 0.21005813876379573,
+      "learning_rate": 1.7374497033042504e-08,
+      "loss": 0.2561,
+      "step": 510
+    },
+    {
+      "epoch": 2.903409090909091,
+      "grad_norm": 0.2557348418435861,
+      "learning_rate": 1.559558532765404e-08,
+      "loss": 0.2289,
+      "step": 511
+    },
+    {
+      "epoch": 2.909090909090909,
+      "grad_norm": 0.23331959502677055,
+      "learning_rate": 1.3912441443896529e-08,
+      "loss": 0.2441,
+      "step": 512
+    },
+    {
+      "epoch": 2.9147727272727275,
+      "grad_norm": 0.19984687683627483,
+      "learning_rate": 1.2325130262923202e-08,
+      "loss": 0.224,
+      "step": 513
+    },
+    {
+      "epoch": 2.9204545454545454,
+      "grad_norm": 0.2743804065799622,
+      "learning_rate": 1.0833712971766442e-08,
+      "loss": 0.2375,
+      "step": 514
+    },
+    {
+      "epoch": 2.9261363636363638,
+      "grad_norm": 0.2224715742517231,
+      "learning_rate": 9.438247060979954e-09,
+      "loss": 0.2487,
+      "step": 515
+    },
+    {
+      "epoch": 2.9318181818181817,
+      "grad_norm": 0.2479834665000146,
+      "learning_rate": 8.13878632242221e-09,
+      "loss": 0.2297,
+      "step": 516
+    },
+    {
+      "epoch": 2.9375,
+      "grad_norm": 0.2886169100995835,
+      "learning_rate": 6.935380847182815e-09,
+      "loss": 0.2415,
+      "step": 517
+    },
+    {
+      "epoch": 2.9431818181818183,
+      "grad_norm": 0.28419909160412293,
+      "learning_rate": 5.828077023651846e-09,
+      "loss": 0.2595,
+      "step": 518
+    },
+    {
+      "epoch": 2.9488636363636362,
+      "grad_norm": 0.21987658336814492,
+      "learning_rate": 4.816917535731547e-09,
+      "loss": 0.2294,
+      "step": 519
+    },
+    {
+      "epoch": 2.9545454545454546,
+      "grad_norm": 0.20939447966816221,
+      "learning_rate": 3.9019413611907084e-09,
+      "loss": 0.2229,
+      "step": 520
+    },
+    {
+      "epoch": 2.9602272727272725,
+      "grad_norm": 0.2845597794271702,
+      "learning_rate": 3.083183770162812e-09,
+      "loss": 0.249,
+      "step": 521
+    },
+    {
+      "epoch": 2.965909090909091,
+      "grad_norm": 0.355389728910669,
+      "learning_rate": 2.360676323786282e-09,
+      "loss": 0.2279,
+      "step": 522
+    },
+    {
+      "epoch": 2.971590909090909,
+      "grad_norm": 0.24621464957062172,
+      "learning_rate": 1.7344468729868502e-09,
+      "loss": 0.2407,
+      "step": 523
+    },
+    {
+      "epoch": 2.9772727272727275,
+      "grad_norm": 0.20550412698336787,
+      "learning_rate": 1.2045195574056367e-09,
+      "loss": 0.2218,
+      "step": 524
+    },
+    {
+      "epoch": 2.9829545454545454,
+      "grad_norm": 0.3911983722494495,
+      "learning_rate": 7.709148044679482e-10,
+      "loss": 0.2403,
+      "step": 525
+    },
+    {
+      "epoch": 2.9886363636363638,
+      "grad_norm": 0.3446613906141255,
+      "learning_rate": 4.3364932859474293e-10,
+      "loss": 0.2473,
+      "step": 526
+    },
+    {
+      "epoch": 2.9943181818181817,
+      "grad_norm": 0.1984549594751942,
+      "learning_rate": 1.9273613056008945e-10,
+      "loss": 0.2491,
+      "step": 527
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.2533625348100663,
+      "learning_rate": 4.8184496989067684e-11,
+      "loss": 0.253,
+      "step": 528
+    },
+    {
+      "epoch": 3.0,
+      "step": 528,
+      "total_flos": 1.2947155177386803e+18,
+      "train_loss": 0.25888053223377827,
+      "train_runtime": 205758.8415,
+      "train_samples_per_second": 0.164,
+      "train_steps_per_second": 0.003
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 528,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.2947155177386803e+18,
+  "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 sha256:3cd1b75c8d1816d01a034bc687c7c7c89c79b66ccbd9a9f864800bcde9b85526
+size 6968