her_nemi_graph_sft_e3_ck174: relabel study: main mix data, batch 64, epoch 2 of 3; eval pass@8 11.1% (20.9/13.4/2.8), pass@1 1.99%

.gitattributes

@@ -62,3 +62,4 @@ graph_nemi_sft_e2/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 smoke_p2v_rl_final/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 her_nemi_graph_sft_e3/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 her_nemi_graph_343_sft_e3_ck300/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+her_nemi_graph_sft_e3_ck174/tokenizer.json filter=lfs diff=lfs merge=lfs -text

her_nemi_graph_sft_e3_ck174/chat_template.jinja

@@ -0,0 +1,120 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {%- if messages[0].content is string %}
+            {{- messages[0].content }}
+        {%- else %}
+            {%- for content in messages[0].content %}
+                {%- if 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' }}
+        {%- if messages[0].content is string %}
+            {{- messages[0].content }}
+        {%- else %}
+            {%- for content in messages[0].content %}
+                {%- if 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set image_count = namespace(value=0) %}
+{%- set video_count = namespace(value=0) %}
+{%- for message in messages %}
+    {%- if message.role == "user" %}
+        {{- '<|im_start|>' + message.role + '\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- else %}
+            {%- for content in message.content %}
+                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
+                    <|vision_start|><|image_pad|><|vision_end|>
+                {%- elif content.type == 'video' or 'video' in content %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
+                    <|vision_start|><|video_pad|><|vision_end|>
+                {%- elif 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role + '\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- else %}
+            {%- for content_item in message.content %}
+                {%- if 'text' in content_item %}
+                    {{- content_item.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and message.content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {%- if message.content is string %}
+            {{- message.content }}
+        {%- else %}
+            {%- for content in message.content %}
+                {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                    {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
+                    <|vision_start|><|image_pad|><|vision_end|>
+                {%- elif content.type == 'video' or 'video' in content %}
+                    {%- set video_count.value = video_count.value + 1 %}
+                    {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
+                    <|vision_start|><|video_pad|><|vision_end|>
+                {%- elif 'text' in content %}
+                    {{- content.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {{- '\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 %}

her_nemi_graph_sft_e3_ck174/config.json

@@ -0,0 +1,70 @@
+{
+  "architectures": [
+    "Qwen3VLForConditionalGeneration"
+  ],
+  "bos_token_id": null,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "hidden_size": 4096,
+  "image_token_id": 151655,
+  "model_type": "qwen3_vl",
+  "pad_token_id": 151643,
+  "text_config": {
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "bos_token_id": 151643,
+    "dtype": "bfloat16",
+    "eos_token_id": 151645,
+    "head_dim": 128,
+    "hidden_act": "silu",
+    "hidden_size": 4096,
+    "initializer_range": 0.02,
+    "intermediate_size": 12288,
+    "max_position_embeddings": 262144,
+    "model_type": "qwen3_vl_text",
+    "num_attention_heads": 32,
+    "num_hidden_layers": 36,
+    "num_key_value_heads": 8,
+    "pad_token_id": null,
+    "rms_norm_eps": 1e-06,
+    "rope_parameters": {
+      "mrope_interleaved": true,
+      "mrope_section": [
+        24,
+        20,
+        20
+      ],
+      "rope_theta": 5000000,
+      "rope_type": "default"
+    },
+    "use_cache": false,
+    "vocab_size": 151936
+  },
+  "tie_word_embeddings": false,
+  "transformers_version": "5.8.0",
+  "use_cache": false,
+  "video_token_id": 151656,
+  "vision_config": {
+    "deepstack_visual_indexes": [
+      8,
+      16,
+      24
+    ],
+    "depth": 27,
+    "dtype": "bfloat16",
+    "hidden_act": "gelu_pytorch_tanh",
+    "hidden_size": 1152,
+    "in_channels": 3,
+    "initializer_range": 0.02,
+    "intermediate_size": 4304,
+    "model_type": "qwen3_vl_vision",
+    "num_heads": 16,
+    "num_position_embeddings": 2304,
+    "out_hidden_size": 4096,
+    "patch_size": 16,
+    "spatial_merge_size": 2,
+    "temporal_patch_size": 2
+  },
+  "vision_end_token_id": 151653,
+  "vision_start_token_id": 151652
+}

her_nemi_graph_sft_e3_ck174/generation_config.json

@@ -0,0 +1,14 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.0,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "5.8.0"
+}

her_nemi_graph_sft_e3_ck174/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8210e7bb6bcd336c8bad76fa4f58403b91788e4ed80bd0308b2213da64b8d3d6
+size 17534340584

her_nemi_graph_sft_e3_ck174/processor_config.json

@@ -0,0 +1,60 @@
+{
+  "image_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_processor_type": "Qwen2VLImageProcessor",
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "merge_size": 2,
+    "patch_size": 16,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "size": {
+      "longest_edge": 16777216,
+      "shortest_edge": 65536
+    },
+    "temporal_patch_size": 2
+  },
+  "processor_class": "Qwen3VLProcessor",
+  "video_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "do_sample_frames": true,
+    "fps": 2,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "max_frames": 768,
+    "merge_size": 2,
+    "min_frames": 4,
+    "patch_size": 16,
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "return_metadata": false,
+    "size": {
+      "longest_edge": 25165824,
+      "shortest_edge": 4096
+    },
+    "temporal_patch_size": 2,
+    "video_processor_type": "Qwen3VLVideoProcessor"
+  }
+}

her_nemi_graph_sft_e3_ck174/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
+size 11422650

her_nemi_graph_sft_e3_ck174/tokenizer_config.json

@@ -0,0 +1,32 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_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|>"
+  ],
+  "is_local": true,
+  "local_files_only": false,
+  "model_max_length": 262144,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "processor_class": "Qwen3VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

her_nemi_graph_sft_e3_ck174/trainer_state.json

@@ -0,0 +1,169 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.07185628742515,
+  "eval_steps": 87,
+  "global_step": 174,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.11976047904191617,
+      "grad_norm": 2.946119673020926,
+      "learning_rate": 9.999614527738882e-06,
+      "loss": 0.9815900802612305,
+      "step": 10
+    },
+    {
+      "epoch": 0.23952095808383234,
+      "grad_norm": 1.773978560773706,
+      "learning_rate": 9.953429730181653e-06,
+      "loss": 0.6094533920288085,
+      "step": 20
+    },
+    {
+      "epoch": 0.3592814371257485,
+      "grad_norm": 1.3316447937757545,
+      "learning_rate": 9.830965649597455e-06,
+      "loss": 0.5093350887298584,
+      "step": 30
+    },
+    {
+      "epoch": 0.47904191616766467,
+      "grad_norm": 1.2747866305703939,
+      "learning_rate": 9.634108145435665e-06,
+      "loss": 0.5005360126495362,
+      "step": 40
+    },
+    {
+      "epoch": 0.5988023952095808,
+      "grad_norm": 1.290897501834079,
+      "learning_rate": 9.365888682780862e-06,
+      "loss": 0.47110352516174314,
+      "step": 50
+    },
+    {
+      "epoch": 0.718562874251497,
+      "grad_norm": 1.2166087417470406,
+      "learning_rate": 9.03043764995379e-06,
+      "loss": 0.46074743270874025,
+      "step": 60
+    },
+    {
+      "epoch": 0.8383233532934131,
+      "grad_norm": 1.3015834547406744,
+      "learning_rate": 8.63292075347872e-06,
+      "loss": 0.48959875106811523,
+      "step": 70
+    },
+    {
+      "epoch": 0.9580838323353293,
+      "grad_norm": 1.1432730868050294,
+      "learning_rate": 8.179459469889269e-06,
+      "loss": 0.4706310272216797,
+      "step": 80
+    },
+    {
+      "epoch": 1.035928143712575,
+      "eval_loss": 0.4270435869693756,
+      "eval_runtime": 16.2772,
+      "eval_samples_per_second": 17.325,
+      "eval_steps_per_second": 2.212,
+      "step": 87
+    },
+    {
+      "epoch": 1.0718562874251496,
+      "grad_norm": 1.072123804546216,
+      "learning_rate": 7.67703677935813e-06,
+      "loss": 0.3950784683227539,
+      "step": 90
+    },
+    {
+      "epoch": 1.1916167664670658,
+      "grad_norm": 1.4140923055253354,
+      "learning_rate": 7.133389632785543e-06,
+      "loss": 0.3176417827606201,
+      "step": 100
+    },
+    {
+      "epoch": 1.311377245508982,
+      "grad_norm": 1.2686477605595623,
+      "learning_rate": 6.5568898082765945e-06,
+      "loss": 0.3139654636383057,
+      "step": 110
+    },
+    {
+      "epoch": 1.4311377245508983,
+      "grad_norm": 1.2258631369271755,
+      "learning_rate": 5.9564149917325845e-06,
+      "loss": 0.3042759895324707,
+      "step": 120
+    },
+    {
+      "epoch": 1.5508982035928143,
+      "grad_norm": 1.3370618334549527,
+      "learning_rate": 5.341212066823356e-06,
+      "loss": 0.2858062028884888,
+      "step": 130
+    },
+    {
+      "epoch": 1.6706586826347305,
+      "grad_norm": 1.2655392766923055,
+      "learning_rate": 4.720754719577448e-06,
+      "loss": 0.30150399208068845,
+      "step": 140
+    },
+    {
+      "epoch": 1.7904191616766467,
+      "grad_norm": 1.2555669121333346,
+      "learning_rate": 4.104597550377776e-06,
+      "loss": 0.28112201690673827,
+      "step": 150
+    },
+    {
+      "epoch": 1.910179640718563,
+      "grad_norm": 1.3403092164219743,
+      "learning_rate": 3.5022289399339933e-06,
+      "loss": 0.29944169521331787,
+      "step": 160
+    },
+    {
+      "epoch": 2.0239520958083834,
+      "grad_norm": 1.042680355861941,
+      "learning_rate": 2.9229249349905686e-06,
+      "loss": 0.2506244421005249,
+      "step": 170
+    },
+    {
+      "epoch": 2.07185628742515,
+      "eval_loss": 0.4412752091884613,
+      "eval_runtime": 15.6964,
+      "eval_samples_per_second": 17.966,
+      "eval_steps_per_second": 2.294,
+      "step": 174
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 261,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 87,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 73649416699904.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}