Upload continuous evaluation checkpoints, batch 2

.gitattributes

@@ -51,3 +51,7 @@ continuous/qwen25/eupe_vit_b/finetune/tokenizer.json filter=lfs diff=lfs merge=l
 continuous/qwen25/eupe_vit_s/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 continuous/qwen25/eupe_vit_t/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 continuous/qwen25/ijepa_vith14/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+continuous/qwen25/mc1_b16_224_2.5b/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+continuous/qwen25/mc1_b16_224_400m/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+continuous/qwen25/mc1_b32_224_2.5b/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+continuous/qwen25/mc1_b32_224_400m/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text

continuous/qwen25/mc1_b16_224_2.5b/finetune/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07089a2232c9358332c65141bfd2654124e8a9f47539f5af2ad4e5dac9bcfab4
+size 3740790376

continuous/qwen25/mc1_b16_224_2.5b/finetune/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
+  }
+}

continuous/qwen25/mc1_b16_224_2.5b/finetune/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

continuous/qwen25/mc1_b16_224_2.5b/finetune/tokenizer_config.json

@@ -0,0 +1,209 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

continuous/qwen25/mc1_b16_224_2.5b/finetune/trainer_state.json

@@ -0,0 +1,1856 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9999037443449803,
+  "eval_steps": 500,
+  "global_step": 2597,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.003850226200789296,
+      "grad_norm": 15.094815254211426,
+      "learning_rate": 1.153846153846154e-06,
+      "loss": 1.8103,
+      "step": 10
+    },
+    {
+      "epoch": 0.007700452401578592,
+      "grad_norm": 6.954693794250488,
+      "learning_rate": 2.435897435897436e-06,
+      "loss": 1.5361,
+      "step": 20
+    },
+    {
+      "epoch": 0.01155067860236789,
+      "grad_norm": 3.3569276332855225,
+      "learning_rate": 3.7179487179487184e-06,
+      "loss": 1.1469,
+      "step": 30
+    },
+    {
+      "epoch": 0.015400904803157185,
+      "grad_norm": 0.7188485264778137,
+      "learning_rate": 5e-06,
+      "loss": 0.9389,
+      "step": 40
+    },
+    {
+      "epoch": 0.019251131003946483,
+      "grad_norm": 0.3255023956298828,
+      "learning_rate": 6.282051282051282e-06,
+      "loss": 0.9078,
+      "step": 50
+    },
+    {
+      "epoch": 0.02310135720473578,
+      "grad_norm": 0.28441500663757324,
+      "learning_rate": 7.564102564102564e-06,
+      "loss": 0.8618,
+      "step": 60
+    },
+    {
+      "epoch": 0.026951583405525073,
+      "grad_norm": 0.23083283007144928,
+      "learning_rate": 8.846153846153847e-06,
+      "loss": 0.8335,
+      "step": 70
+    },
+    {
+      "epoch": 0.03080180960631437,
+      "grad_norm": 0.2355363368988037,
+      "learning_rate": 9.999996111488723e-06,
+      "loss": 0.842,
+      "step": 80
+    },
+    {
+      "epoch": 0.034652035807103666,
+      "grad_norm": 0.21339720487594604,
+      "learning_rate": 9.999529497453782e-06,
+      "loss": 0.8237,
+      "step": 90
+    },
+    {
+      "epoch": 0.038502262007892966,
+      "grad_norm": 0.23083683848381042,
+      "learning_rate": 9.998285264324193e-06,
+      "loss": 0.8025,
+      "step": 100
+    },
+    {
+      "epoch": 0.04235248820868226,
+      "grad_norm": 0.20239479839801788,
+      "learning_rate": 9.996263605626052e-06,
+      "loss": 0.7933,
+      "step": 110
+    },
+    {
+      "epoch": 0.04620271440947156,
+      "grad_norm": 0.22485317289829254,
+      "learning_rate": 9.99346483580503e-06,
+      "loss": 0.805,
+      "step": 120
+    },
+    {
+      "epoch": 0.05005294061026085,
+      "grad_norm": 0.20840834081172943,
+      "learning_rate": 9.989889390177467e-06,
+      "loss": 0.7939,
+      "step": 130
+    },
+    {
+      "epoch": 0.053903166811050146,
+      "grad_norm": 0.23420707881450653,
+      "learning_rate": 9.985537824862647e-06,
+      "loss": 0.7968,
+      "step": 140
+    },
+    {
+      "epoch": 0.057753393011839446,
+      "grad_norm": 0.23247835040092468,
+      "learning_rate": 9.980410816696317e-06,
+      "loss": 0.7774,
+      "step": 150
+    },
+    {
+      "epoch": 0.06160361921262874,
+      "grad_norm": 0.2298012524843216,
+      "learning_rate": 9.97450916312541e-06,
+      "loss": 0.7878,
+      "step": 160
+    },
+    {
+      "epoch": 0.06545384541341803,
+      "grad_norm": 0.2079259306192398,
+      "learning_rate": 9.967833782084004e-06,
+      "loss": 0.7875,
+      "step": 170
+    },
+    {
+      "epoch": 0.06930407161420733,
+      "grad_norm": 0.22917263209819794,
+      "learning_rate": 9.960385711850554e-06,
+      "loss": 0.8132,
+      "step": 180
+    },
+    {
+      "epoch": 0.07315429781499663,
+      "grad_norm": 0.22548286616802216,
+      "learning_rate": 9.952166110886394e-06,
+      "loss": 0.8059,
+      "step": 190
+    },
+    {
+      "epoch": 0.07700452401578593,
+      "grad_norm": 0.25126126408576965,
+      "learning_rate": 9.943176257655567e-06,
+      "loss": 0.7969,
+      "step": 200
+    },
+    {
+      "epoch": 0.08085475021657522,
+      "grad_norm": 0.20955568552017212,
+      "learning_rate": 9.933417550425952e-06,
+      "loss": 0.7906,
+      "step": 210
+    },
+    {
+      "epoch": 0.08470497641736452,
+      "grad_norm": 0.21773022413253784,
+      "learning_rate": 9.9228915070518e-06,
+      "loss": 0.7773,
+      "step": 220
+    },
+    {
+      "epoch": 0.08855520261815382,
+      "grad_norm": 0.2176537811756134,
+      "learning_rate": 9.911599764737632e-06,
+      "loss": 0.78,
+      "step": 230
+    },
+    {
+      "epoch": 0.09240542881894312,
+      "grad_norm": 0.22284938395023346,
+      "learning_rate": 9.899544079783606e-06,
+      "loss": 0.7728,
+      "step": 240
+    },
+    {
+      "epoch": 0.0962556550197324,
+      "grad_norm": 0.2231794148683548,
+      "learning_rate": 9.886726327312338e-06,
+      "loss": 0.7905,
+      "step": 250
+    },
+    {
+      "epoch": 0.1001058812205217,
+      "grad_norm": 0.21960806846618652,
+      "learning_rate": 9.87314850097724e-06,
+      "loss": 0.7871,
+      "step": 260
+    },
+    {
+      "epoch": 0.103956107421311,
+      "grad_norm": 0.20863327383995056,
+      "learning_rate": 9.858812712652452e-06,
+      "loss": 0.7773,
+      "step": 270
+    },
+    {
+      "epoch": 0.10780633362210029,
+      "grad_norm": 0.20630761981010437,
+      "learning_rate": 9.843721192104341e-06,
+      "loss": 0.7747,
+      "step": 280
+    },
+    {
+      "epoch": 0.11165655982288959,
+      "grad_norm": 0.2116522639989853,
+      "learning_rate": 9.827876286644698e-06,
+      "loss": 0.797,
+      "step": 290
+    },
+    {
+      "epoch": 0.11550678602367889,
+      "grad_norm": 0.1993521749973297,
+      "learning_rate": 9.811280460765643e-06,
+      "loss": 0.7585,
+      "step": 300
+    },
+    {
+      "epoch": 0.11935701222446819,
+      "grad_norm": 0.21258966624736786,
+      "learning_rate": 9.793936295756292e-06,
+      "loss": 0.7673,
+      "step": 310
+    },
+    {
+      "epoch": 0.12320723842525748,
+      "grad_norm": 0.21875421702861786,
+      "learning_rate": 9.775846489301279e-06,
+      "loss": 0.7724,
+      "step": 320
+    },
+    {
+      "epoch": 0.12705746462604678,
+      "grad_norm": 0.22189486026763916,
+      "learning_rate": 9.757013855061157e-06,
+      "loss": 0.7697,
+      "step": 330
+    },
+    {
+      "epoch": 0.13090769082683607,
+      "grad_norm": 0.211864173412323,
+      "learning_rate": 9.737441322234758e-06,
+      "loss": 0.7833,
+      "step": 340
+    },
+    {
+      "epoch": 0.13475791702762538,
+      "grad_norm": 0.22496604919433594,
+      "learning_rate": 9.717131935103605e-06,
+      "loss": 0.7825,
+      "step": 350
+    },
+    {
+      "epoch": 0.13860814322841467,
+      "grad_norm": 0.21558573842048645,
+      "learning_rate": 9.696088852558398e-06,
+      "loss": 0.7692,
+      "step": 360
+    },
+    {
+      "epoch": 0.14245836942920398,
+      "grad_norm": 0.21125997602939606,
+      "learning_rate": 9.674315347607686e-06,
+      "loss": 0.7948,
+      "step": 370
+    },
+    {
+      "epoch": 0.14630859562999327,
+      "grad_norm": 0.20958420634269714,
+      "learning_rate": 9.651814806868796e-06,
+      "loss": 0.7605,
+      "step": 380
+    },
+    {
+      "epoch": 0.15015882183078255,
+      "grad_norm": 0.22225898504257202,
+      "learning_rate": 9.62859073004107e-06,
+      "loss": 0.7615,
+      "step": 390
+    },
+    {
+      "epoch": 0.15400904803157187,
+      "grad_norm": 0.19864921271800995,
+      "learning_rate": 9.604646729361547e-06,
+      "loss": 0.767,
+      "step": 400
+    },
+    {
+      "epoch": 0.15785927423236115,
+      "grad_norm": 0.20616960525512695,
+      "learning_rate": 9.579986529043099e-06,
+      "loss": 0.7569,
+      "step": 410
+    },
+    {
+      "epoch": 0.16170950043315044,
+      "grad_norm": 0.21535801887512207,
+      "learning_rate": 9.554613964695189e-06,
+      "loss": 0.7739,
+      "step": 420
+    },
+    {
+      "epoch": 0.16555972663393975,
+      "grad_norm": 0.2227061241865158,
+      "learning_rate": 9.528532982727281e-06,
+      "loss": 0.7663,
+      "step": 430
+    },
+    {
+      "epoch": 0.16940995283472904,
+      "grad_norm": 0.2031794935464859,
+      "learning_rate": 9.501747639735018e-06,
+      "loss": 0.7635,
+      "step": 440
+    },
+    {
+      "epoch": 0.17326017903551832,
+      "grad_norm": 0.21928833425045013,
+      "learning_rate": 9.474262101869272e-06,
+      "loss": 0.7528,
+      "step": 450
+    },
+    {
+      "epoch": 0.17711040523630764,
+      "grad_norm": 0.21432039141654968,
+      "learning_rate": 9.446080644188141e-06,
+      "loss": 0.7587,
+      "step": 460
+    },
+    {
+      "epoch": 0.18096063143709692,
+      "grad_norm": 0.20651133358478546,
+      "learning_rate": 9.417207649992017e-06,
+      "loss": 0.7641,
+      "step": 470
+    },
+    {
+      "epoch": 0.18481085763788624,
+      "grad_norm": 0.21479551494121552,
+      "learning_rate": 9.387647610141816e-06,
+      "loss": 0.7624,
+      "step": 480
+    },
+    {
+      "epoch": 0.18866108383867553,
+      "grad_norm": 0.19629351794719696,
+      "learning_rate": 9.35740512236047e-06,
+      "loss": 0.7499,
+      "step": 490
+    },
+    {
+      "epoch": 0.1925113100394648,
+      "grad_norm": 0.19593378901481628,
+      "learning_rate": 9.326484890517809e-06,
+      "loss": 0.7623,
+      "step": 500
+    },
+    {
+      "epoch": 0.19636153624025413,
+      "grad_norm": 0.22302022576332092,
+      "learning_rate": 9.294891723898926e-06,
+      "loss": 0.7506,
+      "step": 510
+    },
+    {
+      "epoch": 0.2002117624410434,
+      "grad_norm": 0.2225719690322876,
+      "learning_rate": 9.262630536456153e-06,
+      "loss": 0.7815,
+      "step": 520
+    },
+    {
+      "epoch": 0.2040619886418327,
+      "grad_norm": 0.2246273308992386,
+      "learning_rate": 9.229706346044749e-06,
+      "loss": 0.7554,
+      "step": 530
+    },
+    {
+      "epoch": 0.207912214842622,
+      "grad_norm": 0.22572432458400726,
+      "learning_rate": 9.196124273642422e-06,
+      "loss": 0.7663,
+      "step": 540
+    },
+    {
+      "epoch": 0.2117624410434113,
+      "grad_norm": 0.2097289264202118,
+      "learning_rate": 9.161889542552842e-06,
+      "loss": 0.7716,
+      "step": 550
+    },
+    {
+      "epoch": 0.21561266724420058,
+      "grad_norm": 0.2050977498292923,
+      "learning_rate": 9.127007477593193e-06,
+      "loss": 0.773,
+      "step": 560
+    },
+    {
+      "epoch": 0.2194628934449899,
+      "grad_norm": 0.23147499561309814,
+      "learning_rate": 9.091483504265974e-06,
+      "loss": 0.7816,
+      "step": 570
+    },
+    {
+      "epoch": 0.22331311964577918,
+      "grad_norm": 0.22202196717262268,
+      "learning_rate": 9.055323147915116e-06,
+      "loss": 0.7611,
+      "step": 580
+    },
+    {
+      "epoch": 0.2271633458465685,
+      "grad_norm": 0.21739406883716583,
+      "learning_rate": 9.018532032866585e-06,
+      "loss": 0.7719,
+      "step": 590
+    },
+    {
+      "epoch": 0.23101357204735778,
+      "grad_norm": 0.2100285440683365,
+      "learning_rate": 8.981115881553579e-06,
+      "loss": 0.7632,
+      "step": 600
+    },
+    {
+      "epoch": 0.23486379824814707,
+      "grad_norm": 0.23295892775058746,
+      "learning_rate": 8.943080513626476e-06,
+      "loss": 0.7834,
+      "step": 610
+    },
+    {
+      "epoch": 0.23871402444893638,
+      "grad_norm": 0.21928182244300842,
+      "learning_rate": 8.904431845047641e-06,
+      "loss": 0.7686,
+      "step": 620
+    },
+    {
+      "epoch": 0.24256425064972567,
+      "grad_norm": 0.22410830855369568,
+      "learning_rate": 8.86517588717128e-06,
+      "loss": 0.7846,
+      "step": 630
+    },
+    {
+      "epoch": 0.24641447685051496,
+      "grad_norm": 0.20469366014003754,
+      "learning_rate": 8.82531874580844e-06,
+      "loss": 0.7767,
+      "step": 640
+    },
+    {
+      "epoch": 0.25026470305130427,
+      "grad_norm": 0.19215263426303864,
+      "learning_rate": 8.784866620277314e-06,
+      "loss": 0.7338,
+      "step": 650
+    },
+    {
+      "epoch": 0.25411492925209356,
+      "grad_norm": 0.22740741074085236,
+      "learning_rate": 8.743825802439018e-06,
+      "loss": 0.7655,
+      "step": 660
+    },
+    {
+      "epoch": 0.25796515545288284,
+      "grad_norm": 0.2212403416633606,
+      "learning_rate": 8.702202675718956e-06,
+      "loss": 0.775,
+      "step": 670
+    },
+    {
+      "epoch": 0.26181538165367213,
+      "grad_norm": 0.20718887448310852,
+      "learning_rate": 8.660003714113959e-06,
+      "loss": 0.7425,
+      "step": 680
+    },
+    {
+      "epoch": 0.26566560785446147,
+      "grad_norm": 0.21075871586799622,
+      "learning_rate": 8.617235481185323e-06,
+      "loss": 0.7642,
+      "step": 690
+    },
+    {
+      "epoch": 0.26951583405525076,
+      "grad_norm": 0.21888838708400726,
+      "learning_rate": 8.573904629037931e-06,
+      "loss": 0.7708,
+      "step": 700
+    },
+    {
+      "epoch": 0.27336606025604004,
+      "grad_norm": 0.21313002705574036,
+      "learning_rate": 8.530017897285585e-06,
+      "loss": 0.754,
+      "step": 710
+    },
+    {
+      "epoch": 0.27721628645682933,
+      "grad_norm": 0.197549507021904,
+      "learning_rate": 8.485582112002746e-06,
+      "loss": 0.759,
+      "step": 720
+    },
+    {
+      "epoch": 0.2810665126576186,
+      "grad_norm": 0.2119397670030594,
+      "learning_rate": 8.440604184662811e-06,
+      "loss": 0.7607,
+      "step": 730
+    },
+    {
+      "epoch": 0.28491673885840796,
+      "grad_norm": 0.22340957820415497,
+      "learning_rate": 8.395091111063118e-06,
+      "loss": 0.7512,
+      "step": 740
+    },
+    {
+      "epoch": 0.28876696505919724,
+      "grad_norm": 0.19611196219921112,
+      "learning_rate": 8.349049970236822e-06,
+      "loss": 0.7512,
+      "step": 750
+    },
+    {
+      "epoch": 0.29261719125998653,
+      "grad_norm": 0.21055284142494202,
+      "learning_rate": 8.302487923351842e-06,
+      "loss": 0.7435,
+      "step": 760
+    },
+    {
+      "epoch": 0.2964674174607758,
+      "grad_norm": 0.20873229205608368,
+      "learning_rate": 8.255412212597021e-06,
+      "loss": 0.7608,
+      "step": 770
+    },
+    {
+      "epoch": 0.3003176436615651,
+      "grad_norm": 0.20395834743976593,
+      "learning_rate": 8.207830160055688e-06,
+      "loss": 0.7583,
+      "step": 780
+    },
+    {
+      "epoch": 0.3041678698623544,
+      "grad_norm": 0.23574115335941315,
+      "learning_rate": 8.159749166566786e-06,
+      "loss": 0.7445,
+      "step": 790
+    },
+    {
+      "epoch": 0.30801809606314373,
+      "grad_norm": 0.20756986737251282,
+      "learning_rate": 8.11117671057377e-06,
+      "loss": 0.7707,
+      "step": 800
+    },
+    {
+      "epoch": 0.311868322263933,
+      "grad_norm": 0.22355900704860687,
+      "learning_rate": 8.062120346961404e-06,
+      "loss": 0.7628,
+      "step": 810
+    },
+    {
+      "epoch": 0.3157185484647223,
+      "grad_norm": 0.19918036460876465,
+      "learning_rate": 8.012587705880705e-06,
+      "loss": 0.7574,
+      "step": 820
+    },
+    {
+      "epoch": 0.3195687746655116,
+      "grad_norm": 0.21308332681655884,
+      "learning_rate": 7.962586491562146e-06,
+      "loss": 0.7421,
+      "step": 830
+    },
+    {
+      "epoch": 0.3234190008663009,
+      "grad_norm": 0.22615158557891846,
+      "learning_rate": 7.912124481117357e-06,
+      "loss": 0.748,
+      "step": 840
+    },
+    {
+      "epoch": 0.3272692270670902,
+      "grad_norm": 0.20519758760929108,
+      "learning_rate": 7.861209523329494e-06,
+      "loss": 0.7605,
+      "step": 850
+    },
+    {
+      "epoch": 0.3311194532678795,
+      "grad_norm": 0.20955844223499298,
+      "learning_rate": 7.809849537432432e-06,
+      "loss": 0.7434,
+      "step": 860
+    },
+    {
+      "epoch": 0.3349696794686688,
+      "grad_norm": 0.20734281837940216,
+      "learning_rate": 7.758052511879035e-06,
+      "loss": 0.7545,
+      "step": 870
+    },
+    {
+      "epoch": 0.3388199056694581,
+      "grad_norm": 0.20382453501224518,
+      "learning_rate": 7.705826503098642e-06,
+      "loss": 0.7491,
+      "step": 880
+    },
+    {
+      "epoch": 0.34267013187024736,
+      "grad_norm": 0.20242156088352203,
+      "learning_rate": 7.653179634243979e-06,
+      "loss": 0.7488,
+      "step": 890
+    },
+    {
+      "epoch": 0.34652035807103665,
+      "grad_norm": 0.19065499305725098,
+      "learning_rate": 7.600120093927696e-06,
+      "loss": 0.7512,
+      "step": 900
+    },
+    {
+      "epoch": 0.350370584271826,
+      "grad_norm": 0.21800824999809265,
+      "learning_rate": 7.546656134948729e-06,
+      "loss": 0.7376,
+      "step": 910
+    },
+    {
+      "epoch": 0.3542208104726153,
+      "grad_norm": 0.2140025645494461,
+      "learning_rate": 7.492796073008661e-06,
+      "loss": 0.741,
+      "step": 920
+    },
+    {
+      "epoch": 0.35807103667340456,
+      "grad_norm": 0.22619013488292694,
+      "learning_rate": 7.438548285418326e-06,
+      "loss": 0.7299,
+      "step": 930
+    },
+    {
+      "epoch": 0.36192126287419385,
+      "grad_norm": 0.23279012739658356,
+      "learning_rate": 7.3839212097948e-06,
+      "loss": 0.7626,
+      "step": 940
+    },
+    {
+      "epoch": 0.36577148907498314,
+      "grad_norm": 0.21339182555675507,
+      "learning_rate": 7.328923342749039e-06,
+      "loss": 0.7765,
+      "step": 950
+    },
+    {
+      "epoch": 0.3696217152757725,
+      "grad_norm": 0.21681205928325653,
+      "learning_rate": 7.273563238564319e-06,
+      "loss": 0.7633,
+      "step": 960
+    },
+    {
+      "epoch": 0.37347194147656176,
+      "grad_norm": 0.21161167323589325,
+      "learning_rate": 7.217849507865724e-06,
+      "loss": 0.7418,
+      "step": 970
+    },
+    {
+      "epoch": 0.37732216767735105,
+      "grad_norm": 0.21201585233211517,
+      "learning_rate": 7.161790816280858e-06,
+      "loss": 0.7489,
+      "step": 980
+    },
+    {
+      "epoch": 0.38117239387814034,
+      "grad_norm": 0.22004571557044983,
+      "learning_rate": 7.1053958830920125e-06,
+      "loss": 0.7296,
+      "step": 990
+    },
+    {
+      "epoch": 0.3850226200789296,
+      "grad_norm": 0.20815624296665192,
+      "learning_rate": 7.0486734798799715e-06,
+      "loss": 0.7486,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3888728462797189,
+      "grad_norm": 0.2039252668619156,
+      "learning_rate": 6.991632429159708e-06,
+      "loss": 0.7723,
+      "step": 1010
+    },
+    {
+      "epoch": 0.39272307248050825,
+      "grad_norm": 0.2252337485551834,
+      "learning_rate": 6.934281603008132e-06,
+      "loss": 0.7639,
+      "step": 1020
+    },
+    {
+      "epoch": 0.39657329868129754,
+      "grad_norm": 0.2083587795495987,
+      "learning_rate": 6.876629921684143e-06,
+      "loss": 0.7616,
+      "step": 1030
+    },
+    {
+      "epoch": 0.4004235248820868,
+      "grad_norm": 0.21655192971229553,
+      "learning_rate": 6.818686352241199e-06,
+      "loss": 0.7512,
+      "step": 1040
+    },
+    {
+      "epoch": 0.4042737510828761,
+      "grad_norm": 0.19438841938972473,
+      "learning_rate": 6.760459907132573e-06,
+      "loss": 0.7486,
+      "step": 1050
+    },
+    {
+      "epoch": 0.4081239772836654,
+      "grad_norm": 0.1961216777563095,
+      "learning_rate": 6.701959642809589e-06,
+      "loss": 0.7538,
+      "step": 1060
+    },
+    {
+      "epoch": 0.41197420348445474,
+      "grad_norm": 0.2104950100183487,
+      "learning_rate": 6.643194658312987e-06,
+      "loss": 0.7433,
+      "step": 1070
+    },
+    {
+      "epoch": 0.415824429685244,
+      "grad_norm": 0.2062290757894516,
+      "learning_rate": 6.584174093857676e-06,
+      "loss": 0.7684,
+      "step": 1080
+    },
+    {
+      "epoch": 0.4196746558860333,
+      "grad_norm": 0.2361970692873001,
+      "learning_rate": 6.5249071294110735e-06,
+      "loss": 0.7529,
+      "step": 1090
+    },
+    {
+      "epoch": 0.4235248820868226,
+      "grad_norm": 0.21632547676563263,
+      "learning_rate": 6.465402983265274e-06,
+      "loss": 0.7668,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4273751082876119,
+      "grad_norm": 0.21745988726615906,
+      "learning_rate": 6.405670910603243e-06,
+      "loss": 0.7624,
+      "step": 1110
+    },
+    {
+      "epoch": 0.43122533448840117,
+      "grad_norm": 0.22826671600341797,
+      "learning_rate": 6.345720202059291e-06,
+      "loss": 0.7584,
+      "step": 1120
+    },
+    {
+      "epoch": 0.4350755606891905,
+      "grad_norm": 0.21147939562797546,
+      "learning_rate": 6.285560182274008e-06,
+      "loss": 0.7353,
+      "step": 1130
+    },
+    {
+      "epoch": 0.4389257868899798,
+      "grad_norm": 0.20834621787071228,
+      "learning_rate": 6.225200208443936e-06,
+      "loss": 0.7575,
+      "step": 1140
+    },
+    {
+      "epoch": 0.4427760130907691,
+      "grad_norm": 0.20359459519386292,
+      "learning_rate": 6.16464966886616e-06,
+      "loss": 0.7476,
+      "step": 1150
+    },
+    {
+      "epoch": 0.44662623929155837,
+      "grad_norm": 0.2455148547887802,
+      "learning_rate": 6.103917981478065e-06,
+      "loss": 0.7726,
+      "step": 1160
+    },
+    {
+      "epoch": 0.45047646549234766,
+      "grad_norm": 0.21037842333316803,
+      "learning_rate": 6.043014592392489e-06,
+      "loss": 0.783,
+      "step": 1170
+    },
+    {
+      "epoch": 0.454326691693137,
+      "grad_norm": 0.1974606066942215,
+      "learning_rate": 5.981948974428485e-06,
+      "loss": 0.7564,
+      "step": 1180
+    },
+    {
+      "epoch": 0.4581769178939263,
+      "grad_norm": 0.21449510753154755,
+      "learning_rate": 5.920730625637934e-06,
+      "loss": 0.735,
+      "step": 1190
+    },
+    {
+      "epoch": 0.46202714409471557,
+      "grad_norm": 0.22000761330127716,
+      "learning_rate": 5.8593690678282355e-06,
+      "loss": 0.7487,
+      "step": 1200
+    },
+    {
+      "epoch": 0.46587737029550486,
+      "grad_norm": 0.22174695134162903,
+      "learning_rate": 5.797873845081301e-06,
+      "loss": 0.7514,
+      "step": 1210
+    },
+    {
+      "epoch": 0.46972759649629414,
+      "grad_norm": 0.2190488576889038,
+      "learning_rate": 5.736254522269075e-06,
+      "loss": 0.767,
+      "step": 1220
+    },
+    {
+      "epoch": 0.47357782269708343,
+      "grad_norm": 0.2074035108089447,
+      "learning_rate": 5.6745206835658386e-06,
+      "loss": 0.7716,
+      "step": 1230
+    },
+    {
+      "epoch": 0.47742804889787277,
+      "grad_norm": 0.20252758264541626,
+      "learning_rate": 5.612681930957495e-06,
+      "loss": 0.7401,
+      "step": 1240
+    },
+    {
+      "epoch": 0.48127827509866206,
+      "grad_norm": 0.19938316941261292,
+      "learning_rate": 5.550747882748095e-06,
+      "loss": 0.7385,
+      "step": 1250
+    },
+    {
+      "epoch": 0.48512850129945134,
+      "grad_norm": 0.2287740260362625,
+      "learning_rate": 5.4887281720638195e-06,
+      "loss": 0.749,
+      "step": 1260
+    },
+    {
+      "epoch": 0.48897872750024063,
+      "grad_norm": 0.23949089646339417,
+      "learning_rate": 5.42663244535466e-06,
+      "loss": 0.729,
+      "step": 1270
+    },
+    {
+      "epoch": 0.4928289537010299,
+      "grad_norm": 0.2122121900320053,
+      "learning_rate": 5.3644703608940205e-06,
+      "loss": 0.7398,
+      "step": 1280
+    },
+    {
+      "epoch": 0.49667917990181926,
+      "grad_norm": 0.21688058972358704,
+      "learning_rate": 5.302251587276491e-06,
+      "loss": 0.7493,
+      "step": 1290
+    },
+    {
+      "epoch": 0.5005294061026085,
+      "grad_norm": 0.20257171988487244,
+      "learning_rate": 5.2399858019140005e-06,
+      "loss": 0.7513,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5043796323033978,
+      "grad_norm": 0.20717757940292358,
+      "learning_rate": 5.177682689530615e-06,
+      "loss": 0.7639,
+      "step": 1310
+    },
+    {
+      "epoch": 0.5082298585041871,
+      "grad_norm": 0.21947580575942993,
+      "learning_rate": 5.11535194065619e-06,
+      "loss": 0.7343,
+      "step": 1320
+    },
+    {
+      "epoch": 0.5120800847049765,
+      "grad_norm": 0.22963497042655945,
+      "learning_rate": 5.053003250119113e-06,
+      "loss": 0.7426,
+      "step": 1330
+    },
+    {
+      "epoch": 0.5159303109057657,
+      "grad_norm": 0.21908937394618988,
+      "learning_rate": 4.9906463155383925e-06,
+      "loss": 0.7494,
+      "step": 1340
+    },
+    {
+      "epoch": 0.519780537106555,
+      "grad_norm": 0.20958714187145233,
+      "learning_rate": 4.9282908358153046e-06,
+      "loss": 0.7452,
+      "step": 1350
+    },
+    {
+      "epoch": 0.5236307633073443,
+      "grad_norm": 0.24024926126003265,
+      "learning_rate": 4.8659465096248345e-06,
+      "loss": 0.7598,
+      "step": 1360
+    },
+    {
+      "epoch": 0.5274809895081336,
+      "grad_norm": 0.2145129144191742,
+      "learning_rate": 4.803623033907169e-06,
+      "loss": 0.7563,
+      "step": 1370
+    },
+    {
+      "epoch": 0.5313312157089229,
+      "grad_norm": 0.20401081442832947,
+      "learning_rate": 4.741330102359441e-06,
+      "loss": 0.7647,
+      "step": 1380
+    },
+    {
+      "epoch": 0.5351814419097122,
+      "grad_norm": 0.22554157674312592,
+      "learning_rate": 4.679077403927996e-06,
+      "loss": 0.7436,
+      "step": 1390
+    },
+    {
+      "epoch": 0.5390316681105015,
+      "grad_norm": 0.20998014509677887,
+      "learning_rate": 4.616874621301373e-06,
+      "loss": 0.7448,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5428818943112907,
+      "grad_norm": 0.2212584763765335,
+      "learning_rate": 4.554731429404293e-06,
+      "loss": 0.7588,
+      "step": 1410
+    },
+    {
+      "epoch": 0.5467321205120801,
+      "grad_norm": 0.22362466156482697,
+      "learning_rate": 4.492657493892822e-06,
+      "loss": 0.7414,
+      "step": 1420
+    },
+    {
+      "epoch": 0.5505823467128694,
+      "grad_norm": 0.22519551217556,
+      "learning_rate": 4.430662469650999e-06,
+      "loss": 0.7589,
+      "step": 1430
+    },
+    {
+      "epoch": 0.5544325729136587,
+      "grad_norm": 0.22177091240882874,
+      "learning_rate": 4.368755999289118e-06,
+      "loss": 0.7627,
+      "step": 1440
+    },
+    {
+      "epoch": 0.558282799114448,
+      "grad_norm": 0.2202548086643219,
+      "learning_rate": 4.306947711643949e-06,
+      "loss": 0.743,
+      "step": 1450
+    },
+    {
+      "epoch": 0.5621330253152372,
+      "grad_norm": 0.22138161957263947,
+      "learning_rate": 4.245247220281068e-06,
+      "loss": 0.7499,
+      "step": 1460
+    },
+    {
+      "epoch": 0.5659832515160266,
+      "grad_norm": 0.2075446993112564,
+      "learning_rate": 4.183664121999587e-06,
+      "loss": 0.7445,
+      "step": 1470
+    },
+    {
+      "epoch": 0.5698334777168159,
+      "grad_norm": 0.23317070305347443,
+      "learning_rate": 4.122207995339474e-06,
+      "loss": 0.743,
+      "step": 1480
+    },
+    {
+      "epoch": 0.5736837039176051,
+      "grad_norm": 0.20343805849552155,
+      "learning_rate": 4.060888399091734e-06,
+      "loss": 0.7477,
+      "step": 1490
+    },
+    {
+      "epoch": 0.5775339301183945,
+      "grad_norm": 0.2049071043729782,
+      "learning_rate": 3.999714870811639e-06,
+      "loss": 0.7584,
+      "step": 1500
+    },
+    {
+      "epoch": 0.5813841563191837,
+      "grad_norm": 0.21202944219112396,
+      "learning_rate": 3.938696925335275e-06,
+      "loss": 0.7515,
+      "step": 1510
+    },
+    {
+      "epoch": 0.5852343825199731,
+      "grad_norm": 0.2058519572019577,
+      "learning_rate": 3.87784405329962e-06,
+      "loss": 0.76,
+      "step": 1520
+    },
+    {
+      "epoch": 0.5890846087207623,
+      "grad_norm": 0.21359215676784515,
+      "learning_rate": 3.817165719666382e-06,
+      "loss": 0.7594,
+      "step": 1530
+    },
+    {
+      "epoch": 0.5929348349215516,
+      "grad_norm": 0.20771893858909607,
+      "learning_rate": 3.756671362249841e-06,
+      "loss": 0.7662,
+      "step": 1540
+    },
+    {
+      "epoch": 0.596785061122341,
+      "grad_norm": 0.2100876122713089,
+      "learning_rate": 3.6963703902488907e-06,
+      "loss": 0.7518,
+      "step": 1550
+    },
+    {
+      "epoch": 0.6006352873231302,
+      "grad_norm": 0.21102425456047058,
+      "learning_rate": 3.6362721827835655e-06,
+      "loss": 0.7654,
+      "step": 1560
+    },
+    {
+      "epoch": 0.6044855135239195,
+      "grad_norm": 0.2227998971939087,
+      "learning_rate": 3.576386087436212e-06,
+      "loss": 0.7635,
+      "step": 1570
+    },
+    {
+      "epoch": 0.6083357397247088,
+      "grad_norm": 0.22713981568813324,
+      "learning_rate": 3.5167214187975913e-06,
+      "loss": 0.7533,
+      "step": 1580
+    },
+    {
+      "epoch": 0.6121859659254981,
+      "grad_norm": 0.18936499953269958,
+      "learning_rate": 3.457287457018086e-06,
+      "loss": 0.7555,
+      "step": 1590
+    },
+    {
+      "epoch": 0.6160361921262875,
+      "grad_norm": 0.20442752540111542,
+      "learning_rate": 3.3980934463643003e-06,
+      "loss": 0.7442,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6198864183270767,
+      "grad_norm": 0.1965835690498352,
+      "learning_rate": 3.3391485937812053e-06,
+      "loss": 0.7256,
+      "step": 1610
+    },
+    {
+      "epoch": 0.623736644527866,
+      "grad_norm": 0.211113840341568,
+      "learning_rate": 3.280462067460115e-06,
+      "loss": 0.7238,
+      "step": 1620
+    },
+    {
+      "epoch": 0.6275868707286553,
+      "grad_norm": 0.2055027186870575,
+      "learning_rate": 3.222042995412669e-06,
+      "loss": 0.7565,
+      "step": 1630
+    },
+    {
+      "epoch": 0.6314370969294446,
+      "grad_norm": 0.18792873620986938,
+      "learning_rate": 3.163900464051089e-06,
+      "loss": 0.7456,
+      "step": 1640
+    },
+    {
+      "epoch": 0.635287323130234,
+      "grad_norm": 0.2242371290922165,
+      "learning_rate": 3.1060435167748826e-06,
+      "loss": 0.7569,
+      "step": 1650
+    },
+    {
+      "epoch": 0.6391375493310232,
+      "grad_norm": 0.21224342286586761,
+      "learning_rate": 3.048481152564253e-06,
+      "loss": 0.7362,
+      "step": 1660
+    },
+    {
+      "epoch": 0.6429877755318125,
+      "grad_norm": 0.22445707023143768,
+      "learning_rate": 2.9912223245804017e-06,
+      "loss": 0.7577,
+      "step": 1670
+    },
+    {
+      "epoch": 0.6468380017326018,
+      "grad_norm": 0.21960310637950897,
+      "learning_rate": 2.934275938772981e-06,
+      "loss": 0.7516,
+      "step": 1680
+    },
+    {
+      "epoch": 0.6506882279333911,
+      "grad_norm": 0.21108609437942505,
+      "learning_rate": 2.8776508524948676e-06,
+      "loss": 0.7424,
+      "step": 1690
+    },
+    {
+      "epoch": 0.6545384541341804,
+      "grad_norm": 0.1917160451412201,
+      "learning_rate": 2.821355873124503e-06,
+      "loss": 0.7409,
+      "step": 1700
+    },
+    {
+      "epoch": 0.6583886803349697,
+      "grad_norm": 0.21764113008975983,
+      "learning_rate": 2.7653997566960144e-06,
+      "loss": 0.7544,
+      "step": 1710
+    },
+    {
+      "epoch": 0.662238906535759,
+      "grad_norm": 0.21694687008857727,
+      "learning_rate": 2.709791206537309e-06,
+      "loss": 0.7436,
+      "step": 1720
+    },
+    {
+      "epoch": 0.6660891327365482,
+      "grad_norm": 0.21440766751766205,
+      "learning_rate": 2.6545388719163747e-06,
+      "loss": 0.7468,
+      "step": 1730
+    },
+    {
+      "epoch": 0.6699393589373376,
+      "grad_norm": 0.21767599880695343,
+      "learning_rate": 2.599651346695979e-06,
+      "loss": 0.7488,
+      "step": 1740
+    },
+    {
+      "epoch": 0.6737895851381268,
+      "grad_norm": 0.20463621616363525,
+      "learning_rate": 2.545137167997007e-06,
+      "loss": 0.7468,
+      "step": 1750
+    },
+    {
+      "epoch": 0.6776398113389162,
+      "grad_norm": 0.20163628458976746,
+      "learning_rate": 2.4910048148705972e-06,
+      "loss": 0.735,
+      "step": 1760
+    },
+    {
+      "epoch": 0.6814900375397055,
+      "grad_norm": 0.21785108745098114,
+      "learning_rate": 2.4372627069793327e-06,
+      "loss": 0.7797,
+      "step": 1770
+    },
+    {
+      "epoch": 0.6853402637404947,
+      "grad_norm": 0.2241637110710144,
+      "learning_rate": 2.3839192032876486e-06,
+      "loss": 0.7634,
+      "step": 1780
+    },
+    {
+      "epoch": 0.6891904899412841,
+      "grad_norm": 0.21101497113704681,
+      "learning_rate": 2.3309826007617047e-06,
+      "loss": 0.7747,
+      "step": 1790
+    },
+    {
+      "epoch": 0.6930407161420733,
+      "grad_norm": 0.2100452035665512,
+      "learning_rate": 2.278461133078881e-06,
+      "loss": 0.7376,
+      "step": 1800
+    },
+    {
+      "epoch": 0.6968909423428626,
+      "grad_norm": 0.22275123000144958,
+      "learning_rate": 2.2263629693471258e-06,
+      "loss": 0.7299,
+      "step": 1810
+    },
+    {
+      "epoch": 0.700741168543652,
+      "grad_norm": 0.1982683390378952,
+      "learning_rate": 2.1746962128343414e-06,
+      "loss": 0.751,
+      "step": 1820
+    },
+    {
+      "epoch": 0.7045913947444412,
+      "grad_norm": 0.20645099878311157,
+      "learning_rate": 2.1234688997080183e-06,
+      "loss": 0.7426,
+      "step": 1830
+    },
+    {
+      "epoch": 0.7084416209452306,
+      "grad_norm": 0.225324347615242,
+      "learning_rate": 2.0726889977853053e-06,
+      "loss": 0.7485,
+      "step": 1840
+    },
+    {
+      "epoch": 0.7122918471460198,
+      "grad_norm": 0.2218027114868164,
+      "learning_rate": 2.022364405293703e-06,
+      "loss": 0.759,
+      "step": 1850
+    },
+    {
+      "epoch": 0.7161420733468091,
+      "grad_norm": 0.21319644153118134,
+      "learning_rate": 1.9725029496425814e-06,
+      "loss": 0.7543,
+      "step": 1860
+    },
+    {
+      "epoch": 0.7199922995475985,
+      "grad_norm": 0.1935821771621704,
+      "learning_rate": 1.923112386205727e-06,
+      "loss": 0.7485,
+      "step": 1870
+    },
+    {
+      "epoch": 0.7238425257483877,
+      "grad_norm": 0.20326277613639832,
+      "learning_rate": 1.874200397115079e-06,
+      "loss": 0.7573,
+      "step": 1880
+    },
+    {
+      "epoch": 0.727692751949177,
+      "grad_norm": 0.22649911046028137,
+      "learning_rate": 1.825774590065857e-06,
+      "loss": 0.7398,
+      "step": 1890
+    },
+    {
+      "epoch": 0.7315429781499663,
+      "grad_norm": 0.2133716344833374,
+      "learning_rate": 1.7778424971332813e-06,
+      "loss": 0.7501,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7353932043507556,
+      "grad_norm": 0.20128603279590607,
+      "learning_rate": 1.7304115736010407e-06,
+      "loss": 0.7465,
+      "step": 1910
+    },
+    {
+      "epoch": 0.739243430551545,
+      "grad_norm": 0.20871880650520325,
+      "learning_rate": 1.6834891968017108e-06,
+      "loss": 0.7426,
+      "step": 1920
+    },
+    {
+      "epoch": 0.7430936567523342,
+      "grad_norm": 0.21697664260864258,
+      "learning_rate": 1.6370826649692884e-06,
+      "loss": 0.7568,
+      "step": 1930
+    },
+    {
+      "epoch": 0.7469438829531235,
+      "grad_norm": 0.22194243967533112,
+      "learning_rate": 1.5911991961040463e-06,
+      "loss": 0.7641,
+      "step": 1940
+    },
+    {
+      "epoch": 0.7507941091539128,
+      "grad_norm": 0.21561188995838165,
+      "learning_rate": 1.545845926849851e-06,
+      "loss": 0.752,
+      "step": 1950
+    },
+    {
+      "epoch": 0.7546443353547021,
+      "grad_norm": 0.2258695363998413,
+      "learning_rate": 1.5010299113841397e-06,
+      "loss": 0.7689,
+      "step": 1960
+    },
+    {
+      "epoch": 0.7584945615554913,
+      "grad_norm": 0.21041341125965118,
+      "learning_rate": 1.4567581203207238e-06,
+      "loss": 0.7586,
+      "step": 1970
+    },
+    {
+      "epoch": 0.7623447877562807,
+      "grad_norm": 0.21832452714443207,
+      "learning_rate": 1.413037439625593e-06,
+      "loss": 0.7571,
+      "step": 1980
+    },
+    {
+      "epoch": 0.76619501395707,
+      "grad_norm": 0.22595016658306122,
+      "learning_rate": 1.3698746695458847e-06,
+      "loss": 0.7511,
+      "step": 1990
+    },
+    {
+      "epoch": 0.7700452401578592,
+      "grad_norm": 0.20809485018253326,
+      "learning_rate": 1.3272765235521745e-06,
+      "loss": 0.7439,
+      "step": 2000
+    },
+    {
+      "epoch": 0.7738954663586486,
+      "grad_norm": 0.20957212150096893,
+      "learning_rate": 1.2852496272942877e-06,
+      "loss": 0.7526,
+      "step": 2010
+    },
+    {
+      "epoch": 0.7777456925594378,
+      "grad_norm": 0.1973990797996521,
+      "learning_rate": 1.2438005175707396e-06,
+      "loss": 0.7396,
+      "step": 2020
+    },
+    {
+      "epoch": 0.7815959187602272,
+      "grad_norm": 0.20650090277194977,
+      "learning_rate": 1.2029356413120297e-06,
+      "loss": 0.7242,
+      "step": 2030
+    },
+    {
+      "epoch": 0.7854461449610165,
+      "grad_norm": 0.21567979454994202,
+      "learning_rate": 1.1626613545778786e-06,
+      "loss": 0.7413,
+      "step": 2040
+    },
+    {
+      "epoch": 0.7892963711618057,
+      "grad_norm": 0.20169073343276978,
+      "learning_rate": 1.1229839215686294e-06,
+      "loss": 0.7455,
+      "step": 2050
+    },
+    {
+      "epoch": 0.7931465973625951,
+      "grad_norm": 0.19515064358711243,
+      "learning_rate": 1.0839095136509188e-06,
+      "loss": 0.7285,
+      "step": 2060
+    },
+    {
+      "epoch": 0.7969968235633843,
+      "grad_norm": 0.22688184678554535,
+      "learning_rate": 1.045444208397791e-06,
+      "loss": 0.751,
+      "step": 2070
+    },
+    {
+      "epoch": 0.8008470497641736,
+      "grad_norm": 0.1997089833021164,
+      "learning_rate": 1.007593988643402e-06,
+      "loss": 0.7356,
+      "step": 2080
+    },
+    {
+      "epoch": 0.804697275964963,
+      "grad_norm": 0.2073090374469757,
+      "learning_rate": 9.703647415524613e-07,
+      "loss": 0.757,
+      "step": 2090
+    },
+    {
+      "epoch": 0.8085475021657522,
+      "grad_norm": 0.21321067214012146,
+      "learning_rate": 9.337622577045485e-07,
+      "loss": 0.7444,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8123977283665416,
+      "grad_norm": 0.21134330332279205,
+      "learning_rate": 8.977922301934588e-07,
+      "loss": 0.7551,
+      "step": 2110
+    },
+    {
+      "epoch": 0.8162479545673308,
+      "grad_norm": 0.2113180160522461,
+      "learning_rate": 8.624602537417009e-07,
+      "loss": 0.7558,
+      "step": 2120
+    },
+    {
+      "epoch": 0.8200981807681201,
+      "grad_norm": 0.23202139139175415,
+      "learning_rate": 8.277718238303123e-07,
+      "loss": 0.7423,
+      "step": 2130
+    },
+    {
+      "epoch": 0.8239484069689095,
+      "grad_norm": 0.21415174007415771,
+      "learning_rate": 7.937323358440935e-07,
+      "loss": 0.7462,
+      "step": 2140
+    },
+    {
+      "epoch": 0.8277986331696987,
+      "grad_norm": 0.22943955659866333,
+      "learning_rate": 7.603470842324223e-07,
+      "loss": 0.7363,
+      "step": 2150
+    },
+    {
+      "epoch": 0.831648859370488,
+      "grad_norm": 0.20451441407203674,
+      "learning_rate": 7.276212616857575e-07,
+      "loss": 0.7553,
+      "step": 2160
+    },
+    {
+      "epoch": 0.8354990855712773,
+      "grad_norm": 0.20988237857818604,
+      "learning_rate": 6.955599583279837e-07,
+      "loss": 0.7546,
+      "step": 2170
+    },
+    {
+      "epoch": 0.8393493117720666,
+      "grad_norm": 0.20729590952396393,
+      "learning_rate": 6.641681609246981e-07,
+      "loss": 0.7469,
+      "step": 2180
+    },
+    {
+      "epoch": 0.8431995379728559,
+      "grad_norm": 0.2023775428533554,
+      "learning_rate": 6.334507521075711e-07,
+      "loss": 0.7538,
+      "step": 2190
+    },
+    {
+      "epoch": 0.8470497641736452,
+      "grad_norm": 0.22000907361507416,
+      "learning_rate": 6.034125096149179e-07,
+      "loss": 0.7495,
+      "step": 2200
+    },
+    {
+      "epoch": 0.8508999903744345,
+      "grad_norm": 0.21231864392757416,
+      "learning_rate": 5.740581055485722e-07,
+      "loss": 0.7193,
+      "step": 2210
+    },
+    {
+      "epoch": 0.8547502165752238,
+      "grad_norm": 0.1978382170200348,
+      "learning_rate": 5.453921056471961e-07,
+      "loss": 0.7525,
+      "step": 2220
+    },
+    {
+      "epoch": 0.8586004427760131,
+      "grad_norm": 0.21087007224559784,
+      "learning_rate": 5.174189685761288e-07,
+      "loss": 0.7488,
+      "step": 2230
+    },
+    {
+      "epoch": 0.8624506689768023,
+      "grad_norm": 0.19695062935352325,
+      "learning_rate": 4.901430452338979e-07,
+      "loss": 0.7593,
+      "step": 2240
+    },
+    {
+      "epoch": 0.8663008951775917,
+      "grad_norm": 0.2035503387451172,
+      "learning_rate": 4.6356857807548286e-07,
+      "loss": 0.7433,
+      "step": 2250
+    },
+    {
+      "epoch": 0.870151121378381,
+      "grad_norm": 0.20999962091445923,
+      "learning_rate": 4.376997004524541e-07,
+      "loss": 0.7332,
+      "step": 2260
+    },
+    {
+      "epoch": 0.8740013475791703,
+      "grad_norm": 0.20788265764713287,
+      "learning_rate": 4.12540435970073e-07,
+      "loss": 0.7396,
+      "step": 2270
+    },
+    {
+      "epoch": 0.8778515737799596,
+      "grad_norm": 0.2094229757785797,
+      "learning_rate": 3.880946978614703e-07,
+      "loss": 0.753,
+      "step": 2280
+    },
+    {
+      "epoch": 0.8817017999807488,
+      "grad_norm": 0.2088220715522766,
+      "learning_rate": 3.643662883789878e-07,
+      "loss": 0.7545,
+      "step": 2290
+    },
+    {
+      "epoch": 0.8855520261815382,
+      "grad_norm": 0.19726988673210144,
+      "learning_rate": 3.41358898202781e-07,
+      "loss": 0.7349,
+      "step": 2300
+    },
+    {
+      "epoch": 0.8894022523823275,
+      "grad_norm": 0.1956457495689392,
+      "learning_rate": 3.190761058667741e-07,
+      "loss": 0.7459,
+      "step": 2310
+    },
+    {
+      "epoch": 0.8932524785831167,
+      "grad_norm": 0.20350247621536255,
+      "learning_rate": 2.9752137720206463e-07,
+      "loss": 0.7499,
+      "step": 2320
+    },
+    {
+      "epoch": 0.8971027047839061,
+      "grad_norm": 0.20535774528980255,
+      "learning_rate": 2.7669806479785067e-07,
+      "loss": 0.7555,
+      "step": 2330
+    },
+    {
+      "epoch": 0.9009529309846953,
+      "grad_norm": 0.2061716765165329,
+      "learning_rate": 2.566094074799763e-07,
+      "loss": 0.7478,
+      "step": 2340
+    },
+    {
+      "epoch": 0.9048031571854847,
+      "grad_norm": 0.20702125132083893,
+      "learning_rate": 2.3725852980716557e-07,
+      "loss": 0.734,
+      "step": 2350
+    },
+    {
+      "epoch": 0.908653383386274,
+      "grad_norm": 0.2060261368751526,
+      "learning_rate": 2.1864844158503983e-07,
+      "loss": 0.7442,
+      "step": 2360
+    },
+    {
+      "epoch": 0.9125036095870632,
+      "grad_norm": 0.22542691230773926,
+      "learning_rate": 2.007820373979702e-07,
+      "loss": 0.7387,
+      "step": 2370
+    },
+    {
+      "epoch": 0.9163538357878526,
+      "grad_norm": 0.21386562287807465,
+      "learning_rate": 1.8366209615886532e-07,
+      "loss": 0.7465,
+      "step": 2380
+    },
+    {
+      "epoch": 0.9202040619886418,
+      "grad_norm": 0.21342851221561432,
+      "learning_rate": 1.6729128067693613e-07,
+      "loss": 0.7593,
+      "step": 2390
+    },
+    {
+      "epoch": 0.9240542881894311,
+      "grad_norm": 0.2256298065185547,
+      "learning_rate": 1.5167213724353426e-07,
+      "loss": 0.7585,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9279045143902205,
+      "grad_norm": 0.207830548286438,
+      "learning_rate": 1.3680709523610058e-07,
+      "loss": 0.7387,
+      "step": 2410
+    },
+    {
+      "epoch": 0.9317547405910097,
+      "grad_norm": 0.21028558909893036,
+      "learning_rate": 1.2269846674030418e-07,
+      "loss": 0.763,
+      "step": 2420
+    },
+    {
+      "epoch": 0.935604966791799,
+      "grad_norm": 0.22028367221355438,
+      "learning_rate": 1.0934844619042595e-07,
+      "loss": 0.7582,
+      "step": 2430
+    },
+    {
+      "epoch": 0.9394551929925883,
+      "grad_norm": 0.2213229537010193,
+      "learning_rate": 9.675911002803961e-08,
+      "loss": 0.7463,
+      "step": 2440
+    },
+    {
+      "epoch": 0.9433054191933776,
+      "grad_norm": 0.22989563643932343,
+      "learning_rate": 8.493241637904381e-08,
+      "loss": 0.7437,
+      "step": 2450
+    },
+    {
+      "epoch": 0.9471556453941669,
+      "grad_norm": 0.20870694518089294,
+      "learning_rate": 7.387020474909689e-08,
+      "loss": 0.763,
+      "step": 2460
+    },
+    {
+      "epoch": 0.9510058715949562,
+      "grad_norm": 0.22480618953704834,
+      "learning_rate": 6.357419573750856e-08,
+      "loss": 0.7588,
+      "step": 2470
+    },
+    {
+      "epoch": 0.9548560977957455,
+      "grad_norm": 0.21035709977149963,
+      "learning_rate": 5.4045990769613365e-08,
+      "loss": 0.7402,
+      "step": 2480
+    },
+    {
+      "epoch": 0.9587063239965348,
+      "grad_norm": 0.2096334546804428,
+      "learning_rate": 4.5287071847694434e-08,
+      "loss": 0.7471,
+      "step": 2490
+    },
+    {
+      "epoch": 0.9625565501973241,
+      "grad_norm": 0.20947164297103882,
+      "learning_rate": 3.7298801320467817e-08,
+      "loss": 0.744,
+      "step": 2500
+    },
+    {
+      "epoch": 0.9664067763981133,
+      "grad_norm": 0.21082325279712677,
+      "learning_rate": 3.0082421671192576e-08,
+      "loss": 0.7374,
+      "step": 2510
+    },
+    {
+      "epoch": 0.9702570025989027,
+      "grad_norm": 0.20632167160511017,
+      "learning_rate": 2.3639055324410355e-08,
+      "loss": 0.7579,
+      "step": 2520
+    },
+    {
+      "epoch": 0.974107228799692,
+      "grad_norm": 0.2089688777923584,
+      "learning_rate": 1.7969704471369477e-08,
+      "loss": 0.7426,
+      "step": 2530
+    },
+    {
+      "epoch": 0.9779574550004813,
+      "grad_norm": 0.2032756358385086,
+      "learning_rate": 1.3075250914142413e-08,
+      "loss": 0.744,
+      "step": 2540
+    },
+    {
+      "epoch": 0.9818076812012706,
+      "grad_norm": 0.20247232913970947,
+      "learning_rate": 8.956455928476049e-09,
+      "loss": 0.7532,
+      "step": 2550
+    },
+    {
+      "epoch": 0.9856579074020598,
+      "grad_norm": 0.206169992685318,
+      "learning_rate": 5.6139601453786275e-09,
+      "loss": 0.7447,
+      "step": 2560
+    },
+    {
+      "epoch": 0.9895081336028492,
+      "grad_norm": 0.20767617225646973,
+      "learning_rate": 3.0482834514805602e-09,
+      "loss": 0.7356,
+      "step": 2570
+    },
+    {
+      "epoch": 0.9933583598036385,
+      "grad_norm": 0.2012789398431778,
+      "learning_rate": 1.2598249081713366e-09,
+      "loss": 0.7265,
+      "step": 2580
+    },
+    {
+      "epoch": 0.9972085860044277,
+      "grad_norm": 0.2128566950559616,
+      "learning_rate": 2.488626895302826e-10,
+      "loss": 0.7604,
+      "step": 2590
+    },
+    {
+      "epoch": 0.9999037443449803,
+      "step": 2597,
+      "total_flos": 3.5185741077445345e+18,
+      "train_loss": 0.7671444626280469,
+      "train_runtime": 8192.0628,
+      "train_samples_per_second": 81.164,
+      "train_steps_per_second": 0.317
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2597,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.5185741077445345e+18,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/mc1_b16_224_2.5b/finetune/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a06276fc45c01c57458f2d6dd9ba5dac473ae656b60a50fb2188edb3bc9fa8c
+size 6136

continuous/qwen25/mc1_b16_224_2.5b/pretrain/config.json

@@ -0,0 +1,58 @@
+{
+  "_attn_implementation_autoset": true,
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "image_aspect_ratio": "square",
+  "image_crop_resolution": null,
+  "image_grid_pinpoints": null,
+  "image_split_resolution": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 768,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": null,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "mm_tunable_parts": "mm_mlp_adapter",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "open_clip_hub:ViT-B-16",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/mc1/mc1_b16_224_2.5b.pt",
+  "vocab_size": 151936
+}

continuous/qwen25/mc1_b16_224_2.5b/pretrain/mm_projector.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8794197d0ca1e5efe2e5e8aeb3656c02ddc77660ba1a6f78347fd91cadad86c5
+size 14170168

continuous/qwen25/mc1_b16_224_2.5b/pretrain/trainer_state.json

@@ -0,0 +1,421 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 545,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01834862385321101,
+      "grad_norm": 7.864686489105225,
+      "learning_rate": 0.00032142857142857147,
+      "loss": 4.6736,
+      "step": 10
+    },
+    {
+      "epoch": 0.03669724770642202,
+      "grad_norm": 1.6432969570159912,
+      "learning_rate": 0.0006785714285714287,
+      "loss": 3.719,
+      "step": 20
+    },
+    {
+      "epoch": 0.05504587155963303,
+      "grad_norm": 0.8070200681686401,
+      "learning_rate": 0.0009999907688176173,
+      "loss": 3.3153,
+      "step": 30
+    },
+    {
+      "epoch": 0.07339449541284404,
+      "grad_norm": 0.2532338798046112,
+      "learning_rate": 0.0009988834393115766,
+      "loss": 3.2201,
+      "step": 40
+    },
+    {
+      "epoch": 0.09174311926605505,
+      "grad_norm": 0.62339186668396,
+      "learning_rate": 0.0009959345572940772,
+      "loss": 3.189,
+      "step": 50
+    },
+    {
+      "epoch": 0.11009174311926606,
+      "grad_norm": 0.22550037503242493,
+      "learning_rate": 0.0009911550081155983,
+      "loss": 3.1468,
+      "step": 60
+    },
+    {
+      "epoch": 0.12844036697247707,
+      "grad_norm": 0.21378472447395325,
+      "learning_rate": 0.0009845624347566433,
+      "loss": 3.0995,
+      "step": 70
+    },
+    {
+      "epoch": 0.14678899082568808,
+      "grad_norm": 0.31360459327697754,
+      "learning_rate": 0.0009761811727013547,
+      "loss": 3.0641,
+      "step": 80
+    },
+    {
+      "epoch": 0.1651376146788991,
+      "grad_norm": 0.32564228773117065,
+      "learning_rate": 0.0009660421601067666,
+      "loss": 3.0194,
+      "step": 90
+    },
+    {
+      "epoch": 0.1834862385321101,
+      "grad_norm": 0.203284353017807,
+      "learning_rate": 0.0009541828235992926,
+      "loss": 2.9892,
+      "step": 100
+    },
+    {
+      "epoch": 0.2018348623853211,
+      "grad_norm": 0.23422332108020782,
+      "learning_rate": 0.0009406469401200151,
+      "loss": 2.948,
+      "step": 110
+    },
+    {
+      "epoch": 0.22018348623853212,
+      "grad_norm": 0.21101464331150055,
+      "learning_rate": 0.0009254844753287492,
+      "loss": 2.9263,
+      "step": 120
+    },
+    {
+      "epoch": 0.23853211009174313,
+      "grad_norm": 0.22854891419410706,
+      "learning_rate": 0.0009087513991633924,
+      "loss": 2.8797,
+      "step": 130
+    },
+    {
+      "epoch": 0.25688073394495414,
+      "grad_norm": 0.21006974577903748,
+      "learning_rate": 0.0008905094792353917,
+      "loss": 2.8434,
+      "step": 140
+    },
+    {
+      "epoch": 0.27522935779816515,
+      "grad_norm": 0.22404561936855316,
+      "learning_rate": 0.0008708260528239789,
+      "loss": 2.8146,
+      "step": 150
+    },
+    {
+      "epoch": 0.29357798165137616,
+      "grad_norm": 0.2806335389614105,
+      "learning_rate": 0.0008497737783108237,
+      "loss": 2.775,
+      "step": 160
+    },
+    {
+      "epoch": 0.3119266055045872,
+      "grad_norm": 0.2246663123369217,
+      "learning_rate": 0.0008274303669726426,
+      "loss": 2.7513,
+      "step": 170
+    },
+    {
+      "epoch": 0.3302752293577982,
+      "grad_norm": 0.29829517006874084,
+      "learning_rate": 0.0008038782961218137,
+      "loss": 2.734,
+      "step": 180
+    },
+    {
+      "epoch": 0.3486238532110092,
+      "grad_norm": 0.21977505087852478,
+      "learning_rate": 0.000779204504653894,
+      "loss": 2.7211,
+      "step": 190
+    },
+    {
+      "epoch": 0.3669724770642202,
+      "grad_norm": 0.26401564478874207,
+      "learning_rate": 0.0007535000721258791,
+      "loss": 2.694,
+      "step": 200
+    },
+    {
+      "epoch": 0.3853211009174312,
+      "grad_norm": 0.2710863947868347,
+      "learning_rate": 0.00072685988254984,
+      "loss": 2.6882,
+      "step": 210
+    },
+    {
+      "epoch": 0.4036697247706422,
+      "grad_norm": 0.3259151875972748,
+      "learning_rate": 0.0006993822741429906,
+      "loss": 2.6935,
+      "step": 220
+    },
+    {
+      "epoch": 0.42201834862385323,
+      "grad_norm": 0.24984776973724365,
+      "learning_rate": 0.0006711686763270818,
+      "loss": 2.6829,
+      "step": 230
+    },
+    {
+      "epoch": 0.44036697247706424,
+      "grad_norm": 0.2524448335170746,
+      "learning_rate": 0.0006423232353170831,
+      "loss": 2.6696,
+      "step": 240
+    },
+    {
+      "epoch": 0.45871559633027525,
+      "grad_norm": 0.2116161286830902,
+      "learning_rate": 0.0006129524296812335,
+      "loss": 2.6674,
+      "step": 250
+    },
+    {
+      "epoch": 0.47706422018348627,
+      "grad_norm": 0.2032284289598465,
+      "learning_rate": 0.0005831646772915651,
+      "loss": 2.6745,
+      "step": 260
+    },
+    {
+      "epoch": 0.4954128440366973,
+      "grad_norm": 0.19314143061637878,
+      "learning_rate": 0.0005530699351157809,
+      "loss": 2.6519,
+      "step": 270
+    },
+    {
+      "epoch": 0.5137614678899083,
+      "grad_norm": 0.2080564647912979,
+      "learning_rate": 0.0005227792933277943,
+      "loss": 2.6534,
+      "step": 280
+    },
+    {
+      "epoch": 0.5321100917431193,
+      "grad_norm": 0.1877630352973938,
+      "learning_rate": 0.0004924045652352092,
+      "loss": 2.6396,
+      "step": 290
+    },
+    {
+      "epoch": 0.5504587155963303,
+      "grad_norm": 0.23189647495746613,
+      "learning_rate": 0.0004620578745374633,
+      "loss": 2.6418,
+      "step": 300
+    },
+    {
+      "epoch": 0.5688073394495413,
+      "grad_norm": 0.20622986555099487,
+      "learning_rate": 0.0004318512414382059,
+      "loss": 2.6465,
+      "step": 310
+    },
+    {
+      "epoch": 0.5871559633027523,
+      "grad_norm": 0.21966595947742462,
+      "learning_rate": 0.0004018961691397148,
+      "loss": 2.6411,
+      "step": 320
+    },
+    {
+      "epoch": 0.6055045871559633,
+      "grad_norm": 0.24698108434677124,
+      "learning_rate": 0.0003723032322457458,
+      "loss": 2.6284,
+      "step": 330
+    },
+    {
+      "epoch": 0.6238532110091743,
+      "grad_norm": 0.2700939178466797,
+      "learning_rate": 0.0003431816685921625,
+      "loss": 2.6318,
+      "step": 340
+    },
+    {
+      "epoch": 0.6422018348623854,
+      "grad_norm": 0.1821216493844986,
+      "learning_rate": 0.0003146389760120416,
+      "loss": 2.629,
+      "step": 350
+    },
+    {
+      "epoch": 0.6605504587155964,
+      "grad_norm": 0.24043826758861542,
+      "learning_rate": 0.00028678051552373485,
+      "loss": 2.6315,
+      "step": 360
+    },
+    {
+      "epoch": 0.6788990825688074,
+      "grad_norm": 0.1877768188714981,
+      "learning_rate": 0.0002597091224066581,
+      "loss": 2.6259,
+      "step": 370
+    },
+    {
+      "epoch": 0.6972477064220184,
+      "grad_norm": 0.18730507791042328,
+      "learning_rate": 0.0002335247266004629,
+      "loss": 2.6321,
+      "step": 380
+    },
+    {
+      "epoch": 0.7155963302752294,
+      "grad_norm": 0.25525009632110596,
+      "learning_rate": 0.000208323983828828,
+      "loss": 2.6255,
+      "step": 390
+    },
+    {
+      "epoch": 0.7339449541284404,
+      "grad_norm": 0.192392498254776,
+      "learning_rate": 0.0001841999188095224,
+      "loss": 2.6144,
+      "step": 400
+    },
+    {
+      "epoch": 0.7522935779816514,
+      "grad_norm": 0.22381971776485443,
+      "learning_rate": 0.00016124158186777676,
+      "loss": 2.6175,
+      "step": 410
+    },
+    {
+      "epoch": 0.7706422018348624,
+      "grad_norm": 0.18858136236667633,
+      "learning_rate": 0.0001395337202205211,
+      "loss": 2.617,
+      "step": 420
+    },
+    {
+      "epoch": 0.7889908256880734,
+      "grad_norm": 0.18701297044754028,
+      "learning_rate": 0.00011915646514489292,
+      "loss": 2.6124,
+      "step": 430
+    },
+    {
+      "epoch": 0.8073394495412844,
+      "grad_norm": 0.20885521173477173,
+      "learning_rate": 0.00010018503618578818,
+      "loss": 2.6236,
+      "step": 440
+    },
+    {
+      "epoch": 0.8256880733944955,
+      "grad_norm": 0.18807494640350342,
+      "learning_rate": 8.268946349432583e-05,
+      "loss": 2.6163,
+      "step": 450
+    },
+    {
+      "epoch": 0.8440366972477065,
+      "grad_norm": 0.18586969375610352,
+      "learning_rate": 6.673432932217117e-05,
+      "loss": 2.6109,
+      "step": 460
+    },
+    {
+      "epoch": 0.8623853211009175,
+      "grad_norm": 0.17317858338356018,
+      "learning_rate": 5.2378529625954684e-05,
+      "loss": 2.6003,
+      "step": 470
+    },
+    {
+      "epoch": 0.8807339449541285,
+      "grad_norm": 0.1768551617860794,
+      "learning_rate": 3.9675056661785556e-05,
+      "loss": 2.6028,
+      "step": 480
+    },
+    {
+      "epoch": 0.8990825688073395,
+      "grad_norm": 0.17821907997131348,
+      "learning_rate": 2.8670803372379538e-05,
+      "loss": 2.615,
+      "step": 490
+    },
+    {
+      "epoch": 0.9174311926605505,
+      "grad_norm": 0.16098742187023163,
+      "learning_rate": 1.9406390288876586e-05,
+      "loss": 2.6208,
+      "step": 500
+    },
+    {
+      "epoch": 0.9357798165137615,
+      "grad_norm": 0.16520005464553833,
+      "learning_rate": 1.191601558631461e-05,
+      "loss": 2.6096,
+      "step": 510
+    },
+    {
+      "epoch": 0.9541284403669725,
+      "grad_norm": 0.16866159439086914,
+      "learning_rate": 6.2273288462559305e-06,
+      "loss": 2.6078,
+      "step": 520
+    },
+    {
+      "epoch": 0.9724770642201835,
+      "grad_norm": 0.1723398119211197,
+      "learning_rate": 2.361328992552314e-06,
+      "loss": 2.6107,
+      "step": 530
+    },
+    {
+      "epoch": 0.9908256880733946,
+      "grad_norm": 0.17704376578330994,
+      "learning_rate": 3.322867770044202e-07,
+      "loss": 2.603,
+      "step": 540
+    },
+    {
+      "epoch": 1.0,
+      "step": 545,
+      "total_flos": 2.8888413781989786e+17,
+      "train_loss": 2.7956635606398277,
+      "train_runtime": 5331.3938,
+      "train_samples_per_second": 104.687,
+      "train_steps_per_second": 0.102
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 545,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.8888413781989786e+17,
+  "train_batch_size": 128,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/mc1_b16_224_400m/finetune/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

continuous/qwen25/mc1_b16_224_400m/finetune/config.json

@@ -0,0 +1,57 @@
+{
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "LlavaQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "image_aspect_ratio": "square",
+  "image_crop_resolution": null,
+  "image_grid_pinpoints": null,
+  "image_split_resolution": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 768,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": 0.0001,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "open_clip_hub:ViT-B-16",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/mc1/mc1_b16_224_400m.pt",
+  "vocab_size": 151936
+}

continuous/qwen25/mc1_b16_224_400m/finetune/generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.51.3"
+}

continuous/qwen25/mc1_b16_224_400m/finetune/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f80faab3a7bac0d266466dfa21b66d780c69d2ee95ee2376b8eb3a69231ed5f7
+size 3740790376

continuous/qwen25/mc1_b16_224_400m/finetune/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
+  }
+}

continuous/qwen25/mc1_b16_224_400m/finetune/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

continuous/qwen25/mc1_b16_224_400m/finetune/tokenizer_config.json

@@ -0,0 +1,209 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

continuous/qwen25/mc1_b16_224_400m/finetune/trainer_state.json

@@ -0,0 +1,1856 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9999037443449803,
+  "eval_steps": 500,
+  "global_step": 2597,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.003850226200789296,
+      "grad_norm": 18.835752487182617,
+      "learning_rate": 1.153846153846154e-06,
+      "loss": 2.0732,
+      "step": 10
+    },
+    {
+      "epoch": 0.007700452401578592,
+      "grad_norm": 14.589078903198242,
+      "learning_rate": 2.435897435897436e-06,
+      "loss": 1.6833,
+      "step": 20
+    },
+    {
+      "epoch": 0.01155067860236789,
+      "grad_norm": 3.3345439434051514,
+      "learning_rate": 3.7179487179487184e-06,
+      "loss": 1.1279,
+      "step": 30
+    },
+    {
+      "epoch": 0.015400904803157185,
+      "grad_norm": 0.7488361597061157,
+      "learning_rate": 5e-06,
+      "loss": 0.9461,
+      "step": 40
+    },
+    {
+      "epoch": 0.019251131003946483,
+      "grad_norm": 0.34313663840293884,
+      "learning_rate": 6.282051282051282e-06,
+      "loss": 0.9115,
+      "step": 50
+    },
+    {
+      "epoch": 0.02310135720473578,
+      "grad_norm": 0.28679269552230835,
+      "learning_rate": 7.564102564102564e-06,
+      "loss": 0.8619,
+      "step": 60
+    },
+    {
+      "epoch": 0.026951583405525073,
+      "grad_norm": 0.23806780576705933,
+      "learning_rate": 8.846153846153847e-06,
+      "loss": 0.8357,
+      "step": 70
+    },
+    {
+      "epoch": 0.03080180960631437,
+      "grad_norm": 0.23563866317272186,
+      "learning_rate": 9.999996111488723e-06,
+      "loss": 0.844,
+      "step": 80
+    },
+    {
+      "epoch": 0.034652035807103666,
+      "grad_norm": 0.21510663628578186,
+      "learning_rate": 9.999529497453782e-06,
+      "loss": 0.8247,
+      "step": 90
+    },
+    {
+      "epoch": 0.038502262007892966,
+      "grad_norm": 0.23193146288394928,
+      "learning_rate": 9.998285264324193e-06,
+      "loss": 0.803,
+      "step": 100
+    },
+    {
+      "epoch": 0.04235248820868226,
+      "grad_norm": 0.2052900642156601,
+      "learning_rate": 9.996263605626052e-06,
+      "loss": 0.7949,
+      "step": 110
+    },
+    {
+      "epoch": 0.04620271440947156,
+      "grad_norm": 0.23204490542411804,
+      "learning_rate": 9.99346483580503e-06,
+      "loss": 0.8058,
+      "step": 120
+    },
+    {
+      "epoch": 0.05005294061026085,
+      "grad_norm": 0.21052762866020203,
+      "learning_rate": 9.989889390177467e-06,
+      "loss": 0.7949,
+      "step": 130
+    },
+    {
+      "epoch": 0.053903166811050146,
+      "grad_norm": 0.23175011575222015,
+      "learning_rate": 9.985537824862647e-06,
+      "loss": 0.7976,
+      "step": 140
+    },
+    {
+      "epoch": 0.057753393011839446,
+      "grad_norm": 0.2336784303188324,
+      "learning_rate": 9.980410816696317e-06,
+      "loss": 0.7783,
+      "step": 150
+    },
+    {
+      "epoch": 0.06160361921262874,
+      "grad_norm": 0.22231346368789673,
+      "learning_rate": 9.97450916312541e-06,
+      "loss": 0.7888,
+      "step": 160
+    },
+    {
+      "epoch": 0.06545384541341803,
+      "grad_norm": 0.19863121211528778,
+      "learning_rate": 9.967833782084004e-06,
+      "loss": 0.7897,
+      "step": 170
+    },
+    {
+      "epoch": 0.06930407161420733,
+      "grad_norm": 0.23243577778339386,
+      "learning_rate": 9.960385711850554e-06,
+      "loss": 0.814,
+      "step": 180
+    },
+    {
+      "epoch": 0.07315429781499663,
+      "grad_norm": 0.22161060571670532,
+      "learning_rate": 9.952166110886394e-06,
+      "loss": 0.8076,
+      "step": 190
+    },
+    {
+      "epoch": 0.07700452401578593,
+      "grad_norm": 0.24982470273971558,
+      "learning_rate": 9.943176257655567e-06,
+      "loss": 0.7986,
+      "step": 200
+    },
+    {
+      "epoch": 0.08085475021657522,
+      "grad_norm": 0.21747465431690216,
+      "learning_rate": 9.933417550425952e-06,
+      "loss": 0.7926,
+      "step": 210
+    },
+    {
+      "epoch": 0.08470497641736452,
+      "grad_norm": 0.2238951176404953,
+      "learning_rate": 9.9228915070518e-06,
+      "loss": 0.7775,
+      "step": 220
+    },
+    {
+      "epoch": 0.08855520261815382,
+      "grad_norm": 0.21949514746665955,
+      "learning_rate": 9.911599764737632e-06,
+      "loss": 0.7802,
+      "step": 230
+    },
+    {
+      "epoch": 0.09240542881894312,
+      "grad_norm": 0.22720587253570557,
+      "learning_rate": 9.899544079783606e-06,
+      "loss": 0.7738,
+      "step": 240
+    },
+    {
+      "epoch": 0.0962556550197324,
+      "grad_norm": 0.22298425436019897,
+      "learning_rate": 9.886726327312338e-06,
+      "loss": 0.7914,
+      "step": 250
+    },
+    {
+      "epoch": 0.1001058812205217,
+      "grad_norm": 0.2223641723394394,
+      "learning_rate": 9.87314850097724e-06,
+      "loss": 0.7879,
+      "step": 260
+    },
+    {
+      "epoch": 0.103956107421311,
+      "grad_norm": 0.2084207683801651,
+      "learning_rate": 9.858812712652452e-06,
+      "loss": 0.7786,
+      "step": 270
+    },
+    {
+      "epoch": 0.10780633362210029,
+      "grad_norm": 0.20738090574741364,
+      "learning_rate": 9.843721192104341e-06,
+      "loss": 0.7757,
+      "step": 280
+    },
+    {
+      "epoch": 0.11165655982288959,
+      "grad_norm": 0.20937480032444,
+      "learning_rate": 9.827876286644698e-06,
+      "loss": 0.7974,
+      "step": 290
+    },
+    {
+      "epoch": 0.11550678602367889,
+      "grad_norm": 0.20317278802394867,
+      "learning_rate": 9.811280460765643e-06,
+      "loss": 0.7596,
+      "step": 300
+    },
+    {
+      "epoch": 0.11935701222446819,
+      "grad_norm": 0.21985016763210297,
+      "learning_rate": 9.793936295756292e-06,
+      "loss": 0.7683,
+      "step": 310
+    },
+    {
+      "epoch": 0.12320723842525748,
+      "grad_norm": 0.22328853607177734,
+      "learning_rate": 9.775846489301279e-06,
+      "loss": 0.773,
+      "step": 320
+    },
+    {
+      "epoch": 0.12705746462604678,
+      "grad_norm": 0.2253125160932541,
+      "learning_rate": 9.757013855061157e-06,
+      "loss": 0.7704,
+      "step": 330
+    },
+    {
+      "epoch": 0.13090769082683607,
+      "grad_norm": 0.21561259031295776,
+      "learning_rate": 9.737441322234758e-06,
+      "loss": 0.7837,
+      "step": 340
+    },
+    {
+      "epoch": 0.13475791702762538,
+      "grad_norm": 0.22767041623592377,
+      "learning_rate": 9.717131935103605e-06,
+      "loss": 0.7835,
+      "step": 350
+    },
+    {
+      "epoch": 0.13860814322841467,
+      "grad_norm": 0.2360040843486786,
+      "learning_rate": 9.696088852558398e-06,
+      "loss": 0.7699,
+      "step": 360
+    },
+    {
+      "epoch": 0.14245836942920398,
+      "grad_norm": 0.21340255439281464,
+      "learning_rate": 9.674315347607686e-06,
+      "loss": 0.7953,
+      "step": 370
+    },
+    {
+      "epoch": 0.14630859562999327,
+      "grad_norm": 0.21205247938632965,
+      "learning_rate": 9.651814806868796e-06,
+      "loss": 0.7612,
+      "step": 380
+    },
+    {
+      "epoch": 0.15015882183078255,
+      "grad_norm": 0.21794761717319489,
+      "learning_rate": 9.62859073004107e-06,
+      "loss": 0.7619,
+      "step": 390
+    },
+    {
+      "epoch": 0.15400904803157187,
+      "grad_norm": 0.19877320528030396,
+      "learning_rate": 9.604646729361547e-06,
+      "loss": 0.7673,
+      "step": 400
+    },
+    {
+      "epoch": 0.15785927423236115,
+      "grad_norm": 0.21303172409534454,
+      "learning_rate": 9.579986529043099e-06,
+      "loss": 0.758,
+      "step": 410
+    },
+    {
+      "epoch": 0.16170950043315044,
+      "grad_norm": 0.21501217782497406,
+      "learning_rate": 9.554613964695189e-06,
+      "loss": 0.7744,
+      "step": 420
+    },
+    {
+      "epoch": 0.16555972663393975,
+      "grad_norm": 0.21690897643566132,
+      "learning_rate": 9.528532982727281e-06,
+      "loss": 0.7677,
+      "step": 430
+    },
+    {
+      "epoch": 0.16940995283472904,
+      "grad_norm": 0.204963818192482,
+      "learning_rate": 9.501747639735018e-06,
+      "loss": 0.7641,
+      "step": 440
+    },
+    {
+      "epoch": 0.17326017903551832,
+      "grad_norm": 0.21726247668266296,
+      "learning_rate": 9.474262101869272e-06,
+      "loss": 0.7522,
+      "step": 450
+    },
+    {
+      "epoch": 0.17711040523630764,
+      "grad_norm": 0.2178446501493454,
+      "learning_rate": 9.446080644188141e-06,
+      "loss": 0.7596,
+      "step": 460
+    },
+    {
+      "epoch": 0.18096063143709692,
+      "grad_norm": 0.20756758749485016,
+      "learning_rate": 9.417207649992017e-06,
+      "loss": 0.7642,
+      "step": 470
+    },
+    {
+      "epoch": 0.18481085763788624,
+      "grad_norm": 0.21511325240135193,
+      "learning_rate": 9.387647610141816e-06,
+      "loss": 0.7636,
+      "step": 480
+    },
+    {
+      "epoch": 0.18866108383867553,
+      "grad_norm": 0.19718995690345764,
+      "learning_rate": 9.35740512236047e-06,
+      "loss": 0.7505,
+      "step": 490
+    },
+    {
+      "epoch": 0.1925113100394648,
+      "grad_norm": 0.20271191000938416,
+      "learning_rate": 9.326484890517809e-06,
+      "loss": 0.7628,
+      "step": 500
+    },
+    {
+      "epoch": 0.19636153624025413,
+      "grad_norm": 0.21767131984233856,
+      "learning_rate": 9.294891723898926e-06,
+      "loss": 0.7513,
+      "step": 510
+    },
+    {
+      "epoch": 0.2002117624410434,
+      "grad_norm": 0.22505521774291992,
+      "learning_rate": 9.262630536456153e-06,
+      "loss": 0.781,
+      "step": 520
+    },
+    {
+      "epoch": 0.2040619886418327,
+      "grad_norm": 0.22301790118217468,
+      "learning_rate": 9.229706346044749e-06,
+      "loss": 0.7563,
+      "step": 530
+    },
+    {
+      "epoch": 0.207912214842622,
+      "grad_norm": 0.2264462262392044,
+      "learning_rate": 9.196124273642422e-06,
+      "loss": 0.7673,
+      "step": 540
+    },
+    {
+      "epoch": 0.2117624410434113,
+      "grad_norm": 0.21233148872852325,
+      "learning_rate": 9.161889542552842e-06,
+      "loss": 0.7724,
+      "step": 550
+    },
+    {
+      "epoch": 0.21561266724420058,
+      "grad_norm": 0.20597247779369354,
+      "learning_rate": 9.127007477593193e-06,
+      "loss": 0.7731,
+      "step": 560
+    },
+    {
+      "epoch": 0.2194628934449899,
+      "grad_norm": 0.22829784452915192,
+      "learning_rate": 9.091483504265974e-06,
+      "loss": 0.782,
+      "step": 570
+    },
+    {
+      "epoch": 0.22331311964577918,
+      "grad_norm": 0.22144466638565063,
+      "learning_rate": 9.055323147915116e-06,
+      "loss": 0.7616,
+      "step": 580
+    },
+    {
+      "epoch": 0.2271633458465685,
+      "grad_norm": 0.2185191959142685,
+      "learning_rate": 9.018532032866585e-06,
+      "loss": 0.7721,
+      "step": 590
+    },
+    {
+      "epoch": 0.23101357204735778,
+      "grad_norm": 0.20346054434776306,
+      "learning_rate": 8.981115881553579e-06,
+      "loss": 0.7638,
+      "step": 600
+    },
+    {
+      "epoch": 0.23486379824814707,
+      "grad_norm": 0.2268294394016266,
+      "learning_rate": 8.943080513626476e-06,
+      "loss": 0.7851,
+      "step": 610
+    },
+    {
+      "epoch": 0.23871402444893638,
+      "grad_norm": 0.22416383028030396,
+      "learning_rate": 8.904431845047641e-06,
+      "loss": 0.7691,
+      "step": 620
+    },
+    {
+      "epoch": 0.24256425064972567,
+      "grad_norm": 0.2269783318042755,
+      "learning_rate": 8.86517588717128e-06,
+      "loss": 0.7847,
+      "step": 630
+    },
+    {
+      "epoch": 0.24641447685051496,
+      "grad_norm": 0.20742908120155334,
+      "learning_rate": 8.82531874580844e-06,
+      "loss": 0.7773,
+      "step": 640
+    },
+    {
+      "epoch": 0.25026470305130427,
+      "grad_norm": 0.19649049639701843,
+      "learning_rate": 8.784866620277314e-06,
+      "loss": 0.7353,
+      "step": 650
+    },
+    {
+      "epoch": 0.25411492925209356,
+      "grad_norm": 0.22950272262096405,
+      "learning_rate": 8.743825802439018e-06,
+      "loss": 0.766,
+      "step": 660
+    },
+    {
+      "epoch": 0.25796515545288284,
+      "grad_norm": 0.24541804194450378,
+      "learning_rate": 8.702202675718956e-06,
+      "loss": 0.7755,
+      "step": 670
+    },
+    {
+      "epoch": 0.26181538165367213,
+      "grad_norm": 0.2072380781173706,
+      "learning_rate": 8.660003714113959e-06,
+      "loss": 0.7441,
+      "step": 680
+    },
+    {
+      "epoch": 0.26566560785446147,
+      "grad_norm": 0.21945248544216156,
+      "learning_rate": 8.617235481185323e-06,
+      "loss": 0.765,
+      "step": 690
+    },
+    {
+      "epoch": 0.26951583405525076,
+      "grad_norm": 0.22269375622272491,
+      "learning_rate": 8.573904629037931e-06,
+      "loss": 0.7718,
+      "step": 700
+    },
+    {
+      "epoch": 0.27336606025604004,
+      "grad_norm": 0.21736352145671844,
+      "learning_rate": 8.530017897285585e-06,
+      "loss": 0.755,
+      "step": 710
+    },
+    {
+      "epoch": 0.27721628645682933,
+      "grad_norm": 0.1964256763458252,
+      "learning_rate": 8.485582112002746e-06,
+      "loss": 0.7598,
+      "step": 720
+    },
+    {
+      "epoch": 0.2810665126576186,
+      "grad_norm": 0.21001394093036652,
+      "learning_rate": 8.440604184662811e-06,
+      "loss": 0.7612,
+      "step": 730
+    },
+    {
+      "epoch": 0.28491673885840796,
+      "grad_norm": 0.22367703914642334,
+      "learning_rate": 8.395091111063118e-06,
+      "loss": 0.7515,
+      "step": 740
+    },
+    {
+      "epoch": 0.28876696505919724,
+      "grad_norm": 0.20133692026138306,
+      "learning_rate": 8.349049970236822e-06,
+      "loss": 0.7517,
+      "step": 750
+    },
+    {
+      "epoch": 0.29261719125998653,
+      "grad_norm": 0.2130611091852188,
+      "learning_rate": 8.302487923351842e-06,
+      "loss": 0.744,
+      "step": 760
+    },
+    {
+      "epoch": 0.2964674174607758,
+      "grad_norm": 0.21656329929828644,
+      "learning_rate": 8.255412212597021e-06,
+      "loss": 0.7616,
+      "step": 770
+    },
+    {
+      "epoch": 0.3003176436615651,
+      "grad_norm": 0.20560668408870697,
+      "learning_rate": 8.207830160055688e-06,
+      "loss": 0.7585,
+      "step": 780
+    },
+    {
+      "epoch": 0.3041678698623544,
+      "grad_norm": 0.21816343069076538,
+      "learning_rate": 8.159749166566786e-06,
+      "loss": 0.745,
+      "step": 790
+    },
+    {
+      "epoch": 0.30801809606314373,
+      "grad_norm": 0.208966925740242,
+      "learning_rate": 8.11117671057377e-06,
+      "loss": 0.7718,
+      "step": 800
+    },
+    {
+      "epoch": 0.311868322263933,
+      "grad_norm": 0.22284935414791107,
+      "learning_rate": 8.062120346961404e-06,
+      "loss": 0.763,
+      "step": 810
+    },
+    {
+      "epoch": 0.3157185484647223,
+      "grad_norm": 0.19825054705142975,
+      "learning_rate": 8.012587705880705e-06,
+      "loss": 0.7589,
+      "step": 820
+    },
+    {
+      "epoch": 0.3195687746655116,
+      "grad_norm": 0.21199552714824677,
+      "learning_rate": 7.962586491562146e-06,
+      "loss": 0.7423,
+      "step": 830
+    },
+    {
+      "epoch": 0.3234190008663009,
+      "grad_norm": 0.22896474599838257,
+      "learning_rate": 7.912124481117357e-06,
+      "loss": 0.7486,
+      "step": 840
+    },
+    {
+      "epoch": 0.3272692270670902,
+      "grad_norm": 0.21411560475826263,
+      "learning_rate": 7.861209523329494e-06,
+      "loss": 0.7613,
+      "step": 850
+    },
+    {
+      "epoch": 0.3311194532678795,
+      "grad_norm": 0.21603447198867798,
+      "learning_rate": 7.809849537432432e-06,
+      "loss": 0.7441,
+      "step": 860
+    },
+    {
+      "epoch": 0.3349696794686688,
+      "grad_norm": 0.2164251059293747,
+      "learning_rate": 7.758052511879035e-06,
+      "loss": 0.7557,
+      "step": 870
+    },
+    {
+      "epoch": 0.3388199056694581,
+      "grad_norm": 0.20410297811031342,
+      "learning_rate": 7.705826503098642e-06,
+      "loss": 0.7502,
+      "step": 880
+    },
+    {
+      "epoch": 0.34267013187024736,
+      "grad_norm": 0.20623113214969635,
+      "learning_rate": 7.653179634243979e-06,
+      "loss": 0.7495,
+      "step": 890
+    },
+    {
+      "epoch": 0.34652035807103665,
+      "grad_norm": 0.19233396649360657,
+      "learning_rate": 7.600120093927696e-06,
+      "loss": 0.7526,
+      "step": 900
+    },
+    {
+      "epoch": 0.350370584271826,
+      "grad_norm": 0.22144433856010437,
+      "learning_rate": 7.546656134948729e-06,
+      "loss": 0.738,
+      "step": 910
+    },
+    {
+      "epoch": 0.3542208104726153,
+      "grad_norm": 0.21593716740608215,
+      "learning_rate": 7.492796073008661e-06,
+      "loss": 0.7421,
+      "step": 920
+    },
+    {
+      "epoch": 0.35807103667340456,
+      "grad_norm": 0.22990861535072327,
+      "learning_rate": 7.438548285418326e-06,
+      "loss": 0.7305,
+      "step": 930
+    },
+    {
+      "epoch": 0.36192126287419385,
+      "grad_norm": 0.23537759482860565,
+      "learning_rate": 7.3839212097948e-06,
+      "loss": 0.7637,
+      "step": 940
+    },
+    {
+      "epoch": 0.36577148907498314,
+      "grad_norm": 0.22036458551883698,
+      "learning_rate": 7.328923342749039e-06,
+      "loss": 0.7781,
+      "step": 950
+    },
+    {
+      "epoch": 0.3696217152757725,
+      "grad_norm": 0.22246412932872772,
+      "learning_rate": 7.273563238564319e-06,
+      "loss": 0.7639,
+      "step": 960
+    },
+    {
+      "epoch": 0.37347194147656176,
+      "grad_norm": 0.2255546599626541,
+      "learning_rate": 7.217849507865724e-06,
+      "loss": 0.7421,
+      "step": 970
+    },
+    {
+      "epoch": 0.37732216767735105,
+      "grad_norm": 0.20996367931365967,
+      "learning_rate": 7.161790816280858e-06,
+      "loss": 0.7498,
+      "step": 980
+    },
+    {
+      "epoch": 0.38117239387814034,
+      "grad_norm": 0.22872282564640045,
+      "learning_rate": 7.1053958830920125e-06,
+      "loss": 0.7304,
+      "step": 990
+    },
+    {
+      "epoch": 0.3850226200789296,
+      "grad_norm": 0.20718950033187866,
+      "learning_rate": 7.0486734798799715e-06,
+      "loss": 0.7491,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3888728462797189,
+      "grad_norm": 0.2095428854227066,
+      "learning_rate": 6.991632429159708e-06,
+      "loss": 0.7728,
+      "step": 1010
+    },
+    {
+      "epoch": 0.39272307248050825,
+      "grad_norm": 0.22516222298145294,
+      "learning_rate": 6.934281603008132e-06,
+      "loss": 0.7646,
+      "step": 1020
+    },
+    {
+      "epoch": 0.39657329868129754,
+      "grad_norm": 0.20978526771068573,
+      "learning_rate": 6.876629921684143e-06,
+      "loss": 0.7626,
+      "step": 1030
+    },
+    {
+      "epoch": 0.4004235248820868,
+      "grad_norm": 0.21864843368530273,
+      "learning_rate": 6.818686352241199e-06,
+      "loss": 0.7524,
+      "step": 1040
+    },
+    {
+      "epoch": 0.4042737510828761,
+      "grad_norm": 0.19806915521621704,
+      "learning_rate": 6.760459907132573e-06,
+      "loss": 0.7497,
+      "step": 1050
+    },
+    {
+      "epoch": 0.4081239772836654,
+      "grad_norm": 0.19478121399879456,
+      "learning_rate": 6.701959642809589e-06,
+      "loss": 0.7546,
+      "step": 1060
+    },
+    {
+      "epoch": 0.41197420348445474,
+      "grad_norm": 0.21342678368091583,
+      "learning_rate": 6.643194658312987e-06,
+      "loss": 0.744,
+      "step": 1070
+    },
+    {
+      "epoch": 0.415824429685244,
+      "grad_norm": 0.21068835258483887,
+      "learning_rate": 6.584174093857676e-06,
+      "loss": 0.7691,
+      "step": 1080
+    },
+    {
+      "epoch": 0.4196746558860333,
+      "grad_norm": 0.2301860898733139,
+      "learning_rate": 6.5249071294110735e-06,
+      "loss": 0.7538,
+      "step": 1090
+    },
+    {
+      "epoch": 0.4235248820868226,
+      "grad_norm": 0.21837878227233887,
+      "learning_rate": 6.465402983265274e-06,
+      "loss": 0.7676,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4273751082876119,
+      "grad_norm": 0.2188580334186554,
+      "learning_rate": 6.405670910603243e-06,
+      "loss": 0.7631,
+      "step": 1110
+    },
+    {
+      "epoch": 0.43122533448840117,
+      "grad_norm": 0.23615211248397827,
+      "learning_rate": 6.345720202059291e-06,
+      "loss": 0.759,
+      "step": 1120
+    },
+    {
+      "epoch": 0.4350755606891905,
+      "grad_norm": 0.21345657110214233,
+      "learning_rate": 6.285560182274008e-06,
+      "loss": 0.7363,
+      "step": 1130
+    },
+    {
+      "epoch": 0.4389257868899798,
+      "grad_norm": 0.20817747712135315,
+      "learning_rate": 6.225200208443936e-06,
+      "loss": 0.7589,
+      "step": 1140
+    },
+    {
+      "epoch": 0.4427760130907691,
+      "grad_norm": 0.20576463639736176,
+      "learning_rate": 6.16464966886616e-06,
+      "loss": 0.7482,
+      "step": 1150
+    },
+    {
+      "epoch": 0.44662623929155837,
+      "grad_norm": 0.25187379121780396,
+      "learning_rate": 6.103917981478065e-06,
+      "loss": 0.7732,
+      "step": 1160
+    },
+    {
+      "epoch": 0.45047646549234766,
+      "grad_norm": 0.21349893510341644,
+      "learning_rate": 6.043014592392489e-06,
+      "loss": 0.7846,
+      "step": 1170
+    },
+    {
+      "epoch": 0.454326691693137,
+      "grad_norm": 0.19843789935112,
+      "learning_rate": 5.981948974428485e-06,
+      "loss": 0.757,
+      "step": 1180
+    },
+    {
+      "epoch": 0.4581769178939263,
+      "grad_norm": 0.2122807502746582,
+      "learning_rate": 5.920730625637934e-06,
+      "loss": 0.7355,
+      "step": 1190
+    },
+    {
+      "epoch": 0.46202714409471557,
+      "grad_norm": 0.22178325057029724,
+      "learning_rate": 5.8593690678282355e-06,
+      "loss": 0.7497,
+      "step": 1200
+    },
+    {
+      "epoch": 0.46587737029550486,
+      "grad_norm": 0.21961438655853271,
+      "learning_rate": 5.797873845081301e-06,
+      "loss": 0.7524,
+      "step": 1210
+    },
+    {
+      "epoch": 0.46972759649629414,
+      "grad_norm": 0.21854643523693085,
+      "learning_rate": 5.736254522269075e-06,
+      "loss": 0.7678,
+      "step": 1220
+    },
+    {
+      "epoch": 0.47357782269708343,
+      "grad_norm": 0.20794209837913513,
+      "learning_rate": 5.6745206835658386e-06,
+      "loss": 0.7721,
+      "step": 1230
+    },
+    {
+      "epoch": 0.47742804889787277,
+      "grad_norm": 0.20201051235198975,
+      "learning_rate": 5.612681930957495e-06,
+      "loss": 0.7412,
+      "step": 1240
+    },
+    {
+      "epoch": 0.48127827509866206,
+      "grad_norm": 0.19846028089523315,
+      "learning_rate": 5.550747882748095e-06,
+      "loss": 0.7389,
+      "step": 1250
+    },
+    {
+      "epoch": 0.48512850129945134,
+      "grad_norm": 0.22576195001602173,
+      "learning_rate": 5.4887281720638195e-06,
+      "loss": 0.7491,
+      "step": 1260
+    },
+    {
+      "epoch": 0.48897872750024063,
+      "grad_norm": 0.23932309448719025,
+      "learning_rate": 5.42663244535466e-06,
+      "loss": 0.73,
+      "step": 1270
+    },
+    {
+      "epoch": 0.4928289537010299,
+      "grad_norm": 0.21619854867458344,
+      "learning_rate": 5.3644703608940205e-06,
+      "loss": 0.7397,
+      "step": 1280
+    },
+    {
+      "epoch": 0.49667917990181926,
+      "grad_norm": 0.22039608657360077,
+      "learning_rate": 5.302251587276491e-06,
+      "loss": 0.75,
+      "step": 1290
+    },
+    {
+      "epoch": 0.5005294061026085,
+      "grad_norm": 0.20602381229400635,
+      "learning_rate": 5.2399858019140005e-06,
+      "loss": 0.7526,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5043796323033978,
+      "grad_norm": 0.20981571078300476,
+      "learning_rate": 5.177682689530615e-06,
+      "loss": 0.7649,
+      "step": 1310
+    },
+    {
+      "epoch": 0.5082298585041871,
+      "grad_norm": 0.22198666632175446,
+      "learning_rate": 5.11535194065619e-06,
+      "loss": 0.7353,
+      "step": 1320
+    },
+    {
+      "epoch": 0.5120800847049765,
+      "grad_norm": 0.22688715159893036,
+      "learning_rate": 5.053003250119113e-06,
+      "loss": 0.7433,
+      "step": 1330
+    },
+    {
+      "epoch": 0.5159303109057657,
+      "grad_norm": 0.22114503383636475,
+      "learning_rate": 4.9906463155383925e-06,
+      "loss": 0.7508,
+      "step": 1340
+    },
+    {
+      "epoch": 0.519780537106555,
+      "grad_norm": 0.21413393318653107,
+      "learning_rate": 4.9282908358153046e-06,
+      "loss": 0.7459,
+      "step": 1350
+    },
+    {
+      "epoch": 0.5236307633073443,
+      "grad_norm": 0.24493268132209778,
+      "learning_rate": 4.8659465096248345e-06,
+      "loss": 0.7607,
+      "step": 1360
+    },
+    {
+      "epoch": 0.5274809895081336,
+      "grad_norm": 0.21578480303287506,
+      "learning_rate": 4.803623033907169e-06,
+      "loss": 0.7577,
+      "step": 1370
+    },
+    {
+      "epoch": 0.5313312157089229,
+      "grad_norm": 0.20171962678432465,
+      "learning_rate": 4.741330102359441e-06,
+      "loss": 0.7655,
+      "step": 1380
+    },
+    {
+      "epoch": 0.5351814419097122,
+      "grad_norm": 0.22375079989433289,
+      "learning_rate": 4.679077403927996e-06,
+      "loss": 0.7447,
+      "step": 1390
+    },
+    {
+      "epoch": 0.5390316681105015,
+      "grad_norm": 0.21156470477581024,
+      "learning_rate": 4.616874621301373e-06,
+      "loss": 0.7459,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5428818943112907,
+      "grad_norm": 0.22357401251792908,
+      "learning_rate": 4.554731429404293e-06,
+      "loss": 0.7598,
+      "step": 1410
+    },
+    {
+      "epoch": 0.5467321205120801,
+      "grad_norm": 0.22624486684799194,
+      "learning_rate": 4.492657493892822e-06,
+      "loss": 0.7426,
+      "step": 1420
+    },
+    {
+      "epoch": 0.5505823467128694,
+      "grad_norm": 0.22739337384700775,
+      "learning_rate": 4.430662469650999e-06,
+      "loss": 0.7596,
+      "step": 1430
+    },
+    {
+      "epoch": 0.5544325729136587,
+      "grad_norm": 0.2166316658258438,
+      "learning_rate": 4.368755999289118e-06,
+      "loss": 0.7642,
+      "step": 1440
+    },
+    {
+      "epoch": 0.558282799114448,
+      "grad_norm": 0.2227749079465866,
+      "learning_rate": 4.306947711643949e-06,
+      "loss": 0.744,
+      "step": 1450
+    },
+    {
+      "epoch": 0.5621330253152372,
+      "grad_norm": 0.22128525376319885,
+      "learning_rate": 4.245247220281068e-06,
+      "loss": 0.7509,
+      "step": 1460
+    },
+    {
+      "epoch": 0.5659832515160266,
+      "grad_norm": 0.20869411528110504,
+      "learning_rate": 4.183664121999587e-06,
+      "loss": 0.7455,
+      "step": 1470
+    },
+    {
+      "epoch": 0.5698334777168159,
+      "grad_norm": 0.22271610796451569,
+      "learning_rate": 4.122207995339474e-06,
+      "loss": 0.7439,
+      "step": 1480
+    },
+    {
+      "epoch": 0.5736837039176051,
+      "grad_norm": 0.20983220636844635,
+      "learning_rate": 4.060888399091734e-06,
+      "loss": 0.7487,
+      "step": 1490
+    },
+    {
+      "epoch": 0.5775339301183945,
+      "grad_norm": 0.2024894803762436,
+      "learning_rate": 3.999714870811639e-06,
+      "loss": 0.7597,
+      "step": 1500
+    },
+    {
+      "epoch": 0.5813841563191837,
+      "grad_norm": 0.2140243798494339,
+      "learning_rate": 3.938696925335275e-06,
+      "loss": 0.7522,
+      "step": 1510
+    },
+    {
+      "epoch": 0.5852343825199731,
+      "grad_norm": 0.20894376933574677,
+      "learning_rate": 3.87784405329962e-06,
+      "loss": 0.7599,
+      "step": 1520
+    },
+    {
+      "epoch": 0.5890846087207623,
+      "grad_norm": 0.2131212204694748,
+      "learning_rate": 3.817165719666382e-06,
+      "loss": 0.7601,
+      "step": 1530
+    },
+    {
+      "epoch": 0.5929348349215516,
+      "grad_norm": 0.2094799280166626,
+      "learning_rate": 3.756671362249841e-06,
+      "loss": 0.7672,
+      "step": 1540
+    },
+    {
+      "epoch": 0.596785061122341,
+      "grad_norm": 0.20797672867774963,
+      "learning_rate": 3.6963703902488907e-06,
+      "loss": 0.7527,
+      "step": 1550
+    },
+    {
+      "epoch": 0.6006352873231302,
+      "grad_norm": 0.21872657537460327,
+      "learning_rate": 3.6362721827835655e-06,
+      "loss": 0.766,
+      "step": 1560
+    },
+    {
+      "epoch": 0.6044855135239195,
+      "grad_norm": 0.22430601716041565,
+      "learning_rate": 3.576386087436212e-06,
+      "loss": 0.7639,
+      "step": 1570
+    },
+    {
+      "epoch": 0.6083357397247088,
+      "grad_norm": 0.23017176985740662,
+      "learning_rate": 3.5167214187975913e-06,
+      "loss": 0.7539,
+      "step": 1580
+    },
+    {
+      "epoch": 0.6121859659254981,
+      "grad_norm": 0.19031839072704315,
+      "learning_rate": 3.457287457018086e-06,
+      "loss": 0.7561,
+      "step": 1590
+    },
+    {
+      "epoch": 0.6160361921262875,
+      "grad_norm": 0.20836256444454193,
+      "learning_rate": 3.3980934463643003e-06,
+      "loss": 0.7449,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6198864183270767,
+      "grad_norm": 0.19944438338279724,
+      "learning_rate": 3.3391485937812053e-06,
+      "loss": 0.7265,
+      "step": 1610
+    },
+    {
+      "epoch": 0.623736644527866,
+      "grad_norm": 0.21762044727802277,
+      "learning_rate": 3.280462067460115e-06,
+      "loss": 0.7253,
+      "step": 1620
+    },
+    {
+      "epoch": 0.6275868707286553,
+      "grad_norm": 0.21123890578746796,
+      "learning_rate": 3.222042995412669e-06,
+      "loss": 0.7577,
+      "step": 1630
+    },
+    {
+      "epoch": 0.6314370969294446,
+      "grad_norm": 0.19006696343421936,
+      "learning_rate": 3.163900464051089e-06,
+      "loss": 0.7461,
+      "step": 1640
+    },
+    {
+      "epoch": 0.635287323130234,
+      "grad_norm": 0.21360695362091064,
+      "learning_rate": 3.1060435167748826e-06,
+      "loss": 0.7576,
+      "step": 1650
+    },
+    {
+      "epoch": 0.6391375493310232,
+      "grad_norm": 0.21674378216266632,
+      "learning_rate": 3.048481152564253e-06,
+      "loss": 0.7372,
+      "step": 1660
+    },
+    {
+      "epoch": 0.6429877755318125,
+      "grad_norm": 0.22534698247909546,
+      "learning_rate": 2.9912223245804017e-06,
+      "loss": 0.7584,
+      "step": 1670
+    },
+    {
+      "epoch": 0.6468380017326018,
+      "grad_norm": 0.2197372317314148,
+      "learning_rate": 2.934275938772981e-06,
+      "loss": 0.7521,
+      "step": 1680
+    },
+    {
+      "epoch": 0.6506882279333911,
+      "grad_norm": 0.2281721830368042,
+      "learning_rate": 2.8776508524948676e-06,
+      "loss": 0.743,
+      "step": 1690
+    },
+    {
+      "epoch": 0.6545384541341804,
+      "grad_norm": 0.19327642023563385,
+      "learning_rate": 2.821355873124503e-06,
+      "loss": 0.7419,
+      "step": 1700
+    },
+    {
+      "epoch": 0.6583886803349697,
+      "grad_norm": 0.2144935429096222,
+      "learning_rate": 2.7653997566960144e-06,
+      "loss": 0.7553,
+      "step": 1710
+    },
+    {
+      "epoch": 0.662238906535759,
+      "grad_norm": 0.2179148942232132,
+      "learning_rate": 2.709791206537309e-06,
+      "loss": 0.7443,
+      "step": 1720
+    },
+    {
+      "epoch": 0.6660891327365482,
+      "grad_norm": 0.21817097067832947,
+      "learning_rate": 2.6545388719163747e-06,
+      "loss": 0.7474,
+      "step": 1730
+    },
+    {
+      "epoch": 0.6699393589373376,
+      "grad_norm": 0.2147418111562729,
+      "learning_rate": 2.599651346695979e-06,
+      "loss": 0.7498,
+      "step": 1740
+    },
+    {
+      "epoch": 0.6737895851381268,
+      "grad_norm": 0.20563232898712158,
+      "learning_rate": 2.545137167997007e-06,
+      "loss": 0.7479,
+      "step": 1750
+    },
+    {
+      "epoch": 0.6776398113389162,
+      "grad_norm": 0.2030370831489563,
+      "learning_rate": 2.4910048148705972e-06,
+      "loss": 0.7362,
+      "step": 1760
+    },
+    {
+      "epoch": 0.6814900375397055,
+      "grad_norm": 0.2206178456544876,
+      "learning_rate": 2.4372627069793327e-06,
+      "loss": 0.7803,
+      "step": 1770
+    },
+    {
+      "epoch": 0.6853402637404947,
+      "grad_norm": 0.2246749848127365,
+      "learning_rate": 2.3839192032876486e-06,
+      "loss": 0.764,
+      "step": 1780
+    },
+    {
+      "epoch": 0.6891904899412841,
+      "grad_norm": 0.21290338039398193,
+      "learning_rate": 2.3309826007617047e-06,
+      "loss": 0.7756,
+      "step": 1790
+    },
+    {
+      "epoch": 0.6930407161420733,
+      "grad_norm": 0.21262115240097046,
+      "learning_rate": 2.278461133078881e-06,
+      "loss": 0.7386,
+      "step": 1800
+    },
+    {
+      "epoch": 0.6968909423428626,
+      "grad_norm": 0.2239065021276474,
+      "learning_rate": 2.2263629693471258e-06,
+      "loss": 0.7307,
+      "step": 1810
+    },
+    {
+      "epoch": 0.700741168543652,
+      "grad_norm": 0.2005116045475006,
+      "learning_rate": 2.1746962128343414e-06,
+      "loss": 0.752,
+      "step": 1820
+    },
+    {
+      "epoch": 0.7045913947444412,
+      "grad_norm": 0.2104838341474533,
+      "learning_rate": 2.1234688997080183e-06,
+      "loss": 0.7437,
+      "step": 1830
+    },
+    {
+      "epoch": 0.7084416209452306,
+      "grad_norm": 0.225631445646286,
+      "learning_rate": 2.0726889977853053e-06,
+      "loss": 0.749,
+      "step": 1840
+    },
+    {
+      "epoch": 0.7122918471460198,
+      "grad_norm": 0.22998039424419403,
+      "learning_rate": 2.022364405293703e-06,
+      "loss": 0.7605,
+      "step": 1850
+    },
+    {
+      "epoch": 0.7161420733468091,
+      "grad_norm": 0.211993008852005,
+      "learning_rate": 1.9725029496425814e-06,
+      "loss": 0.7552,
+      "step": 1860
+    },
+    {
+      "epoch": 0.7199922995475985,
+      "grad_norm": 0.1955261528491974,
+      "learning_rate": 1.923112386205727e-06,
+      "loss": 0.7493,
+      "step": 1870
+    },
+    {
+      "epoch": 0.7238425257483877,
+      "grad_norm": 0.20740123093128204,
+      "learning_rate": 1.874200397115079e-06,
+      "loss": 0.758,
+      "step": 1880
+    },
+    {
+      "epoch": 0.727692751949177,
+      "grad_norm": 0.2214876115322113,
+      "learning_rate": 1.825774590065857e-06,
+      "loss": 0.7406,
+      "step": 1890
+    },
+    {
+      "epoch": 0.7315429781499663,
+      "grad_norm": 0.22139890491962433,
+      "learning_rate": 1.7778424971332813e-06,
+      "loss": 0.7513,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7353932043507556,
+      "grad_norm": 0.20506468415260315,
+      "learning_rate": 1.7304115736010407e-06,
+      "loss": 0.748,
+      "step": 1910
+    },
+    {
+      "epoch": 0.739243430551545,
+      "grad_norm": 0.21489691734313965,
+      "learning_rate": 1.6834891968017108e-06,
+      "loss": 0.7434,
+      "step": 1920
+    },
+    {
+      "epoch": 0.7430936567523342,
+      "grad_norm": 0.21804192662239075,
+      "learning_rate": 1.6370826649692884e-06,
+      "loss": 0.758,
+      "step": 1930
+    },
+    {
+      "epoch": 0.7469438829531235,
+      "grad_norm": 0.22556059062480927,
+      "learning_rate": 1.5911991961040463e-06,
+      "loss": 0.7647,
+      "step": 1940
+    },
+    {
+      "epoch": 0.7507941091539128,
+      "grad_norm": 0.2177467942237854,
+      "learning_rate": 1.545845926849851e-06,
+      "loss": 0.7527,
+      "step": 1950
+    },
+    {
+      "epoch": 0.7546443353547021,
+      "grad_norm": 0.2318081110715866,
+      "learning_rate": 1.5010299113841397e-06,
+      "loss": 0.7697,
+      "step": 1960
+    },
+    {
+      "epoch": 0.7584945615554913,
+      "grad_norm": 0.21114282310009003,
+      "learning_rate": 1.4567581203207238e-06,
+      "loss": 0.76,
+      "step": 1970
+    },
+    {
+      "epoch": 0.7623447877562807,
+      "grad_norm": 0.22069686651229858,
+      "learning_rate": 1.413037439625593e-06,
+      "loss": 0.7576,
+      "step": 1980
+    },
+    {
+      "epoch": 0.76619501395707,
+      "grad_norm": 0.23069898784160614,
+      "learning_rate": 1.3698746695458847e-06,
+      "loss": 0.7528,
+      "step": 1990
+    },
+    {
+      "epoch": 0.7700452401578592,
+      "grad_norm": 0.21078741550445557,
+      "learning_rate": 1.3272765235521745e-06,
+      "loss": 0.745,
+      "step": 2000
+    },
+    {
+      "epoch": 0.7738954663586486,
+      "grad_norm": 0.21324925124645233,
+      "learning_rate": 1.2852496272942877e-06,
+      "loss": 0.7543,
+      "step": 2010
+    },
+    {
+      "epoch": 0.7777456925594378,
+      "grad_norm": 0.20018094778060913,
+      "learning_rate": 1.2438005175707396e-06,
+      "loss": 0.7401,
+      "step": 2020
+    },
+    {
+      "epoch": 0.7815959187602272,
+      "grad_norm": 0.20591725409030914,
+      "learning_rate": 1.2029356413120297e-06,
+      "loss": 0.725,
+      "step": 2030
+    },
+    {
+      "epoch": 0.7854461449610165,
+      "grad_norm": 0.21032363176345825,
+      "learning_rate": 1.1626613545778786e-06,
+      "loss": 0.7424,
+      "step": 2040
+    },
+    {
+      "epoch": 0.7892963711618057,
+      "grad_norm": 0.20533643662929535,
+      "learning_rate": 1.1229839215686294e-06,
+      "loss": 0.7466,
+      "step": 2050
+    },
+    {
+      "epoch": 0.7931465973625951,
+      "grad_norm": 0.1990252435207367,
+      "learning_rate": 1.0839095136509188e-06,
+      "loss": 0.7293,
+      "step": 2060
+    },
+    {
+      "epoch": 0.7969968235633843,
+      "grad_norm": 0.22310616075992584,
+      "learning_rate": 1.045444208397791e-06,
+      "loss": 0.7521,
+      "step": 2070
+    },
+    {
+      "epoch": 0.8008470497641736,
+      "grad_norm": 0.20535145699977875,
+      "learning_rate": 1.007593988643402e-06,
+      "loss": 0.7369,
+      "step": 2080
+    },
+    {
+      "epoch": 0.804697275964963,
+      "grad_norm": 0.20860625803470612,
+      "learning_rate": 9.703647415524613e-07,
+      "loss": 0.7583,
+      "step": 2090
+    },
+    {
+      "epoch": 0.8085475021657522,
+      "grad_norm": 0.21413582563400269,
+      "learning_rate": 9.337622577045485e-07,
+      "loss": 0.7459,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8123977283665416,
+      "grad_norm": 0.2125559151172638,
+      "learning_rate": 8.977922301934588e-07,
+      "loss": 0.7557,
+      "step": 2110
+    },
+    {
+      "epoch": 0.8162479545673308,
+      "grad_norm": 0.2141505628824234,
+      "learning_rate": 8.624602537417009e-07,
+      "loss": 0.7571,
+      "step": 2120
+    },
+    {
+      "epoch": 0.8200981807681201,
+      "grad_norm": 0.22742731869220734,
+      "learning_rate": 8.277718238303123e-07,
+      "loss": 0.7421,
+      "step": 2130
+    },
+    {
+      "epoch": 0.8239484069689095,
+      "grad_norm": 0.21833297610282898,
+      "learning_rate": 7.937323358440935e-07,
+      "loss": 0.7473,
+      "step": 2140
+    },
+    {
+      "epoch": 0.8277986331696987,
+      "grad_norm": 0.22948749363422394,
+      "learning_rate": 7.603470842324223e-07,
+      "loss": 0.7374,
+      "step": 2150
+    },
+    {
+      "epoch": 0.831648859370488,
+      "grad_norm": 0.20834681391716003,
+      "learning_rate": 7.276212616857575e-07,
+      "loss": 0.7559,
+      "step": 2160
+    },
+    {
+      "epoch": 0.8354990855712773,
+      "grad_norm": 0.20966140925884247,
+      "learning_rate": 6.955599583279837e-07,
+      "loss": 0.7555,
+      "step": 2170
+    },
+    {
+      "epoch": 0.8393493117720666,
+      "grad_norm": 0.2091865837574005,
+      "learning_rate": 6.641681609246981e-07,
+      "loss": 0.7479,
+      "step": 2180
+    },
+    {
+      "epoch": 0.8431995379728559,
+      "grad_norm": 0.20802880823612213,
+      "learning_rate": 6.334507521075711e-07,
+      "loss": 0.7543,
+      "step": 2190
+    },
+    {
+      "epoch": 0.8470497641736452,
+      "grad_norm": 0.22034654021263123,
+      "learning_rate": 6.034125096149179e-07,
+      "loss": 0.7503,
+      "step": 2200
+    },
+    {
+      "epoch": 0.8508999903744345,
+      "grad_norm": 0.2159205973148346,
+      "learning_rate": 5.740581055485722e-07,
+      "loss": 0.7204,
+      "step": 2210
+    },
+    {
+      "epoch": 0.8547502165752238,
+      "grad_norm": 0.1973789930343628,
+      "learning_rate": 5.453921056471961e-07,
+      "loss": 0.7528,
+      "step": 2220
+    },
+    {
+      "epoch": 0.8586004427760131,
+      "grad_norm": 0.21413974463939667,
+      "learning_rate": 5.174189685761288e-07,
+      "loss": 0.75,
+      "step": 2230
+    },
+    {
+      "epoch": 0.8624506689768023,
+      "grad_norm": 0.19621741771697998,
+      "learning_rate": 4.901430452338979e-07,
+      "loss": 0.7605,
+      "step": 2240
+    },
+    {
+      "epoch": 0.8663008951775917,
+      "grad_norm": 0.20545727014541626,
+      "learning_rate": 4.6356857807548286e-07,
+      "loss": 0.744,
+      "step": 2250
+    },
+    {
+      "epoch": 0.870151121378381,
+      "grad_norm": 0.20863088965415955,
+      "learning_rate": 4.376997004524541e-07,
+      "loss": 0.7337,
+      "step": 2260
+    },
+    {
+      "epoch": 0.8740013475791703,
+      "grad_norm": 0.2112920582294464,
+      "learning_rate": 4.12540435970073e-07,
+      "loss": 0.7408,
+      "step": 2270
+    },
+    {
+      "epoch": 0.8778515737799596,
+      "grad_norm": 0.21115165948867798,
+      "learning_rate": 3.880946978614703e-07,
+      "loss": 0.7541,
+      "step": 2280
+    },
+    {
+      "epoch": 0.8817017999807488,
+      "grad_norm": 0.21081185340881348,
+      "learning_rate": 3.643662883789878e-07,
+      "loss": 0.7558,
+      "step": 2290
+    },
+    {
+      "epoch": 0.8855520261815382,
+      "grad_norm": 0.19912992417812347,
+      "learning_rate": 3.41358898202781e-07,
+      "loss": 0.7357,
+      "step": 2300
+    },
+    {
+      "epoch": 0.8894022523823275,
+      "grad_norm": 0.1971040517091751,
+      "learning_rate": 3.190761058667741e-07,
+      "loss": 0.7474,
+      "step": 2310
+    },
+    {
+      "epoch": 0.8932524785831167,
+      "grad_norm": 0.20738078653812408,
+      "learning_rate": 2.9752137720206463e-07,
+      "loss": 0.7499,
+      "step": 2320
+    },
+    {
+      "epoch": 0.8971027047839061,
+      "grad_norm": 0.20806998014450073,
+      "learning_rate": 2.7669806479785067e-07,
+      "loss": 0.7557,
+      "step": 2330
+    },
+    {
+      "epoch": 0.9009529309846953,
+      "grad_norm": 0.2086111307144165,
+      "learning_rate": 2.566094074799763e-07,
+      "loss": 0.7484,
+      "step": 2340
+    },
+    {
+      "epoch": 0.9048031571854847,
+      "grad_norm": 0.21011394262313843,
+      "learning_rate": 2.3725852980716557e-07,
+      "loss": 0.7355,
+      "step": 2350
+    },
+    {
+      "epoch": 0.908653383386274,
+      "grad_norm": 0.20799727737903595,
+      "learning_rate": 2.1864844158503983e-07,
+      "loss": 0.745,
+      "step": 2360
+    },
+    {
+      "epoch": 0.9125036095870632,
+      "grad_norm": 0.22406388819217682,
+      "learning_rate": 2.007820373979702e-07,
+      "loss": 0.74,
+      "step": 2370
+    },
+    {
+      "epoch": 0.9163538357878526,
+      "grad_norm": 0.22140587866306305,
+      "learning_rate": 1.8366209615886532e-07,
+      "loss": 0.7479,
+      "step": 2380
+    },
+    {
+      "epoch": 0.9202040619886418,
+      "grad_norm": 0.21085955202579498,
+      "learning_rate": 1.6729128067693613e-07,
+      "loss": 0.7609,
+      "step": 2390
+    },
+    {
+      "epoch": 0.9240542881894311,
+      "grad_norm": 0.22242023050785065,
+      "learning_rate": 1.5167213724353426e-07,
+      "loss": 0.7597,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9279045143902205,
+      "grad_norm": 0.20664802193641663,
+      "learning_rate": 1.3680709523610058e-07,
+      "loss": 0.7402,
+      "step": 2410
+    },
+    {
+      "epoch": 0.9317547405910097,
+      "grad_norm": 0.20922450721263885,
+      "learning_rate": 1.2269846674030418e-07,
+      "loss": 0.7643,
+      "step": 2420
+    },
+    {
+      "epoch": 0.935604966791799,
+      "grad_norm": 0.22148653864860535,
+      "learning_rate": 1.0934844619042595e-07,
+      "loss": 0.7591,
+      "step": 2430
+    },
+    {
+      "epoch": 0.9394551929925883,
+      "grad_norm": 0.22239552438259125,
+      "learning_rate": 9.675911002803961e-08,
+      "loss": 0.7471,
+      "step": 2440
+    },
+    {
+      "epoch": 0.9433054191933776,
+      "grad_norm": 0.22638718783855438,
+      "learning_rate": 8.493241637904381e-08,
+      "loss": 0.745,
+      "step": 2450
+    },
+    {
+      "epoch": 0.9471556453941669,
+      "grad_norm": 0.21188881993293762,
+      "learning_rate": 7.387020474909689e-08,
+      "loss": 0.7652,
+      "step": 2460
+    },
+    {
+      "epoch": 0.9510058715949562,
+      "grad_norm": 0.2250465750694275,
+      "learning_rate": 6.357419573750856e-08,
+      "loss": 0.7597,
+      "step": 2470
+    },
+    {
+      "epoch": 0.9548560977957455,
+      "grad_norm": 0.2109050154685974,
+      "learning_rate": 5.4045990769613365e-08,
+      "loss": 0.7403,
+      "step": 2480
+    },
+    {
+      "epoch": 0.9587063239965348,
+      "grad_norm": 0.21316084265708923,
+      "learning_rate": 4.5287071847694434e-08,
+      "loss": 0.7474,
+      "step": 2490
+    },
+    {
+      "epoch": 0.9625565501973241,
+      "grad_norm": 0.21073268353939056,
+      "learning_rate": 3.7298801320467817e-08,
+      "loss": 0.7454,
+      "step": 2500
+    },
+    {
+      "epoch": 0.9664067763981133,
+      "grad_norm": 0.211844801902771,
+      "learning_rate": 3.0082421671192576e-08,
+      "loss": 0.7386,
+      "step": 2510
+    },
+    {
+      "epoch": 0.9702570025989027,
+      "grad_norm": 0.20792436599731445,
+      "learning_rate": 2.3639055324410355e-08,
+      "loss": 0.759,
+      "step": 2520
+    },
+    {
+      "epoch": 0.974107228799692,
+      "grad_norm": 0.20976923406124115,
+      "learning_rate": 1.7969704471369477e-08,
+      "loss": 0.7441,
+      "step": 2530
+    },
+    {
+      "epoch": 0.9779574550004813,
+      "grad_norm": 0.2018246352672577,
+      "learning_rate": 1.3075250914142413e-08,
+      "loss": 0.7452,
+      "step": 2540
+    },
+    {
+      "epoch": 0.9818076812012706,
+      "grad_norm": 0.20044636726379395,
+      "learning_rate": 8.956455928476049e-09,
+      "loss": 0.7534,
+      "step": 2550
+    },
+    {
+      "epoch": 0.9856579074020598,
+      "grad_norm": 0.20577794313430786,
+      "learning_rate": 5.6139601453786275e-09,
+      "loss": 0.7456,
+      "step": 2560
+    },
+    {
+      "epoch": 0.9895081336028492,
+      "grad_norm": 0.20805469155311584,
+      "learning_rate": 3.0482834514805602e-09,
+      "loss": 0.7368,
+      "step": 2570
+    },
+    {
+      "epoch": 0.9933583598036385,
+      "grad_norm": 0.2096133530139923,
+      "learning_rate": 1.2598249081713366e-09,
+      "loss": 0.7275,
+      "step": 2580
+    },
+    {
+      "epoch": 0.9972085860044277,
+      "grad_norm": 0.21441590785980225,
+      "learning_rate": 2.488626895302826e-10,
+      "loss": 0.7614,
+      "step": 2590
+    },
+    {
+      "epoch": 0.9999037443449803,
+      "step": 2597,
+      "total_flos": 3.5185741077445345e+18,
+      "train_loss": 0.769550015020242,
+      "train_runtime": 8144.8911,
+      "train_samples_per_second": 81.634,
+      "train_steps_per_second": 0.319
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2597,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.5185741077445345e+18,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/mc1_b16_224_400m/finetune/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1771a043fe2f930e4cd759d38396f06c476a9f1e04e5631eae1a148a1bccfb07
+size 6136

continuous/qwen25/mc1_b16_224_400m/pretrain/config.json

@@ -0,0 +1,58 @@
+{
+  "_attn_implementation_autoset": true,
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "image_aspect_ratio": "square",
+  "image_crop_resolution": null,
+  "image_grid_pinpoints": null,
+  "image_split_resolution": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 768,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": null,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "mm_tunable_parts": "mm_mlp_adapter",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "open_clip_hub:ViT-B-16",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/mc1/mc1_b16_224_400m.pt",
+  "vocab_size": 151936
+}

continuous/qwen25/mc1_b16_224_400m/pretrain/mm_projector.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7be903e6535dd84bc37bb92cc82daa2cc64ad5591d7c489204c4dda4f1a29fdc
+size 14170168

continuous/qwen25/mc1_b16_224_400m/pretrain/trainer_state.json

@@ -0,0 +1,421 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 545,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01834862385321101,
+      "grad_norm": 2.060581922531128,
+      "learning_rate": 0.00032142857142857147,
+      "loss": 4.7295,
+      "step": 10
+    },
+    {
+      "epoch": 0.03669724770642202,
+      "grad_norm": 3.511321783065796,
+      "learning_rate": 0.0006785714285714287,
+      "loss": 3.8069,
+      "step": 20
+    },
+    {
+      "epoch": 0.05504587155963303,
+      "grad_norm": 0.475099116563797,
+      "learning_rate": 0.0009999907688176173,
+      "loss": 3.3457,
+      "step": 30
+    },
+    {
+      "epoch": 0.07339449541284404,
+      "grad_norm": 0.19818420708179474,
+      "learning_rate": 0.0009988834393115766,
+      "loss": 3.2291,
+      "step": 40
+    },
+    {
+      "epoch": 0.09174311926605505,
+      "grad_norm": 0.30398520827293396,
+      "learning_rate": 0.0009959345572940772,
+      "loss": 3.1975,
+      "step": 50
+    },
+    {
+      "epoch": 0.11009174311926606,
+      "grad_norm": 0.3195788860321045,
+      "learning_rate": 0.0009911550081155983,
+      "loss": 3.1536,
+      "step": 60
+    },
+    {
+      "epoch": 0.12844036697247707,
+      "grad_norm": 0.2092699259519577,
+      "learning_rate": 0.0009845624347566433,
+      "loss": 3.091,
+      "step": 70
+    },
+    {
+      "epoch": 0.14678899082568808,
+      "grad_norm": 0.22881554067134857,
+      "learning_rate": 0.0009761811727013547,
+      "loss": 3.0406,
+      "step": 80
+    },
+    {
+      "epoch": 0.1651376146788991,
+      "grad_norm": 0.2676897644996643,
+      "learning_rate": 0.0009660421601067666,
+      "loss": 2.987,
+      "step": 90
+    },
+    {
+      "epoch": 0.1834862385321101,
+      "grad_norm": 0.306372731924057,
+      "learning_rate": 0.0009541828235992926,
+      "loss": 2.9451,
+      "step": 100
+    },
+    {
+      "epoch": 0.2018348623853211,
+      "grad_norm": 0.3045768439769745,
+      "learning_rate": 0.0009406469401200151,
+      "loss": 2.8909,
+      "step": 110
+    },
+    {
+      "epoch": 0.22018348623853212,
+      "grad_norm": 0.25697216391563416,
+      "learning_rate": 0.0009254844753287492,
+      "loss": 2.8583,
+      "step": 120
+    },
+    {
+      "epoch": 0.23853211009174313,
+      "grad_norm": 0.44891321659088135,
+      "learning_rate": 0.0009087513991633924,
+      "loss": 2.8129,
+      "step": 130
+    },
+    {
+      "epoch": 0.25688073394495414,
+      "grad_norm": 0.27873218059539795,
+      "learning_rate": 0.0008905094792353917,
+      "loss": 2.7921,
+      "step": 140
+    },
+    {
+      "epoch": 0.27522935779816515,
+      "grad_norm": 0.2992289364337921,
+      "learning_rate": 0.0008708260528239789,
+      "loss": 2.7758,
+      "step": 150
+    },
+    {
+      "epoch": 0.29357798165137616,
+      "grad_norm": 0.355535089969635,
+      "learning_rate": 0.0008497737783108237,
+      "loss": 2.7535,
+      "step": 160
+    },
+    {
+      "epoch": 0.3119266055045872,
+      "grad_norm": 0.22377924621105194,
+      "learning_rate": 0.0008274303669726426,
+      "loss": 2.7417,
+      "step": 170
+    },
+    {
+      "epoch": 0.3302752293577982,
+      "grad_norm": 0.3818341791629791,
+      "learning_rate": 0.0008038782961218137,
+      "loss": 2.7345,
+      "step": 180
+    },
+    {
+      "epoch": 0.3486238532110092,
+      "grad_norm": 0.2755023241043091,
+      "learning_rate": 0.000779204504653894,
+      "loss": 2.7249,
+      "step": 190
+    },
+    {
+      "epoch": 0.3669724770642202,
+      "grad_norm": 0.2485969364643097,
+      "learning_rate": 0.0007535000721258791,
+      "loss": 2.7026,
+      "step": 200
+    },
+    {
+      "epoch": 0.3853211009174312,
+      "grad_norm": 0.23310555517673492,
+      "learning_rate": 0.00072685988254984,
+      "loss": 2.6966,
+      "step": 210
+    },
+    {
+      "epoch": 0.4036697247706422,
+      "grad_norm": 0.4637734591960907,
+      "learning_rate": 0.0006993822741429906,
+      "loss": 2.7047,
+      "step": 220
+    },
+    {
+      "epoch": 0.42201834862385323,
+      "grad_norm": 0.2551236152648926,
+      "learning_rate": 0.0006711686763270818,
+      "loss": 2.6943,
+      "step": 230
+    },
+    {
+      "epoch": 0.44036697247706424,
+      "grad_norm": 0.25037717819213867,
+      "learning_rate": 0.0006423232353170831,
+      "loss": 2.6835,
+      "step": 240
+    },
+    {
+      "epoch": 0.45871559633027525,
+      "grad_norm": 0.2170170694589615,
+      "learning_rate": 0.0006129524296812335,
+      "loss": 2.6822,
+      "step": 250
+    },
+    {
+      "epoch": 0.47706422018348627,
+      "grad_norm": 0.19993917644023895,
+      "learning_rate": 0.0005831646772915651,
+      "loss": 2.6901,
+      "step": 260
+    },
+    {
+      "epoch": 0.4954128440366973,
+      "grad_norm": 0.33400264382362366,
+      "learning_rate": 0.0005530699351157809,
+      "loss": 2.6679,
+      "step": 270
+    },
+    {
+      "epoch": 0.5137614678899083,
+      "grad_norm": 0.24667854607105255,
+      "learning_rate": 0.0005227792933277943,
+      "loss": 2.6724,
+      "step": 280
+    },
+    {
+      "epoch": 0.5321100917431193,
+      "grad_norm": 0.24577048420906067,
+      "learning_rate": 0.0004924045652352092,
+      "loss": 2.6568,
+      "step": 290
+    },
+    {
+      "epoch": 0.5504587155963303,
+      "grad_norm": 0.2230076938867569,
+      "learning_rate": 0.0004620578745374633,
+      "loss": 2.658,
+      "step": 300
+    },
+    {
+      "epoch": 0.5688073394495413,
+      "grad_norm": 0.3173806965351105,
+      "learning_rate": 0.0004318512414382059,
+      "loss": 2.6638,
+      "step": 310
+    },
+    {
+      "epoch": 0.5871559633027523,
+      "grad_norm": 0.2444869726896286,
+      "learning_rate": 0.0004018961691397148,
+      "loss": 2.6624,
+      "step": 320
+    },
+    {
+      "epoch": 0.6055045871559633,
+      "grad_norm": 0.33505338430404663,
+      "learning_rate": 0.0003723032322457458,
+      "loss": 2.6475,
+      "step": 330
+    },
+    {
+      "epoch": 0.6238532110091743,
+      "grad_norm": 0.4205361008644104,
+      "learning_rate": 0.0003431816685921625,
+      "loss": 2.6518,
+      "step": 340
+    },
+    {
+      "epoch": 0.6422018348623854,
+      "grad_norm": 0.2955787479877472,
+      "learning_rate": 0.0003146389760120416,
+      "loss": 2.6497,
+      "step": 350
+    },
+    {
+      "epoch": 0.6605504587155964,
+      "grad_norm": 0.2778743505477905,
+      "learning_rate": 0.00028678051552373485,
+      "loss": 2.653,
+      "step": 360
+    },
+    {
+      "epoch": 0.6788990825688074,
+      "grad_norm": 0.2387525737285614,
+      "learning_rate": 0.0002597091224066581,
+      "loss": 2.6441,
+      "step": 370
+    },
+    {
+      "epoch": 0.6972477064220184,
+      "grad_norm": 0.19172193109989166,
+      "learning_rate": 0.0002335247266004629,
+      "loss": 2.6513,
+      "step": 380
+    },
+    {
+      "epoch": 0.7155963302752294,
+      "grad_norm": 0.21115437150001526,
+      "learning_rate": 0.000208323983828828,
+      "loss": 2.6432,
+      "step": 390
+    },
+    {
+      "epoch": 0.7339449541284404,
+      "grad_norm": 0.2155393809080124,
+      "learning_rate": 0.0001841999188095224,
+      "loss": 2.6341,
+      "step": 400
+    },
+    {
+      "epoch": 0.7522935779816514,
+      "grad_norm": 0.23280367255210876,
+      "learning_rate": 0.00016124158186777676,
+      "loss": 2.6396,
+      "step": 410
+    },
+    {
+      "epoch": 0.7706422018348624,
+      "grad_norm": 0.24301818013191223,
+      "learning_rate": 0.0001395337202205211,
+      "loss": 2.636,
+      "step": 420
+    },
+    {
+      "epoch": 0.7889908256880734,
+      "grad_norm": 0.17611807584762573,
+      "learning_rate": 0.00011915646514489292,
+      "loss": 2.6319,
+      "step": 430
+    },
+    {
+      "epoch": 0.8073394495412844,
+      "grad_norm": 0.24760229885578156,
+      "learning_rate": 0.00010018503618578818,
+      "loss": 2.6455,
+      "step": 440
+    },
+    {
+      "epoch": 0.8256880733944955,
+      "grad_norm": 0.2205040007829666,
+      "learning_rate": 8.268946349432583e-05,
+      "loss": 2.6353,
+      "step": 450
+    },
+    {
+      "epoch": 0.8440366972477065,
+      "grad_norm": 0.18204469978809357,
+      "learning_rate": 6.673432932217117e-05,
+      "loss": 2.6308,
+      "step": 460
+    },
+    {
+      "epoch": 0.8623853211009175,
+      "grad_norm": 0.20964586734771729,
+      "learning_rate": 5.2378529625954684e-05,
+      "loss": 2.6207,
+      "step": 470
+    },
+    {
+      "epoch": 0.8807339449541285,
+      "grad_norm": 0.23032797873020172,
+      "learning_rate": 3.9675056661785556e-05,
+      "loss": 2.6237,
+      "step": 480
+    },
+    {
+      "epoch": 0.8990825688073395,
+      "grad_norm": 0.2072218656539917,
+      "learning_rate": 2.8670803372379538e-05,
+      "loss": 2.6332,
+      "step": 490
+    },
+    {
+      "epoch": 0.9174311926605505,
+      "grad_norm": 0.17369939386844635,
+      "learning_rate": 1.9406390288876586e-05,
+      "loss": 2.6396,
+      "step": 500
+    },
+    {
+      "epoch": 0.9357798165137615,
+      "grad_norm": 0.18111294507980347,
+      "learning_rate": 1.191601558631461e-05,
+      "loss": 2.6283,
+      "step": 510
+    },
+    {
+      "epoch": 0.9541284403669725,
+      "grad_norm": 0.19162194430828094,
+      "learning_rate": 6.2273288462559305e-06,
+      "loss": 2.6269,
+      "step": 520
+    },
+    {
+      "epoch": 0.9724770642201835,
+      "grad_norm": 0.18139635026454926,
+      "learning_rate": 2.361328992552314e-06,
+      "loss": 2.628,
+      "step": 530
+    },
+    {
+      "epoch": 0.9908256880733946,
+      "grad_norm": 0.17894083261489868,
+      "learning_rate": 3.322867770044202e-07,
+      "loss": 2.6238,
+      "step": 540
+    },
+    {
+      "epoch": 1.0,
+      "step": 545,
+      "total_flos": 2.8888413781989786e+17,
+      "train_loss": 2.8033320698169395,
+      "train_runtime": 5166.9043,
+      "train_samples_per_second": 108.02,
+      "train_steps_per_second": 0.105
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 545,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.8888413781989786e+17,
+  "train_batch_size": 128,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/mc1_b32_224_2.5b/finetune/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

continuous/qwen25/mc1_b32_224_2.5b/finetune/config.json

@@ -0,0 +1,57 @@
+{
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "LlavaQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "image_aspect_ratio": "square",
+  "image_crop_resolution": null,
+  "image_grid_pinpoints": null,
+  "image_split_resolution": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 768,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": 0.0001,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "open_clip_hub:ViT-B-32",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/mc1/mc1_b32_224_2.5b.pt",
+  "vocab_size": 151936
+}

continuous/qwen25/mc1_b32_224_2.5b/finetune/generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.51.3"
+}

continuous/qwen25/mc1_b32_224_2.5b/finetune/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8f0a85627293b1017b1f64df00db505fe4bb3b63cc701496825d3795b9f6b27
+size 3744103520

continuous/qwen25/mc1_b32_224_2.5b/finetune/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
+  }
+}

continuous/qwen25/mc1_b32_224_2.5b/finetune/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

continuous/qwen25/mc1_b32_224_2.5b/finetune/tokenizer_config.json

@@ -0,0 +1,209 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

continuous/qwen25/mc1_b32_224_2.5b/finetune/trainer_state.json

@@ -0,0 +1,1856 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9999037443449803,
+  "eval_steps": 500,
+  "global_step": 2597,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.003850226200789296,
+      "grad_norm": 14.627513885498047,
+      "learning_rate": 1.153846153846154e-06,
+      "loss": 1.8425,
+      "step": 10
+    },
+    {
+      "epoch": 0.007700452401578592,
+      "grad_norm": 14.040750503540039,
+      "learning_rate": 2.435897435897436e-06,
+      "loss": 1.5766,
+      "step": 20
+    },
+    {
+      "epoch": 0.01155067860236789,
+      "grad_norm": 3.52751088142395,
+      "learning_rate": 3.7179487179487184e-06,
+      "loss": 1.1304,
+      "step": 30
+    },
+    {
+      "epoch": 0.015400904803157185,
+      "grad_norm": 0.6807440519332886,
+      "learning_rate": 5e-06,
+      "loss": 0.9649,
+      "step": 40
+    },
+    {
+      "epoch": 0.019251131003946483,
+      "grad_norm": 0.37857431173324585,
+      "learning_rate": 6.282051282051282e-06,
+      "loss": 0.9336,
+      "step": 50
+    },
+    {
+      "epoch": 0.02310135720473578,
+      "grad_norm": 0.281938761472702,
+      "learning_rate": 7.564102564102564e-06,
+      "loss": 0.8892,
+      "step": 60
+    },
+    {
+      "epoch": 0.026951583405525073,
+      "grad_norm": 0.22641639411449432,
+      "learning_rate": 8.846153846153847e-06,
+      "loss": 0.8626,
+      "step": 70
+    },
+    {
+      "epoch": 0.03080180960631437,
+      "grad_norm": 0.2380041778087616,
+      "learning_rate": 9.999996111488723e-06,
+      "loss": 0.8713,
+      "step": 80
+    },
+    {
+      "epoch": 0.034652035807103666,
+      "grad_norm": 0.2176685780286789,
+      "learning_rate": 9.999529497453782e-06,
+      "loss": 0.8515,
+      "step": 90
+    },
+    {
+      "epoch": 0.038502262007892966,
+      "grad_norm": 0.2369755357503891,
+      "learning_rate": 9.998285264324193e-06,
+      "loss": 0.8322,
+      "step": 100
+    },
+    {
+      "epoch": 0.04235248820868226,
+      "grad_norm": 0.20782645046710968,
+      "learning_rate": 9.996263605626052e-06,
+      "loss": 0.821,
+      "step": 110
+    },
+    {
+      "epoch": 0.04620271440947156,
+      "grad_norm": 0.2300320714712143,
+      "learning_rate": 9.99346483580503e-06,
+      "loss": 0.8331,
+      "step": 120
+    },
+    {
+      "epoch": 0.05005294061026085,
+      "grad_norm": 0.2058149129152298,
+      "learning_rate": 9.989889390177467e-06,
+      "loss": 0.8254,
+      "step": 130
+    },
+    {
+      "epoch": 0.053903166811050146,
+      "grad_norm": 0.23310142755508423,
+      "learning_rate": 9.985537824862647e-06,
+      "loss": 0.8253,
+      "step": 140
+    },
+    {
+      "epoch": 0.057753393011839446,
+      "grad_norm": 0.2412598431110382,
+      "learning_rate": 9.980410816696317e-06,
+      "loss": 0.8075,
+      "step": 150
+    },
+    {
+      "epoch": 0.06160361921262874,
+      "grad_norm": 0.225943461060524,
+      "learning_rate": 9.97450916312541e-06,
+      "loss": 0.8147,
+      "step": 160
+    },
+    {
+      "epoch": 0.06545384541341803,
+      "grad_norm": 0.19975495338439941,
+      "learning_rate": 9.967833782084004e-06,
+      "loss": 0.8159,
+      "step": 170
+    },
+    {
+      "epoch": 0.06930407161420733,
+      "grad_norm": 0.22920268774032593,
+      "learning_rate": 9.960385711850554e-06,
+      "loss": 0.843,
+      "step": 180
+    },
+    {
+      "epoch": 0.07315429781499663,
+      "grad_norm": 0.22300903499126434,
+      "learning_rate": 9.952166110886394e-06,
+      "loss": 0.8331,
+      "step": 190
+    },
+    {
+      "epoch": 0.07700452401578593,
+      "grad_norm": 0.24528059363365173,
+      "learning_rate": 9.943176257655567e-06,
+      "loss": 0.8279,
+      "step": 200
+    },
+    {
+      "epoch": 0.08085475021657522,
+      "grad_norm": 0.22386205196380615,
+      "learning_rate": 9.933417550425952e-06,
+      "loss": 0.8169,
+      "step": 210
+    },
+    {
+      "epoch": 0.08470497641736452,
+      "grad_norm": 0.226539745926857,
+      "learning_rate": 9.9228915070518e-06,
+      "loss": 0.8065,
+      "step": 220
+    },
+    {
+      "epoch": 0.08855520261815382,
+      "grad_norm": 0.21960827708244324,
+      "learning_rate": 9.911599764737632e-06,
+      "loss": 0.809,
+      "step": 230
+    },
+    {
+      "epoch": 0.09240542881894312,
+      "grad_norm": 0.2240631878376007,
+      "learning_rate": 9.899544079783606e-06,
+      "loss": 0.8027,
+      "step": 240
+    },
+    {
+      "epoch": 0.0962556550197324,
+      "grad_norm": 0.2270190268754959,
+      "learning_rate": 9.886726327312338e-06,
+      "loss": 0.8169,
+      "step": 250
+    },
+    {
+      "epoch": 0.1001058812205217,
+      "grad_norm": 0.22193707525730133,
+      "learning_rate": 9.87314850097724e-06,
+      "loss": 0.8173,
+      "step": 260
+    },
+    {
+      "epoch": 0.103956107421311,
+      "grad_norm": 0.21485374867916107,
+      "learning_rate": 9.858812712652452e-06,
+      "loss": 0.8061,
+      "step": 270
+    },
+    {
+      "epoch": 0.10780633362210029,
+      "grad_norm": 0.20791712403297424,
+      "learning_rate": 9.843721192104341e-06,
+      "loss": 0.8062,
+      "step": 280
+    },
+    {
+      "epoch": 0.11165655982288959,
+      "grad_norm": 0.22872333228588104,
+      "learning_rate": 9.827876286644698e-06,
+      "loss": 0.8255,
+      "step": 290
+    },
+    {
+      "epoch": 0.11550678602367889,
+      "grad_norm": 0.20580945909023285,
+      "learning_rate": 9.811280460765643e-06,
+      "loss": 0.7857,
+      "step": 300
+    },
+    {
+      "epoch": 0.11935701222446819,
+      "grad_norm": 0.21814164519309998,
+      "learning_rate": 9.793936295756292e-06,
+      "loss": 0.796,
+      "step": 310
+    },
+    {
+      "epoch": 0.12320723842525748,
+      "grad_norm": 0.22327467799186707,
+      "learning_rate": 9.775846489301279e-06,
+      "loss": 0.8,
+      "step": 320
+    },
+    {
+      "epoch": 0.12705746462604678,
+      "grad_norm": 0.22431468963623047,
+      "learning_rate": 9.757013855061157e-06,
+      "loss": 0.7982,
+      "step": 330
+    },
+    {
+      "epoch": 0.13090769082683607,
+      "grad_norm": 0.219771608710289,
+      "learning_rate": 9.737441322234758e-06,
+      "loss": 0.8117,
+      "step": 340
+    },
+    {
+      "epoch": 0.13475791702762538,
+      "grad_norm": 0.22410699725151062,
+      "learning_rate": 9.717131935103605e-06,
+      "loss": 0.8096,
+      "step": 350
+    },
+    {
+      "epoch": 0.13860814322841467,
+      "grad_norm": 0.21965165436267853,
+      "learning_rate": 9.696088852558398e-06,
+      "loss": 0.7962,
+      "step": 360
+    },
+    {
+      "epoch": 0.14245836942920398,
+      "grad_norm": 0.21684831380844116,
+      "learning_rate": 9.674315347607686e-06,
+      "loss": 0.8213,
+      "step": 370
+    },
+    {
+      "epoch": 0.14630859562999327,
+      "grad_norm": 0.21782830357551575,
+      "learning_rate": 9.651814806868796e-06,
+      "loss": 0.7864,
+      "step": 380
+    },
+    {
+      "epoch": 0.15015882183078255,
+      "grad_norm": 0.23439307510852814,
+      "learning_rate": 9.62859073004107e-06,
+      "loss": 0.7915,
+      "step": 390
+    },
+    {
+      "epoch": 0.15400904803157187,
+      "grad_norm": 0.1989804357290268,
+      "learning_rate": 9.604646729361547e-06,
+      "loss": 0.7981,
+      "step": 400
+    },
+    {
+      "epoch": 0.15785927423236115,
+      "grad_norm": 0.2075965851545334,
+      "learning_rate": 9.579986529043099e-06,
+      "loss": 0.7899,
+      "step": 410
+    },
+    {
+      "epoch": 0.16170950043315044,
+      "grad_norm": 0.21945220232009888,
+      "learning_rate": 9.554613964695189e-06,
+      "loss": 0.802,
+      "step": 420
+    },
+    {
+      "epoch": 0.16555972663393975,
+      "grad_norm": 0.2184353619813919,
+      "learning_rate": 9.528532982727281e-06,
+      "loss": 0.7926,
+      "step": 430
+    },
+    {
+      "epoch": 0.16940995283472904,
+      "grad_norm": 0.20605944097042084,
+      "learning_rate": 9.501747639735018e-06,
+      "loss": 0.7925,
+      "step": 440
+    },
+    {
+      "epoch": 0.17326017903551832,
+      "grad_norm": 0.2153046727180481,
+      "learning_rate": 9.474262101869272e-06,
+      "loss": 0.7804,
+      "step": 450
+    },
+    {
+      "epoch": 0.17711040523630764,
+      "grad_norm": 0.21783147752285004,
+      "learning_rate": 9.446080644188141e-06,
+      "loss": 0.786,
+      "step": 460
+    },
+    {
+      "epoch": 0.18096063143709692,
+      "grad_norm": 0.20795570313930511,
+      "learning_rate": 9.417207649992017e-06,
+      "loss": 0.794,
+      "step": 470
+    },
+    {
+      "epoch": 0.18481085763788624,
+      "grad_norm": 0.21840886771678925,
+      "learning_rate": 9.387647610141816e-06,
+      "loss": 0.7933,
+      "step": 480
+    },
+    {
+      "epoch": 0.18866108383867553,
+      "grad_norm": 0.19937139749526978,
+      "learning_rate": 9.35740512236047e-06,
+      "loss": 0.779,
+      "step": 490
+    },
+    {
+      "epoch": 0.1925113100394648,
+      "grad_norm": 0.19466154277324677,
+      "learning_rate": 9.326484890517809e-06,
+      "loss": 0.7881,
+      "step": 500
+    },
+    {
+      "epoch": 0.19636153624025413,
+      "grad_norm": 0.21823608875274658,
+      "learning_rate": 9.294891723898926e-06,
+      "loss": 0.7796,
+      "step": 510
+    },
+    {
+      "epoch": 0.2002117624410434,
+      "grad_norm": 0.23181657493114471,
+      "learning_rate": 9.262630536456153e-06,
+      "loss": 0.8089,
+      "step": 520
+    },
+    {
+      "epoch": 0.2040619886418327,
+      "grad_norm": 0.2261747568845749,
+      "learning_rate": 9.229706346044749e-06,
+      "loss": 0.7842,
+      "step": 530
+    },
+    {
+      "epoch": 0.207912214842622,
+      "grad_norm": 0.2290702611207962,
+      "learning_rate": 9.196124273642422e-06,
+      "loss": 0.7944,
+      "step": 540
+    },
+    {
+      "epoch": 0.2117624410434113,
+      "grad_norm": 0.20900452136993408,
+      "learning_rate": 9.161889542552842e-06,
+      "loss": 0.797,
+      "step": 550
+    },
+    {
+      "epoch": 0.21561266724420058,
+      "grad_norm": 0.2081245332956314,
+      "learning_rate": 9.127007477593193e-06,
+      "loss": 0.8015,
+      "step": 560
+    },
+    {
+      "epoch": 0.2194628934449899,
+      "grad_norm": 0.23261775076389313,
+      "learning_rate": 9.091483504265974e-06,
+      "loss": 0.8132,
+      "step": 570
+    },
+    {
+      "epoch": 0.22331311964577918,
+      "grad_norm": 0.2268608659505844,
+      "learning_rate": 9.055323147915116e-06,
+      "loss": 0.7898,
+      "step": 580
+    },
+    {
+      "epoch": 0.2271633458465685,
+      "grad_norm": 0.2154991328716278,
+      "learning_rate": 9.018532032866585e-06,
+      "loss": 0.8018,
+      "step": 590
+    },
+    {
+      "epoch": 0.23101357204735778,
+      "grad_norm": 0.21540164947509766,
+      "learning_rate": 8.981115881553579e-06,
+      "loss": 0.7928,
+      "step": 600
+    },
+    {
+      "epoch": 0.23486379824814707,
+      "grad_norm": 0.23519639670848846,
+      "learning_rate": 8.943080513626476e-06,
+      "loss": 0.8131,
+      "step": 610
+    },
+    {
+      "epoch": 0.23871402444893638,
+      "grad_norm": 0.21424350142478943,
+      "learning_rate": 8.904431845047641e-06,
+      "loss": 0.7967,
+      "step": 620
+    },
+    {
+      "epoch": 0.24256425064972567,
+      "grad_norm": 0.22372132539749146,
+      "learning_rate": 8.86517588717128e-06,
+      "loss": 0.8137,
+      "step": 630
+    },
+    {
+      "epoch": 0.24641447685051496,
+      "grad_norm": 0.21006228029727936,
+      "learning_rate": 8.82531874580844e-06,
+      "loss": 0.802,
+      "step": 640
+    },
+    {
+      "epoch": 0.25026470305130427,
+      "grad_norm": 0.19713839888572693,
+      "learning_rate": 8.784866620277314e-06,
+      "loss": 0.7608,
+      "step": 650
+    },
+    {
+      "epoch": 0.25411492925209356,
+      "grad_norm": 0.2354169487953186,
+      "learning_rate": 8.743825802439018e-06,
+      "loss": 0.7973,
+      "step": 660
+    },
+    {
+      "epoch": 0.25796515545288284,
+      "grad_norm": 0.2241130918264389,
+      "learning_rate": 8.702202675718956e-06,
+      "loss": 0.801,
+      "step": 670
+    },
+    {
+      "epoch": 0.26181538165367213,
+      "grad_norm": 0.20394828915596008,
+      "learning_rate": 8.660003714113959e-06,
+      "loss": 0.7723,
+      "step": 680
+    },
+    {
+      "epoch": 0.26566560785446147,
+      "grad_norm": 0.2082608938217163,
+      "learning_rate": 8.617235481185323e-06,
+      "loss": 0.7955,
+      "step": 690
+    },
+    {
+      "epoch": 0.26951583405525076,
+      "grad_norm": 0.21591560542583466,
+      "learning_rate": 8.573904629037931e-06,
+      "loss": 0.8001,
+      "step": 700
+    },
+    {
+      "epoch": 0.27336606025604004,
+      "grad_norm": 0.21225883066654205,
+      "learning_rate": 8.530017897285585e-06,
+      "loss": 0.7815,
+      "step": 710
+    },
+    {
+      "epoch": 0.27721628645682933,
+      "grad_norm": 0.19734607636928558,
+      "learning_rate": 8.485582112002746e-06,
+      "loss": 0.7862,
+      "step": 720
+    },
+    {
+      "epoch": 0.2810665126576186,
+      "grad_norm": 0.21192070841789246,
+      "learning_rate": 8.440604184662811e-06,
+      "loss": 0.7899,
+      "step": 730
+    },
+    {
+      "epoch": 0.28491673885840796,
+      "grad_norm": 0.21704880893230438,
+      "learning_rate": 8.395091111063118e-06,
+      "loss": 0.7815,
+      "step": 740
+    },
+    {
+      "epoch": 0.28876696505919724,
+      "grad_norm": 0.20165523886680603,
+      "learning_rate": 8.349049970236822e-06,
+      "loss": 0.7794,
+      "step": 750
+    },
+    {
+      "epoch": 0.29261719125998653,
+      "grad_norm": 0.21149715781211853,
+      "learning_rate": 8.302487923351842e-06,
+      "loss": 0.7722,
+      "step": 760
+    },
+    {
+      "epoch": 0.2964674174607758,
+      "grad_norm": 0.20545819401741028,
+      "learning_rate": 8.255412212597021e-06,
+      "loss": 0.7895,
+      "step": 770
+    },
+    {
+      "epoch": 0.3003176436615651,
+      "grad_norm": 0.2076473981142044,
+      "learning_rate": 8.207830160055688e-06,
+      "loss": 0.7853,
+      "step": 780
+    },
+    {
+      "epoch": 0.3041678698623544,
+      "grad_norm": 0.21878257393836975,
+      "learning_rate": 8.159749166566786e-06,
+      "loss": 0.7725,
+      "step": 790
+    },
+    {
+      "epoch": 0.30801809606314373,
+      "grad_norm": 0.20686431229114532,
+      "learning_rate": 8.11117671057377e-06,
+      "loss": 0.7998,
+      "step": 800
+    },
+    {
+      "epoch": 0.311868322263933,
+      "grad_norm": 0.22066085040569305,
+      "learning_rate": 8.062120346961404e-06,
+      "loss": 0.7927,
+      "step": 810
+    },
+    {
+      "epoch": 0.3157185484647223,
+      "grad_norm": 0.20657578110694885,
+      "learning_rate": 8.012587705880705e-06,
+      "loss": 0.7873,
+      "step": 820
+    },
+    {
+      "epoch": 0.3195687746655116,
+      "grad_norm": 0.21122469007968903,
+      "learning_rate": 7.962586491562146e-06,
+      "loss": 0.7727,
+      "step": 830
+    },
+    {
+      "epoch": 0.3234190008663009,
+      "grad_norm": 0.23014427721500397,
+      "learning_rate": 7.912124481117357e-06,
+      "loss": 0.7771,
+      "step": 840
+    },
+    {
+      "epoch": 0.3272692270670902,
+      "grad_norm": 0.20632344484329224,
+      "learning_rate": 7.861209523329494e-06,
+      "loss": 0.7871,
+      "step": 850
+    },
+    {
+      "epoch": 0.3311194532678795,
+      "grad_norm": 0.21350286900997162,
+      "learning_rate": 7.809849537432432e-06,
+      "loss": 0.7743,
+      "step": 860
+    },
+    {
+      "epoch": 0.3349696794686688,
+      "grad_norm": 0.20935165882110596,
+      "learning_rate": 7.758052511879035e-06,
+      "loss": 0.7825,
+      "step": 870
+    },
+    {
+      "epoch": 0.3388199056694581,
+      "grad_norm": 0.20027987658977509,
+      "learning_rate": 7.705826503098642e-06,
+      "loss": 0.777,
+      "step": 880
+    },
+    {
+      "epoch": 0.34267013187024736,
+      "grad_norm": 0.20396825671195984,
+      "learning_rate": 7.653179634243979e-06,
+      "loss": 0.7769,
+      "step": 890
+    },
+    {
+      "epoch": 0.34652035807103665,
+      "grad_norm": 0.19534096121788025,
+      "learning_rate": 7.600120093927696e-06,
+      "loss": 0.783,
+      "step": 900
+    },
+    {
+      "epoch": 0.350370584271826,
+      "grad_norm": 0.2175610065460205,
+      "learning_rate": 7.546656134948729e-06,
+      "loss": 0.7685,
+      "step": 910
+    },
+    {
+      "epoch": 0.3542208104726153,
+      "grad_norm": 0.22066372632980347,
+      "learning_rate": 7.492796073008661e-06,
+      "loss": 0.7692,
+      "step": 920
+    },
+    {
+      "epoch": 0.35807103667340456,
+      "grad_norm": 0.22744490206241608,
+      "learning_rate": 7.438548285418326e-06,
+      "loss": 0.758,
+      "step": 930
+    },
+    {
+      "epoch": 0.36192126287419385,
+      "grad_norm": 0.2304028570652008,
+      "learning_rate": 7.3839212097948e-06,
+      "loss": 0.7907,
+      "step": 940
+    },
+    {
+      "epoch": 0.36577148907498314,
+      "grad_norm": 0.21917569637298584,
+      "learning_rate": 7.328923342749039e-06,
+      "loss": 0.8043,
+      "step": 950
+    },
+    {
+      "epoch": 0.3696217152757725,
+      "grad_norm": 0.2284258008003235,
+      "learning_rate": 7.273563238564319e-06,
+      "loss": 0.7909,
+      "step": 960
+    },
+    {
+      "epoch": 0.37347194147656176,
+      "grad_norm": 0.2173936665058136,
+      "learning_rate": 7.217849507865724e-06,
+      "loss": 0.7714,
+      "step": 970
+    },
+    {
+      "epoch": 0.37732216767735105,
+      "grad_norm": 0.208927184343338,
+      "learning_rate": 7.161790816280858e-06,
+      "loss": 0.7753,
+      "step": 980
+    },
+    {
+      "epoch": 0.38117239387814034,
+      "grad_norm": 0.22106845676898956,
+      "learning_rate": 7.1053958830920125e-06,
+      "loss": 0.7622,
+      "step": 990
+    },
+    {
+      "epoch": 0.3850226200789296,
+      "grad_norm": 0.20784111320972443,
+      "learning_rate": 7.0486734798799715e-06,
+      "loss": 0.775,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3888728462797189,
+      "grad_norm": 0.20803546905517578,
+      "learning_rate": 6.991632429159708e-06,
+      "loss": 0.8003,
+      "step": 1010
+    },
+    {
+      "epoch": 0.39272307248050825,
+      "grad_norm": 0.22865328192710876,
+      "learning_rate": 6.934281603008132e-06,
+      "loss": 0.7937,
+      "step": 1020
+    },
+    {
+      "epoch": 0.39657329868129754,
+      "grad_norm": 0.2029670923948288,
+      "learning_rate": 6.876629921684143e-06,
+      "loss": 0.7903,
+      "step": 1030
+    },
+    {
+      "epoch": 0.4004235248820868,
+      "grad_norm": 0.21803723275661469,
+      "learning_rate": 6.818686352241199e-06,
+      "loss": 0.7807,
+      "step": 1040
+    },
+    {
+      "epoch": 0.4042737510828761,
+      "grad_norm": 0.19217590987682343,
+      "learning_rate": 6.760459907132573e-06,
+      "loss": 0.7786,
+      "step": 1050
+    },
+    {
+      "epoch": 0.4081239772836654,
+      "grad_norm": 0.19760863482952118,
+      "learning_rate": 6.701959642809589e-06,
+      "loss": 0.7845,
+      "step": 1060
+    },
+    {
+      "epoch": 0.41197420348445474,
+      "grad_norm": 0.2137843668460846,
+      "learning_rate": 6.643194658312987e-06,
+      "loss": 0.7739,
+      "step": 1070
+    },
+    {
+      "epoch": 0.415824429685244,
+      "grad_norm": 0.20635919272899628,
+      "learning_rate": 6.584174093857676e-06,
+      "loss": 0.7984,
+      "step": 1080
+    },
+    {
+      "epoch": 0.4196746558860333,
+      "grad_norm": 0.23427267372608185,
+      "learning_rate": 6.5249071294110735e-06,
+      "loss": 0.7811,
+      "step": 1090
+    },
+    {
+      "epoch": 0.4235248820868226,
+      "grad_norm": 0.2152027189731598,
+      "learning_rate": 6.465402983265274e-06,
+      "loss": 0.7944,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4273751082876119,
+      "grad_norm": 0.22062291204929352,
+      "learning_rate": 6.405670910603243e-06,
+      "loss": 0.7922,
+      "step": 1110
+    },
+    {
+      "epoch": 0.43122533448840117,
+      "grad_norm": 0.23552630841732025,
+      "learning_rate": 6.345720202059291e-06,
+      "loss": 0.7879,
+      "step": 1120
+    },
+    {
+      "epoch": 0.4350755606891905,
+      "grad_norm": 0.21442773938179016,
+      "learning_rate": 6.285560182274008e-06,
+      "loss": 0.7651,
+      "step": 1130
+    },
+    {
+      "epoch": 0.4389257868899798,
+      "grad_norm": 0.20756173133850098,
+      "learning_rate": 6.225200208443936e-06,
+      "loss": 0.7865,
+      "step": 1140
+    },
+    {
+      "epoch": 0.4427760130907691,
+      "grad_norm": 0.21122901141643524,
+      "learning_rate": 6.16464966886616e-06,
+      "loss": 0.779,
+      "step": 1150
+    },
+    {
+      "epoch": 0.44662623929155837,
+      "grad_norm": 0.24453377723693848,
+      "learning_rate": 6.103917981478065e-06,
+      "loss": 0.8016,
+      "step": 1160
+    },
+    {
+      "epoch": 0.45047646549234766,
+      "grad_norm": 0.21272338926792145,
+      "learning_rate": 6.043014592392489e-06,
+      "loss": 0.814,
+      "step": 1170
+    },
+    {
+      "epoch": 0.454326691693137,
+      "grad_norm": 0.19979600608348846,
+      "learning_rate": 5.981948974428485e-06,
+      "loss": 0.7857,
+      "step": 1180
+    },
+    {
+      "epoch": 0.4581769178939263,
+      "grad_norm": 0.21946501731872559,
+      "learning_rate": 5.920730625637934e-06,
+      "loss": 0.7673,
+      "step": 1190
+    },
+    {
+      "epoch": 0.46202714409471557,
+      "grad_norm": 0.2219962328672409,
+      "learning_rate": 5.8593690678282355e-06,
+      "loss": 0.7777,
+      "step": 1200
+    },
+    {
+      "epoch": 0.46587737029550486,
+      "grad_norm": 0.22055117785930634,
+      "learning_rate": 5.797873845081301e-06,
+      "loss": 0.7801,
+      "step": 1210
+    },
+    {
+      "epoch": 0.46972759649629414,
+      "grad_norm": 0.2186819612979889,
+      "learning_rate": 5.736254522269075e-06,
+      "loss": 0.797,
+      "step": 1220
+    },
+    {
+      "epoch": 0.47357782269708343,
+      "grad_norm": 0.20423023402690887,
+      "learning_rate": 5.6745206835658386e-06,
+      "loss": 0.7998,
+      "step": 1230
+    },
+    {
+      "epoch": 0.47742804889787277,
+      "grad_norm": 0.20168128609657288,
+      "learning_rate": 5.612681930957495e-06,
+      "loss": 0.769,
+      "step": 1240
+    },
+    {
+      "epoch": 0.48127827509866206,
+      "grad_norm": 0.20024374127388,
+      "learning_rate": 5.550747882748095e-06,
+      "loss": 0.7684,
+      "step": 1250
+    },
+    {
+      "epoch": 0.48512850129945134,
+      "grad_norm": 0.22639770805835724,
+      "learning_rate": 5.4887281720638195e-06,
+      "loss": 0.7762,
+      "step": 1260
+    },
+    {
+      "epoch": 0.48897872750024063,
+      "grad_norm": 0.26514559984207153,
+      "learning_rate": 5.42663244535466e-06,
+      "loss": 0.7609,
+      "step": 1270
+    },
+    {
+      "epoch": 0.4928289537010299,
+      "grad_norm": 0.21242381632328033,
+      "learning_rate": 5.3644703608940205e-06,
+      "loss": 0.7686,
+      "step": 1280
+    },
+    {
+      "epoch": 0.49667917990181926,
+      "grad_norm": 0.21298331022262573,
+      "learning_rate": 5.302251587276491e-06,
+      "loss": 0.7795,
+      "step": 1290
+    },
+    {
+      "epoch": 0.5005294061026085,
+      "grad_norm": 0.20495860278606415,
+      "learning_rate": 5.2399858019140005e-06,
+      "loss": 0.7807,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5043796323033978,
+      "grad_norm": 0.20415307581424713,
+      "learning_rate": 5.177682689530615e-06,
+      "loss": 0.7914,
+      "step": 1310
+    },
+    {
+      "epoch": 0.5082298585041871,
+      "grad_norm": 0.2150672823190689,
+      "learning_rate": 5.11535194065619e-06,
+      "loss": 0.7643,
+      "step": 1320
+    },
+    {
+      "epoch": 0.5120800847049765,
+      "grad_norm": 0.23607780039310455,
+      "learning_rate": 5.053003250119113e-06,
+      "loss": 0.7739,
+      "step": 1330
+    },
+    {
+      "epoch": 0.5159303109057657,
+      "grad_norm": 0.22037194669246674,
+      "learning_rate": 4.9906463155383925e-06,
+      "loss": 0.7782,
+      "step": 1340
+    },
+    {
+      "epoch": 0.519780537106555,
+      "grad_norm": 0.20937801897525787,
+      "learning_rate": 4.9282908358153046e-06,
+      "loss": 0.7704,
+      "step": 1350
+    },
+    {
+      "epoch": 0.5236307633073443,
+      "grad_norm": 0.24204786121845245,
+      "learning_rate": 4.8659465096248345e-06,
+      "loss": 0.7901,
+      "step": 1360
+    },
+    {
+      "epoch": 0.5274809895081336,
+      "grad_norm": 0.22079528868198395,
+      "learning_rate": 4.803623033907169e-06,
+      "loss": 0.7881,
+      "step": 1370
+    },
+    {
+      "epoch": 0.5313312157089229,
+      "grad_norm": 0.20521317422389984,
+      "learning_rate": 4.741330102359441e-06,
+      "loss": 0.7942,
+      "step": 1380
+    },
+    {
+      "epoch": 0.5351814419097122,
+      "grad_norm": 0.22508221864700317,
+      "learning_rate": 4.679077403927996e-06,
+      "loss": 0.7747,
+      "step": 1390
+    },
+    {
+      "epoch": 0.5390316681105015,
+      "grad_norm": 0.21338631212711334,
+      "learning_rate": 4.616874621301373e-06,
+      "loss": 0.7743,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5428818943112907,
+      "grad_norm": 0.21970006823539734,
+      "learning_rate": 4.554731429404293e-06,
+      "loss": 0.7874,
+      "step": 1410
+    },
+    {
+      "epoch": 0.5467321205120801,
+      "grad_norm": 0.22566400468349457,
+      "learning_rate": 4.492657493892822e-06,
+      "loss": 0.775,
+      "step": 1420
+    },
+    {
+      "epoch": 0.5505823467128694,
+      "grad_norm": 0.22664548456668854,
+      "learning_rate": 4.430662469650999e-06,
+      "loss": 0.7892,
+      "step": 1430
+    },
+    {
+      "epoch": 0.5544325729136587,
+      "grad_norm": 0.21953046321868896,
+      "learning_rate": 4.368755999289118e-06,
+      "loss": 0.7918,
+      "step": 1440
+    },
+    {
+      "epoch": 0.558282799114448,
+      "grad_norm": 0.22127187252044678,
+      "learning_rate": 4.306947711643949e-06,
+      "loss": 0.7738,
+      "step": 1450
+    },
+    {
+      "epoch": 0.5621330253152372,
+      "grad_norm": 0.2230837643146515,
+      "learning_rate": 4.245247220281068e-06,
+      "loss": 0.7796,
+      "step": 1460
+    },
+    {
+      "epoch": 0.5659832515160266,
+      "grad_norm": 0.211197629570961,
+      "learning_rate": 4.183664121999587e-06,
+      "loss": 0.7758,
+      "step": 1470
+    },
+    {
+      "epoch": 0.5698334777168159,
+      "grad_norm": 0.22593528032302856,
+      "learning_rate": 4.122207995339474e-06,
+      "loss": 0.7711,
+      "step": 1480
+    },
+    {
+      "epoch": 0.5736837039176051,
+      "grad_norm": 0.20503315329551697,
+      "learning_rate": 4.060888399091734e-06,
+      "loss": 0.7775,
+      "step": 1490
+    },
+    {
+      "epoch": 0.5775339301183945,
+      "grad_norm": 0.20107236504554749,
+      "learning_rate": 3.999714870811639e-06,
+      "loss": 0.7865,
+      "step": 1500
+    },
+    {
+      "epoch": 0.5813841563191837,
+      "grad_norm": 0.2127494513988495,
+      "learning_rate": 3.938696925335275e-06,
+      "loss": 0.7803,
+      "step": 1510
+    },
+    {
+      "epoch": 0.5852343825199731,
+      "grad_norm": 0.21122244000434875,
+      "learning_rate": 3.87784405329962e-06,
+      "loss": 0.7894,
+      "step": 1520
+    },
+    {
+      "epoch": 0.5890846087207623,
+      "grad_norm": 0.21198853850364685,
+      "learning_rate": 3.817165719666382e-06,
+      "loss": 0.7904,
+      "step": 1530
+    },
+    {
+      "epoch": 0.5929348349215516,
+      "grad_norm": 0.20368939638137817,
+      "learning_rate": 3.756671362249841e-06,
+      "loss": 0.7945,
+      "step": 1540
+    },
+    {
+      "epoch": 0.596785061122341,
+      "grad_norm": 0.21105943620204926,
+      "learning_rate": 3.6963703902488907e-06,
+      "loss": 0.7823,
+      "step": 1550
+    },
+    {
+      "epoch": 0.6006352873231302,
+      "grad_norm": 0.21012234687805176,
+      "learning_rate": 3.6362721827835655e-06,
+      "loss": 0.7979,
+      "step": 1560
+    },
+    {
+      "epoch": 0.6044855135239195,
+      "grad_norm": 0.2217540740966797,
+      "learning_rate": 3.576386087436212e-06,
+      "loss": 0.7934,
+      "step": 1570
+    },
+    {
+      "epoch": 0.6083357397247088,
+      "grad_norm": 0.22914958000183105,
+      "learning_rate": 3.5167214187975913e-06,
+      "loss": 0.7832,
+      "step": 1580
+    },
+    {
+      "epoch": 0.6121859659254981,
+      "grad_norm": 0.19049789011478424,
+      "learning_rate": 3.457287457018086e-06,
+      "loss": 0.7838,
+      "step": 1590
+    },
+    {
+      "epoch": 0.6160361921262875,
+      "grad_norm": 0.20706436038017273,
+      "learning_rate": 3.3980934463643003e-06,
+      "loss": 0.7729,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6198864183270767,
+      "grad_norm": 0.19974836707115173,
+      "learning_rate": 3.3391485937812053e-06,
+      "loss": 0.7547,
+      "step": 1610
+    },
+    {
+      "epoch": 0.623736644527866,
+      "grad_norm": 0.2140933871269226,
+      "learning_rate": 3.280462067460115e-06,
+      "loss": 0.7542,
+      "step": 1620
+    },
+    {
+      "epoch": 0.6275868707286553,
+      "grad_norm": 0.2124594748020172,
+      "learning_rate": 3.222042995412669e-06,
+      "loss": 0.7865,
+      "step": 1630
+    },
+    {
+      "epoch": 0.6314370969294446,
+      "grad_norm": 0.1883147805929184,
+      "learning_rate": 3.163900464051089e-06,
+      "loss": 0.7747,
+      "step": 1640
+    },
+    {
+      "epoch": 0.635287323130234,
+      "grad_norm": 0.21481215953826904,
+      "learning_rate": 3.1060435167748826e-06,
+      "loss": 0.7839,
+      "step": 1650
+    },
+    {
+      "epoch": 0.6391375493310232,
+      "grad_norm": 0.21719808876514435,
+      "learning_rate": 3.048481152564253e-06,
+      "loss": 0.7651,
+      "step": 1660
+    },
+    {
+      "epoch": 0.6429877755318125,
+      "grad_norm": 0.22381862998008728,
+      "learning_rate": 2.9912223245804017e-06,
+      "loss": 0.7845,
+      "step": 1670
+    },
+    {
+      "epoch": 0.6468380017326018,
+      "grad_norm": 0.2199912667274475,
+      "learning_rate": 2.934275938772981e-06,
+      "loss": 0.7821,
+      "step": 1680
+    },
+    {
+      "epoch": 0.6506882279333911,
+      "grad_norm": 0.22058957815170288,
+      "learning_rate": 2.8776508524948676e-06,
+      "loss": 0.7714,
+      "step": 1690
+    },
+    {
+      "epoch": 0.6545384541341804,
+      "grad_norm": 0.19057691097259521,
+      "learning_rate": 2.821355873124503e-06,
+      "loss": 0.772,
+      "step": 1700
+    },
+    {
+      "epoch": 0.6583886803349697,
+      "grad_norm": 0.219065323472023,
+      "learning_rate": 2.7653997566960144e-06,
+      "loss": 0.7825,
+      "step": 1710
+    },
+    {
+      "epoch": 0.662238906535759,
+      "grad_norm": 0.21649852395057678,
+      "learning_rate": 2.709791206537309e-06,
+      "loss": 0.773,
+      "step": 1720
+    },
+    {
+      "epoch": 0.6660891327365482,
+      "grad_norm": 0.21306246519088745,
+      "learning_rate": 2.6545388719163747e-06,
+      "loss": 0.7797,
+      "step": 1730
+    },
+    {
+      "epoch": 0.6699393589373376,
+      "grad_norm": 0.21471701562404633,
+      "learning_rate": 2.599651346695979e-06,
+      "loss": 0.7786,
+      "step": 1740
+    },
+    {
+      "epoch": 0.6737895851381268,
+      "grad_norm": 0.20033955574035645,
+      "learning_rate": 2.545137167997007e-06,
+      "loss": 0.7765,
+      "step": 1750
+    },
+    {
+      "epoch": 0.6776398113389162,
+      "grad_norm": 0.20072396099567413,
+      "learning_rate": 2.4910048148705972e-06,
+      "loss": 0.7661,
+      "step": 1760
+    },
+    {
+      "epoch": 0.6814900375397055,
+      "grad_norm": 0.22002218663692474,
+      "learning_rate": 2.4372627069793327e-06,
+      "loss": 0.8076,
+      "step": 1770
+    },
+    {
+      "epoch": 0.6853402637404947,
+      "grad_norm": 0.22015342116355896,
+      "learning_rate": 2.3839192032876486e-06,
+      "loss": 0.7935,
+      "step": 1780
+    },
+    {
+      "epoch": 0.6891904899412841,
+      "grad_norm": 0.20910853147506714,
+      "learning_rate": 2.3309826007617047e-06,
+      "loss": 0.8041,
+      "step": 1790
+    },
+    {
+      "epoch": 0.6930407161420733,
+      "grad_norm": 0.21569500863552094,
+      "learning_rate": 2.278461133078881e-06,
+      "loss": 0.7668,
+      "step": 1800
+    },
+    {
+      "epoch": 0.6968909423428626,
+      "grad_norm": 0.22438117861747742,
+      "learning_rate": 2.2263629693471258e-06,
+      "loss": 0.7597,
+      "step": 1810
+    },
+    {
+      "epoch": 0.700741168543652,
+      "grad_norm": 0.19963979721069336,
+      "learning_rate": 2.1746962128343414e-06,
+      "loss": 0.7821,
+      "step": 1820
+    },
+    {
+      "epoch": 0.7045913947444412,
+      "grad_norm": 0.20364798605442047,
+      "learning_rate": 2.1234688997080183e-06,
+      "loss": 0.7712,
+      "step": 1830
+    },
+    {
+      "epoch": 0.7084416209452306,
+      "grad_norm": 0.22232916951179504,
+      "learning_rate": 2.0726889977853053e-06,
+      "loss": 0.7779,
+      "step": 1840
+    },
+    {
+      "epoch": 0.7122918471460198,
+      "grad_norm": 0.23047001659870148,
+      "learning_rate": 2.022364405293703e-06,
+      "loss": 0.7872,
+      "step": 1850
+    },
+    {
+      "epoch": 0.7161420733468091,
+      "grad_norm": 0.2134980410337448,
+      "learning_rate": 1.9725029496425814e-06,
+      "loss": 0.7823,
+      "step": 1860
+    },
+    {
+      "epoch": 0.7199922995475985,
+      "grad_norm": 0.19672690331935883,
+      "learning_rate": 1.923112386205727e-06,
+      "loss": 0.779,
+      "step": 1870
+    },
+    {
+      "epoch": 0.7238425257483877,
+      "grad_norm": 0.2036598026752472,
+      "learning_rate": 1.874200397115079e-06,
+      "loss": 0.7858,
+      "step": 1880
+    },
+    {
+      "epoch": 0.727692751949177,
+      "grad_norm": 0.2252436876296997,
+      "learning_rate": 1.825774590065857e-06,
+      "loss": 0.7691,
+      "step": 1890
+    },
+    {
+      "epoch": 0.7315429781499663,
+      "grad_norm": 0.2181544154882431,
+      "learning_rate": 1.7778424971332813e-06,
+      "loss": 0.7788,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7353932043507556,
+      "grad_norm": 0.20285160839557648,
+      "learning_rate": 1.7304115736010407e-06,
+      "loss": 0.7773,
+      "step": 1910
+    },
+    {
+      "epoch": 0.739243430551545,
+      "grad_norm": 0.21418346464633942,
+      "learning_rate": 1.6834891968017108e-06,
+      "loss": 0.7702,
+      "step": 1920
+    },
+    {
+      "epoch": 0.7430936567523342,
+      "grad_norm": 0.2153778225183487,
+      "learning_rate": 1.6370826649692884e-06,
+      "loss": 0.7889,
+      "step": 1930
+    },
+    {
+      "epoch": 0.7469438829531235,
+      "grad_norm": 0.2192072868347168,
+      "learning_rate": 1.5911991961040463e-06,
+      "loss": 0.7938,
+      "step": 1940
+    },
+    {
+      "epoch": 0.7507941091539128,
+      "grad_norm": 0.21680034697055817,
+      "learning_rate": 1.545845926849851e-06,
+      "loss": 0.7792,
+      "step": 1950
+    },
+    {
+      "epoch": 0.7546443353547021,
+      "grad_norm": 0.23316241800785065,
+      "learning_rate": 1.5010299113841397e-06,
+      "loss": 0.7986,
+      "step": 1960
+    },
+    {
+      "epoch": 0.7584945615554913,
+      "grad_norm": 0.21392640471458435,
+      "learning_rate": 1.4567581203207238e-06,
+      "loss": 0.7864,
+      "step": 1970
+    },
+    {
+      "epoch": 0.7623447877562807,
+      "grad_norm": 0.22406625747680664,
+      "learning_rate": 1.413037439625593e-06,
+      "loss": 0.7882,
+      "step": 1980
+    },
+    {
+      "epoch": 0.76619501395707,
+      "grad_norm": 0.22413203120231628,
+      "learning_rate": 1.3698746695458847e-06,
+      "loss": 0.7786,
+      "step": 1990
+    },
+    {
+      "epoch": 0.7700452401578592,
+      "grad_norm": 0.20612797141075134,
+      "learning_rate": 1.3272765235521745e-06,
+      "loss": 0.7719,
+      "step": 2000
+    },
+    {
+      "epoch": 0.7738954663586486,
+      "grad_norm": 0.2124829739332199,
+      "learning_rate": 1.2852496272942877e-06,
+      "loss": 0.7805,
+      "step": 2010
+    },
+    {
+      "epoch": 0.7777456925594378,
+      "grad_norm": 0.19927899539470673,
+      "learning_rate": 1.2438005175707396e-06,
+      "loss": 0.768,
+      "step": 2020
+    },
+    {
+      "epoch": 0.7815959187602272,
+      "grad_norm": 0.2082962840795517,
+      "learning_rate": 1.2029356413120297e-06,
+      "loss": 0.7542,
+      "step": 2030
+    },
+    {
+      "epoch": 0.7854461449610165,
+      "grad_norm": 0.21014700829982758,
+      "learning_rate": 1.1626613545778786e-06,
+      "loss": 0.7706,
+      "step": 2040
+    },
+    {
+      "epoch": 0.7892963711618057,
+      "grad_norm": 0.20060232281684875,
+      "learning_rate": 1.1229839215686294e-06,
+      "loss": 0.7743,
+      "step": 2050
+    },
+    {
+      "epoch": 0.7931465973625951,
+      "grad_norm": 0.194558247923851,
+      "learning_rate": 1.0839095136509188e-06,
+      "loss": 0.7608,
+      "step": 2060
+    },
+    {
+      "epoch": 0.7969968235633843,
+      "grad_norm": 0.22875520586967468,
+      "learning_rate": 1.045444208397791e-06,
+      "loss": 0.7825,
+      "step": 2070
+    },
+    {
+      "epoch": 0.8008470497641736,
+      "grad_norm": 0.1960304081439972,
+      "learning_rate": 1.007593988643402e-06,
+      "loss": 0.7643,
+      "step": 2080
+    },
+    {
+      "epoch": 0.804697275964963,
+      "grad_norm": 0.20116665959358215,
+      "learning_rate": 9.703647415524613e-07,
+      "loss": 0.7857,
+      "step": 2090
+    },
+    {
+      "epoch": 0.8085475021657522,
+      "grad_norm": 0.2121640145778656,
+      "learning_rate": 9.337622577045485e-07,
+      "loss": 0.7745,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8123977283665416,
+      "grad_norm": 0.21365812420845032,
+      "learning_rate": 8.977922301934588e-07,
+      "loss": 0.7818,
+      "step": 2110
+    },
+    {
+      "epoch": 0.8162479545673308,
+      "grad_norm": 0.21028682589530945,
+      "learning_rate": 8.624602537417009e-07,
+      "loss": 0.7835,
+      "step": 2120
+    },
+    {
+      "epoch": 0.8200981807681201,
+      "grad_norm": 0.22910574078559875,
+      "learning_rate": 8.277718238303123e-07,
+      "loss": 0.77,
+      "step": 2130
+    },
+    {
+      "epoch": 0.8239484069689095,
+      "grad_norm": 0.2136904001235962,
+      "learning_rate": 7.937323358440935e-07,
+      "loss": 0.7747,
+      "step": 2140
+    },
+    {
+      "epoch": 0.8277986331696987,
+      "grad_norm": 0.22802819311618805,
+      "learning_rate": 7.603470842324223e-07,
+      "loss": 0.7664,
+      "step": 2150
+    },
+    {
+      "epoch": 0.831648859370488,
+      "grad_norm": 0.20423157513141632,
+      "learning_rate": 7.276212616857575e-07,
+      "loss": 0.7835,
+      "step": 2160
+    },
+    {
+      "epoch": 0.8354990855712773,
+      "grad_norm": 0.20894873142242432,
+      "learning_rate": 6.955599583279837e-07,
+      "loss": 0.7857,
+      "step": 2170
+    },
+    {
+      "epoch": 0.8393493117720666,
+      "grad_norm": 0.20682072639465332,
+      "learning_rate": 6.641681609246981e-07,
+      "loss": 0.7752,
+      "step": 2180
+    },
+    {
+      "epoch": 0.8431995379728559,
+      "grad_norm": 0.20308268070220947,
+      "learning_rate": 6.334507521075711e-07,
+      "loss": 0.7849,
+      "step": 2190
+    },
+    {
+      "epoch": 0.8470497641736452,
+      "grad_norm": 0.22242219746112823,
+      "learning_rate": 6.034125096149179e-07,
+      "loss": 0.778,
+      "step": 2200
+    },
+    {
+      "epoch": 0.8508999903744345,
+      "grad_norm": 0.21589159965515137,
+      "learning_rate": 5.740581055485722e-07,
+      "loss": 0.7485,
+      "step": 2210
+    },
+    {
+      "epoch": 0.8547502165752238,
+      "grad_norm": 0.19380824267864227,
+      "learning_rate": 5.453921056471961e-07,
+      "loss": 0.7826,
+      "step": 2220
+    },
+    {
+      "epoch": 0.8586004427760131,
+      "grad_norm": 0.211557537317276,
+      "learning_rate": 5.174189685761288e-07,
+      "loss": 0.7773,
+      "step": 2230
+    },
+    {
+      "epoch": 0.8624506689768023,
+      "grad_norm": 0.20140938460826874,
+      "learning_rate": 4.901430452338979e-07,
+      "loss": 0.7875,
+      "step": 2240
+    },
+    {
+      "epoch": 0.8663008951775917,
+      "grad_norm": 0.20314903557300568,
+      "learning_rate": 4.6356857807548286e-07,
+      "loss": 0.7731,
+      "step": 2250
+    },
+    {
+      "epoch": 0.870151121378381,
+      "grad_norm": 0.2121332734823227,
+      "learning_rate": 4.376997004524541e-07,
+      "loss": 0.7628,
+      "step": 2260
+    },
+    {
+      "epoch": 0.8740013475791703,
+      "grad_norm": 0.2105894386768341,
+      "learning_rate": 4.12540435970073e-07,
+      "loss": 0.7719,
+      "step": 2270
+    },
+    {
+      "epoch": 0.8778515737799596,
+      "grad_norm": 0.211486354470253,
+      "learning_rate": 3.880946978614703e-07,
+      "loss": 0.779,
+      "step": 2280
+    },
+    {
+      "epoch": 0.8817017999807488,
+      "grad_norm": 0.21407294273376465,
+      "learning_rate": 3.643662883789878e-07,
+      "loss": 0.7857,
+      "step": 2290
+    },
+    {
+      "epoch": 0.8855520261815382,
+      "grad_norm": 0.19690380990505219,
+      "learning_rate": 3.41358898202781e-07,
+      "loss": 0.7643,
+      "step": 2300
+    },
+    {
+      "epoch": 0.8894022523823275,
+      "grad_norm": 0.1978316605091095,
+      "learning_rate": 3.190761058667741e-07,
+      "loss": 0.7774,
+      "step": 2310
+    },
+    {
+      "epoch": 0.8932524785831167,
+      "grad_norm": 0.20489174127578735,
+      "learning_rate": 2.9752137720206463e-07,
+      "loss": 0.7811,
+      "step": 2320
+    },
+    {
+      "epoch": 0.8971027047839061,
+      "grad_norm": 0.20977835357189178,
+      "learning_rate": 2.7669806479785067e-07,
+      "loss": 0.785,
+      "step": 2330
+    },
+    {
+      "epoch": 0.9009529309846953,
+      "grad_norm": 0.20932117104530334,
+      "learning_rate": 2.566094074799763e-07,
+      "loss": 0.7796,
+      "step": 2340
+    },
+    {
+      "epoch": 0.9048031571854847,
+      "grad_norm": 0.20913055539131165,
+      "learning_rate": 2.3725852980716557e-07,
+      "loss": 0.763,
+      "step": 2350
+    },
+    {
+      "epoch": 0.908653383386274,
+      "grad_norm": 0.2064633071422577,
+      "learning_rate": 2.1864844158503983e-07,
+      "loss": 0.775,
+      "step": 2360
+    },
+    {
+      "epoch": 0.9125036095870632,
+      "grad_norm": 0.226271390914917,
+      "learning_rate": 2.007820373979702e-07,
+      "loss": 0.767,
+      "step": 2370
+    },
+    {
+      "epoch": 0.9163538357878526,
+      "grad_norm": 0.21376509964466095,
+      "learning_rate": 1.8366209615886532e-07,
+      "loss": 0.7773,
+      "step": 2380
+    },
+    {
+      "epoch": 0.9202040619886418,
+      "grad_norm": 0.2101443111896515,
+      "learning_rate": 1.6729128067693613e-07,
+      "loss": 0.7855,
+      "step": 2390
+    },
+    {
+      "epoch": 0.9240542881894311,
+      "grad_norm": 0.22688432037830353,
+      "learning_rate": 1.5167213724353426e-07,
+      "loss": 0.7862,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9279045143902205,
+      "grad_norm": 0.20727209746837616,
+      "learning_rate": 1.3680709523610058e-07,
+      "loss": 0.7654,
+      "step": 2410
+    },
+    {
+      "epoch": 0.9317547405910097,
+      "grad_norm": 0.20891693234443665,
+      "learning_rate": 1.2269846674030418e-07,
+      "loss": 0.792,
+      "step": 2420
+    },
+    {
+      "epoch": 0.935604966791799,
+      "grad_norm": 0.22245071828365326,
+      "learning_rate": 1.0934844619042595e-07,
+      "loss": 0.7857,
+      "step": 2430
+    },
+    {
+      "epoch": 0.9394551929925883,
+      "grad_norm": 0.2200334370136261,
+      "learning_rate": 9.675911002803961e-08,
+      "loss": 0.7763,
+      "step": 2440
+    },
+    {
+      "epoch": 0.9433054191933776,
+      "grad_norm": 0.2263546884059906,
+      "learning_rate": 8.493241637904381e-08,
+      "loss": 0.7737,
+      "step": 2450
+    },
+    {
+      "epoch": 0.9471556453941669,
+      "grad_norm": 0.20785392820835114,
+      "learning_rate": 7.387020474909689e-08,
+      "loss": 0.792,
+      "step": 2460
+    },
+    {
+      "epoch": 0.9510058715949562,
+      "grad_norm": 0.2217138260602951,
+      "learning_rate": 6.357419573750856e-08,
+      "loss": 0.788,
+      "step": 2470
+    },
+    {
+      "epoch": 0.9548560977957455,
+      "grad_norm": 0.21275535225868225,
+      "learning_rate": 5.4045990769613365e-08,
+      "loss": 0.7699,
+      "step": 2480
+    },
+    {
+      "epoch": 0.9587063239965348,
+      "grad_norm": 0.2077636420726776,
+      "learning_rate": 4.5287071847694434e-08,
+      "loss": 0.7771,
+      "step": 2490
+    },
+    {
+      "epoch": 0.9625565501973241,
+      "grad_norm": 0.21107785403728485,
+      "learning_rate": 3.7298801320467817e-08,
+      "loss": 0.7732,
+      "step": 2500
+    },
+    {
+      "epoch": 0.9664067763981133,
+      "grad_norm": 0.2079184204339981,
+      "learning_rate": 3.0082421671192576e-08,
+      "loss": 0.7646,
+      "step": 2510
+    },
+    {
+      "epoch": 0.9702570025989027,
+      "grad_norm": 0.22523944079875946,
+      "learning_rate": 2.3639055324410355e-08,
+      "loss": 0.7835,
+      "step": 2520
+    },
+    {
+      "epoch": 0.974107228799692,
+      "grad_norm": 0.21120281517505646,
+      "learning_rate": 1.7969704471369477e-08,
+      "loss": 0.7713,
+      "step": 2530
+    },
+    {
+      "epoch": 0.9779574550004813,
+      "grad_norm": 0.20129713416099548,
+      "learning_rate": 1.3075250914142413e-08,
+      "loss": 0.772,
+      "step": 2540
+    },
+    {
+      "epoch": 0.9818076812012706,
+      "grad_norm": 0.2091480940580368,
+      "learning_rate": 8.956455928476049e-09,
+      "loss": 0.7825,
+      "step": 2550
+    },
+    {
+      "epoch": 0.9856579074020598,
+      "grad_norm": 0.2020321488380432,
+      "learning_rate": 5.6139601453786275e-09,
+      "loss": 0.7756,
+      "step": 2560
+    },
+    {
+      "epoch": 0.9895081336028492,
+      "grad_norm": 0.20928555727005005,
+      "learning_rate": 3.0482834514805602e-09,
+      "loss": 0.7654,
+      "step": 2570
+    },
+    {
+      "epoch": 0.9933583598036385,
+      "grad_norm": 0.20728491246700287,
+      "learning_rate": 1.2598249081713366e-09,
+      "loss": 0.7577,
+      "step": 2580
+    },
+    {
+      "epoch": 0.9972085860044277,
+      "grad_norm": 0.21671123802661896,
+      "learning_rate": 2.488626895302826e-10,
+      "loss": 0.7879,
+      "step": 2590
+    },
+    {
+      "epoch": 0.9999037443449803,
+      "step": 2597,
+      "total_flos": 3.522737300075184e+18,
+      "train_loss": 0.7961161714083055,
+      "train_runtime": 7556.9421,
+      "train_samples_per_second": 87.986,
+      "train_steps_per_second": 0.344
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2597,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.522737300075184e+18,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/mc1_b32_224_2.5b/finetune/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f9ab94fcf483f115e32f57129f899803fa9799c9595ceaf0499c599cf66d7e7
+size 6136

continuous/qwen25/mc1_b32_224_2.5b/pretrain/config.json

@@ -0,0 +1,58 @@
+{
+  "_attn_implementation_autoset": true,
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "image_aspect_ratio": "square",
+  "image_crop_resolution": null,
+  "image_grid_pinpoints": null,
+  "image_split_resolution": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 768,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": null,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "mm_tunable_parts": "mm_mlp_adapter",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "open_clip_hub:ViT-B-32",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/mc1/mc1_b32_224_2.5b.pt",
+  "vocab_size": 151936
+}

continuous/qwen25/mc1_b32_224_2.5b/pretrain/mm_projector.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec0812f6ab5570ece7efc31ee5933086d7d633756d3731fc72ec802a694e0b66
+size 14170168

continuous/qwen25/mc1_b32_224_2.5b/pretrain/trainer_state.json

@@ -0,0 +1,421 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 545,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01834862385321101,
+      "grad_norm": 1.6848793029785156,
+      "learning_rate": 0.00032142857142857147,
+      "loss": 4.9185,
+      "step": 10
+    },
+    {
+      "epoch": 0.03669724770642202,
+      "grad_norm": 2.817183494567871,
+      "learning_rate": 0.0006785714285714287,
+      "loss": 3.9769,
+      "step": 20
+    },
+    {
+      "epoch": 0.05504587155963303,
+      "grad_norm": 0.7975208163261414,
+      "learning_rate": 0.0009999907688176173,
+      "loss": 3.5486,
+      "step": 30
+    },
+    {
+      "epoch": 0.07339449541284404,
+      "grad_norm": 0.3655107617378235,
+      "learning_rate": 0.0009988834393115766,
+      "loss": 3.3266,
+      "step": 40
+    },
+    {
+      "epoch": 0.09174311926605505,
+      "grad_norm": 0.20547817647457123,
+      "learning_rate": 0.0009959345572940772,
+      "loss": 3.2605,
+      "step": 50
+    },
+    {
+      "epoch": 0.11009174311926606,
+      "grad_norm": 0.2803628146648407,
+      "learning_rate": 0.0009911550081155983,
+      "loss": 3.2197,
+      "step": 60
+    },
+    {
+      "epoch": 0.12844036697247707,
+      "grad_norm": 0.19879239797592163,
+      "learning_rate": 0.0009845624347566433,
+      "loss": 3.1871,
+      "step": 70
+    },
+    {
+      "epoch": 0.14678899082568808,
+      "grad_norm": 0.22828316688537598,
+      "learning_rate": 0.0009761811727013547,
+      "loss": 3.1625,
+      "step": 80
+    },
+    {
+      "epoch": 0.1651376146788991,
+      "grad_norm": 0.1625080108642578,
+      "learning_rate": 0.0009660421601067666,
+      "loss": 3.1388,
+      "step": 90
+    },
+    {
+      "epoch": 0.1834862385321101,
+      "grad_norm": 0.1556120365858078,
+      "learning_rate": 0.0009541828235992926,
+      "loss": 3.1236,
+      "step": 100
+    },
+    {
+      "epoch": 0.2018348623853211,
+      "grad_norm": 0.251715749502182,
+      "learning_rate": 0.0009406469401200151,
+      "loss": 3.0947,
+      "step": 110
+    },
+    {
+      "epoch": 0.22018348623853212,
+      "grad_norm": 0.18998011946678162,
+      "learning_rate": 0.0009254844753287492,
+      "loss": 3.0858,
+      "step": 120
+    },
+    {
+      "epoch": 0.23853211009174313,
+      "grad_norm": 0.2999618649482727,
+      "learning_rate": 0.0009087513991633924,
+      "loss": 3.0508,
+      "step": 130
+    },
+    {
+      "epoch": 0.25688073394495414,
+      "grad_norm": 0.2752837538719177,
+      "learning_rate": 0.0008905094792353917,
+      "loss": 3.0388,
+      "step": 140
+    },
+    {
+      "epoch": 0.27522935779816515,
+      "grad_norm": 0.2341361939907074,
+      "learning_rate": 0.0008708260528239789,
+      "loss": 3.0322,
+      "step": 150
+    },
+    {
+      "epoch": 0.29357798165137616,
+      "grad_norm": 0.24553282558918,
+      "learning_rate": 0.0008497737783108237,
+      "loss": 3.0146,
+      "step": 160
+    },
+    {
+      "epoch": 0.3119266055045872,
+      "grad_norm": 0.24441348016262054,
+      "learning_rate": 0.0008274303669726426,
+      "loss": 3.0081,
+      "step": 170
+    },
+    {
+      "epoch": 0.3302752293577982,
+      "grad_norm": 0.22537152469158173,
+      "learning_rate": 0.0008038782961218137,
+      "loss": 2.9923,
+      "step": 180
+    },
+    {
+      "epoch": 0.3486238532110092,
+      "grad_norm": 0.2054811716079712,
+      "learning_rate": 0.000779204504653894,
+      "loss": 2.9839,
+      "step": 190
+    },
+    {
+      "epoch": 0.3669724770642202,
+      "grad_norm": 0.21457339823246002,
+      "learning_rate": 0.0007535000721258791,
+      "loss": 2.9627,
+      "step": 200
+    },
+    {
+      "epoch": 0.3853211009174312,
+      "grad_norm": 0.18931026756763458,
+      "learning_rate": 0.00072685988254984,
+      "loss": 2.9521,
+      "step": 210
+    },
+    {
+      "epoch": 0.4036697247706422,
+      "grad_norm": 0.2476811408996582,
+      "learning_rate": 0.0006993822741429906,
+      "loss": 2.9592,
+      "step": 220
+    },
+    {
+      "epoch": 0.42201834862385323,
+      "grad_norm": 0.18029621243476868,
+      "learning_rate": 0.0006711686763270818,
+      "loss": 2.9574,
+      "step": 230
+    },
+    {
+      "epoch": 0.44036697247706424,
+      "grad_norm": 0.16882188618183136,
+      "learning_rate": 0.0006423232353170831,
+      "loss": 2.9399,
+      "step": 240
+    },
+    {
+      "epoch": 0.45871559633027525,
+      "grad_norm": 0.2525669038295746,
+      "learning_rate": 0.0006129524296812335,
+      "loss": 2.9255,
+      "step": 250
+    },
+    {
+      "epoch": 0.47706422018348627,
+      "grad_norm": 0.18064500391483307,
+      "learning_rate": 0.0005831646772915651,
+      "loss": 2.9364,
+      "step": 260
+    },
+    {
+      "epoch": 0.4954128440366973,
+      "grad_norm": 0.2538670301437378,
+      "learning_rate": 0.0005530699351157809,
+      "loss": 2.9157,
+      "step": 270
+    },
+    {
+      "epoch": 0.5137614678899083,
+      "grad_norm": 0.22056764364242554,
+      "learning_rate": 0.0005227792933277943,
+      "loss": 2.9202,
+      "step": 280
+    },
+    {
+      "epoch": 0.5321100917431193,
+      "grad_norm": 0.18312951922416687,
+      "learning_rate": 0.0004924045652352092,
+      "loss": 2.8985,
+      "step": 290
+    },
+    {
+      "epoch": 0.5504587155963303,
+      "grad_norm": 0.24101364612579346,
+      "learning_rate": 0.0004620578745374633,
+      "loss": 2.8982,
+      "step": 300
+    },
+    {
+      "epoch": 0.5688073394495413,
+      "grad_norm": 0.2000555694103241,
+      "learning_rate": 0.0004318512414382059,
+      "loss": 2.9063,
+      "step": 310
+    },
+    {
+      "epoch": 0.5871559633027523,
+      "grad_norm": 0.2163739800453186,
+      "learning_rate": 0.0004018961691397148,
+      "loss": 2.9018,
+      "step": 320
+    },
+    {
+      "epoch": 0.6055045871559633,
+      "grad_norm": 0.17193381488323212,
+      "learning_rate": 0.0003723032322457458,
+      "loss": 2.8851,
+      "step": 330
+    },
+    {
+      "epoch": 0.6238532110091743,
+      "grad_norm": 0.2436598241329193,
+      "learning_rate": 0.0003431816685921625,
+      "loss": 2.8867,
+      "step": 340
+    },
+    {
+      "epoch": 0.6422018348623854,
+      "grad_norm": 0.23022480309009552,
+      "learning_rate": 0.0003146389760120416,
+      "loss": 2.8917,
+      "step": 350
+    },
+    {
+      "epoch": 0.6605504587155964,
+      "grad_norm": 0.25586044788360596,
+      "learning_rate": 0.00028678051552373485,
+      "loss": 2.8933,
+      "step": 360
+    },
+    {
+      "epoch": 0.6788990825688074,
+      "grad_norm": 0.1990855187177658,
+      "learning_rate": 0.0002597091224066581,
+      "loss": 2.8827,
+      "step": 370
+    },
+    {
+      "epoch": 0.6972477064220184,
+      "grad_norm": 0.17142201960086823,
+      "learning_rate": 0.0002335247266004629,
+      "loss": 2.8947,
+      "step": 380
+    },
+    {
+      "epoch": 0.7155963302752294,
+      "grad_norm": 0.19106601178646088,
+      "learning_rate": 0.000208323983828828,
+      "loss": 2.8799,
+      "step": 390
+    },
+    {
+      "epoch": 0.7339449541284404,
+      "grad_norm": 0.18044452369213104,
+      "learning_rate": 0.0001841999188095224,
+      "loss": 2.8736,
+      "step": 400
+    },
+    {
+      "epoch": 0.7522935779816514,
+      "grad_norm": 0.1736108958721161,
+      "learning_rate": 0.00016124158186777676,
+      "loss": 2.8746,
+      "step": 410
+    },
+    {
+      "epoch": 0.7706422018348624,
+      "grad_norm": 0.19438888132572174,
+      "learning_rate": 0.0001395337202205211,
+      "loss": 2.8704,
+      "step": 420
+    },
+    {
+      "epoch": 0.7889908256880734,
+      "grad_norm": 0.15200117230415344,
+      "learning_rate": 0.00011915646514489292,
+      "loss": 2.8733,
+      "step": 430
+    },
+    {
+      "epoch": 0.8073394495412844,
+      "grad_norm": 0.21938292682170868,
+      "learning_rate": 0.00010018503618578818,
+      "loss": 2.881,
+      "step": 440
+    },
+    {
+      "epoch": 0.8256880733944955,
+      "grad_norm": 0.16492417454719543,
+      "learning_rate": 8.268946349432583e-05,
+      "loss": 2.8779,
+      "step": 450
+    },
+    {
+      "epoch": 0.8440366972477065,
+      "grad_norm": 0.1572408676147461,
+      "learning_rate": 6.673432932217117e-05,
+      "loss": 2.8661,
+      "step": 460
+    },
+    {
+      "epoch": 0.8623853211009175,
+      "grad_norm": 0.2064693421125412,
+      "learning_rate": 5.2378529625954684e-05,
+      "loss": 2.8563,
+      "step": 470
+    },
+    {
+      "epoch": 0.8807339449541285,
+      "grad_norm": 0.17086005210876465,
+      "learning_rate": 3.9675056661785556e-05,
+      "loss": 2.8636,
+      "step": 480
+    },
+    {
+      "epoch": 0.8990825688073395,
+      "grad_norm": 0.1528938263654709,
+      "learning_rate": 2.8670803372379538e-05,
+      "loss": 2.871,
+      "step": 490
+    },
+    {
+      "epoch": 0.9174311926605505,
+      "grad_norm": 0.17105746269226074,
+      "learning_rate": 1.9406390288876586e-05,
+      "loss": 2.8788,
+      "step": 500
+    },
+    {
+      "epoch": 0.9357798165137615,
+      "grad_norm": 0.176364928483963,
+      "learning_rate": 1.191601558631461e-05,
+      "loss": 2.8655,
+      "step": 510
+    },
+    {
+      "epoch": 0.9541284403669725,
+      "grad_norm": 0.14428308606147766,
+      "learning_rate": 6.2273288462559305e-06,
+      "loss": 2.861,
+      "step": 520
+    },
+    {
+      "epoch": 0.9724770642201835,
+      "grad_norm": 0.15799862146377563,
+      "learning_rate": 2.361328992552314e-06,
+      "loss": 2.8653,
+      "step": 530
+    },
+    {
+      "epoch": 0.9908256880733946,
+      "grad_norm": 0.15909910202026367,
+      "learning_rate": 3.322867770044202e-07,
+      "loss": 2.8651,
+      "step": 540
+    },
+    {
+      "epoch": 1.0,
+      "step": 545,
+      "total_flos": 2.8922594581282816e+17,
+      "train_loss": 3.026311083452417,
+      "train_runtime": 5259.6999,
+      "train_samples_per_second": 106.114,
+      "train_steps_per_second": 0.104
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 545,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.8922594581282816e+17,
+  "train_batch_size": 128,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/mc1_b32_224_400m/finetune/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

continuous/qwen25/mc1_b32_224_400m/finetune/config.json

@@ -0,0 +1,57 @@
+{
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "LlavaQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "image_aspect_ratio": "square",
+  "image_crop_resolution": null,
+  "image_grid_pinpoints": null,
+  "image_split_resolution": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 768,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": 0.0001,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "open_clip_hub:ViT-B-32",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/mc1/mc1_b32_224_400m.pt",
+  "vocab_size": 151936
+}

continuous/qwen25/mc1_b32_224_400m/finetune/generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.51.3"
+}

continuous/qwen25/mc1_b32_224_400m/finetune/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b8fa5876b193e68c46810d453f4df7db2447e3a0d9e2700735e2cd9606210a84
+size 3744103520

continuous/qwen25/mc1_b32_224_400m/finetune/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
+  }
+}

continuous/qwen25/mc1_b32_224_400m/finetune/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

continuous/qwen25/mc1_b32_224_400m/finetune/tokenizer_config.json

@@ -0,0 +1,209 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

continuous/qwen25/mc1_b32_224_400m/finetune/trainer_state.json

@@ -0,0 +1,1856 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9999037443449803,
+  "eval_steps": 500,
+  "global_step": 2597,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.003850226200789296,
+      "grad_norm": 14.453398704528809,
+      "learning_rate": 1.153846153846154e-06,
+      "loss": 1.8202,
+      "step": 10
+    },
+    {
+      "epoch": 0.007700452401578592,
+      "grad_norm": 7.41076135635376,
+      "learning_rate": 2.435897435897436e-06,
+      "loss": 1.5345,
+      "step": 20
+    },
+    {
+      "epoch": 0.01155067860236789,
+      "grad_norm": 3.484862804412842,
+      "learning_rate": 3.7179487179487184e-06,
+      "loss": 1.1456,
+      "step": 30
+    },
+    {
+      "epoch": 0.015400904803157185,
+      "grad_norm": 0.7017897367477417,
+      "learning_rate": 5e-06,
+      "loss": 0.9791,
+      "step": 40
+    },
+    {
+      "epoch": 0.019251131003946483,
+      "grad_norm": 0.3567080497741699,
+      "learning_rate": 6.282051282051282e-06,
+      "loss": 0.944,
+      "step": 50
+    },
+    {
+      "epoch": 0.02310135720473578,
+      "grad_norm": 0.3024784028530121,
+      "learning_rate": 7.564102564102564e-06,
+      "loss": 0.8962,
+      "step": 60
+    },
+    {
+      "epoch": 0.026951583405525073,
+      "grad_norm": 0.24065986275672913,
+      "learning_rate": 8.846153846153847e-06,
+      "loss": 0.8681,
+      "step": 70
+    },
+    {
+      "epoch": 0.03080180960631437,
+      "grad_norm": 0.238923117518425,
+      "learning_rate": 9.999996111488723e-06,
+      "loss": 0.8756,
+      "step": 80
+    },
+    {
+      "epoch": 0.034652035807103666,
+      "grad_norm": 0.21641287207603455,
+      "learning_rate": 9.999529497453782e-06,
+      "loss": 0.8546,
+      "step": 90
+    },
+    {
+      "epoch": 0.038502262007892966,
+      "grad_norm": 0.23486867547035217,
+      "learning_rate": 9.998285264324193e-06,
+      "loss": 0.8352,
+      "step": 100
+    },
+    {
+      "epoch": 0.04235248820868226,
+      "grad_norm": 0.20836427807807922,
+      "learning_rate": 9.996263605626052e-06,
+      "loss": 0.8232,
+      "step": 110
+    },
+    {
+      "epoch": 0.04620271440947156,
+      "grad_norm": 0.22640715539455414,
+      "learning_rate": 9.99346483580503e-06,
+      "loss": 0.8355,
+      "step": 120
+    },
+    {
+      "epoch": 0.05005294061026085,
+      "grad_norm": 0.20717664062976837,
+      "learning_rate": 9.989889390177467e-06,
+      "loss": 0.8272,
+      "step": 130
+    },
+    {
+      "epoch": 0.053903166811050146,
+      "grad_norm": 0.2373485118150711,
+      "learning_rate": 9.985537824862647e-06,
+      "loss": 0.827,
+      "step": 140
+    },
+    {
+      "epoch": 0.057753393011839446,
+      "grad_norm": 0.2416808307170868,
+      "learning_rate": 9.980410816696317e-06,
+      "loss": 0.8098,
+      "step": 150
+    },
+    {
+      "epoch": 0.06160361921262874,
+      "grad_norm": 0.22813698649406433,
+      "learning_rate": 9.97450916312541e-06,
+      "loss": 0.8158,
+      "step": 160
+    },
+    {
+      "epoch": 0.06545384541341803,
+      "grad_norm": 0.2023065835237503,
+      "learning_rate": 9.967833782084004e-06,
+      "loss": 0.8178,
+      "step": 170
+    },
+    {
+      "epoch": 0.06930407161420733,
+      "grad_norm": 0.2302059382200241,
+      "learning_rate": 9.960385711850554e-06,
+      "loss": 0.8446,
+      "step": 180
+    },
+    {
+      "epoch": 0.07315429781499663,
+      "grad_norm": 0.22350580990314484,
+      "learning_rate": 9.952166110886394e-06,
+      "loss": 0.835,
+      "step": 190
+    },
+    {
+      "epoch": 0.07700452401578593,
+      "grad_norm": 0.25167757272720337,
+      "learning_rate": 9.943176257655567e-06,
+      "loss": 0.8293,
+      "step": 200
+    },
+    {
+      "epoch": 0.08085475021657522,
+      "grad_norm": 0.2166137993335724,
+      "learning_rate": 9.933417550425952e-06,
+      "loss": 0.8181,
+      "step": 210
+    },
+    {
+      "epoch": 0.08470497641736452,
+      "grad_norm": 0.23414312303066254,
+      "learning_rate": 9.9228915070518e-06,
+      "loss": 0.8079,
+      "step": 220
+    },
+    {
+      "epoch": 0.08855520261815382,
+      "grad_norm": 0.2291792631149292,
+      "learning_rate": 9.911599764737632e-06,
+      "loss": 0.8103,
+      "step": 230
+    },
+    {
+      "epoch": 0.09240542881894312,
+      "grad_norm": 0.23818933963775635,
+      "learning_rate": 9.899544079783606e-06,
+      "loss": 0.8041,
+      "step": 240
+    },
+    {
+      "epoch": 0.0962556550197324,
+      "grad_norm": 0.2235424816608429,
+      "learning_rate": 9.886726327312338e-06,
+      "loss": 0.8192,
+      "step": 250
+    },
+    {
+      "epoch": 0.1001058812205217,
+      "grad_norm": 0.21465884149074554,
+      "learning_rate": 9.87314850097724e-06,
+      "loss": 0.8183,
+      "step": 260
+    },
+    {
+      "epoch": 0.103956107421311,
+      "grad_norm": 0.21066448092460632,
+      "learning_rate": 9.858812712652452e-06,
+      "loss": 0.8073,
+      "step": 270
+    },
+    {
+      "epoch": 0.10780633362210029,
+      "grad_norm": 0.20374567806720734,
+      "learning_rate": 9.843721192104341e-06,
+      "loss": 0.8074,
+      "step": 280
+    },
+    {
+      "epoch": 0.11165655982288959,
+      "grad_norm": 0.23205679655075073,
+      "learning_rate": 9.827876286644698e-06,
+      "loss": 0.8274,
+      "step": 290
+    },
+    {
+      "epoch": 0.11550678602367889,
+      "grad_norm": 0.2049105316400528,
+      "learning_rate": 9.811280460765643e-06,
+      "loss": 0.787,
+      "step": 300
+    },
+    {
+      "epoch": 0.11935701222446819,
+      "grad_norm": 0.2175522893667221,
+      "learning_rate": 9.793936295756292e-06,
+      "loss": 0.7983,
+      "step": 310
+    },
+    {
+      "epoch": 0.12320723842525748,
+      "grad_norm": 0.22981640696525574,
+      "learning_rate": 9.775846489301279e-06,
+      "loss": 0.8018,
+      "step": 320
+    },
+    {
+      "epoch": 0.12705746462604678,
+      "grad_norm": 0.22266274690628052,
+      "learning_rate": 9.757013855061157e-06,
+      "loss": 0.8002,
+      "step": 330
+    },
+    {
+      "epoch": 0.13090769082683607,
+      "grad_norm": 0.21400760114192963,
+      "learning_rate": 9.737441322234758e-06,
+      "loss": 0.8128,
+      "step": 340
+    },
+    {
+      "epoch": 0.13475791702762538,
+      "grad_norm": 0.22386080026626587,
+      "learning_rate": 9.717131935103605e-06,
+      "loss": 0.811,
+      "step": 350
+    },
+    {
+      "epoch": 0.13860814322841467,
+      "grad_norm": 0.21584483981132507,
+      "learning_rate": 9.696088852558398e-06,
+      "loss": 0.7979,
+      "step": 360
+    },
+    {
+      "epoch": 0.14245836942920398,
+      "grad_norm": 0.21524269878864288,
+      "learning_rate": 9.674315347607686e-06,
+      "loss": 0.8225,
+      "step": 370
+    },
+    {
+      "epoch": 0.14630859562999327,
+      "grad_norm": 0.21745565533638,
+      "learning_rate": 9.651814806868796e-06,
+      "loss": 0.7883,
+      "step": 380
+    },
+    {
+      "epoch": 0.15015882183078255,
+      "grad_norm": 0.22399549186229706,
+      "learning_rate": 9.62859073004107e-06,
+      "loss": 0.7928,
+      "step": 390
+    },
+    {
+      "epoch": 0.15400904803157187,
+      "grad_norm": 0.2015438824892044,
+      "learning_rate": 9.604646729361547e-06,
+      "loss": 0.7999,
+      "step": 400
+    },
+    {
+      "epoch": 0.15785927423236115,
+      "grad_norm": 0.21006053686141968,
+      "learning_rate": 9.579986529043099e-06,
+      "loss": 0.7917,
+      "step": 410
+    },
+    {
+      "epoch": 0.16170950043315044,
+      "grad_norm": 0.21839219331741333,
+      "learning_rate": 9.554613964695189e-06,
+      "loss": 0.8038,
+      "step": 420
+    },
+    {
+      "epoch": 0.16555972663393975,
+      "grad_norm": 0.2181902676820755,
+      "learning_rate": 9.528532982727281e-06,
+      "loss": 0.7945,
+      "step": 430
+    },
+    {
+      "epoch": 0.16940995283472904,
+      "grad_norm": 0.20690493285655975,
+      "learning_rate": 9.501747639735018e-06,
+      "loss": 0.7938,
+      "step": 440
+    },
+    {
+      "epoch": 0.17326017903551832,
+      "grad_norm": 0.21885763108730316,
+      "learning_rate": 9.474262101869272e-06,
+      "loss": 0.7824,
+      "step": 450
+    },
+    {
+      "epoch": 0.17711040523630764,
+      "grad_norm": 0.21913309395313263,
+      "learning_rate": 9.446080644188141e-06,
+      "loss": 0.7867,
+      "step": 460
+    },
+    {
+      "epoch": 0.18096063143709692,
+      "grad_norm": 0.21082688868045807,
+      "learning_rate": 9.417207649992017e-06,
+      "loss": 0.7954,
+      "step": 470
+    },
+    {
+      "epoch": 0.18481085763788624,
+      "grad_norm": 0.21544964611530304,
+      "learning_rate": 9.387647610141816e-06,
+      "loss": 0.7946,
+      "step": 480
+    },
+    {
+      "epoch": 0.18866108383867553,
+      "grad_norm": 0.20054034888744354,
+      "learning_rate": 9.35740512236047e-06,
+      "loss": 0.7808,
+      "step": 490
+    },
+    {
+      "epoch": 0.1925113100394648,
+      "grad_norm": 0.19608309864997864,
+      "learning_rate": 9.326484890517809e-06,
+      "loss": 0.7896,
+      "step": 500
+    },
+    {
+      "epoch": 0.19636153624025413,
+      "grad_norm": 0.22134529054164886,
+      "learning_rate": 9.294891723898926e-06,
+      "loss": 0.7812,
+      "step": 510
+    },
+    {
+      "epoch": 0.2002117624410434,
+      "grad_norm": 0.231044739484787,
+      "learning_rate": 9.262630536456153e-06,
+      "loss": 0.811,
+      "step": 520
+    },
+    {
+      "epoch": 0.2040619886418327,
+      "grad_norm": 0.23448783159255981,
+      "learning_rate": 9.229706346044749e-06,
+      "loss": 0.7855,
+      "step": 530
+    },
+    {
+      "epoch": 0.207912214842622,
+      "grad_norm": 0.23248343169689178,
+      "learning_rate": 9.196124273642422e-06,
+      "loss": 0.7962,
+      "step": 540
+    },
+    {
+      "epoch": 0.2117624410434113,
+      "grad_norm": 0.21395684778690338,
+      "learning_rate": 9.161889542552842e-06,
+      "loss": 0.7986,
+      "step": 550
+    },
+    {
+      "epoch": 0.21561266724420058,
+      "grad_norm": 0.2070465236902237,
+      "learning_rate": 9.127007477593193e-06,
+      "loss": 0.8029,
+      "step": 560
+    },
+    {
+      "epoch": 0.2194628934449899,
+      "grad_norm": 0.23257474601268768,
+      "learning_rate": 9.091483504265974e-06,
+      "loss": 0.8143,
+      "step": 570
+    },
+    {
+      "epoch": 0.22331311964577918,
+      "grad_norm": 0.22393366694450378,
+      "learning_rate": 9.055323147915116e-06,
+      "loss": 0.7912,
+      "step": 580
+    },
+    {
+      "epoch": 0.2271633458465685,
+      "grad_norm": 0.22197961807250977,
+      "learning_rate": 9.018532032866585e-06,
+      "loss": 0.8033,
+      "step": 590
+    },
+    {
+      "epoch": 0.23101357204735778,
+      "grad_norm": 0.21754609048366547,
+      "learning_rate": 8.981115881553579e-06,
+      "loss": 0.7941,
+      "step": 600
+    },
+    {
+      "epoch": 0.23486379824814707,
+      "grad_norm": 0.22558744251728058,
+      "learning_rate": 8.943080513626476e-06,
+      "loss": 0.8144,
+      "step": 610
+    },
+    {
+      "epoch": 0.23871402444893638,
+      "grad_norm": 0.2151433378458023,
+      "learning_rate": 8.904431845047641e-06,
+      "loss": 0.7978,
+      "step": 620
+    },
+    {
+      "epoch": 0.24256425064972567,
+      "grad_norm": 0.23146066069602966,
+      "learning_rate": 8.86517588717128e-06,
+      "loss": 0.8149,
+      "step": 630
+    },
+    {
+      "epoch": 0.24641447685051496,
+      "grad_norm": 0.2081734836101532,
+      "learning_rate": 8.82531874580844e-06,
+      "loss": 0.8041,
+      "step": 640
+    },
+    {
+      "epoch": 0.25026470305130427,
+      "grad_norm": 0.19859153032302856,
+      "learning_rate": 8.784866620277314e-06,
+      "loss": 0.7626,
+      "step": 650
+    },
+    {
+      "epoch": 0.25411492925209356,
+      "grad_norm": 0.23843379318714142,
+      "learning_rate": 8.743825802439018e-06,
+      "loss": 0.7983,
+      "step": 660
+    },
+    {
+      "epoch": 0.25796515545288284,
+      "grad_norm": 0.22632353007793427,
+      "learning_rate": 8.702202675718956e-06,
+      "loss": 0.8024,
+      "step": 670
+    },
+    {
+      "epoch": 0.26181538165367213,
+      "grad_norm": 0.20676055550575256,
+      "learning_rate": 8.660003714113959e-06,
+      "loss": 0.7736,
+      "step": 680
+    },
+    {
+      "epoch": 0.26566560785446147,
+      "grad_norm": 0.21125726401805878,
+      "learning_rate": 8.617235481185323e-06,
+      "loss": 0.7969,
+      "step": 690
+    },
+    {
+      "epoch": 0.26951583405525076,
+      "grad_norm": 0.21587277948856354,
+      "learning_rate": 8.573904629037931e-06,
+      "loss": 0.8014,
+      "step": 700
+    },
+    {
+      "epoch": 0.27336606025604004,
+      "grad_norm": 0.21729333698749542,
+      "learning_rate": 8.530017897285585e-06,
+      "loss": 0.7829,
+      "step": 710
+    },
+    {
+      "epoch": 0.27721628645682933,
+      "grad_norm": 0.19784851372241974,
+      "learning_rate": 8.485582112002746e-06,
+      "loss": 0.7874,
+      "step": 720
+    },
+    {
+      "epoch": 0.2810665126576186,
+      "grad_norm": 0.2110746055841446,
+      "learning_rate": 8.440604184662811e-06,
+      "loss": 0.7916,
+      "step": 730
+    },
+    {
+      "epoch": 0.28491673885840796,
+      "grad_norm": 0.22429752349853516,
+      "learning_rate": 8.395091111063118e-06,
+      "loss": 0.7832,
+      "step": 740
+    },
+    {
+      "epoch": 0.28876696505919724,
+      "grad_norm": 0.20070883631706238,
+      "learning_rate": 8.349049970236822e-06,
+      "loss": 0.7806,
+      "step": 750
+    },
+    {
+      "epoch": 0.29261719125998653,
+      "grad_norm": 0.2123759388923645,
+      "learning_rate": 8.302487923351842e-06,
+      "loss": 0.7735,
+      "step": 760
+    },
+    {
+      "epoch": 0.2964674174607758,
+      "grad_norm": 0.20814821124076843,
+      "learning_rate": 8.255412212597021e-06,
+      "loss": 0.7907,
+      "step": 770
+    },
+    {
+      "epoch": 0.3003176436615651,
+      "grad_norm": 0.20113970339298248,
+      "learning_rate": 8.207830160055688e-06,
+      "loss": 0.7867,
+      "step": 780
+    },
+    {
+      "epoch": 0.3041678698623544,
+      "grad_norm": 0.22357168793678284,
+      "learning_rate": 8.159749166566786e-06,
+      "loss": 0.7743,
+      "step": 790
+    },
+    {
+      "epoch": 0.30801809606314373,
+      "grad_norm": 0.2056380659341812,
+      "learning_rate": 8.11117671057377e-06,
+      "loss": 0.8019,
+      "step": 800
+    },
+    {
+      "epoch": 0.311868322263933,
+      "grad_norm": 0.22288499772548676,
+      "learning_rate": 8.062120346961404e-06,
+      "loss": 0.7943,
+      "step": 810
+    },
+    {
+      "epoch": 0.3157185484647223,
+      "grad_norm": 0.2067280262708664,
+      "learning_rate": 8.012587705880705e-06,
+      "loss": 0.7886,
+      "step": 820
+    },
+    {
+      "epoch": 0.3195687746655116,
+      "grad_norm": 0.2079402655363083,
+      "learning_rate": 7.962586491562146e-06,
+      "loss": 0.7742,
+      "step": 830
+    },
+    {
+      "epoch": 0.3234190008663009,
+      "grad_norm": 0.2429584115743637,
+      "learning_rate": 7.912124481117357e-06,
+      "loss": 0.7786,
+      "step": 840
+    },
+    {
+      "epoch": 0.3272692270670902,
+      "grad_norm": 0.20473897457122803,
+      "learning_rate": 7.861209523329494e-06,
+      "loss": 0.7888,
+      "step": 850
+    },
+    {
+      "epoch": 0.3311194532678795,
+      "grad_norm": 0.2175380438566208,
+      "learning_rate": 7.809849537432432e-06,
+      "loss": 0.7759,
+      "step": 860
+    },
+    {
+      "epoch": 0.3349696794686688,
+      "grad_norm": 0.20772261917591095,
+      "learning_rate": 7.758052511879035e-06,
+      "loss": 0.784,
+      "step": 870
+    },
+    {
+      "epoch": 0.3388199056694581,
+      "grad_norm": 0.19892674684524536,
+      "learning_rate": 7.705826503098642e-06,
+      "loss": 0.7786,
+      "step": 880
+    },
+    {
+      "epoch": 0.34267013187024736,
+      "grad_norm": 0.2031731754541397,
+      "learning_rate": 7.653179634243979e-06,
+      "loss": 0.7782,
+      "step": 890
+    },
+    {
+      "epoch": 0.34652035807103665,
+      "grad_norm": 0.1951494812965393,
+      "learning_rate": 7.600120093927696e-06,
+      "loss": 0.7838,
+      "step": 900
+    },
+    {
+      "epoch": 0.350370584271826,
+      "grad_norm": 0.2165498584508896,
+      "learning_rate": 7.546656134948729e-06,
+      "loss": 0.7698,
+      "step": 910
+    },
+    {
+      "epoch": 0.3542208104726153,
+      "grad_norm": 0.21966131031513214,
+      "learning_rate": 7.492796073008661e-06,
+      "loss": 0.7705,
+      "step": 920
+    },
+    {
+      "epoch": 0.35807103667340456,
+      "grad_norm": 0.22888566553592682,
+      "learning_rate": 7.438548285418326e-06,
+      "loss": 0.7596,
+      "step": 930
+    },
+    {
+      "epoch": 0.36192126287419385,
+      "grad_norm": 0.23172657191753387,
+      "learning_rate": 7.3839212097948e-06,
+      "loss": 0.7925,
+      "step": 940
+    },
+    {
+      "epoch": 0.36577148907498314,
+      "grad_norm": 0.22153106331825256,
+      "learning_rate": 7.328923342749039e-06,
+      "loss": 0.8059,
+      "step": 950
+    },
+    {
+      "epoch": 0.3696217152757725,
+      "grad_norm": 0.2266840785741806,
+      "learning_rate": 7.273563238564319e-06,
+      "loss": 0.7924,
+      "step": 960
+    },
+    {
+      "epoch": 0.37347194147656176,
+      "grad_norm": 0.21695195138454437,
+      "learning_rate": 7.217849507865724e-06,
+      "loss": 0.773,
+      "step": 970
+    },
+    {
+      "epoch": 0.37732216767735105,
+      "grad_norm": 0.21033582091331482,
+      "learning_rate": 7.161790816280858e-06,
+      "loss": 0.777,
+      "step": 980
+    },
+    {
+      "epoch": 0.38117239387814034,
+      "grad_norm": 0.22721906006336212,
+      "learning_rate": 7.1053958830920125e-06,
+      "loss": 0.7639,
+      "step": 990
+    },
+    {
+      "epoch": 0.3850226200789296,
+      "grad_norm": 0.20851801335811615,
+      "learning_rate": 7.0486734798799715e-06,
+      "loss": 0.7768,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3888728462797189,
+      "grad_norm": 0.2072061449289322,
+      "learning_rate": 6.991632429159708e-06,
+      "loss": 0.8018,
+      "step": 1010
+    },
+    {
+      "epoch": 0.39272307248050825,
+      "grad_norm": 0.22861555218696594,
+      "learning_rate": 6.934281603008132e-06,
+      "loss": 0.795,
+      "step": 1020
+    },
+    {
+      "epoch": 0.39657329868129754,
+      "grad_norm": 0.2071608603000641,
+      "learning_rate": 6.876629921684143e-06,
+      "loss": 0.7917,
+      "step": 1030
+    },
+    {
+      "epoch": 0.4004235248820868,
+      "grad_norm": 0.21600596606731415,
+      "learning_rate": 6.818686352241199e-06,
+      "loss": 0.7831,
+      "step": 1040
+    },
+    {
+      "epoch": 0.4042737510828761,
+      "grad_norm": 0.19345609843730927,
+      "learning_rate": 6.760459907132573e-06,
+      "loss": 0.7802,
+      "step": 1050
+    },
+    {
+      "epoch": 0.4081239772836654,
+      "grad_norm": 0.19670641422271729,
+      "learning_rate": 6.701959642809589e-06,
+      "loss": 0.7857,
+      "step": 1060
+    },
+    {
+      "epoch": 0.41197420348445474,
+      "grad_norm": 0.21387438476085663,
+      "learning_rate": 6.643194658312987e-06,
+      "loss": 0.7755,
+      "step": 1070
+    },
+    {
+      "epoch": 0.415824429685244,
+      "grad_norm": 0.2051989883184433,
+      "learning_rate": 6.584174093857676e-06,
+      "loss": 0.8006,
+      "step": 1080
+    },
+    {
+      "epoch": 0.4196746558860333,
+      "grad_norm": 0.23363132774829865,
+      "learning_rate": 6.5249071294110735e-06,
+      "loss": 0.7823,
+      "step": 1090
+    },
+    {
+      "epoch": 0.4235248820868226,
+      "grad_norm": 0.21509318053722382,
+      "learning_rate": 6.465402983265274e-06,
+      "loss": 0.7962,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4273751082876119,
+      "grad_norm": 0.222013458609581,
+      "learning_rate": 6.405670910603243e-06,
+      "loss": 0.7934,
+      "step": 1110
+    },
+    {
+      "epoch": 0.43122533448840117,
+      "grad_norm": 0.23090173304080963,
+      "learning_rate": 6.345720202059291e-06,
+      "loss": 0.7892,
+      "step": 1120
+    },
+    {
+      "epoch": 0.4350755606891905,
+      "grad_norm": 0.21809323132038116,
+      "learning_rate": 6.285560182274008e-06,
+      "loss": 0.7668,
+      "step": 1130
+    },
+    {
+      "epoch": 0.4389257868899798,
+      "grad_norm": 0.2095639705657959,
+      "learning_rate": 6.225200208443936e-06,
+      "loss": 0.7882,
+      "step": 1140
+    },
+    {
+      "epoch": 0.4427760130907691,
+      "grad_norm": 0.21081645786762238,
+      "learning_rate": 6.16464966886616e-06,
+      "loss": 0.7809,
+      "step": 1150
+    },
+    {
+      "epoch": 0.44662623929155837,
+      "grad_norm": 0.2454957216978073,
+      "learning_rate": 6.103917981478065e-06,
+      "loss": 0.8031,
+      "step": 1160
+    },
+    {
+      "epoch": 0.45047646549234766,
+      "grad_norm": 0.21493759751319885,
+      "learning_rate": 6.043014592392489e-06,
+      "loss": 0.8158,
+      "step": 1170
+    },
+    {
+      "epoch": 0.454326691693137,
+      "grad_norm": 0.20422640442848206,
+      "learning_rate": 5.981948974428485e-06,
+      "loss": 0.7875,
+      "step": 1180
+    },
+    {
+      "epoch": 0.4581769178939263,
+      "grad_norm": 0.21610763669013977,
+      "learning_rate": 5.920730625637934e-06,
+      "loss": 0.7686,
+      "step": 1190
+    },
+    {
+      "epoch": 0.46202714409471557,
+      "grad_norm": 0.22047767043113708,
+      "learning_rate": 5.8593690678282355e-06,
+      "loss": 0.7794,
+      "step": 1200
+    },
+    {
+      "epoch": 0.46587737029550486,
+      "grad_norm": 0.22001750767230988,
+      "learning_rate": 5.797873845081301e-06,
+      "loss": 0.7817,
+      "step": 1210
+    },
+    {
+      "epoch": 0.46972759649629414,
+      "grad_norm": 0.21941988170146942,
+      "learning_rate": 5.736254522269075e-06,
+      "loss": 0.7981,
+      "step": 1220
+    },
+    {
+      "epoch": 0.47357782269708343,
+      "grad_norm": 0.20622755587100983,
+      "learning_rate": 5.6745206835658386e-06,
+      "loss": 0.8017,
+      "step": 1230
+    },
+    {
+      "epoch": 0.47742804889787277,
+      "grad_norm": 0.2040691375732422,
+      "learning_rate": 5.612681930957495e-06,
+      "loss": 0.7701,
+      "step": 1240
+    },
+    {
+      "epoch": 0.48127827509866206,
+      "grad_norm": 0.2043808549642563,
+      "learning_rate": 5.550747882748095e-06,
+      "loss": 0.7703,
+      "step": 1250
+    },
+    {
+      "epoch": 0.48512850129945134,
+      "grad_norm": 0.22337813675403595,
+      "learning_rate": 5.4887281720638195e-06,
+      "loss": 0.7774,
+      "step": 1260
+    },
+    {
+      "epoch": 0.48897872750024063,
+      "grad_norm": 0.2393646091222763,
+      "learning_rate": 5.42663244535466e-06,
+      "loss": 0.7623,
+      "step": 1270
+    },
+    {
+      "epoch": 0.4928289537010299,
+      "grad_norm": 0.21621346473693848,
+      "learning_rate": 5.3644703608940205e-06,
+      "loss": 0.7702,
+      "step": 1280
+    },
+    {
+      "epoch": 0.49667917990181926,
+      "grad_norm": 0.21759313344955444,
+      "learning_rate": 5.302251587276491e-06,
+      "loss": 0.7809,
+      "step": 1290
+    },
+    {
+      "epoch": 0.5005294061026085,
+      "grad_norm": 0.2060190737247467,
+      "learning_rate": 5.2399858019140005e-06,
+      "loss": 0.7824,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5043796323033978,
+      "grad_norm": 0.20414775609970093,
+      "learning_rate": 5.177682689530615e-06,
+      "loss": 0.7934,
+      "step": 1310
+    },
+    {
+      "epoch": 0.5082298585041871,
+      "grad_norm": 0.21891161799430847,
+      "learning_rate": 5.11535194065619e-06,
+      "loss": 0.7657,
+      "step": 1320
+    },
+    {
+      "epoch": 0.5120800847049765,
+      "grad_norm": 0.23620161414146423,
+      "learning_rate": 5.053003250119113e-06,
+      "loss": 0.7759,
+      "step": 1330
+    },
+    {
+      "epoch": 0.5159303109057657,
+      "grad_norm": 0.22049732506275177,
+      "learning_rate": 4.9906463155383925e-06,
+      "loss": 0.7799,
+      "step": 1340
+    },
+    {
+      "epoch": 0.519780537106555,
+      "grad_norm": 0.21033774316310883,
+      "learning_rate": 4.9282908358153046e-06,
+      "loss": 0.7722,
+      "step": 1350
+    },
+    {
+      "epoch": 0.5236307633073443,
+      "grad_norm": 0.24146153032779694,
+      "learning_rate": 4.8659465096248345e-06,
+      "loss": 0.7915,
+      "step": 1360
+    },
+    {
+      "epoch": 0.5274809895081336,
+      "grad_norm": 0.2195364534854889,
+      "learning_rate": 4.803623033907169e-06,
+      "loss": 0.7894,
+      "step": 1370
+    },
+    {
+      "epoch": 0.5313312157089229,
+      "grad_norm": 0.2017439901828766,
+      "learning_rate": 4.741330102359441e-06,
+      "loss": 0.7956,
+      "step": 1380
+    },
+    {
+      "epoch": 0.5351814419097122,
+      "grad_norm": 0.2286185324192047,
+      "learning_rate": 4.679077403927996e-06,
+      "loss": 0.7757,
+      "step": 1390
+    },
+    {
+      "epoch": 0.5390316681105015,
+      "grad_norm": 0.21812716126441956,
+      "learning_rate": 4.616874621301373e-06,
+      "loss": 0.7755,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5428818943112907,
+      "grad_norm": 0.22189268469810486,
+      "learning_rate": 4.554731429404293e-06,
+      "loss": 0.7893,
+      "step": 1410
+    },
+    {
+      "epoch": 0.5467321205120801,
+      "grad_norm": 0.22893580794334412,
+      "learning_rate": 4.492657493892822e-06,
+      "loss": 0.7762,
+      "step": 1420
+    },
+    {
+      "epoch": 0.5505823467128694,
+      "grad_norm": 0.2304261028766632,
+      "learning_rate": 4.430662469650999e-06,
+      "loss": 0.7909,
+      "step": 1430
+    },
+    {
+      "epoch": 0.5544325729136587,
+      "grad_norm": 0.2249716818332672,
+      "learning_rate": 4.368755999289118e-06,
+      "loss": 0.793,
+      "step": 1440
+    },
+    {
+      "epoch": 0.558282799114448,
+      "grad_norm": 0.22551806271076202,
+      "learning_rate": 4.306947711643949e-06,
+      "loss": 0.775,
+      "step": 1450
+    },
+    {
+      "epoch": 0.5621330253152372,
+      "grad_norm": 0.2211741954088211,
+      "learning_rate": 4.245247220281068e-06,
+      "loss": 0.7807,
+      "step": 1460
+    },
+    {
+      "epoch": 0.5659832515160266,
+      "grad_norm": 0.21645477414131165,
+      "learning_rate": 4.183664121999587e-06,
+      "loss": 0.7775,
+      "step": 1470
+    },
+    {
+      "epoch": 0.5698334777168159,
+      "grad_norm": 0.22381268441677094,
+      "learning_rate": 4.122207995339474e-06,
+      "loss": 0.7726,
+      "step": 1480
+    },
+    {
+      "epoch": 0.5736837039176051,
+      "grad_norm": 0.2069903314113617,
+      "learning_rate": 4.060888399091734e-06,
+      "loss": 0.7793,
+      "step": 1490
+    },
+    {
+      "epoch": 0.5775339301183945,
+      "grad_norm": 0.20204541087150574,
+      "learning_rate": 3.999714870811639e-06,
+      "loss": 0.7883,
+      "step": 1500
+    },
+    {
+      "epoch": 0.5813841563191837,
+      "grad_norm": 0.21819613873958588,
+      "learning_rate": 3.938696925335275e-06,
+      "loss": 0.7824,
+      "step": 1510
+    },
+    {
+      "epoch": 0.5852343825199731,
+      "grad_norm": 0.21351180970668793,
+      "learning_rate": 3.87784405329962e-06,
+      "loss": 0.7908,
+      "step": 1520
+    },
+    {
+      "epoch": 0.5890846087207623,
+      "grad_norm": 0.21121172606945038,
+      "learning_rate": 3.817165719666382e-06,
+      "loss": 0.7922,
+      "step": 1530
+    },
+    {
+      "epoch": 0.5929348349215516,
+      "grad_norm": 0.2046198546886444,
+      "learning_rate": 3.756671362249841e-06,
+      "loss": 0.7962,
+      "step": 1540
+    },
+    {
+      "epoch": 0.596785061122341,
+      "grad_norm": 0.21660271286964417,
+      "learning_rate": 3.6963703902488907e-06,
+      "loss": 0.7842,
+      "step": 1550
+    },
+    {
+      "epoch": 0.6006352873231302,
+      "grad_norm": 0.21457892656326294,
+      "learning_rate": 3.6362721827835655e-06,
+      "loss": 0.8002,
+      "step": 1560
+    },
+    {
+      "epoch": 0.6044855135239195,
+      "grad_norm": 0.2270359843969345,
+      "learning_rate": 3.576386087436212e-06,
+      "loss": 0.7945,
+      "step": 1570
+    },
+    {
+      "epoch": 0.6083357397247088,
+      "grad_norm": 0.22895868122577667,
+      "learning_rate": 3.5167214187975913e-06,
+      "loss": 0.7849,
+      "step": 1580
+    },
+    {
+      "epoch": 0.6121859659254981,
+      "grad_norm": 0.1918918490409851,
+      "learning_rate": 3.457287457018086e-06,
+      "loss": 0.7859,
+      "step": 1590
+    },
+    {
+      "epoch": 0.6160361921262875,
+      "grad_norm": 0.20588071644306183,
+      "learning_rate": 3.3980934463643003e-06,
+      "loss": 0.7749,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6198864183270767,
+      "grad_norm": 0.20313279330730438,
+      "learning_rate": 3.3391485937812053e-06,
+      "loss": 0.7558,
+      "step": 1610
+    },
+    {
+      "epoch": 0.623736644527866,
+      "grad_norm": 0.21566501259803772,
+      "learning_rate": 3.280462067460115e-06,
+      "loss": 0.7561,
+      "step": 1620
+    },
+    {
+      "epoch": 0.6275868707286553,
+      "grad_norm": 0.2129386067390442,
+      "learning_rate": 3.222042995412669e-06,
+      "loss": 0.7876,
+      "step": 1630
+    },
+    {
+      "epoch": 0.6314370969294446,
+      "grad_norm": 0.18700915575027466,
+      "learning_rate": 3.163900464051089e-06,
+      "loss": 0.7764,
+      "step": 1640
+    },
+    {
+      "epoch": 0.635287323130234,
+      "grad_norm": 0.2170291393995285,
+      "learning_rate": 3.1060435167748826e-06,
+      "loss": 0.7857,
+      "step": 1650
+    },
+    {
+      "epoch": 0.6391375493310232,
+      "grad_norm": 0.21767130494117737,
+      "learning_rate": 3.048481152564253e-06,
+      "loss": 0.7667,
+      "step": 1660
+    },
+    {
+      "epoch": 0.6429877755318125,
+      "grad_norm": 0.2259148210287094,
+      "learning_rate": 2.9912223245804017e-06,
+      "loss": 0.7863,
+      "step": 1670
+    },
+    {
+      "epoch": 0.6468380017326018,
+      "grad_norm": 0.2173956036567688,
+      "learning_rate": 2.934275938772981e-06,
+      "loss": 0.7836,
+      "step": 1680
+    },
+    {
+      "epoch": 0.6506882279333911,
+      "grad_norm": 0.22915667295455933,
+      "learning_rate": 2.8776508524948676e-06,
+      "loss": 0.7727,
+      "step": 1690
+    },
+    {
+      "epoch": 0.6545384541341804,
+      "grad_norm": 0.1951102763414383,
+      "learning_rate": 2.821355873124503e-06,
+      "loss": 0.7741,
+      "step": 1700
+    },
+    {
+      "epoch": 0.6583886803349697,
+      "grad_norm": 0.21835941076278687,
+      "learning_rate": 2.7653997566960144e-06,
+      "loss": 0.7841,
+      "step": 1710
+    },
+    {
+      "epoch": 0.662238906535759,
+      "grad_norm": 0.21918807923793793,
+      "learning_rate": 2.709791206537309e-06,
+      "loss": 0.7747,
+      "step": 1720
+    },
+    {
+      "epoch": 0.6660891327365482,
+      "grad_norm": 0.21681544184684753,
+      "learning_rate": 2.6545388719163747e-06,
+      "loss": 0.7814,
+      "step": 1730
+    },
+    {
+      "epoch": 0.6699393589373376,
+      "grad_norm": 0.21708039939403534,
+      "learning_rate": 2.599651346695979e-06,
+      "loss": 0.7805,
+      "step": 1740
+    },
+    {
+      "epoch": 0.6737895851381268,
+      "grad_norm": 0.20250193774700165,
+      "learning_rate": 2.545137167997007e-06,
+      "loss": 0.7781,
+      "step": 1750
+    },
+    {
+      "epoch": 0.6776398113389162,
+      "grad_norm": 0.20104637742042542,
+      "learning_rate": 2.4910048148705972e-06,
+      "loss": 0.7678,
+      "step": 1760
+    },
+    {
+      "epoch": 0.6814900375397055,
+      "grad_norm": 0.2225777804851532,
+      "learning_rate": 2.4372627069793327e-06,
+      "loss": 0.809,
+      "step": 1770
+    },
+    {
+      "epoch": 0.6853402637404947,
+      "grad_norm": 0.22321894764900208,
+      "learning_rate": 2.3839192032876486e-06,
+      "loss": 0.795,
+      "step": 1780
+    },
+    {
+      "epoch": 0.6891904899412841,
+      "grad_norm": 0.21027490496635437,
+      "learning_rate": 2.3309826007617047e-06,
+      "loss": 0.8059,
+      "step": 1790
+    },
+    {
+      "epoch": 0.6930407161420733,
+      "grad_norm": 0.21006019413471222,
+      "learning_rate": 2.278461133078881e-06,
+      "loss": 0.7683,
+      "step": 1800
+    },
+    {
+      "epoch": 0.6968909423428626,
+      "grad_norm": 0.2214488983154297,
+      "learning_rate": 2.2263629693471258e-06,
+      "loss": 0.7608,
+      "step": 1810
+    },
+    {
+      "epoch": 0.700741168543652,
+      "grad_norm": 0.20429213345050812,
+      "learning_rate": 2.1746962128343414e-06,
+      "loss": 0.7834,
+      "step": 1820
+    },
+    {
+      "epoch": 0.7045913947444412,
+      "grad_norm": 0.20274275541305542,
+      "learning_rate": 2.1234688997080183e-06,
+      "loss": 0.7729,
+      "step": 1830
+    },
+    {
+      "epoch": 0.7084416209452306,
+      "grad_norm": 0.2240271419286728,
+      "learning_rate": 2.0726889977853053e-06,
+      "loss": 0.7795,
+      "step": 1840
+    },
+    {
+      "epoch": 0.7122918471460198,
+      "grad_norm": 0.2250552475452423,
+      "learning_rate": 2.022364405293703e-06,
+      "loss": 0.7887,
+      "step": 1850
+    },
+    {
+      "epoch": 0.7161420733468091,
+      "grad_norm": 0.20937807857990265,
+      "learning_rate": 1.9725029496425814e-06,
+      "loss": 0.7839,
+      "step": 1860
+    },
+    {
+      "epoch": 0.7199922995475985,
+      "grad_norm": 0.19602081179618835,
+      "learning_rate": 1.923112386205727e-06,
+      "loss": 0.7803,
+      "step": 1870
+    },
+    {
+      "epoch": 0.7238425257483877,
+      "grad_norm": 0.20575620234012604,
+      "learning_rate": 1.874200397115079e-06,
+      "loss": 0.7873,
+      "step": 1880
+    },
+    {
+      "epoch": 0.727692751949177,
+      "grad_norm": 0.22233130037784576,
+      "learning_rate": 1.825774590065857e-06,
+      "loss": 0.7706,
+      "step": 1890
+    },
+    {
+      "epoch": 0.7315429781499663,
+      "grad_norm": 0.2197222262620926,
+      "learning_rate": 1.7778424971332813e-06,
+      "loss": 0.7808,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7353932043507556,
+      "grad_norm": 0.2048003077507019,
+      "learning_rate": 1.7304115736010407e-06,
+      "loss": 0.7794,
+      "step": 1910
+    },
+    {
+      "epoch": 0.739243430551545,
+      "grad_norm": 0.21641355752944946,
+      "learning_rate": 1.6834891968017108e-06,
+      "loss": 0.772,
+      "step": 1920
+    },
+    {
+      "epoch": 0.7430936567523342,
+      "grad_norm": 0.21949055790901184,
+      "learning_rate": 1.6370826649692884e-06,
+      "loss": 0.7903,
+      "step": 1930
+    },
+    {
+      "epoch": 0.7469438829531235,
+      "grad_norm": 0.22597335278987885,
+      "learning_rate": 1.5911991961040463e-06,
+      "loss": 0.7954,
+      "step": 1940
+    },
+    {
+      "epoch": 0.7507941091539128,
+      "grad_norm": 0.21523839235305786,
+      "learning_rate": 1.545845926849851e-06,
+      "loss": 0.7813,
+      "step": 1950
+    },
+    {
+      "epoch": 0.7546443353547021,
+      "grad_norm": 0.23007658123970032,
+      "learning_rate": 1.5010299113841397e-06,
+      "loss": 0.8006,
+      "step": 1960
+    },
+    {
+      "epoch": 0.7584945615554913,
+      "grad_norm": 0.2090727686882019,
+      "learning_rate": 1.4567581203207238e-06,
+      "loss": 0.7885,
+      "step": 1970
+    },
+    {
+      "epoch": 0.7623447877562807,
+      "grad_norm": 0.22539106011390686,
+      "learning_rate": 1.413037439625593e-06,
+      "loss": 0.7904,
+      "step": 1980
+    },
+    {
+      "epoch": 0.76619501395707,
+      "grad_norm": 0.22490538656711578,
+      "learning_rate": 1.3698746695458847e-06,
+      "loss": 0.7798,
+      "step": 1990
+    },
+    {
+      "epoch": 0.7700452401578592,
+      "grad_norm": 0.2078322172164917,
+      "learning_rate": 1.3272765235521745e-06,
+      "loss": 0.7733,
+      "step": 2000
+    },
+    {
+      "epoch": 0.7738954663586486,
+      "grad_norm": 0.2081754356622696,
+      "learning_rate": 1.2852496272942877e-06,
+      "loss": 0.7819,
+      "step": 2010
+    },
+    {
+      "epoch": 0.7777456925594378,
+      "grad_norm": 0.19935055077075958,
+      "learning_rate": 1.2438005175707396e-06,
+      "loss": 0.7696,
+      "step": 2020
+    },
+    {
+      "epoch": 0.7815959187602272,
+      "grad_norm": 0.20869745314121246,
+      "learning_rate": 1.2029356413120297e-06,
+      "loss": 0.756,
+      "step": 2030
+    },
+    {
+      "epoch": 0.7854461449610165,
+      "grad_norm": 0.21471671760082245,
+      "learning_rate": 1.1626613545778786e-06,
+      "loss": 0.7726,
+      "step": 2040
+    },
+    {
+      "epoch": 0.7892963711618057,
+      "grad_norm": 0.20360571146011353,
+      "learning_rate": 1.1229839215686294e-06,
+      "loss": 0.776,
+      "step": 2050
+    },
+    {
+      "epoch": 0.7931465973625951,
+      "grad_norm": 0.19516447186470032,
+      "learning_rate": 1.0839095136509188e-06,
+      "loss": 0.7616,
+      "step": 2060
+    },
+    {
+      "epoch": 0.7969968235633843,
+      "grad_norm": 0.2299748808145523,
+      "learning_rate": 1.045444208397791e-06,
+      "loss": 0.7839,
+      "step": 2070
+    },
+    {
+      "epoch": 0.8008470497641736,
+      "grad_norm": 0.19916990399360657,
+      "learning_rate": 1.007593988643402e-06,
+      "loss": 0.7664,
+      "step": 2080
+    },
+    {
+      "epoch": 0.804697275964963,
+      "grad_norm": 0.20423093438148499,
+      "learning_rate": 9.703647415524613e-07,
+      "loss": 0.7871,
+      "step": 2090
+    },
+    {
+      "epoch": 0.8085475021657522,
+      "grad_norm": 0.2140628546476364,
+      "learning_rate": 9.337622577045485e-07,
+      "loss": 0.7761,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8123977283665416,
+      "grad_norm": 0.2145545333623886,
+      "learning_rate": 8.977922301934588e-07,
+      "loss": 0.7835,
+      "step": 2110
+    },
+    {
+      "epoch": 0.8162479545673308,
+      "grad_norm": 0.2136877030134201,
+      "learning_rate": 8.624602537417009e-07,
+      "loss": 0.7852,
+      "step": 2120
+    },
+    {
+      "epoch": 0.8200981807681201,
+      "grad_norm": 0.22830118238925934,
+      "learning_rate": 8.277718238303123e-07,
+      "loss": 0.7716,
+      "step": 2130
+    },
+    {
+      "epoch": 0.8239484069689095,
+      "grad_norm": 0.2140616476535797,
+      "learning_rate": 7.937323358440935e-07,
+      "loss": 0.7763,
+      "step": 2140
+    },
+    {
+      "epoch": 0.8277986331696987,
+      "grad_norm": 0.23019348084926605,
+      "learning_rate": 7.603470842324223e-07,
+      "loss": 0.7674,
+      "step": 2150
+    },
+    {
+      "epoch": 0.831648859370488,
+      "grad_norm": 0.20326852798461914,
+      "learning_rate": 7.276212616857575e-07,
+      "loss": 0.7853,
+      "step": 2160
+    },
+    {
+      "epoch": 0.8354990855712773,
+      "grad_norm": 0.2109815776348114,
+      "learning_rate": 6.955599583279837e-07,
+      "loss": 0.7873,
+      "step": 2170
+    },
+    {
+      "epoch": 0.8393493117720666,
+      "grad_norm": 0.20704059302806854,
+      "learning_rate": 6.641681609246981e-07,
+      "loss": 0.7767,
+      "step": 2180
+    },
+    {
+      "epoch": 0.8431995379728559,
+      "grad_norm": 0.20912793278694153,
+      "learning_rate": 6.334507521075711e-07,
+      "loss": 0.7871,
+      "step": 2190
+    },
+    {
+      "epoch": 0.8470497641736452,
+      "grad_norm": 0.2258094996213913,
+      "learning_rate": 6.034125096149179e-07,
+      "loss": 0.7796,
+      "step": 2200
+    },
+    {
+      "epoch": 0.8508999903744345,
+      "grad_norm": 0.21635396778583527,
+      "learning_rate": 5.740581055485722e-07,
+      "loss": 0.75,
+      "step": 2210
+    },
+    {
+      "epoch": 0.8547502165752238,
+      "grad_norm": 0.19561634957790375,
+      "learning_rate": 5.453921056471961e-07,
+      "loss": 0.7837,
+      "step": 2220
+    },
+    {
+      "epoch": 0.8586004427760131,
+      "grad_norm": 0.21199467778205872,
+      "learning_rate": 5.174189685761288e-07,
+      "loss": 0.7784,
+      "step": 2230
+    },
+    {
+      "epoch": 0.8624506689768023,
+      "grad_norm": 0.20541375875473022,
+      "learning_rate": 4.901430452338979e-07,
+      "loss": 0.7891,
+      "step": 2240
+    },
+    {
+      "epoch": 0.8663008951775917,
+      "grad_norm": 0.20293408632278442,
+      "learning_rate": 4.6356857807548286e-07,
+      "loss": 0.7749,
+      "step": 2250
+    },
+    {
+      "epoch": 0.870151121378381,
+      "grad_norm": 0.21144260466098785,
+      "learning_rate": 4.376997004524541e-07,
+      "loss": 0.7644,
+      "step": 2260
+    },
+    {
+      "epoch": 0.8740013475791703,
+      "grad_norm": 0.21344861388206482,
+      "learning_rate": 4.12540435970073e-07,
+      "loss": 0.7735,
+      "step": 2270
+    },
+    {
+      "epoch": 0.8778515737799596,
+      "grad_norm": 0.2116292119026184,
+      "learning_rate": 3.880946978614703e-07,
+      "loss": 0.7805,
+      "step": 2280
+    },
+    {
+      "epoch": 0.8817017999807488,
+      "grad_norm": 0.21220466494560242,
+      "learning_rate": 3.643662883789878e-07,
+      "loss": 0.7875,
+      "step": 2290
+    },
+    {
+      "epoch": 0.8855520261815382,
+      "grad_norm": 0.20233546197414398,
+      "learning_rate": 3.41358898202781e-07,
+      "loss": 0.7658,
+      "step": 2300
+    },
+    {
+      "epoch": 0.8894022523823275,
+      "grad_norm": 0.19838057458400726,
+      "learning_rate": 3.190761058667741e-07,
+      "loss": 0.7792,
+      "step": 2310
+    },
+    {
+      "epoch": 0.8932524785831167,
+      "grad_norm": 0.20751775801181793,
+      "learning_rate": 2.9752137720206463e-07,
+      "loss": 0.7829,
+      "step": 2320
+    },
+    {
+      "epoch": 0.8971027047839061,
+      "grad_norm": 0.20873038470745087,
+      "learning_rate": 2.7669806479785067e-07,
+      "loss": 0.7859,
+      "step": 2330
+    },
+    {
+      "epoch": 0.9009529309846953,
+      "grad_norm": 0.21066701412200928,
+      "learning_rate": 2.566094074799763e-07,
+      "loss": 0.7813,
+      "step": 2340
+    },
+    {
+      "epoch": 0.9048031571854847,
+      "grad_norm": 0.20949746668338776,
+      "learning_rate": 2.3725852980716557e-07,
+      "loss": 0.7642,
+      "step": 2350
+    },
+    {
+      "epoch": 0.908653383386274,
+      "grad_norm": 0.2095063179731369,
+      "learning_rate": 2.1864844158503983e-07,
+      "loss": 0.7764,
+      "step": 2360
+    },
+    {
+      "epoch": 0.9125036095870632,
+      "grad_norm": 0.22941343486309052,
+      "learning_rate": 2.007820373979702e-07,
+      "loss": 0.7688,
+      "step": 2370
+    },
+    {
+      "epoch": 0.9163538357878526,
+      "grad_norm": 0.21286219358444214,
+      "learning_rate": 1.8366209615886532e-07,
+      "loss": 0.7791,
+      "step": 2380
+    },
+    {
+      "epoch": 0.9202040619886418,
+      "grad_norm": 0.20981867611408234,
+      "learning_rate": 1.6729128067693613e-07,
+      "loss": 0.7868,
+      "step": 2390
+    },
+    {
+      "epoch": 0.9240542881894311,
+      "grad_norm": 0.22857896983623505,
+      "learning_rate": 1.5167213724353426e-07,
+      "loss": 0.7881,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9279045143902205,
+      "grad_norm": 0.2102576196193695,
+      "learning_rate": 1.3680709523610058e-07,
+      "loss": 0.7677,
+      "step": 2410
+    },
+    {
+      "epoch": 0.9317547405910097,
+      "grad_norm": 0.21056540310382843,
+      "learning_rate": 1.2269846674030418e-07,
+      "loss": 0.7938,
+      "step": 2420
+    },
+    {
+      "epoch": 0.935604966791799,
+      "grad_norm": 0.21998818218708038,
+      "learning_rate": 1.0934844619042595e-07,
+      "loss": 0.7872,
+      "step": 2430
+    },
+    {
+      "epoch": 0.9394551929925883,
+      "grad_norm": 0.22209565341472626,
+      "learning_rate": 9.675911002803961e-08,
+      "loss": 0.7784,
+      "step": 2440
+    },
+    {
+      "epoch": 0.9433054191933776,
+      "grad_norm": 0.2341625839471817,
+      "learning_rate": 8.493241637904381e-08,
+      "loss": 0.7758,
+      "step": 2450
+    },
+    {
+      "epoch": 0.9471556453941669,
+      "grad_norm": 0.2137785404920578,
+      "learning_rate": 7.387020474909689e-08,
+      "loss": 0.7941,
+      "step": 2460
+    },
+    {
+      "epoch": 0.9510058715949562,
+      "grad_norm": 0.22507698833942413,
+      "learning_rate": 6.357419573750856e-08,
+      "loss": 0.7897,
+      "step": 2470
+    },
+    {
+      "epoch": 0.9548560977957455,
+      "grad_norm": 0.21697045862674713,
+      "learning_rate": 5.4045990769613365e-08,
+      "loss": 0.7715,
+      "step": 2480
+    },
+    {
+      "epoch": 0.9587063239965348,
+      "grad_norm": 0.2092304825782776,
+      "learning_rate": 4.5287071847694434e-08,
+      "loss": 0.7789,
+      "step": 2490
+    },
+    {
+      "epoch": 0.9625565501973241,
+      "grad_norm": 0.21155469119548798,
+      "learning_rate": 3.7298801320467817e-08,
+      "loss": 0.7743,
+      "step": 2500
+    },
+    {
+      "epoch": 0.9664067763981133,
+      "grad_norm": 0.20938760042190552,
+      "learning_rate": 3.0082421671192576e-08,
+      "loss": 0.7662,
+      "step": 2510
+    },
+    {
+      "epoch": 0.9702570025989027,
+      "grad_norm": 0.21425026655197144,
+      "learning_rate": 2.3639055324410355e-08,
+      "loss": 0.7857,
+      "step": 2520
+    },
+    {
+      "epoch": 0.974107228799692,
+      "grad_norm": 0.20930226147174835,
+      "learning_rate": 1.7969704471369477e-08,
+      "loss": 0.7726,
+      "step": 2530
+    },
+    {
+      "epoch": 0.9779574550004813,
+      "grad_norm": 0.20506355166435242,
+      "learning_rate": 1.3075250914142413e-08,
+      "loss": 0.7738,
+      "step": 2540
+    },
+    {
+      "epoch": 0.9818076812012706,
+      "grad_norm": 0.20354686677455902,
+      "learning_rate": 8.956455928476049e-09,
+      "loss": 0.784,
+      "step": 2550
+    },
+    {
+      "epoch": 0.9856579074020598,
+      "grad_norm": 0.20308208465576172,
+      "learning_rate": 5.6139601453786275e-09,
+      "loss": 0.7769,
+      "step": 2560
+    },
+    {
+      "epoch": 0.9895081336028492,
+      "grad_norm": 0.21389342844486237,
+      "learning_rate": 3.0482834514805602e-09,
+      "loss": 0.7667,
+      "step": 2570
+    },
+    {
+      "epoch": 0.9933583598036385,
+      "grad_norm": 0.20575203001499176,
+      "learning_rate": 1.2598249081713366e-09,
+      "loss": 0.7591,
+      "step": 2580
+    },
+    {
+      "epoch": 0.9972085860044277,
+      "grad_norm": 0.21766448020935059,
+      "learning_rate": 2.488626895302826e-10,
+      "loss": 0.7901,
+      "step": 2590
+    },
+    {
+      "epoch": 0.9999037443449803,
+      "step": 2597,
+      "total_flos": 3.522737300075184e+18,
+      "train_loss": 0.7976328194669269,
+      "train_runtime": 7575.7285,
+      "train_samples_per_second": 87.768,
+      "train_steps_per_second": 0.343
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2597,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.522737300075184e+18,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/mc1_b32_224_400m/finetune/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bef65fc3622c36ad518a0f7424922855fa2d3cb2fa29c4bc8399355cd2e2c07c
+size 6136