add checkpoints/codi3b_a0.5_b1.0_g0.5_ls1

.gitattributes

@@ -56,3 +56,9 @@ checkpoints/codi1.5b_a1.0_b1.0_g1.0_ls2/checkpoint-600/tokenizer.json filter=lfs
 checkpoints/codi1.5b_a1.0_b1.0_g1.0_ls2/checkpoint-800/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/codi1.5b_a1.0_b1.0_g1.0_ls2/wandb/offline-run-20260620_130138-bqeecd43/run-bqeecd43.wandb filter=lfs diff=lfs merge=lfs -text
 checkpoints/codi1.5b_a1.0_b1.0_g1.0_ls2/wandb/offline-run-20260621_033730-ue0j104y/run-ue0j104y.wandb filter=lfs diff=lfs merge=lfs -text
+checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-1000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-200/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-400/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-600/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-800/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/wandb/offline-run-20260620_190723-v54ju1cd/run-v54ju1cd.wandb filter=lfs diff=lfs merge=lfs -text

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-1000/added_tokens.json

@@ -0,0 +1,35 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|action_sep|>": 151670,
+  "<|arg_sep|>": 151671,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|call_sep|>": 151666,
+  "<|end_of_text|>": 151673,
+  "<|endoftext|>": 151643,
+  "<|exception_sep|>": 151669,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|frame_sep|>": 151672,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|latent_end|>": 151675,
+  "<|latent_start|>": 151674,
+  "<|line_sep|>": 151667,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|return_sep|>": 151668,
+  "<|trace_context_start|>": 151665,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-1000/chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\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>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-1000/config.json

@@ -0,0 +1,66 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 36,
+  "model_type": "qwen2",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 36,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151676
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-1000/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bffc5827369b4b8aa1fd178a4f98d51f09b0ccb52ae88dc9e6ee21bdbd08f257
+size 6187858991

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-1000/special_tokens_map.json

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

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-1000/thought_projector.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1257681b42584896a62f8422cf41b4bf334eba02a9e5a373d11ceea1e0abe9b1
+size 16788033

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-1000/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83a790d654474f5dfe225f889afd0210313eb1083f942671f2c4b8e95a1c922b
+size 11424004

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-1000/tokenizer_config.json

@@ -0,0 +1,295 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<|trace_context_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151666": {
+      "content": "<|call_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151667": {
+      "content": "<|line_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151668": {
+      "content": "<|return_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151669": {
+      "content": "<|exception_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151670": {
+      "content": "<|action_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151671": {
+      "content": "<|arg_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151672": {
+      "content": "<|frame_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151673": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151674": {
+      "content": "<|latent_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151675": {
+      "content": "<|latent_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-1000/trainer_state.json

@@ -0,0 +1,2046 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.28835063437139563,
+  "eval_steps": 500,
+  "global_step": 1000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0014417531718569781,
+      "grad_norm": 47.75,
+      "kd_loss": 0.2916574478149414,
+      "learning_rate": 1.3333333333333334e-06,
+      "loss": 0.3836,
+      "step": 5,
+      "student_loss": 0.12182186543941498,
+      "teacher_loss": 0.002018272876739502
+    },
+    {
+      "epoch": 0.0028835063437139563,
+      "grad_norm": 22.5,
+      "kd_loss": 0.26723670959472656,
+      "learning_rate": 3e-06,
+      "loss": 0.2911,
+      "step": 10,
+      "student_loss": 0.10971185564994812,
+      "teacher_loss": 0.005489554721862078
+    },
+    {
+      "epoch": 0.004325259515570935,
+      "grad_norm": 6.9375,
+      "kd_loss": 0.21662698686122894,
+      "learning_rate": 4.666666666666667e-06,
+      "loss": 0.1789,
+      "step": 15,
+      "student_loss": 0.0025933366268873215,
+      "teacher_loss": 0.001338644651696086
+    },
+    {
+      "epoch": 0.0057670126874279125,
+      "grad_norm": 3.078125,
+      "kd_loss": 0.13032649457454681,
+      "learning_rate": 6.333333333333333e-06,
+      "loss": 0.1296,
+      "step": 20,
+      "student_loss": 0.023931587114930153,
+      "teacher_loss": 0.0006410324713215232
+    },
+    {
+      "epoch": 0.00720876585928489,
+      "grad_norm": 2.34375,
+      "kd_loss": 0.10749491304159164,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 0.107,
+      "step": 25,
+      "student_loss": 0.0032150023616850376,
+      "teacher_loss": 0.0010774527909234166
+    },
+    {
+      "epoch": 0.00865051903114187,
+      "grad_norm": 1.2890625,
+      "kd_loss": 0.0680936947464943,
+      "learning_rate": 9.666666666666667e-06,
+      "loss": 0.0841,
+      "step": 30,
+      "student_loss": 0.00099717837292701,
+      "teacher_loss": 0.0005822837701998651
+    },
+    {
+      "epoch": 0.010092272202998846,
+      "grad_norm": 2.109375,
+      "kd_loss": 0.06131501868367195,
+      "learning_rate": 9.99958042442916e-06,
+      "loss": 0.0685,
+      "step": 35,
+      "student_loss": 0.0027061135042458773,
+      "teacher_loss": 0.0006183119839988649
+    },
+    {
+      "epoch": 0.011534025374855825,
+      "grad_norm": 1.203125,
+      "kd_loss": 0.05697907507419586,
+      "learning_rate": 9.997876019358083e-06,
+      "loss": 0.0614,
+      "step": 40,
+      "student_loss": 0.029384225606918335,
+      "teacher_loss": 0.025684259831905365
+    },
+    {
+      "epoch": 0.012975778546712802,
+      "grad_norm": 1.296875,
+      "kd_loss": 0.09766291826963425,
+      "learning_rate": 9.99486100792044e-06,
+      "loss": 0.0476,
+      "step": 45,
+      "student_loss": 0.06478797644376755,
+      "teacher_loss": 0.0030884514562785625
+    },
+    {
+      "epoch": 0.01441753171856978,
+      "grad_norm": 1.453125,
+      "kd_loss": 0.03506260737776756,
+      "learning_rate": 9.990536180750724e-06,
+      "loss": 0.0517,
+      "step": 50,
+      "student_loss": 0.02092154324054718,
+      "teacher_loss": 0.01070804987102747
+    },
+    {
+      "epoch": 0.015859284890426758,
+      "grad_norm": 1.7578125,
+      "kd_loss": 0.030483372509479523,
+      "learning_rate": 9.984902671959911e-06,
+      "loss": 0.0461,
+      "step": 55,
+      "student_loss": 0.0011451594764366746,
+      "teacher_loss": 0.0006940242601558566
+    },
+    {
+      "epoch": 0.01730103806228374,
+      "grad_norm": 1.3359375,
+      "kd_loss": 0.030781563371419907,
+      "learning_rate": 9.97796195883804e-06,
+      "loss": 0.0412,
+      "step": 60,
+      "student_loss": 0.003650753991678357,
+      "teacher_loss": 0.002238439628854394
+    },
+    {
+      "epoch": 0.018742791234140715,
+      "grad_norm": 1.5234375,
+      "kd_loss": 0.030627569183707237,
+      "learning_rate": 9.969715861466839e-06,
+      "loss": 0.0425,
+      "step": 65,
+      "student_loss": 0.02980353683233261,
+      "teacher_loss": 0.03997597098350525
+    },
+    {
+      "epoch": 0.020184544405997693,
+      "grad_norm": 0.875,
+      "kd_loss": 0.0586991086602211,
+      "learning_rate": 9.96016654224243e-06,
+      "loss": 0.0466,
+      "step": 70,
+      "student_loss": 0.014155564829707146,
+      "teacher_loss": 0.0020057852379977703
+    },
+    {
+      "epoch": 0.02162629757785467,
+      "grad_norm": 0.796875,
+      "kd_loss": 0.021966831758618355,
+      "learning_rate": 9.94931650530827e-06,
+      "loss": 0.0409,
+      "step": 75,
+      "student_loss": 0.0014096935046836734,
+      "teacher_loss": 0.0008751551504246891
+    },
+    {
+      "epoch": 0.02306805074971165,
+      "grad_norm": 0.64453125,
+      "kd_loss": 0.022734930738806725,
+      "learning_rate": 9.93716859589851e-06,
+      "loss": 0.0362,
+      "step": 80,
+      "student_loss": 0.001171765266917646,
+      "teacher_loss": 0.0007404370117001235
+    },
+    {
+      "epoch": 0.024509803921568627,
+      "grad_norm": 0.765625,
+      "kd_loss": 0.022543881088495255,
+      "learning_rate": 9.923725999591846e-06,
+      "loss": 0.0382,
+      "step": 85,
+      "student_loss": 0.0004507019475568086,
+      "teacher_loss": 0.00039780911174602807
+    },
+    {
+      "epoch": 0.025951557093425604,
+      "grad_norm": 0.69140625,
+      "kd_loss": 0.07131864875555038,
+      "learning_rate": 9.908992241476189e-06,
+      "loss": 0.0387,
+      "step": 90,
+      "student_loss": 0.013878172263503075,
+      "teacher_loss": 0.0030211261473596096
+    },
+    {
+      "epoch": 0.027393310265282585,
+      "grad_norm": 1.1015625,
+      "kd_loss": 0.04375176876783371,
+      "learning_rate": 9.892971185224244e-06,
+      "loss": 0.0342,
+      "step": 95,
+      "student_loss": 0.0081056859344244,
+      "teacher_loss": 0.005189534276723862
+    },
+    {
+      "epoch": 0.02883506343713956,
+      "grad_norm": 0.859375,
+      "kd_loss": 0.021529421210289,
+      "learning_rate": 9.875667032080354e-06,
+      "loss": 0.0311,
+      "step": 100,
+      "student_loss": 0.0005472518969327211,
+      "teacher_loss": 0.000426671642344445
+    },
+    {
+      "epoch": 0.03027681660899654,
+      "grad_norm": 1.3671875,
+      "kd_loss": 0.03554106876254082,
+      "learning_rate": 9.857084319758772e-06,
+      "loss": 0.0342,
+      "step": 105,
+      "student_loss": 0.002422141609713435,
+      "teacher_loss": 0.000269417796516791
+    },
+    {
+      "epoch": 0.031718569780853516,
+      "grad_norm": 1.1953125,
+      "kd_loss": 0.0378132238984108,
+      "learning_rate": 9.837227921253747e-06,
+      "loss": 0.0362,
+      "step": 110,
+      "student_loss": 0.031434450298547745,
+      "teacher_loss": 0.01768127642571926
+    },
+    {
+      "epoch": 0.03316032295271049,
+      "grad_norm": 0.52734375,
+      "kd_loss": 0.03465187922120094,
+      "learning_rate": 9.816103043561648e-06,
+      "loss": 0.0337,
+      "step": 115,
+      "student_loss": 0.000800441310275346,
+      "teacher_loss": 0.0002722400240600109
+    },
+    {
+      "epoch": 0.03460207612456748,
+      "grad_norm": 0.59375,
+      "kd_loss": 0.02393350750207901,
+      "learning_rate": 9.79371522631553e-06,
+      "loss": 0.031,
+      "step": 120,
+      "student_loss": 0.0010254759108647704,
+      "teacher_loss": 0.0006771993357688189
+    },
+    {
+      "epoch": 0.036043829296424454,
+      "grad_norm": 0.734375,
+      "kd_loss": 0.05980119854211807,
+      "learning_rate": 9.770070340332457e-06,
+      "loss": 0.0344,
+      "step": 125,
+      "student_loss": 0.02782793715596199,
+      "teacher_loss": 0.0002460530085954815
+    },
+    {
+      "epoch": 0.03748558246828143,
+      "grad_norm": 0.71875,
+      "kd_loss": 0.02418632246553898,
+      "learning_rate": 9.745174586073982e-06,
+      "loss": 0.0333,
+      "step": 130,
+      "student_loss": 0.0004915493191219866,
+      "teacher_loss": 0.0004100716323591769
+    },
+    {
+      "epoch": 0.03892733564013841,
+      "grad_norm": 0.73046875,
+      "kd_loss": 0.04188523441553116,
+      "learning_rate": 9.719034492020183e-06,
+      "loss": 0.0371,
+      "step": 135,
+      "student_loss": 0.0016317864647135139,
+      "teacher_loss": 0.00047902524238452315
+    },
+    {
+      "epoch": 0.040369088811995385,
+      "grad_norm": 0.65625,
+      "kd_loss": 0.021806886419653893,
+      "learning_rate": 9.691656912957686e-06,
+      "loss": 0.0352,
+      "step": 140,
+      "student_loss": 0.00046953262062743306,
+      "teacher_loss": 0.0003224058309569955
+    },
+    {
+      "epoch": 0.04181084198385236,
+      "grad_norm": 0.443359375,
+      "kd_loss": 0.07382702082395554,
+      "learning_rate": 9.663049028182112e-06,
+      "loss": 0.0314,
+      "step": 145,
+      "student_loss": 0.03231842443346977,
+      "teacher_loss": 0.006043749395757914
+    },
+    {
+      "epoch": 0.04325259515570934,
+      "grad_norm": 0.5703125,
+      "kd_loss": 0.0166649017482996,
+      "learning_rate": 9.633218339615433e-06,
+      "loss": 0.0309,
+      "step": 150,
+      "student_loss": 0.001361644477583468,
+      "teacher_loss": 0.0013295100070536137
+    },
+    {
+      "epoch": 0.04469434832756632,
+      "grad_norm": 0.85546875,
+      "kd_loss": 0.026128817349672318,
+      "learning_rate": 9.602172669838721e-06,
+      "loss": 0.0373,
+      "step": 155,
+      "student_loss": 0.0024132225662469864,
+      "teacher_loss": 0.0007210202165879309
+    },
+    {
+      "epoch": 0.0461361014994233,
+      "grad_norm": 1.4765625,
+      "kd_loss": 0.030131949111819267,
+      "learning_rate": 9.569920160040815e-06,
+      "loss": 0.0339,
+      "step": 160,
+      "student_loss": 0.0748908743262291,
+      "teacher_loss": 0.04347090795636177
+    },
+    {
+      "epoch": 0.04757785467128028,
+      "grad_norm": 0.7109375,
+      "kd_loss": 0.03373745456337929,
+      "learning_rate": 9.536469267883432e-06,
+      "loss": 0.0312,
+      "step": 165,
+      "student_loss": 0.000419136747950688,
+      "teacher_loss": 0.003946289885789156
+    },
+    {
+      "epoch": 0.049019607843137254,
+      "grad_norm": 2.078125,
+      "kd_loss": 0.02021661214530468,
+      "learning_rate": 9.501828765283295e-06,
+      "loss": 0.0331,
+      "step": 170,
+      "student_loss": 0.000686564773786813,
+      "teacher_loss": 0.0005308608524501324
+    },
+    {
+      "epoch": 0.05046136101499423,
+      "grad_norm": 1.4375,
+      "kd_loss": 0.028470534831285477,
+      "learning_rate": 9.466007736111846e-06,
+      "loss": 0.0375,
+      "step": 175,
+      "student_loss": 0.0027788360603153706,
+      "teacher_loss": 0.0002813354367390275
+    },
+    {
+      "epoch": 0.05190311418685121,
+      "grad_norm": 1.2265625,
+      "kd_loss": 0.03539275750517845,
+      "learning_rate": 9.429015573813163e-06,
+      "loss": 0.0298,
+      "step": 180,
+      "student_loss": 0.0011434294283390045,
+      "teacher_loss": 0.00030375414644367993
+    },
+    {
+      "epoch": 0.05334486735870819,
+      "grad_norm": 1.203125,
+      "kd_loss": 0.03853345289826393,
+      "learning_rate": 9.390861978940687e-06,
+      "loss": 0.0354,
+      "step": 185,
+      "student_loss": 0.035494156181812286,
+      "teacher_loss": 0.0002623012405820191
+    },
+    {
+      "epoch": 0.05478662053056517,
+      "grad_norm": 1.046875,
+      "kd_loss": 0.05331602692604065,
+      "learning_rate": 9.351556956613423e-06,
+      "loss": 0.0309,
+      "step": 190,
+      "student_loss": 0.004768161568790674,
+      "teacher_loss": 0.0014110342599451542
+    },
+    {
+      "epoch": 0.056228373702422146,
+      "grad_norm": 1.421875,
+      "kd_loss": 0.053704116493463516,
+      "learning_rate": 9.31111081389227e-06,
+      "loss": 0.0301,
+      "step": 195,
+      "student_loss": 0.006213298067450523,
+      "teacher_loss": 0.0004057004116475582
+    },
+    {
+      "epoch": 0.05767012687427912,
+      "grad_norm": 0.796875,
+      "kd_loss": 0.034625094383955,
+      "learning_rate": 9.269534157077177e-06,
+      "loss": 0.039,
+      "step": 200,
+      "student_loss": 0.0004761361633427441,
+      "teacher_loss": 0.00022407782671507448
+    },
+    {
+      "epoch": 0.0591118800461361,
+      "grad_norm": 1.8515625,
+      "kd_loss": 0.02584153600037098,
+      "learning_rate": 9.226837888925813e-06,
+      "loss": 0.0341,
+      "step": 205,
+      "student_loss": 0.04030226171016693,
+      "teacher_loss": 0.027547728270292282
+    },
+    {
+      "epoch": 0.06055363321799308,
+      "grad_norm": 2.578125,
+      "kd_loss": 0.031695615500211716,
+      "learning_rate": 9.183033205794525e-06,
+      "loss": 0.0296,
+      "step": 210,
+      "student_loss": 0.0005234605050645769,
+      "teacher_loss": 0.00026510158204473555
+    },
+    {
+      "epoch": 0.061995386389850055,
+      "grad_norm": 1.1796875,
+      "kd_loss": 0.03749287873506546,
+      "learning_rate": 9.13813159470227e-06,
+      "loss": 0.0314,
+      "step": 215,
+      "student_loss": 0.0004607264418154955,
+      "teacher_loss": 0.00017995040980167687
+    },
+    {
+      "epoch": 0.06343713956170703,
+      "grad_norm": 2.671875,
+      "kd_loss": 0.04252801090478897,
+      "learning_rate": 9.092144830318357e-06,
+      "loss": 0.0309,
+      "step": 220,
+      "student_loss": 0.03449667617678642,
+      "teacher_loss": 0.010067273862659931
+    },
+    {
+      "epoch": 0.06487889273356401,
+      "grad_norm": 1.1640625,
+      "kd_loss": 0.022738255560398102,
+      "learning_rate": 9.045084971874738e-06,
+      "loss": 0.0371,
+      "step": 225,
+      "student_loss": 0.04431537538766861,
+      "teacher_loss": 0.0017423024401068687
+    },
+    {
+      "epoch": 0.06632064590542099,
+      "grad_norm": 1.1953125,
+      "kd_loss": 0.04256664216518402,
+      "learning_rate": 8.99696436000368e-06,
+      "loss": 0.029,
+      "step": 230,
+      "student_loss": 0.002170365070924163,
+      "teacher_loss": 0.003624099539592862
+    },
+    {
+      "epoch": 0.06776239907727798,
+      "grad_norm": 1.1328125,
+      "kd_loss": 0.07645199447870255,
+      "learning_rate": 8.947795613501658e-06,
+      "loss": 0.0312,
+      "step": 235,
+      "student_loss": 0.009659786708652973,
+      "teacher_loss": 0.00044414011063054204
+    },
+    {
+      "epoch": 0.06920415224913495,
+      "grad_norm": 0.77734375,
+      "kd_loss": 0.03161292523145676,
+      "learning_rate": 8.897591626020284e-06,
+      "loss": 0.0337,
+      "step": 240,
+      "student_loss": 0.009329003281891346,
+      "teacher_loss": 0.0004169123712927103
+    },
+    {
+      "epoch": 0.07064590542099193,
+      "grad_norm": 0.73828125,
+      "kd_loss": 0.016300037503242493,
+      "learning_rate": 8.846365562685178e-06,
+      "loss": 0.0274,
+      "step": 245,
+      "student_loss": 0.0028402707539498806,
+      "teacher_loss": 0.0025411578826606274
+    },
+    {
+      "epoch": 0.07208765859284891,
+      "grad_norm": 1.359375,
+      "kd_loss": 0.020206376910209656,
+      "learning_rate": 8.794130856643635e-06,
+      "loss": 0.0302,
+      "step": 250,
+      "student_loss": 0.0006153527647256851,
+      "teacher_loss": 0.0002888407907448709
+    },
+    {
+      "epoch": 0.07352941176470588,
+      "grad_norm": 0.71875,
+      "kd_loss": 0.01622343435883522,
+      "learning_rate": 8.74090120554202e-06,
+      "loss": 0.031,
+      "step": 255,
+      "student_loss": 0.0003402966831345111,
+      "teacher_loss": 0.00035465144901536405
+    },
+    {
+      "epoch": 0.07497116493656286,
+      "grad_norm": 0.84765625,
+      "kd_loss": 0.028415139764547348,
+      "learning_rate": 8.686690567933803e-06,
+      "loss": 0.033,
+      "step": 260,
+      "student_loss": 0.042195871472358704,
+      "teacher_loss": 0.033024489879608154
+    },
+    {
+      "epoch": 0.07641291810841984,
+      "grad_norm": 1.7890625,
+      "kd_loss": 0.018096741288900375,
+      "learning_rate": 8.63151315961915e-06,
+      "loss": 0.033,
+      "step": 265,
+      "student_loss": 0.0003965997020713985,
+      "teacher_loss": 0.00029857814661227167
+    },
+    {
+      "epoch": 0.07785467128027682,
+      "grad_norm": 0.73828125,
+      "kd_loss": 0.020055225118994713,
+      "learning_rate": 8.575383449917103e-06,
+      "loss": 0.0347,
+      "step": 270,
+      "student_loss": 0.0023262996692210436,
+      "teacher_loss": 0.001756582991220057
+    },
+    {
+      "epoch": 0.07929642445213379,
+      "grad_norm": 1.2109375,
+      "kd_loss": 0.025473114103078842,
+      "learning_rate": 8.518316157871232e-06,
+      "loss": 0.0297,
+      "step": 275,
+      "student_loss": 0.04790982976555824,
+      "teacher_loss": 0.03409208729863167
+    },
+    {
+      "epoch": 0.08073817762399077,
+      "grad_norm": 1.0234375,
+      "kd_loss": 0.035440195351839066,
+      "learning_rate": 8.460326248389825e-06,
+      "loss": 0.0299,
+      "step": 280,
+      "student_loss": 0.0004696878604590893,
+      "teacher_loss": 0.00016561525990255177
+    },
+    {
+      "epoch": 0.08217993079584775,
+      "grad_norm": 1.25,
+      "kd_loss": 0.021167855709791183,
+      "learning_rate": 8.401428928321607e-06,
+      "loss": 0.0331,
+      "step": 285,
+      "student_loss": 0.0006399436388164759,
+      "teacher_loss": 0.0005028079031035304
+    },
+    {
+      "epoch": 0.08362168396770472,
+      "grad_norm": 1.1015625,
+      "kd_loss": 0.017318392172455788,
+      "learning_rate": 8.341639642468002e-06,
+      "loss": 0.0355,
+      "step": 290,
+      "student_loss": 0.031288016587495804,
+      "teacher_loss": 0.028279613703489304
+    },
+    {
+      "epoch": 0.0850634371395617,
+      "grad_norm": 1.2109375,
+      "kd_loss": 0.019709059968590736,
+      "learning_rate": 8.280974069532999e-06,
+      "loss": 0.0329,
+      "step": 295,
+      "student_loss": 0.0007035491289570928,
+      "teacher_loss": 0.0005519128171727061
+    },
+    {
+      "epoch": 0.08650519031141868,
+      "grad_norm": 1.0625,
+      "kd_loss": 0.051209982484579086,
+      "learning_rate": 8.219448118011687e-06,
+      "loss": 0.0308,
+      "step": 300,
+      "student_loss": 0.017601756379008293,
+      "teacher_loss": 0.0016454572323709726
+    },
+    {
+      "epoch": 0.08794694348327567,
+      "grad_norm": 2.28125,
+      "kd_loss": 0.021270615980029106,
+      "learning_rate": 8.157077922018537e-06,
+      "loss": 0.0286,
+      "step": 305,
+      "student_loss": 0.005125237628817558,
+      "teacher_loss": 0.003151164622977376
+    },
+    {
+      "epoch": 0.08938869665513265,
+      "grad_norm": 0.859375,
+      "kd_loss": 0.02120344713330269,
+      "learning_rate": 8.093879837056486e-06,
+      "loss": 0.0302,
+      "step": 310,
+      "student_loss": 0.00047028501285240054,
+      "teacher_loss": 0.0005005286075174809
+    },
+    {
+      "epoch": 0.09083044982698962,
+      "grad_norm": 2.0,
+      "kd_loss": 0.037551358342170715,
+      "learning_rate": 8.029870435728018e-06,
+      "loss": 0.0301,
+      "step": 315,
+      "student_loss": 0.006268176715821028,
+      "teacher_loss": 0.00021872192155569792
+    },
+    {
+      "epoch": 0.0922722029988466,
+      "grad_norm": 1.8828125,
+      "kd_loss": 0.025020649656653404,
+      "learning_rate": 7.965066503389264e-06,
+      "loss": 0.0321,
+      "step": 320,
+      "student_loss": 0.0004365233180578798,
+      "teacher_loss": 0.00025109920534305274
+    },
+    {
+      "epoch": 0.09371395617070358,
+      "grad_norm": 2.390625,
+      "kd_loss": 0.06403576582670212,
+      "learning_rate": 7.89948503374835e-06,
+      "loss": 0.0284,
+      "step": 325,
+      "student_loss": 0.006501263007521629,
+      "teacher_loss": 0.02552512101829052
+    },
+    {
+      "epoch": 0.09515570934256055,
+      "grad_norm": 0.94921875,
+      "kd_loss": 0.048998646438121796,
+      "learning_rate": 7.833143224409076e-06,
+      "loss": 0.03,
+      "step": 330,
+      "student_loss": 0.011290914379060268,
+      "teacher_loss": 0.0012772196205332875
+    },
+    {
+      "epoch": 0.09659746251441753,
+      "grad_norm": 1.2734375,
+      "kd_loss": 0.013905064202845097,
+      "learning_rate": 7.766058472361154e-06,
+      "loss": 0.0266,
+      "step": 335,
+      "student_loss": 0.0008061202825047076,
+      "teacher_loss": 0.0009985697688534856
+    },
+    {
+      "epoch": 0.09803921568627451,
+      "grad_norm": 0.91015625,
+      "kd_loss": 0.025371180847287178,
+      "learning_rate": 7.698248369418146e-06,
+      "loss": 0.0312,
+      "step": 340,
+      "student_loss": 0.010380898602306843,
+      "teacher_loss": 0.007381496950984001
+    },
+    {
+      "epoch": 0.09948096885813149,
+      "grad_norm": 0.9921875,
+      "kd_loss": 0.03340749442577362,
+      "learning_rate": 7.629730697604314e-06,
+      "loss": 0.0326,
+      "step": 345,
+      "student_loss": 0.0104219326749444,
+      "teacher_loss": 0.00029996069497428834
+    },
+    {
+      "epoch": 0.10092272202998846,
+      "grad_norm": 0.8203125,
+      "kd_loss": 0.05953724682331085,
+      "learning_rate": 7.560523424491595e-06,
+      "loss": 0.0265,
+      "step": 350,
+      "student_loss": 0.000867739028763026,
+      "teacher_loss": 0.0006243465468287468
+    },
+    {
+      "epoch": 0.10236447520184544,
+      "grad_norm": 1.734375,
+      "kd_loss": 0.01555018313229084,
+      "learning_rate": 7.490644698487909e-06,
+      "loss": 0.0344,
+      "step": 355,
+      "student_loss": 0.0004700198769569397,
+      "teacher_loss": 0.0004090165311936289
+    },
+    {
+      "epoch": 0.10380622837370242,
+      "grad_norm": 0.88671875,
+      "kd_loss": 0.01545359380543232,
+      "learning_rate": 7.420112844078066e-06,
+      "loss": 0.031,
+      "step": 360,
+      "student_loss": 0.0005492828786373138,
+      "teacher_loss": 0.0005034908535890281
+    },
+    {
+      "epoch": 0.1052479815455594,
+      "grad_norm": 0.875,
+      "kd_loss": 0.029742496088147163,
+      "learning_rate": 7.348946357018479e-06,
+      "loss": 0.0326,
+      "step": 365,
+      "student_loss": 0.043358221650123596,
+      "teacher_loss": 0.029302725568413734
+    },
+    {
+      "epoch": 0.10668973471741638,
+      "grad_norm": 3.75,
+      "kd_loss": 0.05247749388217926,
+      "learning_rate": 7.277163899486975e-06,
+      "loss": 0.0357,
+      "step": 370,
+      "student_loss": 0.020690687000751495,
+      "teacher_loss": 0.000206919910851866
+    },
+    {
+      "epoch": 0.10813148788927336,
+      "grad_norm": 0.6015625,
+      "kd_loss": 0.030279265716671944,
+      "learning_rate": 7.204784295188959e-06,
+      "loss": 0.0291,
+      "step": 375,
+      "student_loss": 0.03159727901220322,
+      "teacher_loss": 0.02069539576768875
+    },
+    {
+      "epoch": 0.10957324106113034,
+      "grad_norm": 0.57421875,
+      "kd_loss": 0.01892879046499729,
+      "learning_rate": 7.1318265244212305e-06,
+      "loss": 0.0302,
+      "step": 380,
+      "student_loss": 0.0004462770011741668,
+      "teacher_loss": 0.000304198736557737
+    },
+    {
+      "epoch": 0.11101499423298732,
+      "grad_norm": 0.62890625,
+      "kd_loss": 0.03064708411693573,
+      "learning_rate": 7.05830971909472e-06,
+      "loss": 0.0284,
+      "step": 385,
+      "student_loss": 0.0003714795457199216,
+      "teacher_loss": 0.0002954486117232591
+    },
+    {
+      "epoch": 0.11245674740484429,
+      "grad_norm": 1.0078125,
+      "kd_loss": 0.01886909268796444,
+      "learning_rate": 6.9842531577174865e-06,
+      "loss": 0.0277,
+      "step": 390,
+      "student_loss": 0.00026233118842355907,
+      "teacher_loss": 0.0002473078202456236
+    },
+    {
+      "epoch": 0.11389850057670127,
+      "grad_norm": 0.7109375,
+      "kd_loss": 0.02147393859922886,
+      "learning_rate": 6.9096762603392595e-06,
+      "loss": 0.031,
+      "step": 395,
+      "student_loss": 0.0006213847082108259,
+      "teacher_loss": 0.000366208259947598
+    },
+    {
+      "epoch": 0.11534025374855825,
+      "grad_norm": 1.7265625,
+      "kd_loss": 0.032356131821870804,
+      "learning_rate": 6.834598583458862e-06,
+      "loss": 0.0277,
+      "step": 400,
+      "student_loss": 0.0007109013968147337,
+      "teacher_loss": 0.0001410621771356091
+    },
+    {
+      "epoch": 0.11678200692041522,
+      "grad_norm": 1.1328125,
+      "kd_loss": 0.03028486669063568,
+      "learning_rate": 6.7590398148958625e-06,
+      "loss": 0.0334,
+      "step": 405,
+      "student_loss": 0.011747114360332489,
+      "teacher_loss": 0.000281431624898687
+    },
+    {
+      "epoch": 0.1182237600922722,
+      "grad_norm": 1.8359375,
+      "kd_loss": 0.03624144569039345,
+      "learning_rate": 6.6830197686277945e-06,
+      "loss": 0.0362,
+      "step": 410,
+      "student_loss": 0.0136904027312994,
+      "teacher_loss": 0.008161901496350765
+    },
+    {
+      "epoch": 0.11966551326412918,
+      "grad_norm": 1.046875,
+      "kd_loss": 0.02301396243274212,
+      "learning_rate": 6.6065583795942625e-06,
+      "loss": 0.0345,
+      "step": 415,
+      "student_loss": 0.004049207549542189,
+      "teacher_loss": 0.003510379930958152
+    },
+    {
+      "epoch": 0.12110726643598616,
+      "grad_norm": 2.828125,
+      "kd_loss": 0.04604360833764076,
+      "learning_rate": 6.52967569846937e-06,
+      "loss": 0.0328,
+      "step": 420,
+      "student_loss": 0.04198378324508667,
+      "teacher_loss": 0.021758265793323517
+    },
+    {
+      "epoch": 0.12254901960784313,
+      "grad_norm": 1.1640625,
+      "kd_loss": 0.012796198017895222,
+      "learning_rate": 6.452391886403767e-06,
+      "loss": 0.0299,
+      "step": 425,
+      "student_loss": 0.0018603568896651268,
+      "teacher_loss": 0.00185443798545748
+    },
+    {
+      "epoch": 0.12399077277970011,
+      "grad_norm": 1.234375,
+      "kd_loss": 0.05164352431893349,
+      "learning_rate": 6.374727209737743e-06,
+      "loss": 0.0362,
+      "step": 430,
+      "student_loss": 0.05342794209718704,
+      "teacher_loss": 0.05994100123643875
+    },
+    {
+      "epoch": 0.1254325259515571,
+      "grad_norm": 0.90625,
+      "kd_loss": 0.015498572029173374,
+      "learning_rate": 6.296702034686726e-06,
+      "loss": 0.0299,
+      "step": 435,
+      "student_loss": 0.0007954059401527047,
+      "teacher_loss": 0.000588005583267659
+    },
+    {
+      "epoch": 0.12687427912341406,
+      "grad_norm": 1.84375,
+      "kd_loss": 0.03917201980948448,
+      "learning_rate": 6.218336822000598e-06,
+      "loss": 0.0315,
+      "step": 440,
+      "student_loss": 0.012786084786057472,
+      "teacher_loss": 0.007422462105751038
+    },
+    {
+      "epoch": 0.12831603229527105,
+      "grad_norm": 1.7578125,
+      "kd_loss": 0.03229602053761482,
+      "learning_rate": 6.139652121598219e-06,
+      "loss": 0.0312,
+      "step": 445,
+      "student_loss": 0.02884654700756073,
+      "teacher_loss": 0.018903804942965508
+    },
+    {
+      "epoch": 0.12975778546712802,
+      "grad_norm": 0.75390625,
+      "kd_loss": 0.035617679357528687,
+      "learning_rate": 6.060668567178561e-06,
+      "loss": 0.0324,
+      "step": 450,
+      "student_loss": 0.0018460770370438695,
+      "teacher_loss": 0.0019348791101947427
+    },
+    {
+      "epoch": 0.131199538638985,
+      "grad_norm": 0.9765625,
+      "kd_loss": 0.0225622970610857,
+      "learning_rate": 5.981406870809889e-06,
+      "loss": 0.0329,
+      "step": 455,
+      "student_loss": 0.0007274725358001888,
+      "teacher_loss": 0.0003756978258024901
+    },
+    {
+      "epoch": 0.13264129181084197,
+      "grad_norm": 1.1171875,
+      "kd_loss": 0.043778520077466965,
+      "learning_rate": 5.9018878174983674e-06,
+      "loss": 0.0285,
+      "step": 460,
+      "student_loss": 0.004358124919235706,
+      "teacher_loss": 0.0003043531905859709
+    },
+    {
+      "epoch": 0.13408304498269896,
+      "grad_norm": 1.078125,
+      "kd_loss": 0.027884893119335175,
+      "learning_rate": 5.822132259737565e-06,
+      "loss": 0.0329,
+      "step": 465,
+      "student_loss": 0.0003728284209500998,
+      "teacher_loss": 0.00016045373922679573
+    },
+    {
+      "epoch": 0.13552479815455595,
+      "grad_norm": 0.69921875,
+      "kd_loss": 0.02432670071721077,
+      "learning_rate": 5.742161112040237e-06,
+      "loss": 0.0301,
+      "step": 470,
+      "student_loss": 0.0002982575388159603,
+      "teacher_loss": 0.00018306604761164635
+    },
+    {
+      "epoch": 0.13696655132641292,
+      "grad_norm": 0.90625,
+      "kd_loss": 0.029901551082730293,
+      "learning_rate": 5.661995345453867e-06,
+      "loss": 0.0271,
+      "step": 475,
+      "student_loss": 0.0004417268792167306,
+      "teacher_loss": 0.0002262392226839438
+    },
+    {
+      "epoch": 0.1384083044982699,
+      "grad_norm": 0.890625,
+      "kd_loss": 0.028335485607385635,
+      "learning_rate": 5.581655982061367e-06,
+      "loss": 0.04,
+      "step": 480,
+      "student_loss": 0.03686260059475899,
+      "teacher_loss": 0.028967848047614098
+    },
+    {
+      "epoch": 0.13985005767012687,
+      "grad_norm": 1.03125,
+      "kd_loss": 0.014099025167524815,
+      "learning_rate": 5.501164089468406e-06,
+      "loss": 0.03,
+      "step": 485,
+      "student_loss": 0.00044251122744753957,
+      "teacher_loss": 0.00045048308675177395
+    },
+    {
+      "epoch": 0.14129181084198386,
+      "grad_norm": 1.0703125,
+      "kd_loss": 0.021259035915136337,
+      "learning_rate": 5.4205407752787884e-06,
+      "loss": 0.0348,
+      "step": 490,
+      "student_loss": 0.00038817411405034363,
+      "teacher_loss": 0.00027494796086102724
+    },
+    {
+      "epoch": 0.14273356401384082,
+      "grad_norm": 1.09375,
+      "kd_loss": 0.03478889912366867,
+      "learning_rate": 5.339807181559359e-06,
+      "loss": 0.0389,
+      "step": 495,
+      "student_loss": 0.0009461977751925588,
+      "teacher_loss": 0.00046065691276453435
+    },
+    {
+      "epoch": 0.14417531718569782,
+      "grad_norm": 0.91796875,
+      "kd_loss": 0.024866795167326927,
+      "learning_rate": 5.258984479295853e-06,
+      "loss": 0.0306,
+      "step": 500,
+      "student_loss": 0.00048162159509956837,
+      "teacher_loss": 0.0003469791554380208
+    },
+    {
+      "epoch": 0.14561707035755478,
+      "grad_norm": 1.046875,
+      "kd_loss": 0.022635649889707565,
+      "learning_rate": 5.1780938628411795e-06,
+      "loss": 0.0377,
+      "step": 505,
+      "student_loss": 0.0004350949893705547,
+      "teacher_loss": 0.0003242486563976854
+    },
+    {
+      "epoch": 0.14705882352941177,
+      "grad_norm": 1.3125,
+      "kd_loss": 0.02024003490805626,
+      "learning_rate": 5.097156544357567e-06,
+      "loss": 0.0313,
+      "step": 510,
+      "student_loss": 0.0002926869783550501,
+      "teacher_loss": 0.0002203087497036904
+    },
+    {
+      "epoch": 0.14850057670126873,
+      "grad_norm": 1.21875,
+      "kd_loss": 0.024929286912083626,
+      "learning_rate": 5.016193748254045e-06,
+      "loss": 0.0286,
+      "step": 515,
+      "student_loss": 0.010843674652278423,
+      "teacher_loss": 0.00017637628479860723
+    },
+    {
+      "epoch": 0.14994232987312572,
+      "grad_norm": 1.4140625,
+      "kd_loss": 0.06931846588850021,
+      "learning_rate": 4.935226705620699e-06,
+      "loss": 0.0338,
+      "step": 520,
+      "student_loss": 0.03976204991340637,
+      "teacher_loss": 0.011458049528300762
+    },
+    {
+      "epoch": 0.1513840830449827,
+      "grad_norm": 0.68359375,
+      "kd_loss": 0.03341987729072571,
+      "learning_rate": 4.8542766486612035e-06,
+      "loss": 0.0333,
+      "step": 525,
+      "student_loss": 0.00037662044633179903,
+      "teacher_loss": 0.00019793077080976218
+    },
+    {
+      "epoch": 0.15282583621683968,
+      "grad_norm": 1.09375,
+      "kd_loss": 0.025850094854831696,
+      "learning_rate": 4.773364805125025e-06,
+      "loss": 0.0312,
+      "step": 530,
+      "student_loss": 0.00043694477062672377,
+      "teacher_loss": 0.00031476584263145924
+    },
+    {
+      "epoch": 0.15426758938869667,
+      "grad_norm": 1.5234375,
+      "kd_loss": 0.03888598829507828,
+      "learning_rate": 4.6925123927408265e-06,
+      "loss": 0.0306,
+      "step": 535,
+      "student_loss": 0.000664425955619663,
+      "teacher_loss": 0.0004196573281660676
+    },
+    {
+      "epoch": 0.15570934256055363,
+      "grad_norm": 0.99609375,
+      "kd_loss": 0.05486282706260681,
+      "learning_rate": 4.611740613652485e-06,
+      "loss": 0.0287,
+      "step": 540,
+      "student_loss": 0.012853885069489479,
+      "teacher_loss": 0.00035151999327354133
+    },
+    {
+      "epoch": 0.15715109573241062,
+      "grad_norm": 1.0390625,
+      "kd_loss": 0.0184871144592762,
+      "learning_rate": 4.531070648859186e-06,
+      "loss": 0.0279,
+      "step": 545,
+      "student_loss": 0.0017356052994728088,
+      "teacher_loss": 0.001496323267929256
+    },
+    {
+      "epoch": 0.15859284890426759,
+      "grad_norm": 0.98828125,
+      "kd_loss": 0.047685761004686356,
+      "learning_rate": 4.450523652661086e-06,
+      "loss": 0.0272,
+      "step": 550,
+      "student_loss": 0.00019285024609416723,
+      "teacher_loss": 0.003556112525984645
+    },
+    {
+      "epoch": 0.16003460207612458,
+      "grad_norm": 1.1328125,
+      "kd_loss": 0.07883454859256744,
+      "learning_rate": 4.370120747111956e-06,
+      "loss": 0.0318,
+      "step": 555,
+      "student_loss": 0.037417806684970856,
+      "teacher_loss": 0.001769757247529924
+    },
+    {
+      "epoch": 0.16147635524798154,
+      "grad_norm": 1.65625,
+      "kd_loss": 0.04116034507751465,
+      "learning_rate": 4.289883016480291e-06,
+      "loss": 0.0342,
+      "step": 560,
+      "student_loss": 0.00150584289804101,
+      "teacher_loss": 0.00032299800659529865
+    },
+    {
+      "epoch": 0.16291810841983853,
+      "grad_norm": 0.8515625,
+      "kd_loss": 0.04062850773334503,
+      "learning_rate": 4.209831501720328e-06,
+      "loss": 0.0322,
+      "step": 565,
+      "student_loss": 0.026969417929649353,
+      "teacher_loss": 0.0193944051861763
+    },
+    {
+      "epoch": 0.1643598615916955,
+      "grad_norm": 1.46875,
+      "kd_loss": 0.032333601266145706,
+      "learning_rate": 4.129987194954421e-06,
+      "loss": 0.0356,
+      "step": 570,
+      "student_loss": 0.010522445663809776,
+      "teacher_loss": 0.007194437552243471
+    },
+    {
+      "epoch": 0.16580161476355249,
+      "grad_norm": 0.8671875,
+      "kd_loss": 0.02821211703121662,
+      "learning_rate": 4.050371033968216e-06,
+      "loss": 0.0326,
+      "step": 575,
+      "student_loss": 0.00044184463331475854,
+      "teacher_loss": 0.00024692591978237033
+    },
+    {
+      "epoch": 0.16724336793540945,
+      "grad_norm": 0.8046875,
+      "kd_loss": 0.02497352659702301,
+      "learning_rate": 3.9710038967200825e-06,
+      "loss": 0.0285,
+      "step": 580,
+      "student_loss": 0.002113678492605686,
+      "teacher_loss": 0.0031359801068902016
+    },
+    {
+      "epoch": 0.16868512110726644,
+      "grad_norm": 1.578125,
+      "kd_loss": 0.030017584562301636,
+      "learning_rate": 3.89190659586623e-06,
+      "loss": 0.0276,
+      "step": 585,
+      "student_loss": 0.0005331283318810165,
+      "teacher_loss": 0.00024908268824219704
+    },
+    {
+      "epoch": 0.1701268742791234,
+      "grad_norm": 1.421875,
+      "kd_loss": 0.025302177295088768,
+      "learning_rate": 3.8130998733029517e-06,
+      "loss": 0.0308,
+      "step": 590,
+      "student_loss": 0.0006629621493630111,
+      "teacher_loss": 0.0010983508545905352
+    },
+    {
+      "epoch": 0.1715686274509804,
+      "grad_norm": 1.4921875,
+      "kd_loss": 0.02107221819460392,
+      "learning_rate": 3.734604394727419e-06,
+      "loss": 0.0315,
+      "step": 595,
+      "student_loss": 0.00039723835652694106,
+      "teacher_loss": 0.0003521108883433044
+    },
+    {
+      "epoch": 0.17301038062283736,
+      "grad_norm": 1.2578125,
+      "kd_loss": 0.03979631885886192,
+      "learning_rate": 3.656440744218464e-06,
+      "loss": 0.0342,
+      "step": 600,
+      "student_loss": 0.04468826204538345,
+      "teacher_loss": 0.0050285086035728455
+    },
+    {
+      "epoch": 0.17445213379469435,
+      "grad_norm": 1.046875,
+      "kd_loss": 0.043974071741104126,
+      "learning_rate": 3.578629418838757e-06,
+      "loss": 0.0337,
+      "step": 605,
+      "student_loss": 0.09002169221639633,
+      "teacher_loss": 0.03622898831963539
+    },
+    {
+      "epoch": 0.17589388696655134,
+      "grad_norm": 1.0,
+      "kd_loss": 0.04286090284585953,
+      "learning_rate": 3.5011908232598124e-06,
+      "loss": 0.0313,
+      "step": 610,
+      "student_loss": 0.009437715634703636,
+      "teacher_loss": 0.0004902155487798154
+    },
+    {
+      "epoch": 0.1773356401384083,
+      "grad_norm": 1.4375,
+      "kd_loss": 0.024805590510368347,
+      "learning_rate": 3.4241452644112085e-06,
+      "loss": 0.0275,
+      "step": 615,
+      "student_loss": 0.0003357514797244221,
+      "teacher_loss": 0.0003598474431782961
+    },
+    {
+      "epoch": 0.1787773933102653,
+      "grad_norm": 1.2109375,
+      "kd_loss": 0.032709866762161255,
+      "learning_rate": 3.3475129461554567e-06,
+      "loss": 0.0383,
+      "step": 620,
+      "student_loss": 0.010687335394322872,
+      "teacher_loss": 0.007166274357587099
+    },
+    {
+      "epoch": 0.18021914648212226,
+      "grad_norm": 1.234375,
+      "kd_loss": 0.03386635705828667,
+      "learning_rate": 3.271313963989886e-06,
+      "loss": 0.0293,
+      "step": 625,
+      "student_loss": 0.014498386532068253,
+      "teacher_loss": 0.004959889687597752
+    },
+    {
+      "epoch": 0.18166089965397925,
+      "grad_norm": 0.96484375,
+      "kd_loss": 0.04790615662932396,
+      "learning_rate": 3.195568299776945e-06,
+      "loss": 0.0314,
+      "step": 630,
+      "student_loss": 0.001984821166843176,
+      "teacher_loss": 0.0007946855621412396
+    },
+    {
+      "epoch": 0.1831026528258362,
+      "grad_norm": 0.89453125,
+      "kd_loss": 0.032921187579631805,
+      "learning_rate": 3.1202958165043053e-06,
+      "loss": 0.0427,
+      "step": 635,
+      "student_loss": 0.002781020710244775,
+      "teacher_loss": 0.00033777387579903007
+    },
+    {
+      "epoch": 0.1845444059976932,
+      "grad_norm": 1.046875,
+      "kd_loss": 0.03445451706647873,
+      "learning_rate": 3.045516253076137e-06,
+      "loss": 0.0285,
+      "step": 640,
+      "student_loss": 0.0006051648524589837,
+      "teacher_loss": 0.00031487070373259485
+    },
+    {
+      "epoch": 0.18598615916955016,
+      "grad_norm": 1.0078125,
+      "kd_loss": 0.032418981194496155,
+      "learning_rate": 2.9712492191369245e-06,
+      "loss": 0.0287,
+      "step": 645,
+      "student_loss": 0.0008554414962418377,
+      "teacher_loss": 0.0009805671870708466
+    },
+    {
+      "epoch": 0.18742791234140715,
+      "grad_norm": 0.97265625,
+      "kd_loss": 0.02188216708600521,
+      "learning_rate": 2.8975141899291777e-06,
+      "loss": 0.029,
+      "step": 650,
+      "student_loss": 0.0008296438027173281,
+      "teacher_loss": 0.0008416228229179978
+    },
+    {
+      "epoch": 0.18886966551326412,
+      "grad_norm": 1.390625,
+      "kd_loss": 0.045378196984529495,
+      "learning_rate": 2.8243305011863843e-06,
+      "loss": 0.0293,
+      "step": 655,
+      "student_loss": 0.000480753427837044,
+      "teacher_loss": 0.0003968694363720715
+    },
+    {
+      "epoch": 0.1903114186851211,
+      "grad_norm": 1.6953125,
+      "kd_loss": 0.02608320489525795,
+      "learning_rate": 2.751717344062552e-06,
+      "loss": 0.0342,
+      "step": 660,
+      "student_loss": 0.0006710074958391488,
+      "teacher_loss": 0.0008470130851492286
+    },
+    {
+      "epoch": 0.19175317185697807,
+      "grad_norm": 2.15625,
+      "kd_loss": 0.06641045957803726,
+      "learning_rate": 2.6796937600996587e-06,
+      "loss": 0.035,
+      "step": 665,
+      "student_loss": 0.0012525359634310007,
+      "teacher_loss": 0.0005023120902478695
+    },
+    {
+      "epoch": 0.19319492502883506,
+      "grad_norm": 1.7421875,
+      "kd_loss": 0.0324208065867424,
+      "learning_rate": 2.6082786362343377e-06,
+      "loss": 0.0351,
+      "step": 670,
+      "student_loss": 0.0069810450077056885,
+      "teacher_loss": 0.0013353026006370783
+    },
+    {
+      "epoch": 0.19463667820069205,
+      "grad_norm": 0.8828125,
+      "kd_loss": 0.018830087035894394,
+      "learning_rate": 2.5374906998451094e-06,
+      "loss": 0.0339,
+      "step": 675,
+      "student_loss": 0.0012051296653226018,
+      "teacher_loss": 0.0009222541702911258
+    },
+    {
+      "epoch": 0.19607843137254902,
+      "grad_norm": 1.1328125,
+      "kd_loss": 0.06835323572158813,
+      "learning_rate": 2.467348513841447e-06,
+      "loss": 0.0284,
+      "step": 680,
+      "student_loss": 0.09790011495351791,
+      "teacher_loss": 0.03948701545596123
+    },
+    {
+      "epoch": 0.197520184544406,
+      "grad_norm": 0.859375,
+      "kd_loss": 0.029812660068273544,
+      "learning_rate": 2.3978704717959777e-06,
+      "loss": 0.0317,
+      "step": 685,
+      "student_loss": 0.037691812962293625,
+      "teacher_loss": 0.03045065887272358
+    },
+    {
+      "epoch": 0.19896193771626297,
+      "grad_norm": 0.84765625,
+      "kd_loss": 0.055110517889261246,
+      "learning_rate": 2.329074793121085e-06,
+      "loss": 0.0326,
+      "step": 690,
+      "student_loss": 0.007291011977940798,
+      "teacher_loss": 0.011587987653911114
+    },
+    {
+      "epoch": 0.20040369088811996,
+      "grad_norm": 1.0078125,
+      "kd_loss": 0.03621595352888107,
+      "learning_rate": 2.260979518291186e-06,
+      "loss": 0.0339,
+      "step": 695,
+      "student_loss": 0.0034336186945438385,
+      "teacher_loss": 0.006187558639794588
+    },
+    {
+      "epoch": 0.20184544405997693,
+      "grad_norm": 2.0,
+      "kd_loss": 0.023546168580651283,
+      "learning_rate": 2.1936025041119268e-06,
+      "loss": 0.0308,
+      "step": 700,
+      "student_loss": 0.00042453769128769636,
+      "teacher_loss": 0.00032098451629281044
+    },
+    {
+      "epoch": 0.20328719723183392,
+      "grad_norm": 2.109375,
+      "kd_loss": 0.019782008603215218,
+      "learning_rate": 2.1269614190375477e-06,
+      "loss": 0.0308,
+      "step": 705,
+      "student_loss": 0.001072320039384067,
+      "teacher_loss": 0.0010817889124155045
+    },
+    {
+      "epoch": 0.20472895040369088,
+      "grad_norm": 0.8046875,
+      "kd_loss": 0.06822223961353302,
+      "learning_rate": 2.061073738537635e-06,
+      "loss": 0.0341,
+      "step": 710,
+      "student_loss": 0.008647946640849113,
+      "teacher_loss": 0.004804851021617651
+    },
+    {
+      "epoch": 0.20617070357554787,
+      "grad_norm": 1.4453125,
+      "kd_loss": 0.025431152433156967,
+      "learning_rate": 1.9959567405144825e-06,
+      "loss": 0.0375,
+      "step": 715,
+      "student_loss": 0.0023363358341157436,
+      "teacher_loss": 0.005615272093564272
+    },
+    {
+      "epoch": 0.20761245674740483,
+      "grad_norm": 0.72265625,
+      "kd_loss": 0.02818908356130123,
+      "learning_rate": 1.931627500772263e-06,
+      "loss": 0.0335,
+      "step": 720,
+      "student_loss": 0.0009902354795485735,
+      "teacher_loss": 0.0034803280141204596
+    },
+    {
+      "epoch": 0.20905420991926182,
+      "grad_norm": 1.5546875,
+      "kd_loss": 0.021861886605620384,
+      "learning_rate": 1.8681028885391905e-06,
+      "loss": 0.0262,
+      "step": 725,
+      "student_loss": 0.00040081748738884926,
+      "teacher_loss": 0.00037730656913481653
+    },
+    {
+      "epoch": 0.2104959630911188,
+      "grad_norm": 1.5625,
+      "kd_loss": 0.037239283323287964,
+      "learning_rate": 1.8053995620438625e-06,
+      "loss": 0.0314,
+      "step": 730,
+      "student_loss": 0.0068473792634904385,
+      "teacher_loss": 0.0015034136595204473
+    },
+    {
+      "epoch": 0.21193771626297578,
+      "grad_norm": 1.1796875,
+      "kd_loss": 0.03357340022921562,
+      "learning_rate": 1.743533964146924e-06,
+      "loss": 0.0364,
+      "step": 735,
+      "student_loss": 0.0006027538329362869,
+      "teacher_loss": 0.00036633480340242386
+    },
+    {
+      "epoch": 0.21337946943483277,
+      "grad_norm": 0.765625,
+      "kd_loss": 0.03631579503417015,
+      "learning_rate": 1.6825223180292138e-06,
+      "loss": 0.0292,
+      "step": 740,
+      "student_loss": 0.0038133456837385893,
+      "teacher_loss": 0.00020591833163052797
+    },
+    {
+      "epoch": 0.21482122260668973,
+      "grad_norm": 0.75,
+      "kd_loss": 0.02808128111064434,
+      "learning_rate": 1.6223806229375182e-06,
+      "loss": 0.0308,
+      "step": 745,
+      "student_loss": 0.03830280900001526,
+      "teacher_loss": 0.018777813762426376
+    },
+    {
+      "epoch": 0.21626297577854672,
+      "grad_norm": 0.8046875,
+      "kd_loss": 0.06128418445587158,
+      "learning_rate": 1.563124649989043e-06,
+      "loss": 0.0301,
+      "step": 750,
+      "student_loss": 0.014656578190624714,
+      "teacher_loss": 0.004599400795996189
+    },
+    {
+      "epoch": 0.2177047289504037,
+      "grad_norm": 1.015625,
+      "kd_loss": 0.028206797316670418,
+      "learning_rate": 1.5047699380357134e-06,
+      "loss": 0.0358,
+      "step": 755,
+      "student_loss": 0.031989697366952896,
+      "teacher_loss": 0.00652338145300746
+    },
+    {
+      "epoch": 0.21914648212226068,
+      "grad_norm": 1.6953125,
+      "kd_loss": 0.06580331921577454,
+      "learning_rate": 1.4473317895893773e-06,
+      "loss": 0.0291,
+      "step": 760,
+      "student_loss": 0.050039730966091156,
+      "teacher_loss": 0.001301902811974287
+    },
+    {
+      "epoch": 0.22058823529411764,
+      "grad_norm": 1.2265625,
+      "kd_loss": 0.05050547048449516,
+      "learning_rate": 1.39082526680899e-06,
+      "loss": 0.0359,
+      "step": 765,
+      "student_loss": 0.002995565999299288,
+      "teacher_loss": 0.0011588622583076358
+    },
+    {
+      "epoch": 0.22202998846597463,
+      "grad_norm": 0.7890625,
+      "kd_loss": 0.05726822093129158,
+      "learning_rate": 1.3352651875508204e-06,
+      "loss": 0.0315,
+      "step": 770,
+      "student_loss": 0.0010233366629108787,
+      "teacher_loss": 0.0013149199075996876
+    },
+    {
+      "epoch": 0.2234717416378316,
+      "grad_norm": 1.3671875,
+      "kd_loss": 0.03695448115468025,
+      "learning_rate": 1.2806661214827286e-06,
+      "loss": 0.0289,
+      "step": 775,
+      "student_loss": 0.0019939893390983343,
+      "teacher_loss": 0.00043505022767931223
+    },
+    {
+      "epoch": 0.22491349480968859,
+      "grad_norm": 1.1796875,
+      "kd_loss": 0.032327134162187576,
+      "learning_rate": 1.2270423862635188e-06,
+      "loss": 0.0323,
+      "step": 780,
+      "student_loss": 0.0004899216000922024,
+      "teacher_loss": 0.00032206837204284966
+    },
+    {
+      "epoch": 0.22635524798154555,
+      "grad_norm": 1.125,
+      "kd_loss": 0.02126060053706169,
+      "learning_rate": 1.1744080437883859e-06,
+      "loss": 0.0289,
+      "step": 785,
+      "student_loss": 0.0004751235246658325,
+      "teacher_loss": 0.0013089554850012064
+    },
+    {
+      "epoch": 0.22779700115340254,
+      "grad_norm": 1.6953125,
+      "kd_loss": 0.018842022866010666,
+      "learning_rate": 1.1227768965014246e-06,
+      "loss": 0.0281,
+      "step": 790,
+      "student_loss": 0.0028977859765291214,
+      "teacher_loss": 0.0014279329916462302
+    },
+    {
+      "epoch": 0.2292387543252595,
+      "grad_norm": 0.87109375,
+      "kd_loss": 0.030527956783771515,
+      "learning_rate": 1.0721624837761768e-06,
+      "loss": 0.0317,
+      "step": 795,
+      "student_loss": 0.0006763353012502193,
+      "teacher_loss": 0.0010857924353331327
+    },
+    {
+      "epoch": 0.2306805074971165,
+      "grad_norm": 1.7890625,
+      "kd_loss": 0.05231979116797447,
+      "learning_rate": 1.0225780783651689e-06,
+      "loss": 0.0364,
+      "step": 800,
+      "student_loss": 0.06148466467857361,
+      "teacher_loss": 0.02855808101594448
+    },
+    {
+      "epoch": 0.23212226066897348,
+      "grad_norm": 1.046875,
+      "kd_loss": 0.023236989974975586,
+      "learning_rate": 9.740366829193587e-07,
+      "loss": 0.0345,
+      "step": 805,
+      "student_loss": 0.0005403392133302987,
+      "teacher_loss": 0.00044607033487409353
+    },
+    {
+      "epoch": 0.23356401384083045,
+      "grad_norm": 34.75,
+      "kd_loss": 0.040367491543293,
+      "learning_rate": 9.265510265784189e-07,
+      "loss": 0.0366,
+      "step": 810,
+      "student_loss": 0.0004765975463669747,
+      "teacher_loss": 0.00021999947784934193
+    },
+    {
+      "epoch": 0.23500576701268744,
+      "grad_norm": 2.578125,
+      "kd_loss": 0.022637881338596344,
+      "learning_rate": 8.801335616327378e-07,
+      "loss": 0.0308,
+      "step": 815,
+      "student_loss": 0.0018720644293352962,
+      "teacher_loss": 0.010451914742588997
+    },
+    {
+      "epoch": 0.2364475201845444,
+      "grad_norm": 1.34375,
+      "kd_loss": 0.037494342774152756,
+      "learning_rate": 8.347964602580245e-07,
+      "loss": 0.0309,
+      "step": 820,
+      "student_loss": 0.007085499819368124,
+      "teacher_loss": 0.0015908918576315045
+    },
+    {
+      "epoch": 0.2378892733564014,
+      "grad_norm": 1.875,
+      "kd_loss": 0.042802467942237854,
+      "learning_rate": 7.905516113233652e-07,
+      "loss": 0.0283,
+      "step": 825,
+      "student_loss": 0.0008434330811724067,
+      "teacher_loss": 0.016496969386935234
+    },
+    {
+      "epoch": 0.23933102652825836,
+      "grad_norm": 0.6484375,
+      "kd_loss": 0.0369848757982254,
+      "learning_rate": 7.474106172735746e-07,
+      "loss": 0.0312,
+      "step": 830,
+      "student_loss": 0.00888098869472742,
+      "teacher_loss": 0.0041339099407196045
+    },
+    {
+      "epoch": 0.24077277970011535,
+      "grad_norm": 1.7578125,
+      "kd_loss": 0.033153776079416275,
+      "learning_rate": 7.053847910866513e-07,
+      "loss": 0.0284,
+      "step": 835,
+      "student_loss": 0.013508547097444534,
+      "teacher_loss": 0.004140567034482956
+    },
+    {
+      "epoch": 0.2422145328719723,
+      "grad_norm": 0.90234375,
+      "kd_loss": 0.03923244774341583,
+      "learning_rate": 6.644851533071556e-07,
+      "loss": 0.0304,
+      "step": 840,
+      "student_loss": 0.00044696227996610105,
+      "teacher_loss": 0.00019727443577721715
+    },
+    {
+      "epoch": 0.2436562860438293,
+      "grad_norm": 0.90625,
+      "kd_loss": 0.027709143236279488,
+      "learning_rate": 6.24722429156251e-07,
+      "loss": 0.0458,
+      "step": 845,
+      "student_loss": 0.000737609516363591,
+      "teacher_loss": 0.0007416712469421327
+    },
+    {
+      "epoch": 0.24509803921568626,
+      "grad_norm": 0.953125,
+      "kd_loss": 0.046404365450143814,
+      "learning_rate": 5.861070457192081e-07,
+      "loss": 0.0332,
+      "step": 850,
+      "student_loss": 0.01347807701677084,
+      "teacher_loss": 0.0005352996522560716
+    },
+    {
+      "epoch": 0.24653979238754326,
+      "grad_norm": 0.828125,
+      "kd_loss": 0.03019327111542225,
+      "learning_rate": 5.486491292110796e-07,
+      "loss": 0.0311,
+      "step": 855,
+      "student_loss": 0.0016365062911063433,
+      "teacher_loss": 0.0002644163032528013
+    },
+    {
+      "epoch": 0.24798154555940022,
+      "grad_norm": 1.15625,
+      "kd_loss": 0.02993069402873516,
+      "learning_rate": 5.123585023212785e-07,
+      "loss": 0.0369,
+      "step": 860,
+      "student_loss": 0.0016339129069820046,
+      "teacher_loss": 0.0011585139436647296
+    },
+    {
+      "epoch": 0.2494232987312572,
+      "grad_norm": 0.84765625,
+      "kd_loss": 0.03818565607070923,
+      "learning_rate": 4.772446816377408e-07,
+      "loss": 0.0331,
+      "step": 865,
+      "student_loss": 0.0003991369449067861,
+      "teacher_loss": 0.0005172351375222206
+    },
+    {
+      "epoch": 0.2508650519031142,
+      "grad_norm": 0.69140625,
+      "kd_loss": 0.038387052714824677,
+      "learning_rate": 4.4331687515137614e-07,
+      "loss": 0.0324,
+      "step": 870,
+      "student_loss": 0.0008950178162194788,
+      "teacher_loss": 0.0005668822559528053
+    },
+    {
+      "epoch": 0.25230680507497116,
+      "grad_norm": 1.359375,
+      "kd_loss": 0.027386674657464027,
+      "learning_rate": 4.1058397984142405e-07,
+      "loss": 0.0313,
+      "step": 875,
+      "student_loss": 0.0002425290731480345,
+      "teacher_loss": 0.00022121419897302985
+    },
+    {
+      "epoch": 0.2537485582468281,
+      "grad_norm": 0.76171875,
+      "kd_loss": 0.037135638296604156,
+      "learning_rate": 3.790545793423761e-07,
+      "loss": 0.033,
+      "step": 880,
+      "student_loss": 0.0007690261118113995,
+      "teacher_loss": 0.00036677613388746977
+    },
+    {
+      "epoch": 0.25519031141868515,
+      "grad_norm": 1.3984375,
+      "kd_loss": 0.07571214437484741,
+      "learning_rate": 3.4873694169306915e-07,
+      "loss": 0.027,
+      "step": 885,
+      "student_loss": 0.00854508951306343,
+      "teacher_loss": 0.050961658358573914
+    },
+    {
+      "epoch": 0.2566320645905421,
+      "grad_norm": 0.6953125,
+      "kd_loss": 0.023151278495788574,
+      "learning_rate": 3.196390171685343e-07,
+      "loss": 0.0329,
+      "step": 890,
+      "student_loss": 0.0004915237659588456,
+      "teacher_loss": 0.00039783472311683
+    },
+    {
+      "epoch": 0.25807381776239907,
+      "grad_norm": 0.85546875,
+      "kd_loss": 0.0425008125603199,
+      "learning_rate": 2.917684361951728e-07,
+      "loss": 0.031,
+      "step": 895,
+      "student_loss": 0.012177797965705395,
+      "teacher_loss": 0.004335513338446617
+    },
+    {
+      "epoch": 0.25951557093425603,
+      "grad_norm": 1.3828125,
+      "kd_loss": 0.03248598799109459,
+      "learning_rate": 2.65132507349814e-07,
+      "loss": 0.0339,
+      "step": 900,
+      "student_loss": 0.028939327225089073,
+      "teacher_loss": 0.02277800440788269
+    },
+    {
+      "epoch": 0.26095732410611305,
+      "grad_norm": 0.86328125,
+      "kd_loss": 0.03494875878095627,
+      "learning_rate": 2.397382154431621e-07,
+      "loss": 0.0301,
+      "step": 905,
+      "student_loss": 0.003344937227666378,
+      "teacher_loss": 0.0022205705754458904
+    },
+    {
+      "epoch": 0.26239907727797,
+      "grad_norm": 1.109375,
+      "kd_loss": 0.01817970909178257,
+      "learning_rate": 2.1559221968815547e-07,
+      "loss": 0.0287,
+      "step": 910,
+      "student_loss": 0.00047651847125962377,
+      "teacher_loss": 0.0005133987870067358
+    },
+    {
+      "epoch": 0.263840830449827,
+      "grad_norm": 1.28125,
+      "kd_loss": 0.051077090203762054,
+      "learning_rate": 1.9270085195370048e-07,
+      "loss": 0.0293,
+      "step": 915,
+      "student_loss": 0.0040283347479999065,
+      "teacher_loss": 0.0008671078830957413
+    },
+    {
+      "epoch": 0.26528258362168394,
+      "grad_norm": 0.9375,
+      "kd_loss": 0.0243892353028059,
+      "learning_rate": 1.7107011510424766e-07,
+      "loss": 0.0308,
+      "step": 920,
+      "student_loss": 0.015063118189573288,
+      "teacher_loss": 0.012549618259072304
+    },
+    {
+      "epoch": 0.26672433679354096,
+      "grad_norm": 0.75390625,
+      "kd_loss": 0.02323193848133087,
+      "learning_rate": 1.5070568142564912e-07,
+      "loss": 0.0293,
+      "step": 925,
+      "student_loss": 0.00041171483462676406,
+      "teacher_loss": 0.0003514218551572412
+    },
+    {
+      "epoch": 0.2681660899653979,
+      "grad_norm": 0.95703125,
+      "kd_loss": 0.032398514449596405,
+      "learning_rate": 1.3161289113769405e-07,
+      "loss": 0.0296,
+      "step": 930,
+      "student_loss": 0.005882553290575743,
+      "teacher_loss": 0.0006708991713821888
+    },
+    {
+      "epoch": 0.2696078431372549,
+      "grad_norm": 1.0078125,
+      "kd_loss": 0.026842886582016945,
+      "learning_rate": 1.1379675099373489e-07,
+      "loss": 0.0287,
+      "step": 935,
+      "student_loss": 0.0020961251575499773,
+      "teacher_loss": 0.0022036894224584103
+    },
+    {
+      "epoch": 0.2710495963091119,
+      "grad_norm": 0.8203125,
+      "kd_loss": 0.0463658906519413,
+      "learning_rate": 9.726193296774767e-08,
+      "loss": 0.0326,
+      "step": 940,
+      "student_loss": 0.007851100526750088,
+      "teacher_loss": 0.0018817407544702291
+    },
+    {
+      "epoch": 0.27249134948096887,
+      "grad_norm": 0.765625,
+      "kd_loss": 0.024322666227817535,
+      "learning_rate": 8.201277302919086e-08,
+      "loss": 0.0335,
+      "step": 945,
+      "student_loss": 0.011676947586238384,
+      "teacher_loss": 0.0026983029674738646
+    },
+    {
+      "epoch": 0.27393310265282583,
+      "grad_norm": 0.66015625,
+      "kd_loss": 0.033283624798059464,
+      "learning_rate": 6.805327000596995e-08,
+      "loss": 0.0287,
+      "step": 950,
+      "student_loss": 0.0003928481601178646,
+      "teacher_loss": 0.00024390498583670706
+    },
+    {
+      "epoch": 0.2753748558246828,
+      "grad_norm": 0.96875,
+      "kd_loss": 0.028704380616545677,
+      "learning_rate": 5.538708453581787e-08,
+      "loss": 0.0305,
+      "step": 955,
+      "student_loss": 0.0011468256125226617,
+      "teacher_loss": 0.0007584833656437695
+    },
+    {
+      "epoch": 0.2768166089965398,
+      "grad_norm": 1.4765625,
+      "kd_loss": 0.03806009516119957,
+      "learning_rate": 4.40175381063529e-08,
+      "loss": 0.0324,
+      "step": 960,
+      "student_loss": 0.009709250181913376,
+      "teacher_loss": 0.006499860435724258
+    },
+    {
+      "epoch": 0.2782583621683968,
+      "grad_norm": 0.90625,
+      "kd_loss": 0.03816923499107361,
+      "learning_rate": 3.394761218407705e-08,
+      "loss": 0.0286,
+      "step": 965,
+      "student_loss": 0.010869469493627548,
+      "teacher_loss": 0.006945203524082899
+    },
+    {
+      "epoch": 0.27970011534025374,
+      "grad_norm": 1.4765625,
+      "kd_loss": 0.03507186844944954,
+      "learning_rate": 2.5179947432540376e-08,
+      "loss": 0.0306,
+      "step": 970,
+      "student_loss": 0.00040590367279946804,
+      "teacher_loss": 0.00021455588284879923
+    },
+    {
+      "epoch": 0.2811418685121107,
+      "grad_norm": 0.81640625,
+      "kd_loss": 0.026400946080684662,
+      "learning_rate": 1.7716843019867646e-08,
+      "loss": 0.0284,
+      "step": 975,
+      "student_loss": 0.014423542656004429,
+      "teacher_loss": 0.0009269384318031371
+    },
+    {
+      "epoch": 0.2825836216839677,
+      "grad_norm": 1.328125,
+      "kd_loss": 0.04043099284172058,
+      "learning_rate": 1.156025601584676e-08,
+      "loss": 0.0323,
+      "step": 980,
+      "student_loss": 0.0008977575344033539,
+      "teacher_loss": 0.0003824856539722532
+    },
+    {
+      "epoch": 0.2840253748558247,
+      "grad_norm": 1.421875,
+      "kd_loss": 0.030134929344058037,
+      "learning_rate": 6.711800878718144e-09,
+      "loss": 0.0317,
+      "step": 985,
+      "student_loss": 0.0003170230193063617,
+      "teacher_loss": 0.0003533938142936677
+    },
+    {
+      "epoch": 0.28546712802768165,
+      "grad_norm": 0.9296875,
+      "kd_loss": 0.028759652748703957,
+      "learning_rate": 3.1727490318111953e-09,
+      "loss": 0.0292,
+      "step": 990,
+      "student_loss": 0.002007333328947425,
+      "teacher_loss": 0.00026892361347563565
+    },
+    {
+      "epoch": 0.2869088811995386,
+      "grad_norm": 1.1484375,
+      "kd_loss": 0.04890036582946777,
+      "learning_rate": 9.440285301370865e-10,
+      "loss": 0.0341,
+      "step": 995,
+      "student_loss": 0.0031924662180244923,
+      "teacher_loss": 0.0016397708095610142
+    },
+    {
+      "epoch": 0.28835063437139563,
+      "grad_norm": 0.94140625,
+      "kd_loss": 0.03925730288028717,
+      "learning_rate": 2.622381702066523e-11,
+      "loss": 0.0292,
+      "step": 1000,
+      "student_loss": 0.006099311169236898,
+      "teacher_loss": 0.0002429906598990783
+    },
+    {
+      "epoch": 0.28835063437139563,
+      "kd_loss": 0.03925730288028717,
+      "step": 1000,
+      "student_loss": 0.006099311169236898,
+      "teacher_loss": 0.0002429906598990783,
+      "total_flos": 0.0,
+      "train_loss": 0.03772550980746746,
+      "train_runtime": 34708.404,
+      "train_samples_per_second": 0.461,
+      "train_steps_per_second": 0.029
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-200/added_tokens.json

@@ -0,0 +1,35 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|action_sep|>": 151670,
+  "<|arg_sep|>": 151671,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|call_sep|>": 151666,
+  "<|end_of_text|>": 151673,
+  "<|endoftext|>": 151643,
+  "<|exception_sep|>": 151669,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|frame_sep|>": 151672,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|latent_end|>": 151675,
+  "<|latent_start|>": 151674,
+  "<|line_sep|>": 151667,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|return_sep|>": 151668,
+  "<|trace_context_start|>": 151665,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-200/chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\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>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-200/config.json

@@ -0,0 +1,66 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 36,
+  "model_type": "qwen2",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 36,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151676
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-200/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14833b99179e6f01afea7848323d6529d63f779fec675d7fba520ac3d94855ac
+size 6187858991

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-200/special_tokens_map.json

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

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-200/thought_projector.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99ab3a4b1822a0586f802d83c96fafb652917eaa55763b113f1564e0d86cfebf
+size 16788033

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-200/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83a790d654474f5dfe225f889afd0210313eb1083f942671f2c4b8e95a1c922b
+size 11424004

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-200/tokenizer_config.json

@@ -0,0 +1,295 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<|trace_context_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151666": {
+      "content": "<|call_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151667": {
+      "content": "<|line_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151668": {
+      "content": "<|return_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151669": {
+      "content": "<|exception_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151670": {
+      "content": "<|action_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151671": {
+      "content": "<|arg_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151672": {
+      "content": "<|frame_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151673": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151674": {
+      "content": "<|latent_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151675": {
+      "content": "<|latent_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-200/trainer_state.json

@@ -0,0 +1,434 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.05767012687427912,
+  "eval_steps": 500,
+  "global_step": 200,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0014417531718569781,
+      "grad_norm": 47.75,
+      "kd_loss": 0.2916574478149414,
+      "learning_rate": 1.3333333333333334e-06,
+      "loss": 0.3836,
+      "step": 5,
+      "student_loss": 0.12182186543941498,
+      "teacher_loss": 0.002018272876739502
+    },
+    {
+      "epoch": 0.0028835063437139563,
+      "grad_norm": 22.5,
+      "kd_loss": 0.26723670959472656,
+      "learning_rate": 3e-06,
+      "loss": 0.2911,
+      "step": 10,
+      "student_loss": 0.10971185564994812,
+      "teacher_loss": 0.005489554721862078
+    },
+    {
+      "epoch": 0.004325259515570935,
+      "grad_norm": 6.9375,
+      "kd_loss": 0.21662698686122894,
+      "learning_rate": 4.666666666666667e-06,
+      "loss": 0.1789,
+      "step": 15,
+      "student_loss": 0.0025933366268873215,
+      "teacher_loss": 0.001338644651696086
+    },
+    {
+      "epoch": 0.0057670126874279125,
+      "grad_norm": 3.078125,
+      "kd_loss": 0.13032649457454681,
+      "learning_rate": 6.333333333333333e-06,
+      "loss": 0.1296,
+      "step": 20,
+      "student_loss": 0.023931587114930153,
+      "teacher_loss": 0.0006410324713215232
+    },
+    {
+      "epoch": 0.00720876585928489,
+      "grad_norm": 2.34375,
+      "kd_loss": 0.10749491304159164,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 0.107,
+      "step": 25,
+      "student_loss": 0.0032150023616850376,
+      "teacher_loss": 0.0010774527909234166
+    },
+    {
+      "epoch": 0.00865051903114187,
+      "grad_norm": 1.2890625,
+      "kd_loss": 0.0680936947464943,
+      "learning_rate": 9.666666666666667e-06,
+      "loss": 0.0841,
+      "step": 30,
+      "student_loss": 0.00099717837292701,
+      "teacher_loss": 0.0005822837701998651
+    },
+    {
+      "epoch": 0.010092272202998846,
+      "grad_norm": 2.109375,
+      "kd_loss": 0.06131501868367195,
+      "learning_rate": 9.99958042442916e-06,
+      "loss": 0.0685,
+      "step": 35,
+      "student_loss": 0.0027061135042458773,
+      "teacher_loss": 0.0006183119839988649
+    },
+    {
+      "epoch": 0.011534025374855825,
+      "grad_norm": 1.203125,
+      "kd_loss": 0.05697907507419586,
+      "learning_rate": 9.997876019358083e-06,
+      "loss": 0.0614,
+      "step": 40,
+      "student_loss": 0.029384225606918335,
+      "teacher_loss": 0.025684259831905365
+    },
+    {
+      "epoch": 0.012975778546712802,
+      "grad_norm": 1.296875,
+      "kd_loss": 0.09766291826963425,
+      "learning_rate": 9.99486100792044e-06,
+      "loss": 0.0476,
+      "step": 45,
+      "student_loss": 0.06478797644376755,
+      "teacher_loss": 0.0030884514562785625
+    },
+    {
+      "epoch": 0.01441753171856978,
+      "grad_norm": 1.453125,
+      "kd_loss": 0.03506260737776756,
+      "learning_rate": 9.990536180750724e-06,
+      "loss": 0.0517,
+      "step": 50,
+      "student_loss": 0.02092154324054718,
+      "teacher_loss": 0.01070804987102747
+    },
+    {
+      "epoch": 0.015859284890426758,
+      "grad_norm": 1.7578125,
+      "kd_loss": 0.030483372509479523,
+      "learning_rate": 9.984902671959911e-06,
+      "loss": 0.0461,
+      "step": 55,
+      "student_loss": 0.0011451594764366746,
+      "teacher_loss": 0.0006940242601558566
+    },
+    {
+      "epoch": 0.01730103806228374,
+      "grad_norm": 1.3359375,
+      "kd_loss": 0.030781563371419907,
+      "learning_rate": 9.97796195883804e-06,
+      "loss": 0.0412,
+      "step": 60,
+      "student_loss": 0.003650753991678357,
+      "teacher_loss": 0.002238439628854394
+    },
+    {
+      "epoch": 0.018742791234140715,
+      "grad_norm": 1.5234375,
+      "kd_loss": 0.030627569183707237,
+      "learning_rate": 9.969715861466839e-06,
+      "loss": 0.0425,
+      "step": 65,
+      "student_loss": 0.02980353683233261,
+      "teacher_loss": 0.03997597098350525
+    },
+    {
+      "epoch": 0.020184544405997693,
+      "grad_norm": 0.875,
+      "kd_loss": 0.0586991086602211,
+      "learning_rate": 9.96016654224243e-06,
+      "loss": 0.0466,
+      "step": 70,
+      "student_loss": 0.014155564829707146,
+      "teacher_loss": 0.0020057852379977703
+    },
+    {
+      "epoch": 0.02162629757785467,
+      "grad_norm": 0.796875,
+      "kd_loss": 0.021966831758618355,
+      "learning_rate": 9.94931650530827e-06,
+      "loss": 0.0409,
+      "step": 75,
+      "student_loss": 0.0014096935046836734,
+      "teacher_loss": 0.0008751551504246891
+    },
+    {
+      "epoch": 0.02306805074971165,
+      "grad_norm": 0.64453125,
+      "kd_loss": 0.022734930738806725,
+      "learning_rate": 9.93716859589851e-06,
+      "loss": 0.0362,
+      "step": 80,
+      "student_loss": 0.001171765266917646,
+      "teacher_loss": 0.0007404370117001235
+    },
+    {
+      "epoch": 0.024509803921568627,
+      "grad_norm": 0.765625,
+      "kd_loss": 0.022543881088495255,
+      "learning_rate": 9.923725999591846e-06,
+      "loss": 0.0382,
+      "step": 85,
+      "student_loss": 0.0004507019475568086,
+      "teacher_loss": 0.00039780911174602807
+    },
+    {
+      "epoch": 0.025951557093425604,
+      "grad_norm": 0.69140625,
+      "kd_loss": 0.07131864875555038,
+      "learning_rate": 9.908992241476189e-06,
+      "loss": 0.0387,
+      "step": 90,
+      "student_loss": 0.013878172263503075,
+      "teacher_loss": 0.0030211261473596096
+    },
+    {
+      "epoch": 0.027393310265282585,
+      "grad_norm": 1.1015625,
+      "kd_loss": 0.04375176876783371,
+      "learning_rate": 9.892971185224244e-06,
+      "loss": 0.0342,
+      "step": 95,
+      "student_loss": 0.0081056859344244,
+      "teacher_loss": 0.005189534276723862
+    },
+    {
+      "epoch": 0.02883506343713956,
+      "grad_norm": 0.859375,
+      "kd_loss": 0.021529421210289,
+      "learning_rate": 9.875667032080354e-06,
+      "loss": 0.0311,
+      "step": 100,
+      "student_loss": 0.0005472518969327211,
+      "teacher_loss": 0.000426671642344445
+    },
+    {
+      "epoch": 0.03027681660899654,
+      "grad_norm": 1.3671875,
+      "kd_loss": 0.03554106876254082,
+      "learning_rate": 9.857084319758772e-06,
+      "loss": 0.0342,
+      "step": 105,
+      "student_loss": 0.002422141609713435,
+      "teacher_loss": 0.000269417796516791
+    },
+    {
+      "epoch": 0.031718569780853516,
+      "grad_norm": 1.1953125,
+      "kd_loss": 0.0378132238984108,
+      "learning_rate": 9.837227921253747e-06,
+      "loss": 0.0362,
+      "step": 110,
+      "student_loss": 0.031434450298547745,
+      "teacher_loss": 0.01768127642571926
+    },
+    {
+      "epoch": 0.03316032295271049,
+      "grad_norm": 0.52734375,
+      "kd_loss": 0.03465187922120094,
+      "learning_rate": 9.816103043561648e-06,
+      "loss": 0.0337,
+      "step": 115,
+      "student_loss": 0.000800441310275346,
+      "teacher_loss": 0.0002722400240600109
+    },
+    {
+      "epoch": 0.03460207612456748,
+      "grad_norm": 0.59375,
+      "kd_loss": 0.02393350750207901,
+      "learning_rate": 9.79371522631553e-06,
+      "loss": 0.031,
+      "step": 120,
+      "student_loss": 0.0010254759108647704,
+      "teacher_loss": 0.0006771993357688189
+    },
+    {
+      "epoch": 0.036043829296424454,
+      "grad_norm": 0.734375,
+      "kd_loss": 0.05980119854211807,
+      "learning_rate": 9.770070340332457e-06,
+      "loss": 0.0344,
+      "step": 125,
+      "student_loss": 0.02782793715596199,
+      "teacher_loss": 0.0002460530085954815
+    },
+    {
+      "epoch": 0.03748558246828143,
+      "grad_norm": 0.71875,
+      "kd_loss": 0.02418632246553898,
+      "learning_rate": 9.745174586073982e-06,
+      "loss": 0.0333,
+      "step": 130,
+      "student_loss": 0.0004915493191219866,
+      "teacher_loss": 0.0004100716323591769
+    },
+    {
+      "epoch": 0.03892733564013841,
+      "grad_norm": 0.73046875,
+      "kd_loss": 0.04188523441553116,
+      "learning_rate": 9.719034492020183e-06,
+      "loss": 0.0371,
+      "step": 135,
+      "student_loss": 0.0016317864647135139,
+      "teacher_loss": 0.00047902524238452315
+    },
+    {
+      "epoch": 0.040369088811995385,
+      "grad_norm": 0.65625,
+      "kd_loss": 0.021806886419653893,
+      "learning_rate": 9.691656912957686e-06,
+      "loss": 0.0352,
+      "step": 140,
+      "student_loss": 0.00046953262062743306,
+      "teacher_loss": 0.0003224058309569955
+    },
+    {
+      "epoch": 0.04181084198385236,
+      "grad_norm": 0.443359375,
+      "kd_loss": 0.07382702082395554,
+      "learning_rate": 9.663049028182112e-06,
+      "loss": 0.0314,
+      "step": 145,
+      "student_loss": 0.03231842443346977,
+      "teacher_loss": 0.006043749395757914
+    },
+    {
+      "epoch": 0.04325259515570934,
+      "grad_norm": 0.5703125,
+      "kd_loss": 0.0166649017482996,
+      "learning_rate": 9.633218339615433e-06,
+      "loss": 0.0309,
+      "step": 150,
+      "student_loss": 0.001361644477583468,
+      "teacher_loss": 0.0013295100070536137
+    },
+    {
+      "epoch": 0.04469434832756632,
+      "grad_norm": 0.85546875,
+      "kd_loss": 0.026128817349672318,
+      "learning_rate": 9.602172669838721e-06,
+      "loss": 0.0373,
+      "step": 155,
+      "student_loss": 0.0024132225662469864,
+      "teacher_loss": 0.0007210202165879309
+    },
+    {
+      "epoch": 0.0461361014994233,
+      "grad_norm": 1.4765625,
+      "kd_loss": 0.030131949111819267,
+      "learning_rate": 9.569920160040815e-06,
+      "loss": 0.0339,
+      "step": 160,
+      "student_loss": 0.0748908743262291,
+      "teacher_loss": 0.04347090795636177
+    },
+    {
+      "epoch": 0.04757785467128028,
+      "grad_norm": 0.7109375,
+      "kd_loss": 0.03373745456337929,
+      "learning_rate": 9.536469267883432e-06,
+      "loss": 0.0312,
+      "step": 165,
+      "student_loss": 0.000419136747950688,
+      "teacher_loss": 0.003946289885789156
+    },
+    {
+      "epoch": 0.049019607843137254,
+      "grad_norm": 2.078125,
+      "kd_loss": 0.02021661214530468,
+      "learning_rate": 9.501828765283295e-06,
+      "loss": 0.0331,
+      "step": 170,
+      "student_loss": 0.000686564773786813,
+      "teacher_loss": 0.0005308608524501324
+    },
+    {
+      "epoch": 0.05046136101499423,
+      "grad_norm": 1.4375,
+      "kd_loss": 0.028470534831285477,
+      "learning_rate": 9.466007736111846e-06,
+      "loss": 0.0375,
+      "step": 175,
+      "student_loss": 0.0027788360603153706,
+      "teacher_loss": 0.0002813354367390275
+    },
+    {
+      "epoch": 0.05190311418685121,
+      "grad_norm": 1.2265625,
+      "kd_loss": 0.03539275750517845,
+      "learning_rate": 9.429015573813163e-06,
+      "loss": 0.0298,
+      "step": 180,
+      "student_loss": 0.0011434294283390045,
+      "teacher_loss": 0.00030375414644367993
+    },
+    {
+      "epoch": 0.05334486735870819,
+      "grad_norm": 1.203125,
+      "kd_loss": 0.03853345289826393,
+      "learning_rate": 9.390861978940687e-06,
+      "loss": 0.0354,
+      "step": 185,
+      "student_loss": 0.035494156181812286,
+      "teacher_loss": 0.0002623012405820191
+    },
+    {
+      "epoch": 0.05478662053056517,
+      "grad_norm": 1.046875,
+      "kd_loss": 0.05331602692604065,
+      "learning_rate": 9.351556956613423e-06,
+      "loss": 0.0309,
+      "step": 190,
+      "student_loss": 0.004768161568790674,
+      "teacher_loss": 0.0014110342599451542
+    },
+    {
+      "epoch": 0.056228373702422146,
+      "grad_norm": 1.421875,
+      "kd_loss": 0.053704116493463516,
+      "learning_rate": 9.31111081389227e-06,
+      "loss": 0.0301,
+      "step": 195,
+      "student_loss": 0.006213298067450523,
+      "teacher_loss": 0.0004057004116475582
+    },
+    {
+      "epoch": 0.05767012687427912,
+      "grad_norm": 0.796875,
+      "kd_loss": 0.034625094383955,
+      "learning_rate": 9.269534157077177e-06,
+      "loss": 0.039,
+      "step": 200,
+      "student_loss": 0.0004761361633427441,
+      "teacher_loss": 0.00022407782671507448
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-400/added_tokens.json

@@ -0,0 +1,35 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|action_sep|>": 151670,
+  "<|arg_sep|>": 151671,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|call_sep|>": 151666,
+  "<|end_of_text|>": 151673,
+  "<|endoftext|>": 151643,
+  "<|exception_sep|>": 151669,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|frame_sep|>": 151672,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|latent_end|>": 151675,
+  "<|latent_start|>": 151674,
+  "<|line_sep|>": 151667,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|return_sep|>": 151668,
+  "<|trace_context_start|>": 151665,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-400/chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\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>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-400/config.json

@@ -0,0 +1,66 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 36,
+  "model_type": "qwen2",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 36,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151676
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-400/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ce9b99fe951a25f2d18b02be8c7bb17716ff783b7666692a7437096e5e45a98
+size 6187858991

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-400/special_tokens_map.json

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

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-400/thought_projector.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4bcc3a480186b6817c4955332aa209acda79bc25d65605a3161816b5918a8cd
+size 16788033

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-400/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83a790d654474f5dfe225f889afd0210313eb1083f942671f2c4b8e95a1c922b
+size 11424004

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-400/tokenizer_config.json

@@ -0,0 +1,295 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<|trace_context_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151666": {
+      "content": "<|call_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151667": {
+      "content": "<|line_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151668": {
+      "content": "<|return_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151669": {
+      "content": "<|exception_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151670": {
+      "content": "<|action_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151671": {
+      "content": "<|arg_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151672": {
+      "content": "<|frame_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151673": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151674": {
+      "content": "<|latent_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151675": {
+      "content": "<|latent_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-400/trainer_state.json

@@ -0,0 +1,834 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.11534025374855825,
+  "eval_steps": 500,
+  "global_step": 400,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0014417531718569781,
+      "grad_norm": 47.75,
+      "kd_loss": 0.2916574478149414,
+      "learning_rate": 1.3333333333333334e-06,
+      "loss": 0.3836,
+      "step": 5,
+      "student_loss": 0.12182186543941498,
+      "teacher_loss": 0.002018272876739502
+    },
+    {
+      "epoch": 0.0028835063437139563,
+      "grad_norm": 22.5,
+      "kd_loss": 0.26723670959472656,
+      "learning_rate": 3e-06,
+      "loss": 0.2911,
+      "step": 10,
+      "student_loss": 0.10971185564994812,
+      "teacher_loss": 0.005489554721862078
+    },
+    {
+      "epoch": 0.004325259515570935,
+      "grad_norm": 6.9375,
+      "kd_loss": 0.21662698686122894,
+      "learning_rate": 4.666666666666667e-06,
+      "loss": 0.1789,
+      "step": 15,
+      "student_loss": 0.0025933366268873215,
+      "teacher_loss": 0.001338644651696086
+    },
+    {
+      "epoch": 0.0057670126874279125,
+      "grad_norm": 3.078125,
+      "kd_loss": 0.13032649457454681,
+      "learning_rate": 6.333333333333333e-06,
+      "loss": 0.1296,
+      "step": 20,
+      "student_loss": 0.023931587114930153,
+      "teacher_loss": 0.0006410324713215232
+    },
+    {
+      "epoch": 0.00720876585928489,
+      "grad_norm": 2.34375,
+      "kd_loss": 0.10749491304159164,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 0.107,
+      "step": 25,
+      "student_loss": 0.0032150023616850376,
+      "teacher_loss": 0.0010774527909234166
+    },
+    {
+      "epoch": 0.00865051903114187,
+      "grad_norm": 1.2890625,
+      "kd_loss": 0.0680936947464943,
+      "learning_rate": 9.666666666666667e-06,
+      "loss": 0.0841,
+      "step": 30,
+      "student_loss": 0.00099717837292701,
+      "teacher_loss": 0.0005822837701998651
+    },
+    {
+      "epoch": 0.010092272202998846,
+      "grad_norm": 2.109375,
+      "kd_loss": 0.06131501868367195,
+      "learning_rate": 9.99958042442916e-06,
+      "loss": 0.0685,
+      "step": 35,
+      "student_loss": 0.0027061135042458773,
+      "teacher_loss": 0.0006183119839988649
+    },
+    {
+      "epoch": 0.011534025374855825,
+      "grad_norm": 1.203125,
+      "kd_loss": 0.05697907507419586,
+      "learning_rate": 9.997876019358083e-06,
+      "loss": 0.0614,
+      "step": 40,
+      "student_loss": 0.029384225606918335,
+      "teacher_loss": 0.025684259831905365
+    },
+    {
+      "epoch": 0.012975778546712802,
+      "grad_norm": 1.296875,
+      "kd_loss": 0.09766291826963425,
+      "learning_rate": 9.99486100792044e-06,
+      "loss": 0.0476,
+      "step": 45,
+      "student_loss": 0.06478797644376755,
+      "teacher_loss": 0.0030884514562785625
+    },
+    {
+      "epoch": 0.01441753171856978,
+      "grad_norm": 1.453125,
+      "kd_loss": 0.03506260737776756,
+      "learning_rate": 9.990536180750724e-06,
+      "loss": 0.0517,
+      "step": 50,
+      "student_loss": 0.02092154324054718,
+      "teacher_loss": 0.01070804987102747
+    },
+    {
+      "epoch": 0.015859284890426758,
+      "grad_norm": 1.7578125,
+      "kd_loss": 0.030483372509479523,
+      "learning_rate": 9.984902671959911e-06,
+      "loss": 0.0461,
+      "step": 55,
+      "student_loss": 0.0011451594764366746,
+      "teacher_loss": 0.0006940242601558566
+    },
+    {
+      "epoch": 0.01730103806228374,
+      "grad_norm": 1.3359375,
+      "kd_loss": 0.030781563371419907,
+      "learning_rate": 9.97796195883804e-06,
+      "loss": 0.0412,
+      "step": 60,
+      "student_loss": 0.003650753991678357,
+      "teacher_loss": 0.002238439628854394
+    },
+    {
+      "epoch": 0.018742791234140715,
+      "grad_norm": 1.5234375,
+      "kd_loss": 0.030627569183707237,
+      "learning_rate": 9.969715861466839e-06,
+      "loss": 0.0425,
+      "step": 65,
+      "student_loss": 0.02980353683233261,
+      "teacher_loss": 0.03997597098350525
+    },
+    {
+      "epoch": 0.020184544405997693,
+      "grad_norm": 0.875,
+      "kd_loss": 0.0586991086602211,
+      "learning_rate": 9.96016654224243e-06,
+      "loss": 0.0466,
+      "step": 70,
+      "student_loss": 0.014155564829707146,
+      "teacher_loss": 0.0020057852379977703
+    },
+    {
+      "epoch": 0.02162629757785467,
+      "grad_norm": 0.796875,
+      "kd_loss": 0.021966831758618355,
+      "learning_rate": 9.94931650530827e-06,
+      "loss": 0.0409,
+      "step": 75,
+      "student_loss": 0.0014096935046836734,
+      "teacher_loss": 0.0008751551504246891
+    },
+    {
+      "epoch": 0.02306805074971165,
+      "grad_norm": 0.64453125,
+      "kd_loss": 0.022734930738806725,
+      "learning_rate": 9.93716859589851e-06,
+      "loss": 0.0362,
+      "step": 80,
+      "student_loss": 0.001171765266917646,
+      "teacher_loss": 0.0007404370117001235
+    },
+    {
+      "epoch": 0.024509803921568627,
+      "grad_norm": 0.765625,
+      "kd_loss": 0.022543881088495255,
+      "learning_rate": 9.923725999591846e-06,
+      "loss": 0.0382,
+      "step": 85,
+      "student_loss": 0.0004507019475568086,
+      "teacher_loss": 0.00039780911174602807
+    },
+    {
+      "epoch": 0.025951557093425604,
+      "grad_norm": 0.69140625,
+      "kd_loss": 0.07131864875555038,
+      "learning_rate": 9.908992241476189e-06,
+      "loss": 0.0387,
+      "step": 90,
+      "student_loss": 0.013878172263503075,
+      "teacher_loss": 0.0030211261473596096
+    },
+    {
+      "epoch": 0.027393310265282585,
+      "grad_norm": 1.1015625,
+      "kd_loss": 0.04375176876783371,
+      "learning_rate": 9.892971185224244e-06,
+      "loss": 0.0342,
+      "step": 95,
+      "student_loss": 0.0081056859344244,
+      "teacher_loss": 0.005189534276723862
+    },
+    {
+      "epoch": 0.02883506343713956,
+      "grad_norm": 0.859375,
+      "kd_loss": 0.021529421210289,
+      "learning_rate": 9.875667032080354e-06,
+      "loss": 0.0311,
+      "step": 100,
+      "student_loss": 0.0005472518969327211,
+      "teacher_loss": 0.000426671642344445
+    },
+    {
+      "epoch": 0.03027681660899654,
+      "grad_norm": 1.3671875,
+      "kd_loss": 0.03554106876254082,
+      "learning_rate": 9.857084319758772e-06,
+      "loss": 0.0342,
+      "step": 105,
+      "student_loss": 0.002422141609713435,
+      "teacher_loss": 0.000269417796516791
+    },
+    {
+      "epoch": 0.031718569780853516,
+      "grad_norm": 1.1953125,
+      "kd_loss": 0.0378132238984108,
+      "learning_rate": 9.837227921253747e-06,
+      "loss": 0.0362,
+      "step": 110,
+      "student_loss": 0.031434450298547745,
+      "teacher_loss": 0.01768127642571926
+    },
+    {
+      "epoch": 0.03316032295271049,
+      "grad_norm": 0.52734375,
+      "kd_loss": 0.03465187922120094,
+      "learning_rate": 9.816103043561648e-06,
+      "loss": 0.0337,
+      "step": 115,
+      "student_loss": 0.000800441310275346,
+      "teacher_loss": 0.0002722400240600109
+    },
+    {
+      "epoch": 0.03460207612456748,
+      "grad_norm": 0.59375,
+      "kd_loss": 0.02393350750207901,
+      "learning_rate": 9.79371522631553e-06,
+      "loss": 0.031,
+      "step": 120,
+      "student_loss": 0.0010254759108647704,
+      "teacher_loss": 0.0006771993357688189
+    },
+    {
+      "epoch": 0.036043829296424454,
+      "grad_norm": 0.734375,
+      "kd_loss": 0.05980119854211807,
+      "learning_rate": 9.770070340332457e-06,
+      "loss": 0.0344,
+      "step": 125,
+      "student_loss": 0.02782793715596199,
+      "teacher_loss": 0.0002460530085954815
+    },
+    {
+      "epoch": 0.03748558246828143,
+      "grad_norm": 0.71875,
+      "kd_loss": 0.02418632246553898,
+      "learning_rate": 9.745174586073982e-06,
+      "loss": 0.0333,
+      "step": 130,
+      "student_loss": 0.0004915493191219866,
+      "teacher_loss": 0.0004100716323591769
+    },
+    {
+      "epoch": 0.03892733564013841,
+      "grad_norm": 0.73046875,
+      "kd_loss": 0.04188523441553116,
+      "learning_rate": 9.719034492020183e-06,
+      "loss": 0.0371,
+      "step": 135,
+      "student_loss": 0.0016317864647135139,
+      "teacher_loss": 0.00047902524238452315
+    },
+    {
+      "epoch": 0.040369088811995385,
+      "grad_norm": 0.65625,
+      "kd_loss": 0.021806886419653893,
+      "learning_rate": 9.691656912957686e-06,
+      "loss": 0.0352,
+      "step": 140,
+      "student_loss": 0.00046953262062743306,
+      "teacher_loss": 0.0003224058309569955
+    },
+    {
+      "epoch": 0.04181084198385236,
+      "grad_norm": 0.443359375,
+      "kd_loss": 0.07382702082395554,
+      "learning_rate": 9.663049028182112e-06,
+      "loss": 0.0314,
+      "step": 145,
+      "student_loss": 0.03231842443346977,
+      "teacher_loss": 0.006043749395757914
+    },
+    {
+      "epoch": 0.04325259515570934,
+      "grad_norm": 0.5703125,
+      "kd_loss": 0.0166649017482996,
+      "learning_rate": 9.633218339615433e-06,
+      "loss": 0.0309,
+      "step": 150,
+      "student_loss": 0.001361644477583468,
+      "teacher_loss": 0.0013295100070536137
+    },
+    {
+      "epoch": 0.04469434832756632,
+      "grad_norm": 0.85546875,
+      "kd_loss": 0.026128817349672318,
+      "learning_rate": 9.602172669838721e-06,
+      "loss": 0.0373,
+      "step": 155,
+      "student_loss": 0.0024132225662469864,
+      "teacher_loss": 0.0007210202165879309
+    },
+    {
+      "epoch": 0.0461361014994233,
+      "grad_norm": 1.4765625,
+      "kd_loss": 0.030131949111819267,
+      "learning_rate": 9.569920160040815e-06,
+      "loss": 0.0339,
+      "step": 160,
+      "student_loss": 0.0748908743262291,
+      "teacher_loss": 0.04347090795636177
+    },
+    {
+      "epoch": 0.04757785467128028,
+      "grad_norm": 0.7109375,
+      "kd_loss": 0.03373745456337929,
+      "learning_rate": 9.536469267883432e-06,
+      "loss": 0.0312,
+      "step": 165,
+      "student_loss": 0.000419136747950688,
+      "teacher_loss": 0.003946289885789156
+    },
+    {
+      "epoch": 0.049019607843137254,
+      "grad_norm": 2.078125,
+      "kd_loss": 0.02021661214530468,
+      "learning_rate": 9.501828765283295e-06,
+      "loss": 0.0331,
+      "step": 170,
+      "student_loss": 0.000686564773786813,
+      "teacher_loss": 0.0005308608524501324
+    },
+    {
+      "epoch": 0.05046136101499423,
+      "grad_norm": 1.4375,
+      "kd_loss": 0.028470534831285477,
+      "learning_rate": 9.466007736111846e-06,
+      "loss": 0.0375,
+      "step": 175,
+      "student_loss": 0.0027788360603153706,
+      "teacher_loss": 0.0002813354367390275
+    },
+    {
+      "epoch": 0.05190311418685121,
+      "grad_norm": 1.2265625,
+      "kd_loss": 0.03539275750517845,
+      "learning_rate": 9.429015573813163e-06,
+      "loss": 0.0298,
+      "step": 180,
+      "student_loss": 0.0011434294283390045,
+      "teacher_loss": 0.00030375414644367993
+    },
+    {
+      "epoch": 0.05334486735870819,
+      "grad_norm": 1.203125,
+      "kd_loss": 0.03853345289826393,
+      "learning_rate": 9.390861978940687e-06,
+      "loss": 0.0354,
+      "step": 185,
+      "student_loss": 0.035494156181812286,
+      "teacher_loss": 0.0002623012405820191
+    },
+    {
+      "epoch": 0.05478662053056517,
+      "grad_norm": 1.046875,
+      "kd_loss": 0.05331602692604065,
+      "learning_rate": 9.351556956613423e-06,
+      "loss": 0.0309,
+      "step": 190,
+      "student_loss": 0.004768161568790674,
+      "teacher_loss": 0.0014110342599451542
+    },
+    {
+      "epoch": 0.056228373702422146,
+      "grad_norm": 1.421875,
+      "kd_loss": 0.053704116493463516,
+      "learning_rate": 9.31111081389227e-06,
+      "loss": 0.0301,
+      "step": 195,
+      "student_loss": 0.006213298067450523,
+      "teacher_loss": 0.0004057004116475582
+    },
+    {
+      "epoch": 0.05767012687427912,
+      "grad_norm": 0.796875,
+      "kd_loss": 0.034625094383955,
+      "learning_rate": 9.269534157077177e-06,
+      "loss": 0.039,
+      "step": 200,
+      "student_loss": 0.0004761361633427441,
+      "teacher_loss": 0.00022407782671507448
+    },
+    {
+      "epoch": 0.0591118800461361,
+      "grad_norm": 1.8515625,
+      "kd_loss": 0.02584153600037098,
+      "learning_rate": 9.226837888925813e-06,
+      "loss": 0.0341,
+      "step": 205,
+      "student_loss": 0.04030226171016693,
+      "teacher_loss": 0.027547728270292282
+    },
+    {
+      "epoch": 0.06055363321799308,
+      "grad_norm": 2.578125,
+      "kd_loss": 0.031695615500211716,
+      "learning_rate": 9.183033205794525e-06,
+      "loss": 0.0296,
+      "step": 210,
+      "student_loss": 0.0005234605050645769,
+      "teacher_loss": 0.00026510158204473555
+    },
+    {
+      "epoch": 0.061995386389850055,
+      "grad_norm": 1.1796875,
+      "kd_loss": 0.03749287873506546,
+      "learning_rate": 9.13813159470227e-06,
+      "loss": 0.0314,
+      "step": 215,
+      "student_loss": 0.0004607264418154955,
+      "teacher_loss": 0.00017995040980167687
+    },
+    {
+      "epoch": 0.06343713956170703,
+      "grad_norm": 2.671875,
+      "kd_loss": 0.04252801090478897,
+      "learning_rate": 9.092144830318357e-06,
+      "loss": 0.0309,
+      "step": 220,
+      "student_loss": 0.03449667617678642,
+      "teacher_loss": 0.010067273862659931
+    },
+    {
+      "epoch": 0.06487889273356401,
+      "grad_norm": 1.1640625,
+      "kd_loss": 0.022738255560398102,
+      "learning_rate": 9.045084971874738e-06,
+      "loss": 0.0371,
+      "step": 225,
+      "student_loss": 0.04431537538766861,
+      "teacher_loss": 0.0017423024401068687
+    },
+    {
+      "epoch": 0.06632064590542099,
+      "grad_norm": 1.1953125,
+      "kd_loss": 0.04256664216518402,
+      "learning_rate": 8.99696436000368e-06,
+      "loss": 0.029,
+      "step": 230,
+      "student_loss": 0.002170365070924163,
+      "teacher_loss": 0.003624099539592862
+    },
+    {
+      "epoch": 0.06776239907727798,
+      "grad_norm": 1.1328125,
+      "kd_loss": 0.07645199447870255,
+      "learning_rate": 8.947795613501658e-06,
+      "loss": 0.0312,
+      "step": 235,
+      "student_loss": 0.009659786708652973,
+      "teacher_loss": 0.00044414011063054204
+    },
+    {
+      "epoch": 0.06920415224913495,
+      "grad_norm": 0.77734375,
+      "kd_loss": 0.03161292523145676,
+      "learning_rate": 8.897591626020284e-06,
+      "loss": 0.0337,
+      "step": 240,
+      "student_loss": 0.009329003281891346,
+      "teacher_loss": 0.0004169123712927103
+    },
+    {
+      "epoch": 0.07064590542099193,
+      "grad_norm": 0.73828125,
+      "kd_loss": 0.016300037503242493,
+      "learning_rate": 8.846365562685178e-06,
+      "loss": 0.0274,
+      "step": 245,
+      "student_loss": 0.0028402707539498806,
+      "teacher_loss": 0.0025411578826606274
+    },
+    {
+      "epoch": 0.07208765859284891,
+      "grad_norm": 1.359375,
+      "kd_loss": 0.020206376910209656,
+      "learning_rate": 8.794130856643635e-06,
+      "loss": 0.0302,
+      "step": 250,
+      "student_loss": 0.0006153527647256851,
+      "teacher_loss": 0.0002888407907448709
+    },
+    {
+      "epoch": 0.07352941176470588,
+      "grad_norm": 0.71875,
+      "kd_loss": 0.01622343435883522,
+      "learning_rate": 8.74090120554202e-06,
+      "loss": 0.031,
+      "step": 255,
+      "student_loss": 0.0003402966831345111,
+      "teacher_loss": 0.00035465144901536405
+    },
+    {
+      "epoch": 0.07497116493656286,
+      "grad_norm": 0.84765625,
+      "kd_loss": 0.028415139764547348,
+      "learning_rate": 8.686690567933803e-06,
+      "loss": 0.033,
+      "step": 260,
+      "student_loss": 0.042195871472358704,
+      "teacher_loss": 0.033024489879608154
+    },
+    {
+      "epoch": 0.07641291810841984,
+      "grad_norm": 1.7890625,
+      "kd_loss": 0.018096741288900375,
+      "learning_rate": 8.63151315961915e-06,
+      "loss": 0.033,
+      "step": 265,
+      "student_loss": 0.0003965997020713985,
+      "teacher_loss": 0.00029857814661227167
+    },
+    {
+      "epoch": 0.07785467128027682,
+      "grad_norm": 0.73828125,
+      "kd_loss": 0.020055225118994713,
+      "learning_rate": 8.575383449917103e-06,
+      "loss": 0.0347,
+      "step": 270,
+      "student_loss": 0.0023262996692210436,
+      "teacher_loss": 0.001756582991220057
+    },
+    {
+      "epoch": 0.07929642445213379,
+      "grad_norm": 1.2109375,
+      "kd_loss": 0.025473114103078842,
+      "learning_rate": 8.518316157871232e-06,
+      "loss": 0.0297,
+      "step": 275,
+      "student_loss": 0.04790982976555824,
+      "teacher_loss": 0.03409208729863167
+    },
+    {
+      "epoch": 0.08073817762399077,
+      "grad_norm": 1.0234375,
+      "kd_loss": 0.035440195351839066,
+      "learning_rate": 8.460326248389825e-06,
+      "loss": 0.0299,
+      "step": 280,
+      "student_loss": 0.0004696878604590893,
+      "teacher_loss": 0.00016561525990255177
+    },
+    {
+      "epoch": 0.08217993079584775,
+      "grad_norm": 1.25,
+      "kd_loss": 0.021167855709791183,
+      "learning_rate": 8.401428928321607e-06,
+      "loss": 0.0331,
+      "step": 285,
+      "student_loss": 0.0006399436388164759,
+      "teacher_loss": 0.0005028079031035304
+    },
+    {
+      "epoch": 0.08362168396770472,
+      "grad_norm": 1.1015625,
+      "kd_loss": 0.017318392172455788,
+      "learning_rate": 8.341639642468002e-06,
+      "loss": 0.0355,
+      "step": 290,
+      "student_loss": 0.031288016587495804,
+      "teacher_loss": 0.028279613703489304
+    },
+    {
+      "epoch": 0.0850634371395617,
+      "grad_norm": 1.2109375,
+      "kd_loss": 0.019709059968590736,
+      "learning_rate": 8.280974069532999e-06,
+      "loss": 0.0329,
+      "step": 295,
+      "student_loss": 0.0007035491289570928,
+      "teacher_loss": 0.0005519128171727061
+    },
+    {
+      "epoch": 0.08650519031141868,
+      "grad_norm": 1.0625,
+      "kd_loss": 0.051209982484579086,
+      "learning_rate": 8.219448118011687e-06,
+      "loss": 0.0308,
+      "step": 300,
+      "student_loss": 0.017601756379008293,
+      "teacher_loss": 0.0016454572323709726
+    },
+    {
+      "epoch": 0.08794694348327567,
+      "grad_norm": 2.28125,
+      "kd_loss": 0.021270615980029106,
+      "learning_rate": 8.157077922018537e-06,
+      "loss": 0.0286,
+      "step": 305,
+      "student_loss": 0.005125237628817558,
+      "teacher_loss": 0.003151164622977376
+    },
+    {
+      "epoch": 0.08938869665513265,
+      "grad_norm": 0.859375,
+      "kd_loss": 0.02120344713330269,
+      "learning_rate": 8.093879837056486e-06,
+      "loss": 0.0302,
+      "step": 310,
+      "student_loss": 0.00047028501285240054,
+      "teacher_loss": 0.0005005286075174809
+    },
+    {
+      "epoch": 0.09083044982698962,
+      "grad_norm": 2.0,
+      "kd_loss": 0.037551358342170715,
+      "learning_rate": 8.029870435728018e-06,
+      "loss": 0.0301,
+      "step": 315,
+      "student_loss": 0.006268176715821028,
+      "teacher_loss": 0.00021872192155569792
+    },
+    {
+      "epoch": 0.0922722029988466,
+      "grad_norm": 1.8828125,
+      "kd_loss": 0.025020649656653404,
+      "learning_rate": 7.965066503389264e-06,
+      "loss": 0.0321,
+      "step": 320,
+      "student_loss": 0.0004365233180578798,
+      "teacher_loss": 0.00025109920534305274
+    },
+    {
+      "epoch": 0.09371395617070358,
+      "grad_norm": 2.390625,
+      "kd_loss": 0.06403576582670212,
+      "learning_rate": 7.89948503374835e-06,
+      "loss": 0.0284,
+      "step": 325,
+      "student_loss": 0.006501263007521629,
+      "teacher_loss": 0.02552512101829052
+    },
+    {
+      "epoch": 0.09515570934256055,
+      "grad_norm": 0.94921875,
+      "kd_loss": 0.048998646438121796,
+      "learning_rate": 7.833143224409076e-06,
+      "loss": 0.03,
+      "step": 330,
+      "student_loss": 0.011290914379060268,
+      "teacher_loss": 0.0012772196205332875
+    },
+    {
+      "epoch": 0.09659746251441753,
+      "grad_norm": 1.2734375,
+      "kd_loss": 0.013905064202845097,
+      "learning_rate": 7.766058472361154e-06,
+      "loss": 0.0266,
+      "step": 335,
+      "student_loss": 0.0008061202825047076,
+      "teacher_loss": 0.0009985697688534856
+    },
+    {
+      "epoch": 0.09803921568627451,
+      "grad_norm": 0.91015625,
+      "kd_loss": 0.025371180847287178,
+      "learning_rate": 7.698248369418146e-06,
+      "loss": 0.0312,
+      "step": 340,
+      "student_loss": 0.010380898602306843,
+      "teacher_loss": 0.007381496950984001
+    },
+    {
+      "epoch": 0.09948096885813149,
+      "grad_norm": 0.9921875,
+      "kd_loss": 0.03340749442577362,
+      "learning_rate": 7.629730697604314e-06,
+      "loss": 0.0326,
+      "step": 345,
+      "student_loss": 0.0104219326749444,
+      "teacher_loss": 0.00029996069497428834
+    },
+    {
+      "epoch": 0.10092272202998846,
+      "grad_norm": 0.8203125,
+      "kd_loss": 0.05953724682331085,
+      "learning_rate": 7.560523424491595e-06,
+      "loss": 0.0265,
+      "step": 350,
+      "student_loss": 0.000867739028763026,
+      "teacher_loss": 0.0006243465468287468
+    },
+    {
+      "epoch": 0.10236447520184544,
+      "grad_norm": 1.734375,
+      "kd_loss": 0.01555018313229084,
+      "learning_rate": 7.490644698487909e-06,
+      "loss": 0.0344,
+      "step": 355,
+      "student_loss": 0.0004700198769569397,
+      "teacher_loss": 0.0004090165311936289
+    },
+    {
+      "epoch": 0.10380622837370242,
+      "grad_norm": 0.88671875,
+      "kd_loss": 0.01545359380543232,
+      "learning_rate": 7.420112844078066e-06,
+      "loss": 0.031,
+      "step": 360,
+      "student_loss": 0.0005492828786373138,
+      "teacher_loss": 0.0005034908535890281
+    },
+    {
+      "epoch": 0.1052479815455594,
+      "grad_norm": 0.875,
+      "kd_loss": 0.029742496088147163,
+      "learning_rate": 7.348946357018479e-06,
+      "loss": 0.0326,
+      "step": 365,
+      "student_loss": 0.043358221650123596,
+      "teacher_loss": 0.029302725568413734
+    },
+    {
+      "epoch": 0.10668973471741638,
+      "grad_norm": 3.75,
+      "kd_loss": 0.05247749388217926,
+      "learning_rate": 7.277163899486975e-06,
+      "loss": 0.0357,
+      "step": 370,
+      "student_loss": 0.020690687000751495,
+      "teacher_loss": 0.000206919910851866
+    },
+    {
+      "epoch": 0.10813148788927336,
+      "grad_norm": 0.6015625,
+      "kd_loss": 0.030279265716671944,
+      "learning_rate": 7.204784295188959e-06,
+      "loss": 0.0291,
+      "step": 375,
+      "student_loss": 0.03159727901220322,
+      "teacher_loss": 0.02069539576768875
+    },
+    {
+      "epoch": 0.10957324106113034,
+      "grad_norm": 0.57421875,
+      "kd_loss": 0.01892879046499729,
+      "learning_rate": 7.1318265244212305e-06,
+      "loss": 0.0302,
+      "step": 380,
+      "student_loss": 0.0004462770011741668,
+      "teacher_loss": 0.000304198736557737
+    },
+    {
+      "epoch": 0.11101499423298732,
+      "grad_norm": 0.62890625,
+      "kd_loss": 0.03064708411693573,
+      "learning_rate": 7.05830971909472e-06,
+      "loss": 0.0284,
+      "step": 385,
+      "student_loss": 0.0003714795457199216,
+      "teacher_loss": 0.0002954486117232591
+    },
+    {
+      "epoch": 0.11245674740484429,
+      "grad_norm": 1.0078125,
+      "kd_loss": 0.01886909268796444,
+      "learning_rate": 6.9842531577174865e-06,
+      "loss": 0.0277,
+      "step": 390,
+      "student_loss": 0.00026233118842355907,
+      "teacher_loss": 0.0002473078202456236
+    },
+    {
+      "epoch": 0.11389850057670127,
+      "grad_norm": 0.7109375,
+      "kd_loss": 0.02147393859922886,
+      "learning_rate": 6.9096762603392595e-06,
+      "loss": 0.031,
+      "step": 395,
+      "student_loss": 0.0006213847082108259,
+      "teacher_loss": 0.000366208259947598
+    },
+    {
+      "epoch": 0.11534025374855825,
+      "grad_norm": 1.7265625,
+      "kd_loss": 0.032356131821870804,
+      "learning_rate": 6.834598583458862e-06,
+      "loss": 0.0277,
+      "step": 400,
+      "student_loss": 0.0007109013968147337,
+      "teacher_loss": 0.0001410621771356091
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-600/added_tokens.json

@@ -0,0 +1,35 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|action_sep|>": 151670,
+  "<|arg_sep|>": 151671,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|call_sep|>": 151666,
+  "<|end_of_text|>": 151673,
+  "<|endoftext|>": 151643,
+  "<|exception_sep|>": 151669,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|frame_sep|>": 151672,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|latent_end|>": 151675,
+  "<|latent_start|>": 151674,
+  "<|line_sep|>": 151667,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|return_sep|>": 151668,
+  "<|trace_context_start|>": 151665,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-600/chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\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>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-600/config.json

@@ -0,0 +1,66 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 36,
+  "model_type": "qwen2",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 36,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151676
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-600/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af5bad19db88c44f2a49f76ead45082c8bdbcc4a1e131f4cc1e2e67e1db620f7
+size 6187858991

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-600/special_tokens_map.json

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

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-600/thought_projector.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c41a9feeb78fba186674320bccdb26d3136951ecfcf69dcc0afc2483ebf55d5
+size 16788033

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-600/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83a790d654474f5dfe225f889afd0210313eb1083f942671f2c4b8e95a1c922b
+size 11424004

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-600/tokenizer_config.json

@@ -0,0 +1,295 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<|trace_context_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151666": {
+      "content": "<|call_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151667": {
+      "content": "<|line_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151668": {
+      "content": "<|return_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151669": {
+      "content": "<|exception_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151670": {
+      "content": "<|action_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151671": {
+      "content": "<|arg_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151672": {
+      "content": "<|frame_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151673": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151674": {
+      "content": "<|latent_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151675": {
+      "content": "<|latent_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-600/trainer_state.json

@@ -0,0 +1,1234 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.17301038062283736,
+  "eval_steps": 500,
+  "global_step": 600,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0014417531718569781,
+      "grad_norm": 47.75,
+      "kd_loss": 0.2916574478149414,
+      "learning_rate": 1.3333333333333334e-06,
+      "loss": 0.3836,
+      "step": 5,
+      "student_loss": 0.12182186543941498,
+      "teacher_loss": 0.002018272876739502
+    },
+    {
+      "epoch": 0.0028835063437139563,
+      "grad_norm": 22.5,
+      "kd_loss": 0.26723670959472656,
+      "learning_rate": 3e-06,
+      "loss": 0.2911,
+      "step": 10,
+      "student_loss": 0.10971185564994812,
+      "teacher_loss": 0.005489554721862078
+    },
+    {
+      "epoch": 0.004325259515570935,
+      "grad_norm": 6.9375,
+      "kd_loss": 0.21662698686122894,
+      "learning_rate": 4.666666666666667e-06,
+      "loss": 0.1789,
+      "step": 15,
+      "student_loss": 0.0025933366268873215,
+      "teacher_loss": 0.001338644651696086
+    },
+    {
+      "epoch": 0.0057670126874279125,
+      "grad_norm": 3.078125,
+      "kd_loss": 0.13032649457454681,
+      "learning_rate": 6.333333333333333e-06,
+      "loss": 0.1296,
+      "step": 20,
+      "student_loss": 0.023931587114930153,
+      "teacher_loss": 0.0006410324713215232
+    },
+    {
+      "epoch": 0.00720876585928489,
+      "grad_norm": 2.34375,
+      "kd_loss": 0.10749491304159164,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 0.107,
+      "step": 25,
+      "student_loss": 0.0032150023616850376,
+      "teacher_loss": 0.0010774527909234166
+    },
+    {
+      "epoch": 0.00865051903114187,
+      "grad_norm": 1.2890625,
+      "kd_loss": 0.0680936947464943,
+      "learning_rate": 9.666666666666667e-06,
+      "loss": 0.0841,
+      "step": 30,
+      "student_loss": 0.00099717837292701,
+      "teacher_loss": 0.0005822837701998651
+    },
+    {
+      "epoch": 0.010092272202998846,
+      "grad_norm": 2.109375,
+      "kd_loss": 0.06131501868367195,
+      "learning_rate": 9.99958042442916e-06,
+      "loss": 0.0685,
+      "step": 35,
+      "student_loss": 0.0027061135042458773,
+      "teacher_loss": 0.0006183119839988649
+    },
+    {
+      "epoch": 0.011534025374855825,
+      "grad_norm": 1.203125,
+      "kd_loss": 0.05697907507419586,
+      "learning_rate": 9.997876019358083e-06,
+      "loss": 0.0614,
+      "step": 40,
+      "student_loss": 0.029384225606918335,
+      "teacher_loss": 0.025684259831905365
+    },
+    {
+      "epoch": 0.012975778546712802,
+      "grad_norm": 1.296875,
+      "kd_loss": 0.09766291826963425,
+      "learning_rate": 9.99486100792044e-06,
+      "loss": 0.0476,
+      "step": 45,
+      "student_loss": 0.06478797644376755,
+      "teacher_loss": 0.0030884514562785625
+    },
+    {
+      "epoch": 0.01441753171856978,
+      "grad_norm": 1.453125,
+      "kd_loss": 0.03506260737776756,
+      "learning_rate": 9.990536180750724e-06,
+      "loss": 0.0517,
+      "step": 50,
+      "student_loss": 0.02092154324054718,
+      "teacher_loss": 0.01070804987102747
+    },
+    {
+      "epoch": 0.015859284890426758,
+      "grad_norm": 1.7578125,
+      "kd_loss": 0.030483372509479523,
+      "learning_rate": 9.984902671959911e-06,
+      "loss": 0.0461,
+      "step": 55,
+      "student_loss": 0.0011451594764366746,
+      "teacher_loss": 0.0006940242601558566
+    },
+    {
+      "epoch": 0.01730103806228374,
+      "grad_norm": 1.3359375,
+      "kd_loss": 0.030781563371419907,
+      "learning_rate": 9.97796195883804e-06,
+      "loss": 0.0412,
+      "step": 60,
+      "student_loss": 0.003650753991678357,
+      "teacher_loss": 0.002238439628854394
+    },
+    {
+      "epoch": 0.018742791234140715,
+      "grad_norm": 1.5234375,
+      "kd_loss": 0.030627569183707237,
+      "learning_rate": 9.969715861466839e-06,
+      "loss": 0.0425,
+      "step": 65,
+      "student_loss": 0.02980353683233261,
+      "teacher_loss": 0.03997597098350525
+    },
+    {
+      "epoch": 0.020184544405997693,
+      "grad_norm": 0.875,
+      "kd_loss": 0.0586991086602211,
+      "learning_rate": 9.96016654224243e-06,
+      "loss": 0.0466,
+      "step": 70,
+      "student_loss": 0.014155564829707146,
+      "teacher_loss": 0.0020057852379977703
+    },
+    {
+      "epoch": 0.02162629757785467,
+      "grad_norm": 0.796875,
+      "kd_loss": 0.021966831758618355,
+      "learning_rate": 9.94931650530827e-06,
+      "loss": 0.0409,
+      "step": 75,
+      "student_loss": 0.0014096935046836734,
+      "teacher_loss": 0.0008751551504246891
+    },
+    {
+      "epoch": 0.02306805074971165,
+      "grad_norm": 0.64453125,
+      "kd_loss": 0.022734930738806725,
+      "learning_rate": 9.93716859589851e-06,
+      "loss": 0.0362,
+      "step": 80,
+      "student_loss": 0.001171765266917646,
+      "teacher_loss": 0.0007404370117001235
+    },
+    {
+      "epoch": 0.024509803921568627,
+      "grad_norm": 0.765625,
+      "kd_loss": 0.022543881088495255,
+      "learning_rate": 9.923725999591846e-06,
+      "loss": 0.0382,
+      "step": 85,
+      "student_loss": 0.0004507019475568086,
+      "teacher_loss": 0.00039780911174602807
+    },
+    {
+      "epoch": 0.025951557093425604,
+      "grad_norm": 0.69140625,
+      "kd_loss": 0.07131864875555038,
+      "learning_rate": 9.908992241476189e-06,
+      "loss": 0.0387,
+      "step": 90,
+      "student_loss": 0.013878172263503075,
+      "teacher_loss": 0.0030211261473596096
+    },
+    {
+      "epoch": 0.027393310265282585,
+      "grad_norm": 1.1015625,
+      "kd_loss": 0.04375176876783371,
+      "learning_rate": 9.892971185224244e-06,
+      "loss": 0.0342,
+      "step": 95,
+      "student_loss": 0.0081056859344244,
+      "teacher_loss": 0.005189534276723862
+    },
+    {
+      "epoch": 0.02883506343713956,
+      "grad_norm": 0.859375,
+      "kd_loss": 0.021529421210289,
+      "learning_rate": 9.875667032080354e-06,
+      "loss": 0.0311,
+      "step": 100,
+      "student_loss": 0.0005472518969327211,
+      "teacher_loss": 0.000426671642344445
+    },
+    {
+      "epoch": 0.03027681660899654,
+      "grad_norm": 1.3671875,
+      "kd_loss": 0.03554106876254082,
+      "learning_rate": 9.857084319758772e-06,
+      "loss": 0.0342,
+      "step": 105,
+      "student_loss": 0.002422141609713435,
+      "teacher_loss": 0.000269417796516791
+    },
+    {
+      "epoch": 0.031718569780853516,
+      "grad_norm": 1.1953125,
+      "kd_loss": 0.0378132238984108,
+      "learning_rate": 9.837227921253747e-06,
+      "loss": 0.0362,
+      "step": 110,
+      "student_loss": 0.031434450298547745,
+      "teacher_loss": 0.01768127642571926
+    },
+    {
+      "epoch": 0.03316032295271049,
+      "grad_norm": 0.52734375,
+      "kd_loss": 0.03465187922120094,
+      "learning_rate": 9.816103043561648e-06,
+      "loss": 0.0337,
+      "step": 115,
+      "student_loss": 0.000800441310275346,
+      "teacher_loss": 0.0002722400240600109
+    },
+    {
+      "epoch": 0.03460207612456748,
+      "grad_norm": 0.59375,
+      "kd_loss": 0.02393350750207901,
+      "learning_rate": 9.79371522631553e-06,
+      "loss": 0.031,
+      "step": 120,
+      "student_loss": 0.0010254759108647704,
+      "teacher_loss": 0.0006771993357688189
+    },
+    {
+      "epoch": 0.036043829296424454,
+      "grad_norm": 0.734375,
+      "kd_loss": 0.05980119854211807,
+      "learning_rate": 9.770070340332457e-06,
+      "loss": 0.0344,
+      "step": 125,
+      "student_loss": 0.02782793715596199,
+      "teacher_loss": 0.0002460530085954815
+    },
+    {
+      "epoch": 0.03748558246828143,
+      "grad_norm": 0.71875,
+      "kd_loss": 0.02418632246553898,
+      "learning_rate": 9.745174586073982e-06,
+      "loss": 0.0333,
+      "step": 130,
+      "student_loss": 0.0004915493191219866,
+      "teacher_loss": 0.0004100716323591769
+    },
+    {
+      "epoch": 0.03892733564013841,
+      "grad_norm": 0.73046875,
+      "kd_loss": 0.04188523441553116,
+      "learning_rate": 9.719034492020183e-06,
+      "loss": 0.0371,
+      "step": 135,
+      "student_loss": 0.0016317864647135139,
+      "teacher_loss": 0.00047902524238452315
+    },
+    {
+      "epoch": 0.040369088811995385,
+      "grad_norm": 0.65625,
+      "kd_loss": 0.021806886419653893,
+      "learning_rate": 9.691656912957686e-06,
+      "loss": 0.0352,
+      "step": 140,
+      "student_loss": 0.00046953262062743306,
+      "teacher_loss": 0.0003224058309569955
+    },
+    {
+      "epoch": 0.04181084198385236,
+      "grad_norm": 0.443359375,
+      "kd_loss": 0.07382702082395554,
+      "learning_rate": 9.663049028182112e-06,
+      "loss": 0.0314,
+      "step": 145,
+      "student_loss": 0.03231842443346977,
+      "teacher_loss": 0.006043749395757914
+    },
+    {
+      "epoch": 0.04325259515570934,
+      "grad_norm": 0.5703125,
+      "kd_loss": 0.0166649017482996,
+      "learning_rate": 9.633218339615433e-06,
+      "loss": 0.0309,
+      "step": 150,
+      "student_loss": 0.001361644477583468,
+      "teacher_loss": 0.0013295100070536137
+    },
+    {
+      "epoch": 0.04469434832756632,
+      "grad_norm": 0.85546875,
+      "kd_loss": 0.026128817349672318,
+      "learning_rate": 9.602172669838721e-06,
+      "loss": 0.0373,
+      "step": 155,
+      "student_loss": 0.0024132225662469864,
+      "teacher_loss": 0.0007210202165879309
+    },
+    {
+      "epoch": 0.0461361014994233,
+      "grad_norm": 1.4765625,
+      "kd_loss": 0.030131949111819267,
+      "learning_rate": 9.569920160040815e-06,
+      "loss": 0.0339,
+      "step": 160,
+      "student_loss": 0.0748908743262291,
+      "teacher_loss": 0.04347090795636177
+    },
+    {
+      "epoch": 0.04757785467128028,
+      "grad_norm": 0.7109375,
+      "kd_loss": 0.03373745456337929,
+      "learning_rate": 9.536469267883432e-06,
+      "loss": 0.0312,
+      "step": 165,
+      "student_loss": 0.000419136747950688,
+      "teacher_loss": 0.003946289885789156
+    },
+    {
+      "epoch": 0.049019607843137254,
+      "grad_norm": 2.078125,
+      "kd_loss": 0.02021661214530468,
+      "learning_rate": 9.501828765283295e-06,
+      "loss": 0.0331,
+      "step": 170,
+      "student_loss": 0.000686564773786813,
+      "teacher_loss": 0.0005308608524501324
+    },
+    {
+      "epoch": 0.05046136101499423,
+      "grad_norm": 1.4375,
+      "kd_loss": 0.028470534831285477,
+      "learning_rate": 9.466007736111846e-06,
+      "loss": 0.0375,
+      "step": 175,
+      "student_loss": 0.0027788360603153706,
+      "teacher_loss": 0.0002813354367390275
+    },
+    {
+      "epoch": 0.05190311418685121,
+      "grad_norm": 1.2265625,
+      "kd_loss": 0.03539275750517845,
+      "learning_rate": 9.429015573813163e-06,
+      "loss": 0.0298,
+      "step": 180,
+      "student_loss": 0.0011434294283390045,
+      "teacher_loss": 0.00030375414644367993
+    },
+    {
+      "epoch": 0.05334486735870819,
+      "grad_norm": 1.203125,
+      "kd_loss": 0.03853345289826393,
+      "learning_rate": 9.390861978940687e-06,
+      "loss": 0.0354,
+      "step": 185,
+      "student_loss": 0.035494156181812286,
+      "teacher_loss": 0.0002623012405820191
+    },
+    {
+      "epoch": 0.05478662053056517,
+      "grad_norm": 1.046875,
+      "kd_loss": 0.05331602692604065,
+      "learning_rate": 9.351556956613423e-06,
+      "loss": 0.0309,
+      "step": 190,
+      "student_loss": 0.004768161568790674,
+      "teacher_loss": 0.0014110342599451542
+    },
+    {
+      "epoch": 0.056228373702422146,
+      "grad_norm": 1.421875,
+      "kd_loss": 0.053704116493463516,
+      "learning_rate": 9.31111081389227e-06,
+      "loss": 0.0301,
+      "step": 195,
+      "student_loss": 0.006213298067450523,
+      "teacher_loss": 0.0004057004116475582
+    },
+    {
+      "epoch": 0.05767012687427912,
+      "grad_norm": 0.796875,
+      "kd_loss": 0.034625094383955,
+      "learning_rate": 9.269534157077177e-06,
+      "loss": 0.039,
+      "step": 200,
+      "student_loss": 0.0004761361633427441,
+      "teacher_loss": 0.00022407782671507448
+    },
+    {
+      "epoch": 0.0591118800461361,
+      "grad_norm": 1.8515625,
+      "kd_loss": 0.02584153600037098,
+      "learning_rate": 9.226837888925813e-06,
+      "loss": 0.0341,
+      "step": 205,
+      "student_loss": 0.04030226171016693,
+      "teacher_loss": 0.027547728270292282
+    },
+    {
+      "epoch": 0.06055363321799308,
+      "grad_norm": 2.578125,
+      "kd_loss": 0.031695615500211716,
+      "learning_rate": 9.183033205794525e-06,
+      "loss": 0.0296,
+      "step": 210,
+      "student_loss": 0.0005234605050645769,
+      "teacher_loss": 0.00026510158204473555
+    },
+    {
+      "epoch": 0.061995386389850055,
+      "grad_norm": 1.1796875,
+      "kd_loss": 0.03749287873506546,
+      "learning_rate": 9.13813159470227e-06,
+      "loss": 0.0314,
+      "step": 215,
+      "student_loss": 0.0004607264418154955,
+      "teacher_loss": 0.00017995040980167687
+    },
+    {
+      "epoch": 0.06343713956170703,
+      "grad_norm": 2.671875,
+      "kd_loss": 0.04252801090478897,
+      "learning_rate": 9.092144830318357e-06,
+      "loss": 0.0309,
+      "step": 220,
+      "student_loss": 0.03449667617678642,
+      "teacher_loss": 0.010067273862659931
+    },
+    {
+      "epoch": 0.06487889273356401,
+      "grad_norm": 1.1640625,
+      "kd_loss": 0.022738255560398102,
+      "learning_rate": 9.045084971874738e-06,
+      "loss": 0.0371,
+      "step": 225,
+      "student_loss": 0.04431537538766861,
+      "teacher_loss": 0.0017423024401068687
+    },
+    {
+      "epoch": 0.06632064590542099,
+      "grad_norm": 1.1953125,
+      "kd_loss": 0.04256664216518402,
+      "learning_rate": 8.99696436000368e-06,
+      "loss": 0.029,
+      "step": 230,
+      "student_loss": 0.002170365070924163,
+      "teacher_loss": 0.003624099539592862
+    },
+    {
+      "epoch": 0.06776239907727798,
+      "grad_norm": 1.1328125,
+      "kd_loss": 0.07645199447870255,
+      "learning_rate": 8.947795613501658e-06,
+      "loss": 0.0312,
+      "step": 235,
+      "student_loss": 0.009659786708652973,
+      "teacher_loss": 0.00044414011063054204
+    },
+    {
+      "epoch": 0.06920415224913495,
+      "grad_norm": 0.77734375,
+      "kd_loss": 0.03161292523145676,
+      "learning_rate": 8.897591626020284e-06,
+      "loss": 0.0337,
+      "step": 240,
+      "student_loss": 0.009329003281891346,
+      "teacher_loss": 0.0004169123712927103
+    },
+    {
+      "epoch": 0.07064590542099193,
+      "grad_norm": 0.73828125,
+      "kd_loss": 0.016300037503242493,
+      "learning_rate": 8.846365562685178e-06,
+      "loss": 0.0274,
+      "step": 245,
+      "student_loss": 0.0028402707539498806,
+      "teacher_loss": 0.0025411578826606274
+    },
+    {
+      "epoch": 0.07208765859284891,
+      "grad_norm": 1.359375,
+      "kd_loss": 0.020206376910209656,
+      "learning_rate": 8.794130856643635e-06,
+      "loss": 0.0302,
+      "step": 250,
+      "student_loss": 0.0006153527647256851,
+      "teacher_loss": 0.0002888407907448709
+    },
+    {
+      "epoch": 0.07352941176470588,
+      "grad_norm": 0.71875,
+      "kd_loss": 0.01622343435883522,
+      "learning_rate": 8.74090120554202e-06,
+      "loss": 0.031,
+      "step": 255,
+      "student_loss": 0.0003402966831345111,
+      "teacher_loss": 0.00035465144901536405
+    },
+    {
+      "epoch": 0.07497116493656286,
+      "grad_norm": 0.84765625,
+      "kd_loss": 0.028415139764547348,
+      "learning_rate": 8.686690567933803e-06,
+      "loss": 0.033,
+      "step": 260,
+      "student_loss": 0.042195871472358704,
+      "teacher_loss": 0.033024489879608154
+    },
+    {
+      "epoch": 0.07641291810841984,
+      "grad_norm": 1.7890625,
+      "kd_loss": 0.018096741288900375,
+      "learning_rate": 8.63151315961915e-06,
+      "loss": 0.033,
+      "step": 265,
+      "student_loss": 0.0003965997020713985,
+      "teacher_loss": 0.00029857814661227167
+    },
+    {
+      "epoch": 0.07785467128027682,
+      "grad_norm": 0.73828125,
+      "kd_loss": 0.020055225118994713,
+      "learning_rate": 8.575383449917103e-06,
+      "loss": 0.0347,
+      "step": 270,
+      "student_loss": 0.0023262996692210436,
+      "teacher_loss": 0.001756582991220057
+    },
+    {
+      "epoch": 0.07929642445213379,
+      "grad_norm": 1.2109375,
+      "kd_loss": 0.025473114103078842,
+      "learning_rate": 8.518316157871232e-06,
+      "loss": 0.0297,
+      "step": 275,
+      "student_loss": 0.04790982976555824,
+      "teacher_loss": 0.03409208729863167
+    },
+    {
+      "epoch": 0.08073817762399077,
+      "grad_norm": 1.0234375,
+      "kd_loss": 0.035440195351839066,
+      "learning_rate": 8.460326248389825e-06,
+      "loss": 0.0299,
+      "step": 280,
+      "student_loss": 0.0004696878604590893,
+      "teacher_loss": 0.00016561525990255177
+    },
+    {
+      "epoch": 0.08217993079584775,
+      "grad_norm": 1.25,
+      "kd_loss": 0.021167855709791183,
+      "learning_rate": 8.401428928321607e-06,
+      "loss": 0.0331,
+      "step": 285,
+      "student_loss": 0.0006399436388164759,
+      "teacher_loss": 0.0005028079031035304
+    },
+    {
+      "epoch": 0.08362168396770472,
+      "grad_norm": 1.1015625,
+      "kd_loss": 0.017318392172455788,
+      "learning_rate": 8.341639642468002e-06,
+      "loss": 0.0355,
+      "step": 290,
+      "student_loss": 0.031288016587495804,
+      "teacher_loss": 0.028279613703489304
+    },
+    {
+      "epoch": 0.0850634371395617,
+      "grad_norm": 1.2109375,
+      "kd_loss": 0.019709059968590736,
+      "learning_rate": 8.280974069532999e-06,
+      "loss": 0.0329,
+      "step": 295,
+      "student_loss": 0.0007035491289570928,
+      "teacher_loss": 0.0005519128171727061
+    },
+    {
+      "epoch": 0.08650519031141868,
+      "grad_norm": 1.0625,
+      "kd_loss": 0.051209982484579086,
+      "learning_rate": 8.219448118011687e-06,
+      "loss": 0.0308,
+      "step": 300,
+      "student_loss": 0.017601756379008293,
+      "teacher_loss": 0.0016454572323709726
+    },
+    {
+      "epoch": 0.08794694348327567,
+      "grad_norm": 2.28125,
+      "kd_loss": 0.021270615980029106,
+      "learning_rate": 8.157077922018537e-06,
+      "loss": 0.0286,
+      "step": 305,
+      "student_loss": 0.005125237628817558,
+      "teacher_loss": 0.003151164622977376
+    },
+    {
+      "epoch": 0.08938869665513265,
+      "grad_norm": 0.859375,
+      "kd_loss": 0.02120344713330269,
+      "learning_rate": 8.093879837056486e-06,
+      "loss": 0.0302,
+      "step": 310,
+      "student_loss": 0.00047028501285240054,
+      "teacher_loss": 0.0005005286075174809
+    },
+    {
+      "epoch": 0.09083044982698962,
+      "grad_norm": 2.0,
+      "kd_loss": 0.037551358342170715,
+      "learning_rate": 8.029870435728018e-06,
+      "loss": 0.0301,
+      "step": 315,
+      "student_loss": 0.006268176715821028,
+      "teacher_loss": 0.00021872192155569792
+    },
+    {
+      "epoch": 0.0922722029988466,
+      "grad_norm": 1.8828125,
+      "kd_loss": 0.025020649656653404,
+      "learning_rate": 7.965066503389264e-06,
+      "loss": 0.0321,
+      "step": 320,
+      "student_loss": 0.0004365233180578798,
+      "teacher_loss": 0.00025109920534305274
+    },
+    {
+      "epoch": 0.09371395617070358,
+      "grad_norm": 2.390625,
+      "kd_loss": 0.06403576582670212,
+      "learning_rate": 7.89948503374835e-06,
+      "loss": 0.0284,
+      "step": 325,
+      "student_loss": 0.006501263007521629,
+      "teacher_loss": 0.02552512101829052
+    },
+    {
+      "epoch": 0.09515570934256055,
+      "grad_norm": 0.94921875,
+      "kd_loss": 0.048998646438121796,
+      "learning_rate": 7.833143224409076e-06,
+      "loss": 0.03,
+      "step": 330,
+      "student_loss": 0.011290914379060268,
+      "teacher_loss": 0.0012772196205332875
+    },
+    {
+      "epoch": 0.09659746251441753,
+      "grad_norm": 1.2734375,
+      "kd_loss": 0.013905064202845097,
+      "learning_rate": 7.766058472361154e-06,
+      "loss": 0.0266,
+      "step": 335,
+      "student_loss": 0.0008061202825047076,
+      "teacher_loss": 0.0009985697688534856
+    },
+    {
+      "epoch": 0.09803921568627451,
+      "grad_norm": 0.91015625,
+      "kd_loss": 0.025371180847287178,
+      "learning_rate": 7.698248369418146e-06,
+      "loss": 0.0312,
+      "step": 340,
+      "student_loss": 0.010380898602306843,
+      "teacher_loss": 0.007381496950984001
+    },
+    {
+      "epoch": 0.09948096885813149,
+      "grad_norm": 0.9921875,
+      "kd_loss": 0.03340749442577362,
+      "learning_rate": 7.629730697604314e-06,
+      "loss": 0.0326,
+      "step": 345,
+      "student_loss": 0.0104219326749444,
+      "teacher_loss": 0.00029996069497428834
+    },
+    {
+      "epoch": 0.10092272202998846,
+      "grad_norm": 0.8203125,
+      "kd_loss": 0.05953724682331085,
+      "learning_rate": 7.560523424491595e-06,
+      "loss": 0.0265,
+      "step": 350,
+      "student_loss": 0.000867739028763026,
+      "teacher_loss": 0.0006243465468287468
+    },
+    {
+      "epoch": 0.10236447520184544,
+      "grad_norm": 1.734375,
+      "kd_loss": 0.01555018313229084,
+      "learning_rate": 7.490644698487909e-06,
+      "loss": 0.0344,
+      "step": 355,
+      "student_loss": 0.0004700198769569397,
+      "teacher_loss": 0.0004090165311936289
+    },
+    {
+      "epoch": 0.10380622837370242,
+      "grad_norm": 0.88671875,
+      "kd_loss": 0.01545359380543232,
+      "learning_rate": 7.420112844078066e-06,
+      "loss": 0.031,
+      "step": 360,
+      "student_loss": 0.0005492828786373138,
+      "teacher_loss": 0.0005034908535890281
+    },
+    {
+      "epoch": 0.1052479815455594,
+      "grad_norm": 0.875,
+      "kd_loss": 0.029742496088147163,
+      "learning_rate": 7.348946357018479e-06,
+      "loss": 0.0326,
+      "step": 365,
+      "student_loss": 0.043358221650123596,
+      "teacher_loss": 0.029302725568413734
+    },
+    {
+      "epoch": 0.10668973471741638,
+      "grad_norm": 3.75,
+      "kd_loss": 0.05247749388217926,
+      "learning_rate": 7.277163899486975e-06,
+      "loss": 0.0357,
+      "step": 370,
+      "student_loss": 0.020690687000751495,
+      "teacher_loss": 0.000206919910851866
+    },
+    {
+      "epoch": 0.10813148788927336,
+      "grad_norm": 0.6015625,
+      "kd_loss": 0.030279265716671944,
+      "learning_rate": 7.204784295188959e-06,
+      "loss": 0.0291,
+      "step": 375,
+      "student_loss": 0.03159727901220322,
+      "teacher_loss": 0.02069539576768875
+    },
+    {
+      "epoch": 0.10957324106113034,
+      "grad_norm": 0.57421875,
+      "kd_loss": 0.01892879046499729,
+      "learning_rate": 7.1318265244212305e-06,
+      "loss": 0.0302,
+      "step": 380,
+      "student_loss": 0.0004462770011741668,
+      "teacher_loss": 0.000304198736557737
+    },
+    {
+      "epoch": 0.11101499423298732,
+      "grad_norm": 0.62890625,
+      "kd_loss": 0.03064708411693573,
+      "learning_rate": 7.05830971909472e-06,
+      "loss": 0.0284,
+      "step": 385,
+      "student_loss": 0.0003714795457199216,
+      "teacher_loss": 0.0002954486117232591
+    },
+    {
+      "epoch": 0.11245674740484429,
+      "grad_norm": 1.0078125,
+      "kd_loss": 0.01886909268796444,
+      "learning_rate": 6.9842531577174865e-06,
+      "loss": 0.0277,
+      "step": 390,
+      "student_loss": 0.00026233118842355907,
+      "teacher_loss": 0.0002473078202456236
+    },
+    {
+      "epoch": 0.11389850057670127,
+      "grad_norm": 0.7109375,
+      "kd_loss": 0.02147393859922886,
+      "learning_rate": 6.9096762603392595e-06,
+      "loss": 0.031,
+      "step": 395,
+      "student_loss": 0.0006213847082108259,
+      "teacher_loss": 0.000366208259947598
+    },
+    {
+      "epoch": 0.11534025374855825,
+      "grad_norm": 1.7265625,
+      "kd_loss": 0.032356131821870804,
+      "learning_rate": 6.834598583458862e-06,
+      "loss": 0.0277,
+      "step": 400,
+      "student_loss": 0.0007109013968147337,
+      "teacher_loss": 0.0001410621771356091
+    },
+    {
+      "epoch": 0.11678200692041522,
+      "grad_norm": 1.1328125,
+      "kd_loss": 0.03028486669063568,
+      "learning_rate": 6.7590398148958625e-06,
+      "loss": 0.0334,
+      "step": 405,
+      "student_loss": 0.011747114360332489,
+      "teacher_loss": 0.000281431624898687
+    },
+    {
+      "epoch": 0.1182237600922722,
+      "grad_norm": 1.8359375,
+      "kd_loss": 0.03624144569039345,
+      "learning_rate": 6.6830197686277945e-06,
+      "loss": 0.0362,
+      "step": 410,
+      "student_loss": 0.0136904027312994,
+      "teacher_loss": 0.008161901496350765
+    },
+    {
+      "epoch": 0.11966551326412918,
+      "grad_norm": 1.046875,
+      "kd_loss": 0.02301396243274212,
+      "learning_rate": 6.6065583795942625e-06,
+      "loss": 0.0345,
+      "step": 415,
+      "student_loss": 0.004049207549542189,
+      "teacher_loss": 0.003510379930958152
+    },
+    {
+      "epoch": 0.12110726643598616,
+      "grad_norm": 2.828125,
+      "kd_loss": 0.04604360833764076,
+      "learning_rate": 6.52967569846937e-06,
+      "loss": 0.0328,
+      "step": 420,
+      "student_loss": 0.04198378324508667,
+      "teacher_loss": 0.021758265793323517
+    },
+    {
+      "epoch": 0.12254901960784313,
+      "grad_norm": 1.1640625,
+      "kd_loss": 0.012796198017895222,
+      "learning_rate": 6.452391886403767e-06,
+      "loss": 0.0299,
+      "step": 425,
+      "student_loss": 0.0018603568896651268,
+      "teacher_loss": 0.00185443798545748
+    },
+    {
+      "epoch": 0.12399077277970011,
+      "grad_norm": 1.234375,
+      "kd_loss": 0.05164352431893349,
+      "learning_rate": 6.374727209737743e-06,
+      "loss": 0.0362,
+      "step": 430,
+      "student_loss": 0.05342794209718704,
+      "teacher_loss": 0.05994100123643875
+    },
+    {
+      "epoch": 0.1254325259515571,
+      "grad_norm": 0.90625,
+      "kd_loss": 0.015498572029173374,
+      "learning_rate": 6.296702034686726e-06,
+      "loss": 0.0299,
+      "step": 435,
+      "student_loss": 0.0007954059401527047,
+      "teacher_loss": 0.000588005583267659
+    },
+    {
+      "epoch": 0.12687427912341406,
+      "grad_norm": 1.84375,
+      "kd_loss": 0.03917201980948448,
+      "learning_rate": 6.218336822000598e-06,
+      "loss": 0.0315,
+      "step": 440,
+      "student_loss": 0.012786084786057472,
+      "teacher_loss": 0.007422462105751038
+    },
+    {
+      "epoch": 0.12831603229527105,
+      "grad_norm": 1.7578125,
+      "kd_loss": 0.03229602053761482,
+      "learning_rate": 6.139652121598219e-06,
+      "loss": 0.0312,
+      "step": 445,
+      "student_loss": 0.02884654700756073,
+      "teacher_loss": 0.018903804942965508
+    },
+    {
+      "epoch": 0.12975778546712802,
+      "grad_norm": 0.75390625,
+      "kd_loss": 0.035617679357528687,
+      "learning_rate": 6.060668567178561e-06,
+      "loss": 0.0324,
+      "step": 450,
+      "student_loss": 0.0018460770370438695,
+      "teacher_loss": 0.0019348791101947427
+    },
+    {
+      "epoch": 0.131199538638985,
+      "grad_norm": 0.9765625,
+      "kd_loss": 0.0225622970610857,
+      "learning_rate": 5.981406870809889e-06,
+      "loss": 0.0329,
+      "step": 455,
+      "student_loss": 0.0007274725358001888,
+      "teacher_loss": 0.0003756978258024901
+    },
+    {
+      "epoch": 0.13264129181084197,
+      "grad_norm": 1.1171875,
+      "kd_loss": 0.043778520077466965,
+      "learning_rate": 5.9018878174983674e-06,
+      "loss": 0.0285,
+      "step": 460,
+      "student_loss": 0.004358124919235706,
+      "teacher_loss": 0.0003043531905859709
+    },
+    {
+      "epoch": 0.13408304498269896,
+      "grad_norm": 1.078125,
+      "kd_loss": 0.027884893119335175,
+      "learning_rate": 5.822132259737565e-06,
+      "loss": 0.0329,
+      "step": 465,
+      "student_loss": 0.0003728284209500998,
+      "teacher_loss": 0.00016045373922679573
+    },
+    {
+      "epoch": 0.13552479815455595,
+      "grad_norm": 0.69921875,
+      "kd_loss": 0.02432670071721077,
+      "learning_rate": 5.742161112040237e-06,
+      "loss": 0.0301,
+      "step": 470,
+      "student_loss": 0.0002982575388159603,
+      "teacher_loss": 0.00018306604761164635
+    },
+    {
+      "epoch": 0.13696655132641292,
+      "grad_norm": 0.90625,
+      "kd_loss": 0.029901551082730293,
+      "learning_rate": 5.661995345453867e-06,
+      "loss": 0.0271,
+      "step": 475,
+      "student_loss": 0.0004417268792167306,
+      "teacher_loss": 0.0002262392226839438
+    },
+    {
+      "epoch": 0.1384083044982699,
+      "grad_norm": 0.890625,
+      "kd_loss": 0.028335485607385635,
+      "learning_rate": 5.581655982061367e-06,
+      "loss": 0.04,
+      "step": 480,
+      "student_loss": 0.03686260059475899,
+      "teacher_loss": 0.028967848047614098
+    },
+    {
+      "epoch": 0.13985005767012687,
+      "grad_norm": 1.03125,
+      "kd_loss": 0.014099025167524815,
+      "learning_rate": 5.501164089468406e-06,
+      "loss": 0.03,
+      "step": 485,
+      "student_loss": 0.00044251122744753957,
+      "teacher_loss": 0.00045048308675177395
+    },
+    {
+      "epoch": 0.14129181084198386,
+      "grad_norm": 1.0703125,
+      "kd_loss": 0.021259035915136337,
+      "learning_rate": 5.4205407752787884e-06,
+      "loss": 0.0348,
+      "step": 490,
+      "student_loss": 0.00038817411405034363,
+      "teacher_loss": 0.00027494796086102724
+    },
+    {
+      "epoch": 0.14273356401384082,
+      "grad_norm": 1.09375,
+      "kd_loss": 0.03478889912366867,
+      "learning_rate": 5.339807181559359e-06,
+      "loss": 0.0389,
+      "step": 495,
+      "student_loss": 0.0009461977751925588,
+      "teacher_loss": 0.00046065691276453435
+    },
+    {
+      "epoch": 0.14417531718569782,
+      "grad_norm": 0.91796875,
+      "kd_loss": 0.024866795167326927,
+      "learning_rate": 5.258984479295853e-06,
+      "loss": 0.0306,
+      "step": 500,
+      "student_loss": 0.00048162159509956837,
+      "teacher_loss": 0.0003469791554380208
+    },
+    {
+      "epoch": 0.14561707035755478,
+      "grad_norm": 1.046875,
+      "kd_loss": 0.022635649889707565,
+      "learning_rate": 5.1780938628411795e-06,
+      "loss": 0.0377,
+      "step": 505,
+      "student_loss": 0.0004350949893705547,
+      "teacher_loss": 0.0003242486563976854
+    },
+    {
+      "epoch": 0.14705882352941177,
+      "grad_norm": 1.3125,
+      "kd_loss": 0.02024003490805626,
+      "learning_rate": 5.097156544357567e-06,
+      "loss": 0.0313,
+      "step": 510,
+      "student_loss": 0.0002926869783550501,
+      "teacher_loss": 0.0002203087497036904
+    },
+    {
+      "epoch": 0.14850057670126873,
+      "grad_norm": 1.21875,
+      "kd_loss": 0.024929286912083626,
+      "learning_rate": 5.016193748254045e-06,
+      "loss": 0.0286,
+      "step": 515,
+      "student_loss": 0.010843674652278423,
+      "teacher_loss": 0.00017637628479860723
+    },
+    {
+      "epoch": 0.14994232987312572,
+      "grad_norm": 1.4140625,
+      "kd_loss": 0.06931846588850021,
+      "learning_rate": 4.935226705620699e-06,
+      "loss": 0.0338,
+      "step": 520,
+      "student_loss": 0.03976204991340637,
+      "teacher_loss": 0.011458049528300762
+    },
+    {
+      "epoch": 0.1513840830449827,
+      "grad_norm": 0.68359375,
+      "kd_loss": 0.03341987729072571,
+      "learning_rate": 4.8542766486612035e-06,
+      "loss": 0.0333,
+      "step": 525,
+      "student_loss": 0.00037662044633179903,
+      "teacher_loss": 0.00019793077080976218
+    },
+    {
+      "epoch": 0.15282583621683968,
+      "grad_norm": 1.09375,
+      "kd_loss": 0.025850094854831696,
+      "learning_rate": 4.773364805125025e-06,
+      "loss": 0.0312,
+      "step": 530,
+      "student_loss": 0.00043694477062672377,
+      "teacher_loss": 0.00031476584263145924
+    },
+    {
+      "epoch": 0.15426758938869667,
+      "grad_norm": 1.5234375,
+      "kd_loss": 0.03888598829507828,
+      "learning_rate": 4.6925123927408265e-06,
+      "loss": 0.0306,
+      "step": 535,
+      "student_loss": 0.000664425955619663,
+      "teacher_loss": 0.0004196573281660676
+    },
+    {
+      "epoch": 0.15570934256055363,
+      "grad_norm": 0.99609375,
+      "kd_loss": 0.05486282706260681,
+      "learning_rate": 4.611740613652485e-06,
+      "loss": 0.0287,
+      "step": 540,
+      "student_loss": 0.012853885069489479,
+      "teacher_loss": 0.00035151999327354133
+    },
+    {
+      "epoch": 0.15715109573241062,
+      "grad_norm": 1.0390625,
+      "kd_loss": 0.0184871144592762,
+      "learning_rate": 4.531070648859186e-06,
+      "loss": 0.0279,
+      "step": 545,
+      "student_loss": 0.0017356052994728088,
+      "teacher_loss": 0.001496323267929256
+    },
+    {
+      "epoch": 0.15859284890426759,
+      "grad_norm": 0.98828125,
+      "kd_loss": 0.047685761004686356,
+      "learning_rate": 4.450523652661086e-06,
+      "loss": 0.0272,
+      "step": 550,
+      "student_loss": 0.00019285024609416723,
+      "teacher_loss": 0.003556112525984645
+    },
+    {
+      "epoch": 0.16003460207612458,
+      "grad_norm": 1.1328125,
+      "kd_loss": 0.07883454859256744,
+      "learning_rate": 4.370120747111956e-06,
+      "loss": 0.0318,
+      "step": 555,
+      "student_loss": 0.037417806684970856,
+      "teacher_loss": 0.001769757247529924
+    },
+    {
+      "epoch": 0.16147635524798154,
+      "grad_norm": 1.65625,
+      "kd_loss": 0.04116034507751465,
+      "learning_rate": 4.289883016480291e-06,
+      "loss": 0.0342,
+      "step": 560,
+      "student_loss": 0.00150584289804101,
+      "teacher_loss": 0.00032299800659529865
+    },
+    {
+      "epoch": 0.16291810841983853,
+      "grad_norm": 0.8515625,
+      "kd_loss": 0.04062850773334503,
+      "learning_rate": 4.209831501720328e-06,
+      "loss": 0.0322,
+      "step": 565,
+      "student_loss": 0.026969417929649353,
+      "teacher_loss": 0.0193944051861763
+    },
+    {
+      "epoch": 0.1643598615916955,
+      "grad_norm": 1.46875,
+      "kd_loss": 0.032333601266145706,
+      "learning_rate": 4.129987194954421e-06,
+      "loss": 0.0356,
+      "step": 570,
+      "student_loss": 0.010522445663809776,
+      "teacher_loss": 0.007194437552243471
+    },
+    {
+      "epoch": 0.16580161476355249,
+      "grad_norm": 0.8671875,
+      "kd_loss": 0.02821211703121662,
+      "learning_rate": 4.050371033968216e-06,
+      "loss": 0.0326,
+      "step": 575,
+      "student_loss": 0.00044184463331475854,
+      "teacher_loss": 0.00024692591978237033
+    },
+    {
+      "epoch": 0.16724336793540945,
+      "grad_norm": 0.8046875,
+      "kd_loss": 0.02497352659702301,
+      "learning_rate": 3.9710038967200825e-06,
+      "loss": 0.0285,
+      "step": 580,
+      "student_loss": 0.002113678492605686,
+      "teacher_loss": 0.0031359801068902016
+    },
+    {
+      "epoch": 0.16868512110726644,
+      "grad_norm": 1.578125,
+      "kd_loss": 0.030017584562301636,
+      "learning_rate": 3.89190659586623e-06,
+      "loss": 0.0276,
+      "step": 585,
+      "student_loss": 0.0005331283318810165,
+      "teacher_loss": 0.00024908268824219704
+    },
+    {
+      "epoch": 0.1701268742791234,
+      "grad_norm": 1.421875,
+      "kd_loss": 0.025302177295088768,
+      "learning_rate": 3.8130998733029517e-06,
+      "loss": 0.0308,
+      "step": 590,
+      "student_loss": 0.0006629621493630111,
+      "teacher_loss": 0.0010983508545905352
+    },
+    {
+      "epoch": 0.1715686274509804,
+      "grad_norm": 1.4921875,
+      "kd_loss": 0.02107221819460392,
+      "learning_rate": 3.734604394727419e-06,
+      "loss": 0.0315,
+      "step": 595,
+      "student_loss": 0.00039723835652694106,
+      "teacher_loss": 0.0003521108883433044
+    },
+    {
+      "epoch": 0.17301038062283736,
+      "grad_norm": 1.2578125,
+      "kd_loss": 0.03979631885886192,
+      "learning_rate": 3.656440744218464e-06,
+      "loss": 0.0342,
+      "step": 600,
+      "student_loss": 0.04468826204538345,
+      "teacher_loss": 0.0050285086035728455
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-800/added_tokens.json

@@ -0,0 +1,35 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|action_sep|>": 151670,
+  "<|arg_sep|>": 151671,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|call_sep|>": 151666,
+  "<|end_of_text|>": 151673,
+  "<|endoftext|>": 151643,
+  "<|exception_sep|>": 151669,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|frame_sep|>": 151672,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|latent_end|>": 151675,
+  "<|latent_start|>": 151674,
+  "<|line_sep|>": 151667,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|return_sep|>": 151668,
+  "<|trace_context_start|>": 151665,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-800/chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\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>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-800/config.json

@@ -0,0 +1,66 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151643,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 36,
+  "model_type": "qwen2",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 36,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151676
+}

checkpoints/codi3b_a0.5_b1.0_g0.5_ls1/checkpoint-800/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:866734fa84f877fac014927a18ddcc1cd7654cd577a28392c4fa53ce5aebd35c
+size 6187858991