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
+}

config.json

@@ -0,0 +1,38 @@
+{
+  "architectures": [
+    "Qwen3MoeForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "decoder_sparse_step": 1,
+  "eos_token_id": 151643,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 6144,
+  "max_position_embeddings": 32768,
+  "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,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "router_aux_loss_coef": 0.001,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "bos_token_id": 151643,
+  "eos_token_id": 151643,
+  "max_new_tokens": 2048,
+  "transformers_version": "4.51.3"
+}

model-00001-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16edc6251d9c9bd9ff3c7849f502361f9fe07a68852b3a2bffe14439e2766421
+size 4997184968

model-00002-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a0c6bda61dc4e754b28b053e8afca42c18c3a1bd589a0ba014ffdf3fbc4ca53
+size 4997741608

model-00003-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64b1d59c4bee396b7901e13de3c9346b3701087bd32d78ccb8fe1732020bd596
+size 4997742208

model-00004-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:816b2874514236f1ff41d5b4ee72c83ad2be8f70b362eebebce7713e6004cd3d
+size 4997743184

model-00005-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1bdf8ef44b0daf86e0b736cbc5f38242c94719755ddc3f524d32ddaf995cac42
+size 4997743184

model-00006-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5aa04d60ae3b8904fd5f55a4929ee2ea3b3031d572d02e67a85fb200df80d674
+size 4997743184

model-00007-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:499b0c62c1340ce4319fb2a4a8d99e1e949636604bc5e48aab2c3e86795e0259
+size 4997743184

model-00008-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e934c332065c83f1a5a3f79de616b0ee0328d3ec28fb1675901efdd2eeeaa1a
+size 4997743184

model-00009-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e81b01429b2ba9439c60fc15ec429b9e7ea8690db66a3a6046d2559b89e87178
+size 4997743184

model-00010-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb968b901e7ff636847c3c8dd581b1ee9d1c7ad57aa210d49a71894fe72425ae
+size 4997743184

model-00011-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4cdba9e98d6d91ef18229e0ca273596264d6831c1e456e8cb0ebf43dcbd454b7
+size 4997743184

model-00012-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:994722f3776aa8c9de3883fa1e0e2436a4b9904ccaad75fb171b17fe38efa8d8
+size 4997743184

model-00013-of-00013.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f12abfc30695a9d5a34d85bdfeb87b1fc3be5cfd1e9b2e74f81891f7807496d
+size 1094220288

special_tokens_map.json

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

tokenizer.json

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

tokenizer_config.json

@@ -0,0 +1,241 @@
+{
+  "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": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0].role == 'system' %}\n        {{- messages[0].content + '\\n\\n' }}\n    {%- endif %}\n    {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0].role == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0].content + '<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}\n{%- for message in messages[::-1] %}\n    {%- set index = (messages|length - 1) - loop.index0 %}\n    {%- if ns.multi_step_tool and message.role == \"user\" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}\n        {%- set ns.multi_step_tool = false %}\n        {%- set ns.last_query_index = index %}\n    {%- endif %}\n{%- endfor %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {%- set content = message.content %}\n        {%- set reasoning_content = '' %}\n        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}\n            {%- set reasoning_content = message.reasoning_content %}\n        {%- else %}\n            {%- if '</think>' in message.content %}\n                {%- set content = message.content.split('</think>')[-1].lstrip('\\n') %}\n                {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\\n').split('<think>')[-1].lstrip('\\n') %}\n            {%- endif %}\n        {%- endif %}\n        {%- if loop.index0 > ns.last_query_index %}\n            {%- if loop.last or (not loop.last and reasoning_content) %}\n                {{- '<|im_start|>' + message.role + '\\n<think>\\n' + reasoning_content.strip('\\n') + '\\n</think>\\n\\n' + content.lstrip('\\n') }}\n            {%- else %}\n                {{- '<|im_start|>' + message.role + '\\n' + content }}\n            {%- endif %}\n        {%- else %}\n            {{- '<|im_start|>' + message.role + '\\n' + content }}\n        {%- endif %}\n        {%- if message.tool_calls %}\n            {%- for tool_call in message.tool_calls %}\n                {%- if (loop.first and content) or (not loop.first) %}\n                    {{- '\\n' }}\n                {%- endif %}\n                {%- if tool_call.function %}\n                    {%- set tool_call = tool_call.function %}\n                {%- endif %}\n                {{- '<tool_call>\\n{\"name\": \"' }}\n                {{- tool_call.name }}\n                {{- '\", \"arguments\": ' }}\n                {%- if tool_call.arguments is string %}\n                    {{- tool_call.arguments }}\n                {%- else %}\n                    {{- tool_call.arguments | tojson }}\n                {%- endif %}\n                {{- '}\\n</tool_call>' }}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n    {%- if enable_thinking is defined and enable_thinking is false %}\n        {{- '<think>\\n\\n</think>\\n\\n' }}\n    {%- endif %}\n{%- endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

trainer_state.json

@@ -0,0 +1,3534 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9033423667570009,
+  "eval_steps": 500,
+  "global_step": 500,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0018066847335140017,
+      "grad_norm": 2.856340142212096,
+      "learning_rate": 0.0,
+      "loss": 0.8251,
+      "step": 1
+    },
+    {
+      "epoch": 0.0036133694670280035,
+      "grad_norm": 2.8951332517054396,
+      "learning_rate": 1.0714285714285714e-06,
+      "loss": 0.8284,
+      "step": 2
+    },
+    {
+      "epoch": 0.005420054200542005,
+      "grad_norm": 2.9032377880415146,
+      "learning_rate": 2.1428571428571427e-06,
+      "loss": 0.8428,
+      "step": 3
+    },
+    {
+      "epoch": 0.007226738934056007,
+      "grad_norm": 2.8338573193945784,
+      "learning_rate": 3.2142857142857143e-06,
+      "loss": 0.8377,
+      "step": 4
+    },
+    {
+      "epoch": 0.009033423667570008,
+      "grad_norm": 2.5960999216862874,
+      "learning_rate": 4.2857142857142855e-06,
+      "loss": 0.8226,
+      "step": 5
+    },
+    {
+      "epoch": 0.01084010840108401,
+      "grad_norm": 2.089743525705263,
+      "learning_rate": 5.357142857142857e-06,
+      "loss": 0.8095,
+      "step": 6
+    },
+    {
+      "epoch": 0.012646793134598013,
+      "grad_norm": 1.5138563523925936,
+      "learning_rate": 6.428571428571429e-06,
+      "loss": 0.774,
+      "step": 7
+    },
+    {
+      "epoch": 0.014453477868112014,
+      "grad_norm": 1.4040706089231183,
+      "learning_rate": 7.5e-06,
+      "loss": 0.7734,
+      "step": 8
+    },
+    {
+      "epoch": 0.016260162601626018,
+      "grad_norm": 1.3582517891600783,
+      "learning_rate": 8.571428571428571e-06,
+      "loss": 0.7658,
+      "step": 9
+    },
+    {
+      "epoch": 0.018066847335140017,
+      "grad_norm": 2.377418251405277,
+      "learning_rate": 9.642857142857144e-06,
+      "loss": 0.7504,
+      "step": 10
+    },
+    {
+      "epoch": 0.01987353206865402,
+      "grad_norm": 2.4549416029918376,
+      "learning_rate": 1.0714285714285714e-05,
+      "loss": 0.7366,
+      "step": 11
+    },
+    {
+      "epoch": 0.02168021680216802,
+      "grad_norm": 2.2555108109041293,
+      "learning_rate": 1.1785714285714286e-05,
+      "loss": 0.7499,
+      "step": 12
+    },
+    {
+      "epoch": 0.023486901535682024,
+      "grad_norm": 1.8326851058754405,
+      "learning_rate": 1.2857142857142857e-05,
+      "loss": 0.7266,
+      "step": 13
+    },
+    {
+      "epoch": 0.025293586269196026,
+      "grad_norm": 1.8415475690845629,
+      "learning_rate": 1.3928571428571429e-05,
+      "loss": 0.7305,
+      "step": 14
+    },
+    {
+      "epoch": 0.02710027100271003,
+      "grad_norm": 1.4353926861015165,
+      "learning_rate": 1.5e-05,
+      "loss": 0.7186,
+      "step": 15
+    },
+    {
+      "epoch": 0.028906955736224028,
+      "grad_norm": 1.206228081083532,
+      "learning_rate": 1.6071428571428572e-05,
+      "loss": 0.7025,
+      "step": 16
+    },
+    {
+      "epoch": 0.03071364046973803,
+      "grad_norm": 0.8451068109607998,
+      "learning_rate": 1.7142857142857142e-05,
+      "loss": 0.7002,
+      "step": 17
+    },
+    {
+      "epoch": 0.032520325203252036,
+      "grad_norm": 0.8127082507274842,
+      "learning_rate": 1.8214285714285712e-05,
+      "loss": 0.6888,
+      "step": 18
+    },
+    {
+      "epoch": 0.03432700993676603,
+      "grad_norm": 0.7964889459374392,
+      "learning_rate": 1.928571428571429e-05,
+      "loss": 0.6827,
+      "step": 19
+    },
+    {
+      "epoch": 0.036133694670280034,
+      "grad_norm": 0.7187445808436193,
+      "learning_rate": 2.0357142857142858e-05,
+      "loss": 0.6596,
+      "step": 20
+    },
+    {
+      "epoch": 0.037940379403794036,
+      "grad_norm": 0.6276308058155495,
+      "learning_rate": 2.1428571428571428e-05,
+      "loss": 0.6654,
+      "step": 21
+    },
+    {
+      "epoch": 0.03974706413730804,
+      "grad_norm": 0.7008229118584881,
+      "learning_rate": 2.25e-05,
+      "loss": 0.6546,
+      "step": 22
+    },
+    {
+      "epoch": 0.04155374887082204,
+      "grad_norm": 0.5985175146313038,
+      "learning_rate": 2.357142857142857e-05,
+      "loss": 0.6549,
+      "step": 23
+    },
+    {
+      "epoch": 0.04336043360433604,
+      "grad_norm": 0.5391125438001545,
+      "learning_rate": 2.464285714285714e-05,
+      "loss": 0.66,
+      "step": 24
+    },
+    {
+      "epoch": 0.045167118337850046,
+      "grad_norm": 0.5912058816850287,
+      "learning_rate": 2.5714285714285714e-05,
+      "loss": 0.642,
+      "step": 25
+    },
+    {
+      "epoch": 0.04697380307136405,
+      "grad_norm": 0.4985965458263971,
+      "learning_rate": 2.6785714285714288e-05,
+      "loss": 0.6561,
+      "step": 26
+    },
+    {
+      "epoch": 0.04878048780487805,
+      "grad_norm": 0.5030750883133794,
+      "learning_rate": 2.7857142857142858e-05,
+      "loss": 0.6342,
+      "step": 27
+    },
+    {
+      "epoch": 0.05058717253839205,
+      "grad_norm": 0.5876044454362082,
+      "learning_rate": 2.892857142857143e-05,
+      "loss": 0.6372,
+      "step": 28
+    },
+    {
+      "epoch": 0.052393857271906055,
+      "grad_norm": 0.47044355779770175,
+      "learning_rate": 3e-05,
+      "loss": 0.6487,
+      "step": 29
+    },
+    {
+      "epoch": 0.05420054200542006,
+      "grad_norm": 0.4243036163988648,
+      "learning_rate": 2.9999731440137413e-05,
+      "loss": 0.6491,
+      "step": 30
+    },
+    {
+      "epoch": 0.05600722673893405,
+      "grad_norm": 0.3791575064900639,
+      "learning_rate": 2.9998925770166232e-05,
+      "loss": 0.6382,
+      "step": 31
+    },
+    {
+      "epoch": 0.057813911472448055,
+      "grad_norm": 0.5117054209056439,
+      "learning_rate": 2.9997583018935875e-05,
+      "loss": 0.6264,
+      "step": 32
+    },
+    {
+      "epoch": 0.05962059620596206,
+      "grad_norm": 0.3987355292978457,
+      "learning_rate": 2.9995703234527553e-05,
+      "loss": 0.6362,
+      "step": 33
+    },
+    {
+      "epoch": 0.06142728093947606,
+      "grad_norm": 0.36554956455062115,
+      "learning_rate": 2.999328648425255e-05,
+      "loss": 0.6239,
+      "step": 34
+    },
+    {
+      "epoch": 0.06323396567299007,
+      "grad_norm": 0.3685727924627649,
+      "learning_rate": 2.999033285464982e-05,
+      "loss": 0.6292,
+      "step": 35
+    },
+    {
+      "epoch": 0.06504065040650407,
+      "grad_norm": 0.4087523520863666,
+      "learning_rate": 2.9986842451482876e-05,
+      "loss": 0.6251,
+      "step": 36
+    },
+    {
+      "epoch": 0.06684733514001806,
+      "grad_norm": 0.3360266188354773,
+      "learning_rate": 2.9982815399736008e-05,
+      "loss": 0.6298,
+      "step": 37
+    },
+    {
+      "epoch": 0.06865401987353206,
+      "grad_norm": 0.36908008650337826,
+      "learning_rate": 2.9978251843609816e-05,
+      "loss": 0.6223,
+      "step": 38
+    },
+    {
+      "epoch": 0.07046070460704607,
+      "grad_norm": 0.4099167738856685,
+      "learning_rate": 2.9973151946516027e-05,
+      "loss": 0.6245,
+      "step": 39
+    },
+    {
+      "epoch": 0.07226738934056007,
+      "grad_norm": 0.33906987451657594,
+      "learning_rate": 2.996751589107167e-05,
+      "loss": 0.6126,
+      "step": 40
+    },
+    {
+      "epoch": 0.07407407407407407,
+      "grad_norm": 0.34433260301645086,
+      "learning_rate": 2.9961343879092512e-05,
+      "loss": 0.6188,
+      "step": 41
+    },
+    {
+      "epoch": 0.07588075880758807,
+      "grad_norm": 0.3499268147032493,
+      "learning_rate": 2.9954636131585845e-05,
+      "loss": 0.6122,
+      "step": 42
+    },
+    {
+      "epoch": 0.07768744354110207,
+      "grad_norm": 0.32793614431417295,
+      "learning_rate": 2.9947392888742566e-05,
+      "loss": 0.6118,
+      "step": 43
+    },
+    {
+      "epoch": 0.07949412827461608,
+      "grad_norm": 0.30973333856649427,
+      "learning_rate": 2.993961440992859e-05,
+      "loss": 0.6061,
+      "step": 44
+    },
+    {
+      "epoch": 0.08130081300813008,
+      "grad_norm": 0.3345295896786868,
+      "learning_rate": 2.993130097367553e-05,
+      "loss": 0.6135,
+      "step": 45
+    },
+    {
+      "epoch": 0.08310749774164408,
+      "grad_norm": 0.32719294127854426,
+      "learning_rate": 2.9922452877670775e-05,
+      "loss": 0.6148,
+      "step": 46
+    },
+    {
+      "epoch": 0.08491418247515808,
+      "grad_norm": 0.3224587604777241,
+      "learning_rate": 2.991307043874677e-05,
+      "loss": 0.6177,
+      "step": 47
+    },
+    {
+      "epoch": 0.08672086720867209,
+      "grad_norm": 0.33185222511919255,
+      "learning_rate": 2.9903153992869734e-05,
+      "loss": 0.6041,
+      "step": 48
+    },
+    {
+      "epoch": 0.08852755194218609,
+      "grad_norm": 0.3191136824919923,
+      "learning_rate": 2.989270389512756e-05,
+      "loss": 0.6112,
+      "step": 49
+    },
+    {
+      "epoch": 0.09033423667570009,
+      "grad_norm": 0.3262397814668044,
+      "learning_rate": 2.988172051971717e-05,
+      "loss": 0.6104,
+      "step": 50
+    },
+    {
+      "epoch": 0.0921409214092141,
+      "grad_norm": 0.3018631505024317,
+      "learning_rate": 2.9870204259931062e-05,
+      "loss": 0.6,
+      "step": 51
+    },
+    {
+      "epoch": 0.0939476061427281,
+      "grad_norm": 0.3233492233788076,
+      "learning_rate": 2.9858155528143256e-05,
+      "loss": 0.6175,
+      "step": 52
+    },
+    {
+      "epoch": 0.0957542908762421,
+      "grad_norm": 1.1675547386030114,
+      "learning_rate": 2.9845574755794522e-05,
+      "loss": 0.6032,
+      "step": 53
+    },
+    {
+      "epoch": 0.0975609756097561,
+      "grad_norm": 0.8614103747959223,
+      "learning_rate": 2.9832462393376926e-05,
+      "loss": 0.5931,
+      "step": 54
+    },
+    {
+      "epoch": 0.0993676603432701,
+      "grad_norm": 0.36866939977273266,
+      "learning_rate": 2.9818818910417706e-05,
+      "loss": 0.6088,
+      "step": 55
+    },
+    {
+      "epoch": 0.1011743450767841,
+      "grad_norm": 0.34023461156160906,
+      "learning_rate": 2.9804644795462437e-05,
+      "loss": 0.6133,
+      "step": 56
+    },
+    {
+      "epoch": 0.10298102981029811,
+      "grad_norm": 0.3238116268454224,
+      "learning_rate": 2.9789940556057574e-05,
+      "loss": 0.5994,
+      "step": 57
+    },
+    {
+      "epoch": 0.10478771454381211,
+      "grad_norm": 0.36171821384392444,
+      "learning_rate": 2.9774706718732255e-05,
+      "loss": 0.6074,
+      "step": 58
+    },
+    {
+      "epoch": 0.10659439927732611,
+      "grad_norm": 0.3036739995769363,
+      "learning_rate": 2.9758943828979444e-05,
+      "loss": 0.5936,
+      "step": 59
+    },
+    {
+      "epoch": 0.10840108401084012,
+      "grad_norm": 0.39561147640744193,
+      "learning_rate": 2.9742652451236414e-05,
+      "loss": 0.5951,
+      "step": 60
+    },
+    {
+      "epoch": 0.1102077687443541,
+      "grad_norm": 0.3195408550704984,
+      "learning_rate": 2.972583316886451e-05,
+      "loss": 0.5987,
+      "step": 61
+    },
+    {
+      "epoch": 0.1120144534778681,
+      "grad_norm": 0.36256179053996285,
+      "learning_rate": 2.9708486584128303e-05,
+      "loss": 0.6029,
+      "step": 62
+    },
+    {
+      "epoch": 0.11382113821138211,
+      "grad_norm": 0.3168276663836238,
+      "learning_rate": 2.9690613318173966e-05,
+      "loss": 0.6028,
+      "step": 63
+    },
+    {
+      "epoch": 0.11562782294489611,
+      "grad_norm": 0.3385072279584169,
+      "learning_rate": 2.9672214011007087e-05,
+      "loss": 0.6016,
+      "step": 64
+    },
+    {
+      "epoch": 0.11743450767841011,
+      "grad_norm": 0.3045076493642425,
+      "learning_rate": 2.9653289321469715e-05,
+      "loss": 0.5954,
+      "step": 65
+    },
+    {
+      "epoch": 0.11924119241192412,
+      "grad_norm": 0.28236729220672974,
+      "learning_rate": 2.9633839927216793e-05,
+      "loss": 0.5987,
+      "step": 66
+    },
+    {
+      "epoch": 0.12104787714543812,
+      "grad_norm": 0.32676542492238664,
+      "learning_rate": 2.9613866524691867e-05,
+      "loss": 0.589,
+      "step": 67
+    },
+    {
+      "epoch": 0.12285456187895212,
+      "grad_norm": 0.30030271908067696,
+      "learning_rate": 2.9593369829102173e-05,
+      "loss": 0.6044,
+      "step": 68
+    },
+    {
+      "epoch": 0.12466124661246612,
+      "grad_norm": 0.307464810600275,
+      "learning_rate": 2.957235057439301e-05,
+      "loss": 0.6059,
+      "step": 69
+    },
+    {
+      "epoch": 0.12646793134598014,
+      "grad_norm": 0.29750919621617494,
+      "learning_rate": 2.955080951322147e-05,
+      "loss": 0.6103,
+      "step": 70
+    },
+    {
+      "epoch": 0.12827461607949414,
+      "grad_norm": 0.46327257258765353,
+      "learning_rate": 2.9528747416929467e-05,
+      "loss": 0.5833,
+      "step": 71
+    },
+    {
+      "epoch": 0.13008130081300814,
+      "grad_norm": 0.30034640119560296,
+      "learning_rate": 2.9506165075516148e-05,
+      "loss": 0.5999,
+      "step": 72
+    },
+    {
+      "epoch": 0.13188798554652212,
+      "grad_norm": 0.3533019791208094,
+      "learning_rate": 2.9483063297609577e-05,
+      "loss": 0.5987,
+      "step": 73
+    },
+    {
+      "epoch": 0.13369467028003612,
+      "grad_norm": 0.3133821834377155,
+      "learning_rate": 2.9459442910437798e-05,
+      "loss": 0.591,
+      "step": 74
+    },
+    {
+      "epoch": 0.13550135501355012,
+      "grad_norm": 0.341056306207014,
+      "learning_rate": 2.94353047597992e-05,
+      "loss": 0.5898,
+      "step": 75
+    },
+    {
+      "epoch": 0.13730803974706413,
+      "grad_norm": 0.30089700024847743,
+      "learning_rate": 2.941064971003224e-05,
+      "loss": 0.5854,
+      "step": 76
+    },
+    {
+      "epoch": 0.13911472448057813,
+      "grad_norm": 0.34563853674055495,
+      "learning_rate": 2.9385478643984484e-05,
+      "loss": 0.5809,
+      "step": 77
+    },
+    {
+      "epoch": 0.14092140921409213,
+      "grad_norm": 0.3699790225189987,
+      "learning_rate": 2.9359792462981007e-05,
+      "loss": 0.5917,
+      "step": 78
+    },
+    {
+      "epoch": 0.14272809394760613,
+      "grad_norm": 0.47587430322144614,
+      "learning_rate": 2.9333592086792113e-05,
+      "loss": 0.5997,
+      "step": 79
+    },
+    {
+      "epoch": 0.14453477868112014,
+      "grad_norm": 0.37885442839342814,
+      "learning_rate": 2.9306878453600382e-05,
+      "loss": 0.6088,
+      "step": 80
+    },
+    {
+      "epoch": 0.14634146341463414,
+      "grad_norm": 0.33085997734962097,
+      "learning_rate": 2.9279652519967105e-05,
+      "loss": 0.5936,
+      "step": 81
+    },
+    {
+      "epoch": 0.14814814814814814,
+      "grad_norm": 0.39964165806810825,
+      "learning_rate": 2.9251915260798024e-05,
+      "loss": 0.586,
+      "step": 82
+    },
+    {
+      "epoch": 0.14995483288166214,
+      "grad_norm": 0.3476351246427051,
+      "learning_rate": 2.9223667669308395e-05,
+      "loss": 0.5959,
+      "step": 83
+    },
+    {
+      "epoch": 0.15176151761517614,
+      "grad_norm": 0.2849744592308525,
+      "learning_rate": 2.9194910756987464e-05,
+      "loss": 0.5898,
+      "step": 84
+    },
+    {
+      "epoch": 0.15356820234869015,
+      "grad_norm": 0.33683023117097444,
+      "learning_rate": 2.9165645553562215e-05,
+      "loss": 0.5925,
+      "step": 85
+    },
+    {
+      "epoch": 0.15537488708220415,
+      "grad_norm": 0.3377829007429357,
+      "learning_rate": 2.9135873106960525e-05,
+      "loss": 0.5914,
+      "step": 86
+    },
+    {
+      "epoch": 0.15718157181571815,
+      "grad_norm": 0.33675095070586347,
+      "learning_rate": 2.9105594483273603e-05,
+      "loss": 0.5941,
+      "step": 87
+    },
+    {
+      "epoch": 0.15898825654923215,
+      "grad_norm": 0.3299786139881529,
+      "learning_rate": 2.9074810766717865e-05,
+      "loss": 0.5983,
+      "step": 88
+    },
+    {
+      "epoch": 0.16079494128274616,
+      "grad_norm": 0.3317530707365067,
+      "learning_rate": 2.904352305959606e-05,
+      "loss": 0.5954,
+      "step": 89
+    },
+    {
+      "epoch": 0.16260162601626016,
+      "grad_norm": 0.3817908479661208,
+      "learning_rate": 2.9011732482257835e-05,
+      "loss": 0.587,
+      "step": 90
+    },
+    {
+      "epoch": 0.16440831074977416,
+      "grad_norm": 0.39134883533048737,
+      "learning_rate": 2.89794401730596e-05,
+      "loss": 0.5845,
+      "step": 91
+    },
+    {
+      "epoch": 0.16621499548328816,
+      "grad_norm": 0.35216735680799904,
+      "learning_rate": 2.894664728832377e-05,
+      "loss": 0.5891,
+      "step": 92
+    },
+    {
+      "epoch": 0.16802168021680217,
+      "grad_norm": 0.2876418130508674,
+      "learning_rate": 2.8913355002297367e-05,
+      "loss": 0.5947,
+      "step": 93
+    },
+    {
+      "epoch": 0.16982836495031617,
+      "grad_norm": 0.34255080300627916,
+      "learning_rate": 2.887956450710995e-05,
+      "loss": 0.5947,
+      "step": 94
+    },
+    {
+      "epoch": 0.17163504968383017,
+      "grad_norm": 0.36287002873794244,
+      "learning_rate": 2.8845277012730963e-05,
+      "loss": 0.5868,
+      "step": 95
+    },
+    {
+      "epoch": 0.17344173441734417,
+      "grad_norm": 0.2777633081272573,
+      "learning_rate": 2.8810493746926364e-05,
+      "loss": 0.5861,
+      "step": 96
+    },
+    {
+      "epoch": 0.17524841915085818,
+      "grad_norm": 0.3344975549131989,
+      "learning_rate": 2.87752159552147e-05,
+      "loss": 0.5922,
+      "step": 97
+    },
+    {
+      "epoch": 0.17705510388437218,
+      "grad_norm": 0.2767648422009983,
+      "learning_rate": 2.87394449008225e-05,
+      "loss": 0.5846,
+      "step": 98
+    },
+    {
+      "epoch": 0.17886178861788618,
+      "grad_norm": 0.34145685485506944,
+      "learning_rate": 2.8703181864639013e-05,
+      "loss": 0.5982,
+      "step": 99
+    },
+    {
+      "epoch": 0.18066847335140018,
+      "grad_norm": 0.3283253672178832,
+      "learning_rate": 2.8666428145170385e-05,
+      "loss": 0.571,
+      "step": 100
+    },
+    {
+      "epoch": 0.18247515808491419,
+      "grad_norm": 0.7591732431618092,
+      "learning_rate": 2.8629185058493116e-05,
+      "loss": 0.5872,
+      "step": 101
+    },
+    {
+      "epoch": 0.1842818428184282,
+      "grad_norm": 0.3744655373953689,
+      "learning_rate": 2.8591453938206985e-05,
+      "loss": 0.6056,
+      "step": 102
+    },
+    {
+      "epoch": 0.1860885275519422,
+      "grad_norm": 0.2943651931913048,
+      "learning_rate": 2.8553236135387247e-05,
+      "loss": 0.5837,
+      "step": 103
+    },
+    {
+      "epoch": 0.1878952122854562,
+      "grad_norm": 0.36771572937915975,
+      "learning_rate": 2.8514533018536286e-05,
+      "loss": 0.5821,
+      "step": 104
+    },
+    {
+      "epoch": 0.1897018970189702,
+      "grad_norm": 0.2933240647845247,
+      "learning_rate": 2.8475345973534605e-05,
+      "loss": 0.575,
+      "step": 105
+    },
+    {
+      "epoch": 0.1915085817524842,
+      "grad_norm": 0.33474489976765504,
+      "learning_rate": 2.8435676403591193e-05,
+      "loss": 0.5777,
+      "step": 106
+    },
+    {
+      "epoch": 0.1933152664859982,
+      "grad_norm": 0.28810788343158517,
+      "learning_rate": 2.8395525729193284e-05,
+      "loss": 0.5837,
+      "step": 107
+    },
+    {
+      "epoch": 0.1951219512195122,
+      "grad_norm": 0.2798402180713021,
+      "learning_rate": 2.835489538805548e-05,
+      "loss": 0.5817,
+      "step": 108
+    },
+    {
+      "epoch": 0.1969286359530262,
+      "grad_norm": 0.3076540058939344,
+      "learning_rate": 2.8313786835068314e-05,
+      "loss": 0.5905,
+      "step": 109
+    },
+    {
+      "epoch": 0.1987353206865402,
+      "grad_norm": 0.29718835850752623,
+      "learning_rate": 2.8272201542246077e-05,
+      "loss": 0.5725,
+      "step": 110
+    },
+    {
+      "epoch": 0.2005420054200542,
+      "grad_norm": 0.3061355603306112,
+      "learning_rate": 2.8230140998674185e-05,
+      "loss": 0.5917,
+      "step": 111
+    },
+    {
+      "epoch": 0.2023486901535682,
+      "grad_norm": 0.2821188737882899,
+      "learning_rate": 2.8187606710455807e-05,
+      "loss": 0.585,
+      "step": 112
+    },
+    {
+      "epoch": 0.2041553748870822,
+      "grad_norm": 0.3033302866342871,
+      "learning_rate": 2.8144600200657953e-05,
+      "loss": 0.589,
+      "step": 113
+    },
+    {
+      "epoch": 0.20596205962059622,
+      "grad_norm": 0.3314322092338695,
+      "learning_rate": 2.8101123009256946e-05,
+      "loss": 0.5722,
+      "step": 114
+    },
+    {
+      "epoch": 0.20776874435411022,
+      "grad_norm": 0.28608928204987905,
+      "learning_rate": 2.8057176693083253e-05,
+      "loss": 0.5882,
+      "step": 115
+    },
+    {
+      "epoch": 0.20957542908762422,
+      "grad_norm": 0.3057108753865152,
+      "learning_rate": 2.8012762825765763e-05,
+      "loss": 0.5818,
+      "step": 116
+    },
+    {
+      "epoch": 0.21138211382113822,
+      "grad_norm": 0.3209910234174254,
+      "learning_rate": 2.7967882997675424e-05,
+      "loss": 0.5928,
+      "step": 117
+    },
+    {
+      "epoch": 0.21318879855465223,
+      "grad_norm": 0.30437467545397623,
+      "learning_rate": 2.7922538815868287e-05,
+      "loss": 0.58,
+      "step": 118
+    },
+    {
+      "epoch": 0.21499548328816623,
+      "grad_norm": 0.2801009874570902,
+      "learning_rate": 2.7876731904027994e-05,
+      "loss": 0.5709,
+      "step": 119
+    },
+    {
+      "epoch": 0.21680216802168023,
+      "grad_norm": 0.3103706652157662,
+      "learning_rate": 2.78304639024076e-05,
+      "loss": 0.5895,
+      "step": 120
+    },
+    {
+      "epoch": 0.2186088527551942,
+      "grad_norm": 0.27937342188480463,
+      "learning_rate": 2.7783736467770863e-05,
+      "loss": 0.585,
+      "step": 121
+    },
+    {
+      "epoch": 0.2204155374887082,
+      "grad_norm": 0.2997029543835232,
+      "learning_rate": 2.7736551273332908e-05,
+      "loss": 0.5894,
+      "step": 122
+    },
+    {
+      "epoch": 0.2222222222222222,
+      "grad_norm": 0.34384629773558706,
+      "learning_rate": 2.7688910008700305e-05,
+      "loss": 0.5796,
+      "step": 123
+    },
+    {
+      "epoch": 0.2240289069557362,
+      "grad_norm": 0.2664105289861669,
+      "learning_rate": 2.764081437981059e-05,
+      "loss": 0.5659,
+      "step": 124
+    },
+    {
+      "epoch": 0.22583559168925021,
+      "grad_norm": 0.3186538282497774,
+      "learning_rate": 2.7592266108871158e-05,
+      "loss": 0.5734,
+      "step": 125
+    },
+    {
+      "epoch": 0.22764227642276422,
+      "grad_norm": 0.32445318752564883,
+      "learning_rate": 2.754326693429761e-05,
+      "loss": 0.5857,
+      "step": 126
+    },
+    {
+      "epoch": 0.22944896115627822,
+      "grad_norm": 0.29475841183859425,
+      "learning_rate": 2.7493818610651493e-05,
+      "loss": 0.5784,
+      "step": 127
+    },
+    {
+      "epoch": 0.23125564588979222,
+      "grad_norm": 0.2921391135668834,
+      "learning_rate": 2.744392290857747e-05,
+      "loss": 0.5776,
+      "step": 128
+    },
+    {
+      "epoch": 0.23306233062330622,
+      "grad_norm": 0.3087235604232491,
+      "learning_rate": 2.7393581614739924e-05,
+      "loss": 0.5726,
+      "step": 129
+    },
+    {
+      "epoch": 0.23486901535682023,
+      "grad_norm": 0.31066353921877315,
+      "learning_rate": 2.7342796531758984e-05,
+      "loss": 0.5909,
+      "step": 130
+    },
+    {
+      "epoch": 0.23667570009033423,
+      "grad_norm": 0.29890929195605337,
+      "learning_rate": 2.729156947814598e-05,
+      "loss": 0.5853,
+      "step": 131
+    },
+    {
+      "epoch": 0.23848238482384823,
+      "grad_norm": 0.28472433105870404,
+      "learning_rate": 2.7239902288238297e-05,
+      "loss": 0.5646,
+      "step": 132
+    },
+    {
+      "epoch": 0.24028906955736223,
+      "grad_norm": 0.2736200531503188,
+      "learning_rate": 2.7187796812133733e-05,
+      "loss": 0.5796,
+      "step": 133
+    },
+    {
+      "epoch": 0.24209575429087624,
+      "grad_norm": 0.3070080436261309,
+      "learning_rate": 2.7135254915624213e-05,
+      "loss": 0.5809,
+      "step": 134
+    },
+    {
+      "epoch": 0.24390243902439024,
+      "grad_norm": 0.3265703089577213,
+      "learning_rate": 2.708227848012901e-05,
+      "loss": 0.578,
+      "step": 135
+    },
+    {
+      "epoch": 0.24570912375790424,
+      "grad_norm": 0.30391687833637804,
+      "learning_rate": 2.7028869402627357e-05,
+      "loss": 0.5813,
+      "step": 136
+    },
+    {
+      "epoch": 0.24751580849141824,
+      "grad_norm": 0.2744127434629191,
+      "learning_rate": 2.6975029595590523e-05,
+      "loss": 0.5652,
+      "step": 137
+    },
+    {
+      "epoch": 0.24932249322493225,
+      "grad_norm": 0.29860630701447144,
+      "learning_rate": 2.6920760986913332e-05,
+      "loss": 0.5822,
+      "step": 138
+    },
+    {
+      "epoch": 0.25112917795844625,
+      "grad_norm": 0.288217397793013,
+      "learning_rate": 2.6866065519845124e-05,
+      "loss": 0.5852,
+      "step": 139
+    },
+    {
+      "epoch": 0.2529358626919603,
+      "grad_norm": 0.3102177204255471,
+      "learning_rate": 2.681094515292018e-05,
+      "loss": 0.5839,
+      "step": 140
+    },
+    {
+      "epoch": 0.25474254742547425,
+      "grad_norm": 0.25876938510867553,
+      "learning_rate": 2.6755401859887598e-05,
+      "loss": 0.5652,
+      "step": 141
+    },
+    {
+      "epoch": 0.2565492321589883,
+      "grad_norm": 0.35200793877127573,
+      "learning_rate": 2.6699437629640595e-05,
+      "loss": 0.5851,
+      "step": 142
+    },
+    {
+      "epoch": 0.25835591689250226,
+      "grad_norm": 0.3366233594400498,
+      "learning_rate": 2.6643054466145297e-05,
+      "loss": 0.5769,
+      "step": 143
+    },
+    {
+      "epoch": 0.2601626016260163,
+      "grad_norm": 0.2744652872976245,
+      "learning_rate": 2.6586254388368995e-05,
+      "loss": 0.5651,
+      "step": 144
+    },
+    {
+      "epoch": 0.26196928635953026,
+      "grad_norm": 0.2819069426916813,
+      "learning_rate": 2.652903943020783e-05,
+      "loss": 0.5782,
+      "step": 145
+    },
+    {
+      "epoch": 0.26377597109304424,
+      "grad_norm": 0.28188684980039447,
+      "learning_rate": 2.647141164041398e-05,
+      "loss": 0.5832,
+      "step": 146
+    },
+    {
+      "epoch": 0.26558265582655827,
+      "grad_norm": 0.3022789575136876,
+      "learning_rate": 2.641337308252228e-05,
+      "loss": 0.5723,
+      "step": 147
+    },
+    {
+      "epoch": 0.26738934056007224,
+      "grad_norm": 0.2969708316880134,
+      "learning_rate": 2.6354925834776346e-05,
+      "loss": 0.5737,
+      "step": 148
+    },
+    {
+      "epoch": 0.26919602529358627,
+      "grad_norm": 0.3174068946369213,
+      "learning_rate": 2.6296071990054167e-05,
+      "loss": 0.5715,
+      "step": 149
+    },
+    {
+      "epoch": 0.27100271002710025,
+      "grad_norm": 0.3030181528315971,
+      "learning_rate": 2.6236813655793123e-05,
+      "loss": 0.5811,
+      "step": 150
+    },
+    {
+      "epoch": 0.2728093947606143,
+      "grad_norm": 0.2899938347916164,
+      "learning_rate": 2.617715295391457e-05,
+      "loss": 0.5626,
+      "step": 151
+    },
+    {
+      "epoch": 0.27461607949412825,
+      "grad_norm": 0.27779649164699266,
+      "learning_rate": 2.6117092020747824e-05,
+      "loss": 0.5707,
+      "step": 152
+    },
+    {
+      "epoch": 0.2764227642276423,
+      "grad_norm": 0.2926021527795552,
+      "learning_rate": 2.6056633006953677e-05,
+      "loss": 0.5708,
+      "step": 153
+    },
+    {
+      "epoch": 0.27822944896115626,
+      "grad_norm": 0.26963336820522915,
+      "learning_rate": 2.5995778077447393e-05,
+      "loss": 0.5745,
+      "step": 154
+    },
+    {
+      "epoch": 0.2800361336946703,
+      "grad_norm": 0.3349132892798814,
+      "learning_rate": 2.5934529411321174e-05,
+      "loss": 0.5672,
+      "step": 155
+    },
+    {
+      "epoch": 0.28184281842818426,
+      "grad_norm": 0.2590271467673413,
+      "learning_rate": 2.587288920176613e-05,
+      "loss": 0.5669,
+      "step": 156
+    },
+    {
+      "epoch": 0.2836495031616983,
+      "grad_norm": 0.29630089545374866,
+      "learning_rate": 2.581085965599375e-05,
+      "loss": 0.5804,
+      "step": 157
+    },
+    {
+      "epoch": 0.28545618789521227,
+      "grad_norm": 0.2702307351115828,
+      "learning_rate": 2.5748442995156882e-05,
+      "loss": 0.5773,
+      "step": 158
+    },
+    {
+      "epoch": 0.2872628726287263,
+      "grad_norm": 0.29324539904768665,
+      "learning_rate": 2.5685641454270172e-05,
+      "loss": 0.5588,
+      "step": 159
+    },
+    {
+      "epoch": 0.28906955736224027,
+      "grad_norm": 0.31279263255156625,
+      "learning_rate": 2.5622457282130046e-05,
+      "loss": 0.5679,
+      "step": 160
+    },
+    {
+      "epoch": 0.2908762420957543,
+      "grad_norm": 0.2751393817619885,
+      "learning_rate": 2.5558892741234173e-05,
+      "loss": 0.5755,
+      "step": 161
+    },
+    {
+      "epoch": 0.2926829268292683,
+      "grad_norm": 0.30386224893133906,
+      "learning_rate": 2.5494950107700482e-05,
+      "loss": 0.5614,
+      "step": 162
+    },
+    {
+      "epoch": 0.2944896115627823,
+      "grad_norm": 0.28104662190654356,
+      "learning_rate": 2.5430631671185616e-05,
+      "loss": 0.5696,
+      "step": 163
+    },
+    {
+      "epoch": 0.2962962962962963,
+      "grad_norm": 0.28742135786623646,
+      "learning_rate": 2.5365939734802973e-05,
+      "loss": 0.5793,
+      "step": 164
+    },
+    {
+      "epoch": 0.2981029810298103,
+      "grad_norm": 0.27235295379008123,
+      "learning_rate": 2.5300876615040223e-05,
+      "loss": 0.5701,
+      "step": 165
+    },
+    {
+      "epoch": 0.2999096657633243,
+      "grad_norm": 0.3188704256651444,
+      "learning_rate": 2.523544464167637e-05,
+      "loss": 0.5765,
+      "step": 166
+    },
+    {
+      "epoch": 0.3017163504968383,
+      "grad_norm": 0.26343489233453454,
+      "learning_rate": 2.5169646157698313e-05,
+      "loss": 0.5749,
+      "step": 167
+    },
+    {
+      "epoch": 0.3035230352303523,
+      "grad_norm": 0.2865611587869202,
+      "learning_rate": 2.5103483519216964e-05,
+      "loss": 0.5736,
+      "step": 168
+    },
+    {
+      "epoch": 0.3053297199638663,
+      "grad_norm": 0.26728258410063105,
+      "learning_rate": 2.5036959095382875e-05,
+      "loss": 0.566,
+      "step": 169
+    },
+    {
+      "epoch": 0.3071364046973803,
+      "grad_norm": 0.2820790941687026,
+      "learning_rate": 2.4970075268301388e-05,
+      "loss": 0.5796,
+      "step": 170
+    },
+    {
+      "epoch": 0.3089430894308943,
+      "grad_norm": 0.2700876452038959,
+      "learning_rate": 2.4902834432947353e-05,
+      "loss": 0.5629,
+      "step": 171
+    },
+    {
+      "epoch": 0.3107497741644083,
+      "grad_norm": 0.2729250561913214,
+      "learning_rate": 2.4835238997079382e-05,
+      "loss": 0.5695,
+      "step": 172
+    },
+    {
+      "epoch": 0.31255645889792233,
+      "grad_norm": 0.2834762697483736,
+      "learning_rate": 2.4767291381153603e-05,
+      "loss": 0.573,
+      "step": 173
+    },
+    {
+      "epoch": 0.3143631436314363,
+      "grad_norm": 0.24895246096135648,
+      "learning_rate": 2.4698994018236994e-05,
+      "loss": 0.5639,
+      "step": 174
+    },
+    {
+      "epoch": 0.31616982836495033,
+      "grad_norm": 0.2795685572559788,
+      "learning_rate": 2.4630349353920284e-05,
+      "loss": 0.5684,
+      "step": 175
+    },
+    {
+      "epoch": 0.3179765130984643,
+      "grad_norm": 0.29024431988696586,
+      "learning_rate": 2.4561359846230346e-05,
+      "loss": 0.5823,
+      "step": 176
+    },
+    {
+      "epoch": 0.31978319783197834,
+      "grad_norm": 0.258992859302709,
+      "learning_rate": 2.4492027965542217e-05,
+      "loss": 0.5581,
+      "step": 177
+    },
+    {
+      "epoch": 0.3215898825654923,
+      "grad_norm": 0.276483347231218,
+      "learning_rate": 2.44223561944906e-05,
+      "loss": 0.565,
+      "step": 178
+    },
+    {
+      "epoch": 0.32339656729900634,
+      "grad_norm": 0.2633075704342389,
+      "learning_rate": 2.4352347027881003e-05,
+      "loss": 0.5744,
+      "step": 179
+    },
+    {
+      "epoch": 0.3252032520325203,
+      "grad_norm": 0.2754735044492125,
+      "learning_rate": 2.4282002972600382e-05,
+      "loss": 0.5825,
+      "step": 180
+    },
+    {
+      "epoch": 0.32700993676603435,
+      "grad_norm": 0.2621633378275852,
+      "learning_rate": 2.4211326547527377e-05,
+      "loss": 0.5611,
+      "step": 181
+    },
+    {
+      "epoch": 0.3288166214995483,
+      "grad_norm": 0.2808746811690711,
+      "learning_rate": 2.4140320283442125e-05,
+      "loss": 0.571,
+      "step": 182
+    },
+    {
+      "epoch": 0.33062330623306235,
+      "grad_norm": 0.2887210711912606,
+      "learning_rate": 2.4068986722935625e-05,
+      "loss": 0.5816,
+      "step": 183
+    },
+    {
+      "epoch": 0.3324299909665763,
+      "grad_norm": 0.2857284132894835,
+      "learning_rate": 2.3997328420318705e-05,
+      "loss": 0.578,
+      "step": 184
+    },
+    {
+      "epoch": 0.33423667570009036,
+      "grad_norm": 0.2979433369387901,
+      "learning_rate": 2.3925347941530556e-05,
+      "loss": 0.5744,
+      "step": 185
+    },
+    {
+      "epoch": 0.33604336043360433,
+      "grad_norm": 0.2615907243391146,
+      "learning_rate": 2.3853047864046843e-05,
+      "loss": 0.5693,
+      "step": 186
+    },
+    {
+      "epoch": 0.33785004516711836,
+      "grad_norm": 0.2915356887177567,
+      "learning_rate": 2.3780430776787413e-05,
+      "loss": 0.5663,
+      "step": 187
+    },
+    {
+      "epoch": 0.33965672990063234,
+      "grad_norm": 0.24760068221876105,
+      "learning_rate": 2.3707499280023604e-05,
+      "loss": 0.5666,
+      "step": 188
+    },
+    {
+      "epoch": 0.34146341463414637,
+      "grad_norm": 0.29126927804076114,
+      "learning_rate": 2.3634255985285104e-05,
+      "loss": 0.5739,
+      "step": 189
+    },
+    {
+      "epoch": 0.34327009936766034,
+      "grad_norm": 0.267270750792911,
+      "learning_rate": 2.356070351526648e-05,
+      "loss": 0.5628,
+      "step": 190
+    },
+    {
+      "epoch": 0.34507678410117437,
+      "grad_norm": 0.2721749843090162,
+      "learning_rate": 2.348684450373322e-05,
+      "loss": 0.5605,
+      "step": 191
+    },
+    {
+      "epoch": 0.34688346883468835,
+      "grad_norm": 0.263555551581403,
+      "learning_rate": 2.3412681595427467e-05,
+      "loss": 0.5682,
+      "step": 192
+    },
+    {
+      "epoch": 0.3486901535682023,
+      "grad_norm": 0.28601230100868225,
+      "learning_rate": 2.3338217445973268e-05,
+      "loss": 0.5696,
+      "step": 193
+    },
+    {
+      "epoch": 0.35049683830171635,
+      "grad_norm": 0.2742823730620319,
+      "learning_rate": 2.3263454721781537e-05,
+      "loss": 0.5714,
+      "step": 194
+    },
+    {
+      "epoch": 0.3523035230352303,
+      "grad_norm": 0.28725351491141005,
+      "learning_rate": 2.318839609995453e-05,
+      "loss": 0.5549,
+      "step": 195
+    },
+    {
+      "epoch": 0.35411020776874436,
+      "grad_norm": 0.26198310789623336,
+      "learning_rate": 2.3113044268189995e-05,
+      "loss": 0.5675,
+      "step": 196
+    },
+    {
+      "epoch": 0.35591689250225833,
+      "grad_norm": 0.2514021936041488,
+      "learning_rate": 2.303740192468495e-05,
+      "loss": 0.5676,
+      "step": 197
+    },
+    {
+      "epoch": 0.35772357723577236,
+      "grad_norm": 0.25544153506945566,
+      "learning_rate": 2.2961471778039045e-05,
+      "loss": 0.5446,
+      "step": 198
+    },
+    {
+      "epoch": 0.35953026196928634,
+      "grad_norm": 0.26725221085105316,
+      "learning_rate": 2.288525654715757e-05,
+      "loss": 0.5677,
+      "step": 199
+    },
+    {
+      "epoch": 0.36133694670280037,
+      "grad_norm": 0.2710926729712142,
+      "learning_rate": 2.280875896115413e-05,
+      "loss": 0.5566,
+      "step": 200
+    },
+    {
+      "epoch": 0.36314363143631434,
+      "grad_norm": 0.2743559460193731,
+      "learning_rate": 2.2731981759252876e-05,
+      "loss": 0.5512,
+      "step": 201
+    },
+    {
+      "epoch": 0.36495031616982837,
+      "grad_norm": 0.23972000569515187,
+      "learning_rate": 2.2654927690690445e-05,
+      "loss": 0.578,
+      "step": 202
+    },
+    {
+      "epoch": 0.36675700090334235,
+      "grad_norm": 0.28424099186211327,
+      "learning_rate": 2.257759951461752e-05,
+      "loss": 0.5652,
+      "step": 203
+    },
+    {
+      "epoch": 0.3685636856368564,
+      "grad_norm": 0.28275959894008174,
+      "learning_rate": 2.25e-05,
+      "loss": 0.5777,
+      "step": 204
+    },
+    {
+      "epoch": 0.37037037037037035,
+      "grad_norm": 0.25257610551688425,
+      "learning_rate": 2.24221319255199e-05,
+      "loss": 0.5681,
+      "step": 205
+    },
+    {
+      "epoch": 0.3721770551038844,
+      "grad_norm": 0.2733254950534001,
+      "learning_rate": 2.234399807947579e-05,
+      "loss": 0.5552,
+      "step": 206
+    },
+    {
+      "epoch": 0.37398373983739835,
+      "grad_norm": 0.2935291399566541,
+      "learning_rate": 2.2265601259683e-05,
+      "loss": 0.5633,
+      "step": 207
+    },
+    {
+      "epoch": 0.3757904245709124,
+      "grad_norm": 0.2647620630872173,
+      "learning_rate": 2.2186944273373426e-05,
+      "loss": 0.551,
+      "step": 208
+    },
+    {
+      "epoch": 0.37759710930442636,
+      "grad_norm": 0.26404141294185235,
+      "learning_rate": 2.210802993709498e-05,
+      "loss": 0.5723,
+      "step": 209
+    },
+    {
+      "epoch": 0.3794037940379404,
+      "grad_norm": 0.2794808048034287,
+      "learning_rate": 2.202886107661078e-05,
+      "loss": 0.5685,
+      "step": 210
+    },
+    {
+      "epoch": 0.38121047877145436,
+      "grad_norm": 0.24698040415270853,
+      "learning_rate": 2.1949440526797928e-05,
+      "loss": 0.5677,
+      "step": 211
+    },
+    {
+      "epoch": 0.3830171635049684,
+      "grad_norm": 0.26493352241096035,
+      "learning_rate": 2.1869771131546015e-05,
+      "loss": 0.5754,
+      "step": 212
+    },
+    {
+      "epoch": 0.38482384823848237,
+      "grad_norm": 0.27005961064526746,
+      "learning_rate": 2.178985574365529e-05,
+      "loss": 0.5761,
+      "step": 213
+    },
+    {
+      "epoch": 0.3866305329719964,
+      "grad_norm": 0.26417433576636645,
+      "learning_rate": 2.170969722473449e-05,
+      "loss": 0.5617,
+      "step": 214
+    },
+    {
+      "epoch": 0.3884372177055104,
+      "grad_norm": 0.3423329599625978,
+      "learning_rate": 2.1629298445098403e-05,
+      "loss": 0.5888,
+      "step": 215
+    },
+    {
+      "epoch": 0.3902439024390244,
+      "grad_norm": 0.25349686334946275,
+      "learning_rate": 2.154866228366505e-05,
+      "loss": 0.577,
+      "step": 216
+    },
+    {
+      "epoch": 0.3920505871725384,
+      "grad_norm": 0.2667613165609595,
+      "learning_rate": 2.146779162785263e-05,
+      "loss": 0.5773,
+      "step": 217
+    },
+    {
+      "epoch": 0.3938572719060524,
+      "grad_norm": 0.2748791533458348,
+      "learning_rate": 2.138668937347609e-05,
+      "loss": 0.5695,
+      "step": 218
+    },
+    {
+      "epoch": 0.3956639566395664,
+      "grad_norm": 0.2584383208693147,
+      "learning_rate": 2.1305358424643484e-05,
+      "loss": 0.5644,
+      "step": 219
+    },
+    {
+      "epoch": 0.3974706413730804,
+      "grad_norm": 0.24548945436183006,
+      "learning_rate": 2.1223801693651927e-05,
+      "loss": 0.5663,
+      "step": 220
+    },
+    {
+      "epoch": 0.3992773261065944,
+      "grad_norm": 0.26923305044108836,
+      "learning_rate": 2.114202210088336e-05,
+      "loss": 0.5555,
+      "step": 221
+    },
+    {
+      "epoch": 0.4010840108401084,
+      "grad_norm": 0.2481503736019401,
+      "learning_rate": 2.106002257469993e-05,
+      "loss": 0.5727,
+      "step": 222
+    },
+    {
+      "epoch": 0.4028906955736224,
+      "grad_norm": 0.2438014817553578,
+      "learning_rate": 2.0977806051339172e-05,
+      "loss": 0.5555,
+      "step": 223
+    },
+    {
+      "epoch": 0.4046973803071364,
+      "grad_norm": 0.24661736205016646,
+      "learning_rate": 2.0895375474808857e-05,
+      "loss": 0.5599,
+      "step": 224
+    },
+    {
+      "epoch": 0.4065040650406504,
+      "grad_norm": 0.28507710205686154,
+      "learning_rate": 2.0812733796781544e-05,
+      "loss": 0.5652,
+      "step": 225
+    },
+    {
+      "epoch": 0.4083107497741644,
+      "grad_norm": 0.2345546061331579,
+      "learning_rate": 2.0729883976488936e-05,
+      "loss": 0.5711,
+      "step": 226
+    },
+    {
+      "epoch": 0.4101174345076784,
+      "grad_norm": 0.2753853952012803,
+      "learning_rate": 2.064682898061588e-05,
+      "loss": 0.5753,
+      "step": 227
+    },
+    {
+      "epoch": 0.41192411924119243,
+      "grad_norm": 0.2665502889667531,
+      "learning_rate": 2.0563571783194146e-05,
+      "loss": 0.562,
+      "step": 228
+    },
+    {
+      "epoch": 0.4137308039747064,
+      "grad_norm": 0.24377225460049187,
+      "learning_rate": 2.0480115365495928e-05,
+      "loss": 0.5641,
+      "step": 229
+    },
+    {
+      "epoch": 0.41553748870822044,
+      "grad_norm": 0.2647925611744176,
+      "learning_rate": 2.0396462715927107e-05,
+      "loss": 0.5664,
+      "step": 230
+    },
+    {
+      "epoch": 0.4173441734417344,
+      "grad_norm": 0.342102177343815,
+      "learning_rate": 2.0312616829920222e-05,
+      "loss": 0.5659,
+      "step": 231
+    },
+    {
+      "epoch": 0.41915085817524844,
+      "grad_norm": 0.243969430790238,
+      "learning_rate": 2.022858070982723e-05,
+      "loss": 0.5687,
+      "step": 232
+    },
+    {
+      "epoch": 0.4209575429087624,
+      "grad_norm": 0.2300953816194087,
+      "learning_rate": 2.0144357364811973e-05,
+      "loss": 0.5504,
+      "step": 233
+    },
+    {
+      "epoch": 0.42276422764227645,
+      "grad_norm": 0.25714355821754453,
+      "learning_rate": 2.0059949810742452e-05,
+      "loss": 0.5659,
+      "step": 234
+    },
+    {
+      "epoch": 0.4245709123757904,
+      "grad_norm": 0.22585242580872408,
+      "learning_rate": 1.997536107008281e-05,
+      "loss": 0.5613,
+      "step": 235
+    },
+    {
+      "epoch": 0.42637759710930445,
+      "grad_norm": 0.24695641665189788,
+      "learning_rate": 1.9890594171785128e-05,
+      "loss": 0.5726,
+      "step": 236
+    },
+    {
+      "epoch": 0.4281842818428184,
+      "grad_norm": 0.29309439020784595,
+      "learning_rate": 1.9805652151180945e-05,
+      "loss": 0.5724,
+      "step": 237
+    },
+    {
+      "epoch": 0.42999096657633246,
+      "grad_norm": 0.23918338870842643,
+      "learning_rate": 1.972053804987258e-05,
+      "loss": 0.5805,
+      "step": 238
+    },
+    {
+      "epoch": 0.43179765130984643,
+      "grad_norm": 0.24359713017650977,
+      "learning_rate": 1.963525491562421e-05,
+      "loss": 0.5617,
+      "step": 239
+    },
+    {
+      "epoch": 0.43360433604336046,
+      "grad_norm": 0.2460747441245982,
+      "learning_rate": 1.954980580225275e-05,
+      "loss": 0.5619,
+      "step": 240
+    },
+    {
+      "epoch": 0.43541102077687444,
+      "grad_norm": 0.23922583066140365,
+      "learning_rate": 1.946419376951848e-05,
+      "loss": 0.5608,
+      "step": 241
+    },
+    {
+      "epoch": 0.4372177055103884,
+      "grad_norm": 0.23168259157144178,
+      "learning_rate": 1.9378421883015505e-05,
+      "loss": 0.5614,
+      "step": 242
+    },
+    {
+      "epoch": 0.43902439024390244,
+      "grad_norm": 0.26698136393959426,
+      "learning_rate": 1.9292493214061953e-05,
+      "loss": 0.566,
+      "step": 243
+    },
+    {
+      "epoch": 0.4408310749774164,
+      "grad_norm": 0.262835231624255,
+      "learning_rate": 1.9206410839590042e-05,
+      "loss": 0.5726,
+      "step": 244
+    },
+    {
+      "epoch": 0.44263775971093045,
+      "grad_norm": 0.2699198680731764,
+      "learning_rate": 1.9120177842035853e-05,
+      "loss": 0.5647,
+      "step": 245
+    },
+    {
+      "epoch": 0.4444444444444444,
+      "grad_norm": 0.25317122033414935,
+      "learning_rate": 1.9033797309228984e-05,
+      "loss": 0.5713,
+      "step": 246
+    },
+    {
+      "epoch": 0.44625112917795845,
+      "grad_norm": 0.26315048476885655,
+      "learning_rate": 1.8947272334281977e-05,
+      "loss": 0.5531,
+      "step": 247
+    },
+    {
+      "epoch": 0.4480578139114724,
+      "grad_norm": 0.2342696943629009,
+      "learning_rate": 1.8860606015479537e-05,
+      "loss": 0.5716,
+      "step": 248
+    },
+    {
+      "epoch": 0.44986449864498645,
+      "grad_norm": 0.2667760175408812,
+      "learning_rate": 1.877380145616763e-05,
+      "loss": 0.5553,
+      "step": 249
+    },
+    {
+      "epoch": 0.45167118337850043,
+      "grad_norm": 0.2646112181455277,
+      "learning_rate": 1.868686176464232e-05,
+      "loss": 0.5703,
+      "step": 250
+    },
+    {
+      "epoch": 0.45347786811201446,
+      "grad_norm": 0.24355396922430822,
+      "learning_rate": 1.8599790054038487e-05,
+      "loss": 0.5575,
+      "step": 251
+    },
+    {
+      "epoch": 0.45528455284552843,
+      "grad_norm": 0.2649879029310733,
+      "learning_rate": 1.8512589442218358e-05,
+      "loss": 0.5394,
+      "step": 252
+    },
+    {
+      "epoch": 0.45709123757904246,
+      "grad_norm": 0.25888050455028316,
+      "learning_rate": 1.8425263051659838e-05,
+      "loss": 0.5625,
+      "step": 253
+    },
+    {
+      "epoch": 0.45889792231255644,
+      "grad_norm": 0.3051654207330716,
+      "learning_rate": 1.8337814009344716e-05,
+      "loss": 0.5612,
+      "step": 254
+    },
+    {
+      "epoch": 0.46070460704607047,
+      "grad_norm": 0.29056971135960513,
+      "learning_rate": 1.8250245446646707e-05,
+      "loss": 0.5551,
+      "step": 255
+    },
+    {
+      "epoch": 0.46251129177958444,
+      "grad_norm": 0.2723843462217241,
+      "learning_rate": 1.8162560499219286e-05,
+      "loss": 0.5542,
+      "step": 256
+    },
+    {
+      "epoch": 0.4643179765130985,
+      "grad_norm": 0.2917329215307656,
+      "learning_rate": 1.807476230688346e-05,
+      "loss": 0.5645,
+      "step": 257
+    },
+    {
+      "epoch": 0.46612466124661245,
+      "grad_norm": 0.28092280024605654,
+      "learning_rate": 1.7986854013515274e-05,
+      "loss": 0.5603,
+      "step": 258
+    },
+    {
+      "epoch": 0.4679313459801265,
+      "grad_norm": 0.26756992910153876,
+      "learning_rate": 1.78988387669333e-05,
+      "loss": 0.5637,
+      "step": 259
+    },
+    {
+      "epoch": 0.46973803071364045,
+      "grad_norm": 0.30386996041687137,
+      "learning_rate": 1.781071971878587e-05,
+      "loss": 0.5772,
+      "step": 260
+    },
+    {
+      "epoch": 0.4715447154471545,
+      "grad_norm": 0.2591945941304491,
+      "learning_rate": 1.7722500024438244e-05,
+      "loss": 0.5682,
+      "step": 261
+    },
+    {
+      "epoch": 0.47335140018066846,
+      "grad_norm": 0.2847344325665358,
+      "learning_rate": 1.7634182842859628e-05,
+      "loss": 0.5565,
+      "step": 262
+    },
+    {
+      "epoch": 0.4751580849141825,
+      "grad_norm": 0.27145604581476696,
+      "learning_rate": 1.7545771336510033e-05,
+      "loss": 0.5614,
+      "step": 263
+    },
+    {
+      "epoch": 0.47696476964769646,
+      "grad_norm": 0.26988472049564605,
+      "learning_rate": 1.7457268671227067e-05,
+      "loss": 0.5671,
+      "step": 264
+    },
+    {
+      "epoch": 0.4787714543812105,
+      "grad_norm": 0.2814724976804236,
+      "learning_rate": 1.736867801611254e-05,
+      "loss": 0.5493,
+      "step": 265
+    },
+    {
+      "epoch": 0.48057813911472447,
+      "grad_norm": 0.2670683615450742,
+      "learning_rate": 1.728000254341901e-05,
+      "loss": 0.5729,
+      "step": 266
+    },
+    {
+      "epoch": 0.4823848238482385,
+      "grad_norm": 0.25597481791147625,
+      "learning_rate": 1.7191245428436175e-05,
+      "loss": 0.5577,
+      "step": 267
+    },
+    {
+      "epoch": 0.48419150858175247,
+      "grad_norm": 0.24797038870315238,
+      "learning_rate": 1.7102409849377188e-05,
+      "loss": 0.5573,
+      "step": 268
+    },
+    {
+      "epoch": 0.4859981933152665,
+      "grad_norm": 0.3010209021459523,
+      "learning_rate": 1.7013498987264832e-05,
+      "loss": 0.5643,
+      "step": 269
+    },
+    {
+      "epoch": 0.4878048780487805,
+      "grad_norm": 0.23652166032681626,
+      "learning_rate": 1.6924516025817636e-05,
+      "loss": 0.5463,
+      "step": 270
+    },
+    {
+      "epoch": 0.4896115627822945,
+      "grad_norm": 0.2559329783768578,
+      "learning_rate": 1.683546415133584e-05,
+      "loss": 0.5674,
+      "step": 271
+    },
+    {
+      "epoch": 0.4914182475158085,
+      "grad_norm": 0.2709870470141194,
+      "learning_rate": 1.6746346552587342e-05,
+      "loss": 0.5626,
+      "step": 272
+    },
+    {
+      "epoch": 0.4932249322493225,
+      "grad_norm": 0.2528039966576236,
+      "learning_rate": 1.665716642069349e-05,
+      "loss": 0.5699,
+      "step": 273
+    },
+    {
+      "epoch": 0.4950316169828365,
+      "grad_norm": 0.2744611375336476,
+      "learning_rate": 1.6567926949014805e-05,
+      "loss": 0.5579,
+      "step": 274
+    },
+    {
+      "epoch": 0.4968383017163505,
+      "grad_norm": 0.2404834130103736,
+      "learning_rate": 1.6478631333036655e-05,
+      "loss": 0.5521,
+      "step": 275
+    },
+    {
+      "epoch": 0.4986449864498645,
+      "grad_norm": 0.24903679080344832,
+      "learning_rate": 1.638928277025482e-05,
+      "loss": 0.5439,
+      "step": 276
+    },
+    {
+      "epoch": 0.5004516711833785,
+      "grad_norm": 0.25060914067945456,
+      "learning_rate": 1.6299884460061005e-05,
+      "loss": 0.5558,
+      "step": 277
+    },
+    {
+      "epoch": 0.5022583559168925,
+      "grad_norm": 0.24166390605958826,
+      "learning_rate": 1.621043960362826e-05,
+      "loss": 0.5431,
+      "step": 278
+    },
+    {
+      "epoch": 0.5040650406504065,
+      "grad_norm": 0.25206084683101343,
+      "learning_rate": 1.6120951403796367e-05,
+      "loss": 0.5577,
+      "step": 279
+    },
+    {
+      "epoch": 0.5058717253839206,
+      "grad_norm": 0.2478974752282698,
+      "learning_rate": 1.603142306495714e-05,
+      "loss": 0.5539,
+      "step": 280
+    },
+    {
+      "epoch": 0.5076784101174345,
+      "grad_norm": 0.22643026345349437,
+      "learning_rate": 1.5941857792939702e-05,
+      "loss": 0.5548,
+      "step": 281
+    },
+    {
+      "epoch": 0.5094850948509485,
+      "grad_norm": 0.23862222490609036,
+      "learning_rate": 1.585225879489567e-05,
+      "loss": 0.5493,
+      "step": 282
+    },
+    {
+      "epoch": 0.5112917795844625,
+      "grad_norm": 0.24226834868489433,
+      "learning_rate": 1.5762629279184326e-05,
+      "loss": 0.5643,
+      "step": 283
+    },
+    {
+      "epoch": 0.5130984643179766,
+      "grad_norm": 0.22354027179310576,
+      "learning_rate": 1.5672972455257726e-05,
+      "loss": 0.5528,
+      "step": 284
+    },
+    {
+      "epoch": 0.5149051490514905,
+      "grad_norm": 0.23633207917558613,
+      "learning_rate": 1.5583291533545775e-05,
+      "loss": 0.5477,
+      "step": 285
+    },
+    {
+      "epoch": 0.5167118337850045,
+      "grad_norm": 0.25686758225980666,
+      "learning_rate": 1.549358972534128e-05,
+      "loss": 0.545,
+      "step": 286
+    },
+    {
+      "epoch": 0.5185185185185185,
+      "grad_norm": 0.23037131466509797,
+      "learning_rate": 1.5403870242684942e-05,
+      "loss": 0.5668,
+      "step": 287
+    },
+    {
+      "epoch": 0.5203252032520326,
+      "grad_norm": 0.24547397538099036,
+      "learning_rate": 1.5314136298250355e-05,
+      "loss": 0.5639,
+      "step": 288
+    },
+    {
+      "epoch": 0.5221318879855466,
+      "grad_norm": 0.23059235409573473,
+      "learning_rate": 1.5224391105228956e-05,
+      "loss": 0.5482,
+      "step": 289
+    },
+    {
+      "epoch": 0.5239385727190605,
+      "grad_norm": 0.2225942146950027,
+      "learning_rate": 1.5134637877214968e-05,
+      "loss": 0.5598,
+      "step": 290
+    },
+    {
+      "epoch": 0.5257452574525745,
+      "grad_norm": 0.2549489142286338,
+      "learning_rate": 1.5044879828090346e-05,
+      "loss": 0.5641,
+      "step": 291
+    },
+    {
+      "epoch": 0.5275519421860885,
+      "grad_norm": 0.23439530930013888,
+      "learning_rate": 1.4955120171909658e-05,
+      "loss": 0.5679,
+      "step": 292
+    },
+    {
+      "epoch": 0.5293586269196026,
+      "grad_norm": 0.2116950796453125,
+      "learning_rate": 1.4865362122785031e-05,
+      "loss": 0.561,
+      "step": 293
+    },
+    {
+      "epoch": 0.5311653116531165,
+      "grad_norm": 0.2671877529948715,
+      "learning_rate": 1.4775608894771048e-05,
+      "loss": 0.5529,
+      "step": 294
+    },
+    {
+      "epoch": 0.5329719963866305,
+      "grad_norm": 0.25069170413138997,
+      "learning_rate": 1.4685863701749648e-05,
+      "loss": 0.5507,
+      "step": 295
+    },
+    {
+      "epoch": 0.5347786811201445,
+      "grad_norm": 0.2647161016797911,
+      "learning_rate": 1.4596129757315062e-05,
+      "loss": 0.5547,
+      "step": 296
+    },
+    {
+      "epoch": 0.5365853658536586,
+      "grad_norm": 0.2404787664917223,
+      "learning_rate": 1.4506410274658718e-05,
+      "loss": 0.5596,
+      "step": 297
+    },
+    {
+      "epoch": 0.5383920505871725,
+      "grad_norm": 0.2370171785729201,
+      "learning_rate": 1.441670846645423e-05,
+      "loss": 0.5547,
+      "step": 298
+    },
+    {
+      "epoch": 0.5401987353206865,
+      "grad_norm": 0.24169300307232816,
+      "learning_rate": 1.4327027544742281e-05,
+      "loss": 0.5699,
+      "step": 299
+    },
+    {
+      "epoch": 0.5420054200542005,
+      "grad_norm": 0.22653197687002416,
+      "learning_rate": 1.4237370720815675e-05,
+      "loss": 0.5575,
+      "step": 300
+    },
+    {
+      "epoch": 0.5438121047877146,
+      "grad_norm": 0.21798372306013064,
+      "learning_rate": 1.4147741205104336e-05,
+      "loss": 0.5599,
+      "step": 301
+    },
+    {
+      "epoch": 0.5456187895212286,
+      "grad_norm": 0.21125535901643153,
+      "learning_rate": 1.40581422070603e-05,
+      "loss": 0.5712,
+      "step": 302
+    },
+    {
+      "epoch": 0.5474254742547425,
+      "grad_norm": 0.25304414304531125,
+      "learning_rate": 1.3968576935042864e-05,
+      "loss": 0.5528,
+      "step": 303
+    },
+    {
+      "epoch": 0.5492321589882565,
+      "grad_norm": 0.2291646315525619,
+      "learning_rate": 1.3879048596203637e-05,
+      "loss": 0.548,
+      "step": 304
+    },
+    {
+      "epoch": 0.5510388437217706,
+      "grad_norm": 0.234299277071628,
+      "learning_rate": 1.3789560396371742e-05,
+      "loss": 0.5544,
+      "step": 305
+    },
+    {
+      "epoch": 0.5528455284552846,
+      "grad_norm": 0.23156086692778918,
+      "learning_rate": 1.3700115539938995e-05,
+      "loss": 0.5522,
+      "step": 306
+    },
+    {
+      "epoch": 0.5546522131887985,
+      "grad_norm": 0.24098434342394212,
+      "learning_rate": 1.3610717229745183e-05,
+      "loss": 0.5691,
+      "step": 307
+    },
+    {
+      "epoch": 0.5564588979223125,
+      "grad_norm": 0.22029204252195114,
+      "learning_rate": 1.3521368666963348e-05,
+      "loss": 0.5485,
+      "step": 308
+    },
+    {
+      "epoch": 0.5582655826558266,
+      "grad_norm": 0.2252564996199875,
+      "learning_rate": 1.3432073050985201e-05,
+      "loss": 0.558,
+      "step": 309
+    },
+    {
+      "epoch": 0.5600722673893406,
+      "grad_norm": 0.54269816685778,
+      "learning_rate": 1.3342833579306511e-05,
+      "loss": 0.5405,
+      "step": 310
+    },
+    {
+      "epoch": 0.5618789521228545,
+      "grad_norm": 0.22524996151253962,
+      "learning_rate": 1.3253653447412658e-05,
+      "loss": 0.5596,
+      "step": 311
+    },
+    {
+      "epoch": 0.5636856368563685,
+      "grad_norm": 0.22490802419520534,
+      "learning_rate": 1.3164535848664167e-05,
+      "loss": 0.5587,
+      "step": 312
+    },
+    {
+      "epoch": 0.5654923215898826,
+      "grad_norm": 0.22235081843769053,
+      "learning_rate": 1.3075483974182368e-05,
+      "loss": 0.5597,
+      "step": 313
+    },
+    {
+      "epoch": 0.5672990063233966,
+      "grad_norm": 0.2104992086470039,
+      "learning_rate": 1.2986501012735174e-05,
+      "loss": 0.5525,
+      "step": 314
+    },
+    {
+      "epoch": 0.5691056910569106,
+      "grad_norm": 0.2078910179139799,
+      "learning_rate": 1.2897590150622812e-05,
+      "loss": 0.5527,
+      "step": 315
+    },
+    {
+      "epoch": 0.5709123757904245,
+      "grad_norm": 0.2272482276824017,
+      "learning_rate": 1.2808754571563827e-05,
+      "loss": 0.5483,
+      "step": 316
+    },
+    {
+      "epoch": 0.5727190605239386,
+      "grad_norm": 0.22061016014341045,
+      "learning_rate": 1.2719997456580993e-05,
+      "loss": 0.5444,
+      "step": 317
+    },
+    {
+      "epoch": 0.5745257452574526,
+      "grad_norm": 0.22297379952628166,
+      "learning_rate": 1.2631321983887464e-05,
+      "loss": 0.5571,
+      "step": 318
+    },
+    {
+      "epoch": 0.5763324299909666,
+      "grad_norm": 0.2511013826033874,
+      "learning_rate": 1.2542731328772936e-05,
+      "loss": 0.5663,
+      "step": 319
+    },
+    {
+      "epoch": 0.5781391147244805,
+      "grad_norm": 0.23825224343188256,
+      "learning_rate": 1.2454228663489967e-05,
+      "loss": 0.5484,
+      "step": 320
+    },
+    {
+      "epoch": 0.5799457994579946,
+      "grad_norm": 0.23105290779278828,
+      "learning_rate": 1.2365817157140373e-05,
+      "loss": 0.5458,
+      "step": 321
+    },
+    {
+      "epoch": 0.5817524841915086,
+      "grad_norm": 0.23085267620316915,
+      "learning_rate": 1.2277499975561755e-05,
+      "loss": 0.5668,
+      "step": 322
+    },
+    {
+      "epoch": 0.5835591689250226,
+      "grad_norm": 0.23185477482530575,
+      "learning_rate": 1.2189280281214128e-05,
+      "loss": 0.553,
+      "step": 323
+    },
+    {
+      "epoch": 0.5853658536585366,
+      "grad_norm": 0.22547107943816214,
+      "learning_rate": 1.2101161233066704e-05,
+      "loss": 0.5335,
+      "step": 324
+    },
+    {
+      "epoch": 0.5871725383920506,
+      "grad_norm": 0.2225322417388763,
+      "learning_rate": 1.201314598648473e-05,
+      "loss": 0.5512,
+      "step": 325
+    },
+    {
+      "epoch": 0.5889792231255646,
+      "grad_norm": 0.24571430304257608,
+      "learning_rate": 1.1925237693116546e-05,
+      "loss": 0.5606,
+      "step": 326
+    },
+    {
+      "epoch": 0.5907859078590786,
+      "grad_norm": 0.23385259271752673,
+      "learning_rate": 1.1837439500780718e-05,
+      "loss": 0.547,
+      "step": 327
+    },
+    {
+      "epoch": 0.5925925925925926,
+      "grad_norm": 0.24708072043988052,
+      "learning_rate": 1.1749754553353297e-05,
+      "loss": 0.5535,
+      "step": 328
+    },
+    {
+      "epoch": 0.5943992773261066,
+      "grad_norm": 0.26008103232321017,
+      "learning_rate": 1.1662185990655285e-05,
+      "loss": 0.5436,
+      "step": 329
+    },
+    {
+      "epoch": 0.5962059620596206,
+      "grad_norm": 0.22767889432823338,
+      "learning_rate": 1.1574736948340163e-05,
+      "loss": 0.5512,
+      "step": 330
+    },
+    {
+      "epoch": 0.5980126467931346,
+      "grad_norm": 0.3321594311081388,
+      "learning_rate": 1.1487410557781642e-05,
+      "loss": 0.5549,
+      "step": 331
+    },
+    {
+      "epoch": 0.5998193315266486,
+      "grad_norm": 0.26329291651497444,
+      "learning_rate": 1.140020994596151e-05,
+      "loss": 0.5382,
+      "step": 332
+    },
+    {
+      "epoch": 0.6016260162601627,
+      "grad_norm": 0.20737071717618447,
+      "learning_rate": 1.1313138235357683e-05,
+      "loss": 0.5592,
+      "step": 333
+    },
+    {
+      "epoch": 0.6034327009936766,
+      "grad_norm": 0.2344184829076728,
+      "learning_rate": 1.1226198543832373e-05,
+      "loss": 0.5712,
+      "step": 334
+    },
+    {
+      "epoch": 0.6052393857271906,
+      "grad_norm": 0.2654227286303454,
+      "learning_rate": 1.1139393984520467e-05,
+      "loss": 0.5547,
+      "step": 335
+    },
+    {
+      "epoch": 0.6070460704607046,
+      "grad_norm": 0.21797005899046726,
+      "learning_rate": 1.1052727665718027e-05,
+      "loss": 0.5514,
+      "step": 336
+    },
+    {
+      "epoch": 0.6088527551942186,
+      "grad_norm": 0.21944637468999736,
+      "learning_rate": 1.0966202690771015e-05,
+      "loss": 0.5525,
+      "step": 337
+    },
+    {
+      "epoch": 0.6106594399277326,
+      "grad_norm": 0.2494773325418584,
+      "learning_rate": 1.087982215796415e-05,
+      "loss": 0.5486,
+      "step": 338
+    },
+    {
+      "epoch": 0.6124661246612466,
+      "grad_norm": 0.23296628301776853,
+      "learning_rate": 1.079358916040996e-05,
+      "loss": 0.5574,
+      "step": 339
+    },
+    {
+      "epoch": 0.6142728093947606,
+      "grad_norm": 0.2281765830312914,
+      "learning_rate": 1.0707506785938052e-05,
+      "loss": 0.5522,
+      "step": 340
+    },
+    {
+      "epoch": 0.6160794941282746,
+      "grad_norm": 0.19775912996938846,
+      "learning_rate": 1.0621578116984501e-05,
+      "loss": 0.5513,
+      "step": 341
+    },
+    {
+      "epoch": 0.6178861788617886,
+      "grad_norm": 0.22005369077907766,
+      "learning_rate": 1.0535806230481525e-05,
+      "loss": 0.5539,
+      "step": 342
+    },
+    {
+      "epoch": 0.6196928635953026,
+      "grad_norm": 0.21312133655782012,
+      "learning_rate": 1.0450194197747252e-05,
+      "loss": 0.5485,
+      "step": 343
+    },
+    {
+      "epoch": 0.6214995483288166,
+      "grad_norm": 0.2009774823522313,
+      "learning_rate": 1.036474508437579e-05,
+      "loss": 0.5379,
+      "step": 344
+    },
+    {
+      "epoch": 0.6233062330623306,
+      "grad_norm": 0.24236962291409364,
+      "learning_rate": 1.0279461950127422e-05,
+      "loss": 0.5516,
+      "step": 345
+    },
+    {
+      "epoch": 0.6251129177958447,
+      "grad_norm": 0.23348877792022393,
+      "learning_rate": 1.0194347848819055e-05,
+      "loss": 0.5537,
+      "step": 346
+    },
+    {
+      "epoch": 0.6269196025293586,
+      "grad_norm": 0.20738282623870158,
+      "learning_rate": 1.0109405828214872e-05,
+      "loss": 0.5395,
+      "step": 347
+    },
+    {
+      "epoch": 0.6287262872628726,
+      "grad_norm": 0.2111195446000444,
+      "learning_rate": 1.0024638929917192e-05,
+      "loss": 0.5405,
+      "step": 348
+    },
+    {
+      "epoch": 0.6305329719963866,
+      "grad_norm": 0.20840492602401303,
+      "learning_rate": 9.940050189257552e-06,
+      "loss": 0.5441,
+      "step": 349
+    },
+    {
+      "epoch": 0.6323396567299007,
+      "grad_norm": 0.19875429668478228,
+      "learning_rate": 9.85564263518803e-06,
+      "loss": 0.5558,
+      "step": 350
+    },
+    {
+      "epoch": 0.6341463414634146,
+      "grad_norm": 0.2134423124227348,
+      "learning_rate": 9.771419290172776e-06,
+      "loss": 0.5454,
+      "step": 351
+    },
+    {
+      "epoch": 0.6359530261969286,
+      "grad_norm": 0.2117582878529125,
+      "learning_rate": 9.68738317007978e-06,
+      "loss": 0.5594,
+      "step": 352
+    },
+    {
+      "epoch": 0.6377597109304426,
+      "grad_norm": 0.20952140719342904,
+      "learning_rate": 9.6035372840729e-06,
+      "loss": 0.5603,
+      "step": 353
+    },
+    {
+      "epoch": 0.6395663956639567,
+      "grad_norm": 0.2112668403044356,
+      "learning_rate": 9.519884634504074e-06,
+      "loss": 0.555,
+      "step": 354
+    },
+    {
+      "epoch": 0.6413730803974707,
+      "grad_norm": 0.2396477976508343,
+      "learning_rate": 9.436428216805862e-06,
+      "loss": 0.5434,
+      "step": 355
+    },
+    {
+      "epoch": 0.6431797651309846,
+      "grad_norm": 0.19281125097595814,
+      "learning_rate": 9.35317101938412e-06,
+      "loss": 0.5573,
+      "step": 356
+    },
+    {
+      "epoch": 0.6449864498644986,
+      "grad_norm": 0.23133791674565504,
+      "learning_rate": 9.270116023511063e-06,
+      "loss": 0.559,
+      "step": 357
+    },
+    {
+      "epoch": 0.6467931345980127,
+      "grad_norm": 0.2314539721640054,
+      "learning_rate": 9.187266203218457e-06,
+      "loss": 0.5645,
+      "step": 358
+    },
+    {
+      "epoch": 0.6485998193315267,
+      "grad_norm": 0.20554990129737805,
+      "learning_rate": 9.104624525191147e-06,
+      "loss": 0.5376,
+      "step": 359
+    },
+    {
+      "epoch": 0.6504065040650406,
+      "grad_norm": 0.18800898611182573,
+      "learning_rate": 9.022193948660825e-06,
+      "loss": 0.5463,
+      "step": 360
+    },
+    {
+      "epoch": 0.6522131887985546,
+      "grad_norm": 0.22018788975804213,
+      "learning_rate": 8.939977425300072e-06,
+      "loss": 0.5515,
+      "step": 361
+    },
+    {
+      "epoch": 0.6540198735320687,
+      "grad_norm": 0.21350063459455076,
+      "learning_rate": 8.857977899116645e-06,
+      "loss": 0.5627,
+      "step": 362
+    },
+    {
+      "epoch": 0.6558265582655827,
+      "grad_norm": 0.23529884866342432,
+      "learning_rate": 8.776198306348072e-06,
+      "loss": 0.5532,
+      "step": 363
+    },
+    {
+      "epoch": 0.6576332429990966,
+      "grad_norm": 0.21785824771559895,
+      "learning_rate": 8.69464157535652e-06,
+      "loss": 0.5529,
+      "step": 364
+    },
+    {
+      "epoch": 0.6594399277326106,
+      "grad_norm": 0.19604204196426292,
+      "learning_rate": 8.61331062652391e-06,
+      "loss": 0.5444,
+      "step": 365
+    },
+    {
+      "epoch": 0.6612466124661247,
+      "grad_norm": 0.21901286813907167,
+      "learning_rate": 8.532208372147376e-06,
+      "loss": 0.5658,
+      "step": 366
+    },
+    {
+      "epoch": 0.6630532971996387,
+      "grad_norm": 0.21070560052739,
+      "learning_rate": 8.451337716334953e-06,
+      "loss": 0.5471,
+      "step": 367
+    },
+    {
+      "epoch": 0.6648599819331527,
+      "grad_norm": 0.20758358513307357,
+      "learning_rate": 8.370701554901603e-06,
+      "loss": 0.5587,
+      "step": 368
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 0.21507352661412657,
+      "learning_rate": 8.290302775265509e-06,
+      "loss": 0.5594,
+      "step": 369
+    },
+    {
+      "epoch": 0.6684733514001807,
+      "grad_norm": 0.2043695834839199,
+      "learning_rate": 8.210144256344714e-06,
+      "loss": 0.5584,
+      "step": 370
+    },
+    {
+      "epoch": 0.6702800361336947,
+      "grad_norm": 0.22276926331354296,
+      "learning_rate": 8.130228868453989e-06,
+      "loss": 0.546,
+      "step": 371
+    },
+    {
+      "epoch": 0.6720867208672087,
+      "grad_norm": 0.20718485157930183,
+      "learning_rate": 8.050559473202078e-06,
+      "loss": 0.5635,
+      "step": 372
+    },
+    {
+      "epoch": 0.6738934056007226,
+      "grad_norm": 0.21821823323387163,
+      "learning_rate": 7.971138923389221e-06,
+      "loss": 0.5494,
+      "step": 373
+    },
+    {
+      "epoch": 0.6757000903342367,
+      "grad_norm": 0.20198041756355264,
+      "learning_rate": 7.89197006290502e-06,
+      "loss": 0.5391,
+      "step": 374
+    },
+    {
+      "epoch": 0.6775067750677507,
+      "grad_norm": 0.662160537969569,
+      "learning_rate": 7.813055726626579e-06,
+      "loss": 0.5529,
+      "step": 375
+    },
+    {
+      "epoch": 0.6793134598012647,
+      "grad_norm": 0.20082813122266513,
+      "learning_rate": 7.734398740316999e-06,
+      "loss": 0.5381,
+      "step": 376
+    },
+    {
+      "epoch": 0.6811201445347786,
+      "grad_norm": 0.20640835193534096,
+      "learning_rate": 7.65600192052422e-06,
+      "loss": 0.554,
+      "step": 377
+    },
+    {
+      "epoch": 0.6829268292682927,
+      "grad_norm": 0.2081855645945218,
+      "learning_rate": 7.577868074480107e-06,
+      "loss": 0.5551,
+      "step": 378
+    },
+    {
+      "epoch": 0.6847335140018067,
+      "grad_norm": 0.22143850372886462,
+      "learning_rate": 7.500000000000004e-06,
+      "loss": 0.558,
+      "step": 379
+    },
+    {
+      "epoch": 0.6865401987353207,
+      "grad_norm": 0.19601847546745818,
+      "learning_rate": 7.422400485382484e-06,
+      "loss": 0.5393,
+      "step": 380
+    },
+    {
+      "epoch": 0.6883468834688347,
+      "grad_norm": 0.21231904921309738,
+      "learning_rate": 7.345072309309557e-06,
+      "loss": 0.5554,
+      "step": 381
+    },
+    {
+      "epoch": 0.6901535682023487,
+      "grad_norm": 0.21547150171318516,
+      "learning_rate": 7.268018240747129e-06,
+      "loss": 0.5426,
+      "step": 382
+    },
+    {
+      "epoch": 0.6919602529358627,
+      "grad_norm": 0.20632440068252555,
+      "learning_rate": 7.191241038845876e-06,
+      "loss": 0.5617,
+      "step": 383
+    },
+    {
+      "epoch": 0.6937669376693767,
+      "grad_norm": 0.20396330319640896,
+      "learning_rate": 7.114743452842428e-06,
+      "loss": 0.5587,
+      "step": 384
+    },
+    {
+      "epoch": 0.6955736224028907,
+      "grad_norm": 0.1954447016465696,
+      "learning_rate": 7.038528221960959e-06,
+      "loss": 0.5418,
+      "step": 385
+    },
+    {
+      "epoch": 0.6973803071364046,
+      "grad_norm": 0.2027643331648993,
+      "learning_rate": 6.962598075315047e-06,
+      "loss": 0.5531,
+      "step": 386
+    },
+    {
+      "epoch": 0.6991869918699187,
+      "grad_norm": 0.26130739064414765,
+      "learning_rate": 6.886955731810009e-06,
+      "loss": 0.5535,
+      "step": 387
+    },
+    {
+      "epoch": 0.7009936766034327,
+      "grad_norm": 0.21585210183249268,
+      "learning_rate": 6.811603900045477e-06,
+      "loss": 0.5527,
+      "step": 388
+    },
+    {
+      "epoch": 0.7028003613369467,
+      "grad_norm": 0.190777040998849,
+      "learning_rate": 6.736545278218464e-06,
+      "loss": 0.5551,
+      "step": 389
+    },
+    {
+      "epoch": 0.7046070460704607,
+      "grad_norm": 0.19836100360350714,
+      "learning_rate": 6.661782554026735e-06,
+      "loss": 0.5497,
+      "step": 390
+    },
+    {
+      "epoch": 0.7064137308039747,
+      "grad_norm": 0.21643464865120854,
+      "learning_rate": 6.587318404572537e-06,
+      "loss": 0.549,
+      "step": 391
+    },
+    {
+      "epoch": 0.7082204155374887,
+      "grad_norm": 0.1993977853648827,
+      "learning_rate": 6.513155496266783e-06,
+      "loss": 0.5556,
+      "step": 392
+    },
+    {
+      "epoch": 0.7100271002710027,
+      "grad_norm": 0.21028221055918736,
+      "learning_rate": 6.439296484733526e-06,
+      "loss": 0.5601,
+      "step": 393
+    },
+    {
+      "epoch": 0.7118337850045167,
+      "grad_norm": 0.3375716495970427,
+      "learning_rate": 6.3657440147149e-06,
+      "loss": 0.5629,
+      "step": 394
+    },
+    {
+      "epoch": 0.7136404697380307,
+      "grad_norm": 0.23687864959416866,
+      "learning_rate": 6.292500719976397e-06,
+      "loss": 0.5572,
+      "step": 395
+    },
+    {
+      "epoch": 0.7154471544715447,
+      "grad_norm": 0.18881298296034638,
+      "learning_rate": 6.219569223212585e-06,
+      "loss": 0.5511,
+      "step": 396
+    },
+    {
+      "epoch": 0.7172538392050587,
+      "grad_norm": 0.20470029831189293,
+      "learning_rate": 6.146952135953158e-06,
+      "loss": 0.5512,
+      "step": 397
+    },
+    {
+      "epoch": 0.7190605239385727,
+      "grad_norm": 0.19576998582854852,
+      "learning_rate": 6.074652058469447e-06,
+      "loss": 0.5524,
+      "step": 398
+    },
+    {
+      "epoch": 0.7208672086720868,
+      "grad_norm": 0.20611373253291534,
+      "learning_rate": 6.0026715796812945e-06,
+      "loss": 0.5553,
+      "step": 399
+    },
+    {
+      "epoch": 0.7226738934056007,
+      "grad_norm": 0.23471404781984317,
+      "learning_rate": 5.931013277064377e-06,
+      "loss": 0.5524,
+      "step": 400
+    },
+    {
+      "epoch": 0.7244805781391147,
+      "grad_norm": 0.19110919938350365,
+      "learning_rate": 5.859679716557879e-06,
+      "loss": 0.5543,
+      "step": 401
+    },
+    {
+      "epoch": 0.7262872628726287,
+      "grad_norm": 0.211937906141866,
+      "learning_rate": 5.788673452472622e-06,
+      "loss": 0.561,
+      "step": 402
+    },
+    {
+      "epoch": 0.7280939476061428,
+      "grad_norm": 0.20802338936823334,
+      "learning_rate": 5.717997027399624e-06,
+      "loss": 0.5426,
+      "step": 403
+    },
+    {
+      "epoch": 0.7299006323396567,
+      "grad_norm": 0.20723368523348565,
+      "learning_rate": 5.647652972118998e-06,
+      "loss": 0.5484,
+      "step": 404
+    },
+    {
+      "epoch": 0.7317073170731707,
+      "grad_norm": 0.20267313877420606,
+      "learning_rate": 5.577643805509404e-06,
+      "loss": 0.5464,
+      "step": 405
+    },
+    {
+      "epoch": 0.7335140018066847,
+      "grad_norm": 0.21939184955076324,
+      "learning_rate": 5.507972034457786e-06,
+      "loss": 0.5522,
+      "step": 406
+    },
+    {
+      "epoch": 0.7353206865401988,
+      "grad_norm": 0.18909160425846863,
+      "learning_rate": 5.438640153769654e-06,
+      "loss": 0.5446,
+      "step": 407
+    },
+    {
+      "epoch": 0.7371273712737128,
+      "grad_norm": 0.21881744668955247,
+      "learning_rate": 5.369650646079719e-06,
+      "loss": 0.5381,
+      "step": 408
+    },
+    {
+      "epoch": 0.7389340560072267,
+      "grad_norm": 0.20756435822062427,
+      "learning_rate": 5.301005981763007e-06,
+      "loss": 0.5579,
+      "step": 409
+    },
+    {
+      "epoch": 0.7407407407407407,
+      "grad_norm": 0.2078186876803581,
+      "learning_rate": 5.2327086188463964e-06,
+      "loss": 0.5581,
+      "step": 410
+    },
+    {
+      "epoch": 0.7425474254742548,
+      "grad_norm": 0.2084081536752294,
+      "learning_rate": 5.164761002920615e-06,
+      "loss": 0.5607,
+      "step": 411
+    },
+    {
+      "epoch": 0.7443541102077688,
+      "grad_norm": 0.19586330960019913,
+      "learning_rate": 5.097165567052646e-06,
+      "loss": 0.5336,
+      "step": 412
+    },
+    {
+      "epoch": 0.7461607949412827,
+      "grad_norm": 0.20608558252347248,
+      "learning_rate": 5.029924731698619e-06,
+      "loss": 0.5499,
+      "step": 413
+    },
+    {
+      "epoch": 0.7479674796747967,
+      "grad_norm": 0.1987787234405567,
+      "learning_rate": 4.963040904617131e-06,
+      "loss": 0.5384,
+      "step": 414
+    },
+    {
+      "epoch": 0.7497741644083108,
+      "grad_norm": 0.19067064134400263,
+      "learning_rate": 4.8965164807830355e-06,
+      "loss": 0.5588,
+      "step": 415
+    },
+    {
+      "epoch": 0.7515808491418248,
+      "grad_norm": 0.1881350869130335,
+      "learning_rate": 4.83035384230169e-06,
+      "loss": 0.5457,
+      "step": 416
+    },
+    {
+      "epoch": 0.7533875338753387,
+      "grad_norm": 0.21158969911561606,
+      "learning_rate": 4.764555358323632e-06,
+      "loss": 0.5667,
+      "step": 417
+    },
+    {
+      "epoch": 0.7551942186088527,
+      "grad_norm": 0.2452094550267619,
+      "learning_rate": 4.699123384959783e-06,
+      "loss": 0.5484,
+      "step": 418
+    },
+    {
+      "epoch": 0.7570009033423668,
+      "grad_norm": 0.1753065680730648,
+      "learning_rate": 4.6340602651970304e-06,
+      "loss": 0.55,
+      "step": 419
+    },
+    {
+      "epoch": 0.7588075880758808,
+      "grad_norm": 0.2038591820649658,
+      "learning_rate": 4.5693683288143884e-06,
+      "loss": 0.5639,
+      "step": 420
+    },
+    {
+      "epoch": 0.7606142728093948,
+      "grad_norm": 0.1884902431578462,
+      "learning_rate": 4.505049892299517e-06,
+      "loss": 0.5459,
+      "step": 421
+    },
+    {
+      "epoch": 0.7624209575429087,
+      "grad_norm": 0.18108880162202162,
+      "learning_rate": 4.441107258765827e-06,
+      "loss": 0.5493,
+      "step": 422
+    },
+    {
+      "epoch": 0.7642276422764228,
+      "grad_norm": 0.20213211715192836,
+      "learning_rate": 4.377542717869959e-06,
+      "loss": 0.558,
+      "step": 423
+    },
+    {
+      "epoch": 0.7660343270099368,
+      "grad_norm": 0.20199802307009435,
+      "learning_rate": 4.31435854572983e-06,
+      "loss": 0.5614,
+      "step": 424
+    },
+    {
+      "epoch": 0.7678410117434508,
+      "grad_norm": 0.1907226474034031,
+      "learning_rate": 4.251557004843116e-06,
+      "loss": 0.5402,
+      "step": 425
+    },
+    {
+      "epoch": 0.7696476964769647,
+      "grad_norm": 0.18804170518073093,
+      "learning_rate": 4.1891403440062495e-06,
+      "loss": 0.5603,
+      "step": 426
+    },
+    {
+      "epoch": 0.7714543812104788,
+      "grad_norm": 0.1803906532902956,
+      "learning_rate": 4.1271107982338745e-06,
+      "loss": 0.5323,
+      "step": 427
+    },
+    {
+      "epoch": 0.7732610659439928,
+      "grad_norm": 0.26949831229585547,
+      "learning_rate": 4.06547058867883e-06,
+      "loss": 0.5526,
+      "step": 428
+    },
+    {
+      "epoch": 0.7750677506775068,
+      "grad_norm": 0.1947858647850873,
+      "learning_rate": 4.004221922552608e-06,
+      "loss": 0.5476,
+      "step": 429
+    },
+    {
+      "epoch": 0.7768744354110207,
+      "grad_norm": 0.176842627630576,
+      "learning_rate": 3.943366993046323e-06,
+      "loss": 0.5488,
+      "step": 430
+    },
+    {
+      "epoch": 0.7786811201445348,
+      "grad_norm": 0.293863357388934,
+      "learning_rate": 3.88290797925218e-06,
+      "loss": 0.5575,
+      "step": 431
+    },
+    {
+      "epoch": 0.7804878048780488,
+      "grad_norm": 0.17901888035257343,
+      "learning_rate": 3.82284704608543e-06,
+      "loss": 0.5529,
+      "step": 432
+    },
+    {
+      "epoch": 0.7822944896115628,
+      "grad_norm": 0.19332621955868448,
+      "learning_rate": 3.763186344206877e-06,
+      "loss": 0.5607,
+      "step": 433
+    },
+    {
+      "epoch": 0.7841011743450768,
+      "grad_norm": 0.18496082102672784,
+      "learning_rate": 3.7039280099458373e-06,
+      "loss": 0.5588,
+      "step": 434
+    },
+    {
+      "epoch": 0.7859078590785907,
+      "grad_norm": 0.18993724697897982,
+      "learning_rate": 3.645074165223656e-06,
+      "loss": 0.5375,
+      "step": 435
+    },
+    {
+      "epoch": 0.7877145438121048,
+      "grad_norm": 0.19268146724531712,
+      "learning_rate": 3.586626917477724e-06,
+      "loss": 0.551,
+      "step": 436
+    },
+    {
+      "epoch": 0.7895212285456188,
+      "grad_norm": 0.18103708715973252,
+      "learning_rate": 3.528588359586024e-06,
+      "loss": 0.5499,
+      "step": 437
+    },
+    {
+      "epoch": 0.7913279132791328,
+      "grad_norm": 0.20985265964985653,
+      "learning_rate": 3.4709605697921735e-06,
+      "loss": 0.5416,
+      "step": 438
+    },
+    {
+      "epoch": 0.7931345980126467,
+      "grad_norm": 0.21561474422323904,
+      "learning_rate": 3.413745611631009e-06,
+      "loss": 0.5583,
+      "step": 439
+    },
+    {
+      "epoch": 0.7949412827461608,
+      "grad_norm": 0.17490103602265775,
+      "learning_rate": 3.356945533854708e-06,
+      "loss": 0.5496,
+      "step": 440
+    },
+    {
+      "epoch": 0.7967479674796748,
+      "grad_norm": 0.16466411873700984,
+      "learning_rate": 3.3005623703594082e-06,
+      "loss": 0.5354,
+      "step": 441
+    },
+    {
+      "epoch": 0.7985546522131888,
+      "grad_norm": 0.18135794783104248,
+      "learning_rate": 3.244598140112404e-06,
+      "loss": 0.5441,
+      "step": 442
+    },
+    {
+      "epoch": 0.8003613369467028,
+      "grad_norm": 0.20748633625312018,
+      "learning_rate": 3.1890548470798196e-06,
+      "loss": 0.552,
+      "step": 443
+    },
+    {
+      "epoch": 0.8021680216802168,
+      "grad_norm": 0.1884179425899726,
+      "learning_rate": 3.1339344801548854e-06,
+      "loss": 0.5573,
+      "step": 444
+    },
+    {
+      "epoch": 0.8039747064137308,
+      "grad_norm": 0.19495474378106878,
+      "learning_rate": 3.079239013086673e-06,
+      "loss": 0.5454,
+      "step": 445
+    },
+    {
+      "epoch": 0.8057813911472448,
+      "grad_norm": 0.18268990190252204,
+      "learning_rate": 3.02497040440948e-06,
+      "loss": 0.5434,
+      "step": 446
+    },
+    {
+      "epoch": 0.8075880758807588,
+      "grad_norm": 0.20059027999185472,
+      "learning_rate": 2.9711305973726404e-06,
+      "loss": 0.5479,
+      "step": 447
+    },
+    {
+      "epoch": 0.8093947606142728,
+      "grad_norm": 0.22292779858686448,
+      "learning_rate": 2.91772151987099e-06,
+      "loss": 0.5508,
+      "step": 448
+    },
+    {
+      "epoch": 0.8112014453477868,
+      "grad_norm": 0.18941532454561305,
+      "learning_rate": 2.86474508437579e-06,
+      "loss": 0.545,
+      "step": 449
+    },
+    {
+      "epoch": 0.8130081300813008,
+      "grad_norm": 0.17978750905710117,
+      "learning_rate": 2.812203187866272e-06,
+      "loss": 0.556,
+      "step": 450
+    },
+    {
+      "epoch": 0.8148148148148148,
+      "grad_norm": 0.21165503368414987,
+      "learning_rate": 2.7600977117617027e-06,
+      "loss": 0.5567,
+      "step": 451
+    },
+    {
+      "epoch": 0.8166214995483289,
+      "grad_norm": 0.25166342627264704,
+      "learning_rate": 2.7084305218540217e-06,
+      "loss": 0.5542,
+      "step": 452
+    },
+    {
+      "epoch": 0.8184281842818428,
+      "grad_norm": 0.17912582965998358,
+      "learning_rate": 2.6572034682410144e-06,
+      "loss": 0.5456,
+      "step": 453
+    },
+    {
+      "epoch": 0.8202348690153568,
+      "grad_norm": 0.21041229136704262,
+      "learning_rate": 2.60641838526008e-06,
+      "loss": 0.5303,
+      "step": 454
+    },
+    {
+      "epoch": 0.8220415537488708,
+      "grad_norm": 0.17116821534893534,
+      "learning_rate": 2.5560770914225356e-06,
+      "loss": 0.5533,
+      "step": 455
+    },
+    {
+      "epoch": 0.8238482384823849,
+      "grad_norm": 0.1830634128552793,
+      "learning_rate": 2.5061813893485087e-06,
+      "loss": 0.5463,
+      "step": 456
+    },
+    {
+      "epoch": 0.8256549232158988,
+      "grad_norm": 0.18808524644151434,
+      "learning_rate": 2.45673306570239e-06,
+      "loss": 0.5599,
+      "step": 457
+    },
+    {
+      "epoch": 0.8274616079494128,
+      "grad_norm": 0.17427150355266183,
+      "learning_rate": 2.407733891128841e-06,
+      "loss": 0.5519,
+      "step": 458
+    },
+    {
+      "epoch": 0.8292682926829268,
+      "grad_norm": 0.18728720554040562,
+      "learning_rate": 2.3591856201894124e-06,
+      "loss": 0.5422,
+      "step": 459
+    },
+    {
+      "epoch": 0.8310749774164409,
+      "grad_norm": 0.18544489209732046,
+      "learning_rate": 2.3110899912996993e-06,
+      "loss": 0.551,
+      "step": 460
+    },
+    {
+      "epoch": 0.8328816621499548,
+      "grad_norm": 0.17977343031514503,
+      "learning_rate": 2.263448726667098e-06,
+      "loss": 0.5393,
+      "step": 461
+    },
+    {
+      "epoch": 0.8346883468834688,
+      "grad_norm": 0.17921097737693858,
+      "learning_rate": 2.2162635322291365e-06,
+      "loss": 0.5478,
+      "step": 462
+    },
+    {
+      "epoch": 0.8364950316169828,
+      "grad_norm": 0.20070476223578554,
+      "learning_rate": 2.169536097592401e-06,
+      "loss": 0.5512,
+      "step": 463
+    },
+    {
+      "epoch": 0.8383017163504969,
+      "grad_norm": 0.27692731072187426,
+      "learning_rate": 2.1232680959720085e-06,
+      "loss": 0.5538,
+      "step": 464
+    },
+    {
+      "epoch": 0.8401084010840109,
+      "grad_norm": 0.19399701632530203,
+      "learning_rate": 2.0774611841317164e-06,
+      "loss": 0.5635,
+      "step": 465
+    },
+    {
+      "epoch": 0.8419150858175248,
+      "grad_norm": 0.17874665419769412,
+      "learning_rate": 2.0321170023245835e-06,
+      "loss": 0.5466,
+      "step": 466
+    },
+    {
+      "epoch": 0.8437217705510388,
+      "grad_norm": 0.19389405602159773,
+      "learning_rate": 1.9872371742342366e-06,
+      "loss": 0.5436,
+      "step": 467
+    },
+    {
+      "epoch": 0.8455284552845529,
+      "grad_norm": 0.5663378302294984,
+      "learning_rate": 1.942823306916746e-06,
+      "loss": 0.5539,
+      "step": 468
+    },
+    {
+      "epoch": 0.8473351400180669,
+      "grad_norm": 0.19569469058843655,
+      "learning_rate": 1.8988769907430553e-06,
+      "loss": 0.5433,
+      "step": 469
+    },
+    {
+      "epoch": 0.8491418247515808,
+      "grad_norm": 0.236538997514183,
+      "learning_rate": 1.8553997993420495e-06,
+      "loss": 0.5395,
+      "step": 470
+    },
+    {
+      "epoch": 0.8509485094850948,
+      "grad_norm": 0.2566860235526064,
+      "learning_rate": 1.8123932895441953e-06,
+      "loss": 0.5688,
+      "step": 471
+    },
+    {
+      "epoch": 0.8527551942186089,
+      "grad_norm": 0.19382358330061378,
+      "learning_rate": 1.7698590013258187e-06,
+      "loss": 0.5397,
+      "step": 472
+    },
+    {
+      "epoch": 0.8545618789521229,
+      "grad_norm": 0.1712718137071858,
+      "learning_rate": 1.7277984577539224e-06,
+      "loss": 0.5586,
+      "step": 473
+    },
+    {
+      "epoch": 0.8563685636856369,
+      "grad_norm": 0.16962484799926925,
+      "learning_rate": 1.6862131649316885e-06,
+      "loss": 0.5543,
+      "step": 474
+    },
+    {
+      "epoch": 0.8581752484191508,
+      "grad_norm": 0.4519901995615342,
+      "learning_rate": 1.6451046119445174e-06,
+      "loss": 0.5443,
+      "step": 475
+    },
+    {
+      "epoch": 0.8599819331526649,
+      "grad_norm": 0.18306941536442994,
+      "learning_rate": 1.6044742708067206e-06,
+      "loss": 0.5502,
+      "step": 476
+    },
+    {
+      "epoch": 0.8617886178861789,
+      "grad_norm": 0.18409536525414097,
+      "learning_rate": 1.5643235964088065e-06,
+      "loss": 0.548,
+      "step": 477
+    },
+    {
+      "epoch": 0.8635953026196929,
+      "grad_norm": 0.196534805012363,
+      "learning_rate": 1.524654026465394e-06,
+      "loss": 0.5579,
+      "step": 478
+    },
+    {
+      "epoch": 0.8654019873532068,
+      "grad_norm": 0.17811685192063484,
+      "learning_rate": 1.4854669814637145e-06,
+      "loss": 0.5438,
+      "step": 479
+    },
+    {
+      "epoch": 0.8672086720867209,
+      "grad_norm": 0.17346684568559728,
+      "learning_rate": 1.4467638646127551e-06,
+      "loss": 0.5424,
+      "step": 480
+    },
+    {
+      "epoch": 0.8690153568202349,
+      "grad_norm": 0.16852517344675289,
+      "learning_rate": 1.4085460617930196e-06,
+      "loss": 0.5533,
+      "step": 481
+    },
+    {
+      "epoch": 0.8708220415537489,
+      "grad_norm": 0.1802396650429446,
+      "learning_rate": 1.3708149415068827e-06,
+      "loss": 0.5575,
+      "step": 482
+    },
+    {
+      "epoch": 0.8726287262872628,
+      "grad_norm": 0.20760583978000396,
+      "learning_rate": 1.3335718548296194e-06,
+      "loss": 0.5425,
+      "step": 483
+    },
+    {
+      "epoch": 0.8744354110207768,
+      "grad_norm": 0.17775906373949274,
+      "learning_rate": 1.2968181353609854e-06,
+      "loss": 0.5386,
+      "step": 484
+    },
+    {
+      "epoch": 0.8762420957542909,
+      "grad_norm": 0.2086584073138206,
+      "learning_rate": 1.2605550991775033e-06,
+      "loss": 0.5486,
+      "step": 485
+    },
+    {
+      "epoch": 0.8780487804878049,
+      "grad_norm": 0.18285260902918074,
+      "learning_rate": 1.2247840447852993e-06,
+      "loss": 0.5548,
+      "step": 486
+    },
+    {
+      "epoch": 0.8798554652213189,
+      "grad_norm": 0.1752253781052565,
+      "learning_rate": 1.1895062530736412e-06,
+      "loss": 0.5423,
+      "step": 487
+    },
+    {
+      "epoch": 0.8816621499548328,
+      "grad_norm": 0.17033890975977392,
+      "learning_rate": 1.1547229872690407e-06,
+      "loss": 0.5417,
+      "step": 488
+    },
+    {
+      "epoch": 0.8834688346883469,
+      "grad_norm": 0.16626635367640144,
+      "learning_rate": 1.1204354928900496e-06,
+      "loss": 0.5508,
+      "step": 489
+    },
+    {
+      "epoch": 0.8852755194218609,
+      "grad_norm": 0.17403069224138948,
+      "learning_rate": 1.086644997702635e-06,
+      "loss": 0.5484,
+      "step": 490
+    },
+    {
+      "epoch": 0.8870822041553749,
+      "grad_norm": 0.1696438985371448,
+      "learning_rate": 1.0533527116762298e-06,
+      "loss": 0.5393,
+      "step": 491
+    },
+    {
+      "epoch": 0.8888888888888888,
+      "grad_norm": 0.1787883314589599,
+      "learning_rate": 1.0205598269404026e-06,
+      "loss": 0.5412,
+      "step": 492
+    },
+    {
+      "epoch": 0.8906955736224029,
+      "grad_norm": 0.17339051853647025,
+      "learning_rate": 9.882675177421662e-07,
+      "loss": 0.5356,
+      "step": 493
+    },
+    {
+      "epoch": 0.8925022583559169,
+      "grad_norm": 0.1702531815864387,
+      "learning_rate": 9.56476940403942e-07,
+      "loss": 0.5524,
+      "step": 494
+    },
+    {
+      "epoch": 0.8943089430894309,
+      "grad_norm": 0.17509542948768647,
+      "learning_rate": 9.251892332821388e-07,
+      "loss": 0.5492,
+      "step": 495
+    },
+    {
+      "epoch": 0.8961156278229448,
+      "grad_norm": 0.17446109688291325,
+      "learning_rate": 8.944055167263982e-07,
+      "loss": 0.5614,
+      "step": 496
+    },
+    {
+      "epoch": 0.8979223125564589,
+      "grad_norm": 0.19555682227139368,
+      "learning_rate": 8.641268930394791e-07,
+      "loss": 0.549,
+      "step": 497
+    },
+    {
+      "epoch": 0.8997289972899729,
+      "grad_norm": 0.18714322043208526,
+      "learning_rate": 8.34354446437785e-07,
+      "loss": 0.5482,
+      "step": 498
+    },
+    {
+      "epoch": 0.9015356820234869,
+      "grad_norm": 0.20117371163389913,
+      "learning_rate": 8.050892430125362e-07,
+      "loss": 0.5572,
+      "step": 499
+    },
+    {
+      "epoch": 0.9033423667570009,
+      "grad_norm": 0.17690610631546144,
+      "learning_rate": 7.763323306916037e-07,
+      "loss": 0.5558,
+      "step": 500
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 553,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 579500543967232.0,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:98adbbb8c05c24cdf16e5fa8134fc92b30728f6a7326bb4cbcbfc91e21242dd0
+size 8145