mjf-su commited on
Commit
defe877
·
verified ·
1 Parent(s): 2e71eb8

Upload folder using huggingface_hub

Browse files
.gitattributes CHANGED
@@ -33,3 +33,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ checkpoint-1000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
37
+ checkpoint-2000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
38
+ checkpoint-2901/tokenizer.json filter=lfs diff=lfs merge=lfs -text
39
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
chat_template.jinja ADDED
@@ -0,0 +1,120 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0].role == 'system' %}
4
+ {%- if messages[0].content is string %}
5
+ {{- messages[0].content }}
6
+ {%- else %}
7
+ {%- for content in messages[0].content %}
8
+ {%- if 'text' in content %}
9
+ {{- content.text }}
10
+ {%- endif %}
11
+ {%- endfor %}
12
+ {%- endif %}
13
+ {{- '\n\n' }}
14
+ {%- endif %}
15
+ {{- "# 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>" }}
16
+ {%- for tool in tools %}
17
+ {{- "\n" }}
18
+ {{- tool | tojson }}
19
+ {%- endfor %}
20
+ {{- "\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" }}
21
+ {%- else %}
22
+ {%- if messages[0].role == 'system' %}
23
+ {{- '<|im_start|>system\n' }}
24
+ {%- if messages[0].content is string %}
25
+ {{- messages[0].content }}
26
+ {%- else %}
27
+ {%- for content in messages[0].content %}
28
+ {%- if 'text' in content %}
29
+ {{- content.text }}
30
+ {%- endif %}
31
+ {%- endfor %}
32
+ {%- endif %}
33
+ {{- '<|im_end|>\n' }}
34
+ {%- endif %}
35
+ {%- endif %}
36
+ {%- set image_count = namespace(value=0) %}
37
+ {%- set video_count = namespace(value=0) %}
38
+ {%- for message in messages %}
39
+ {%- if message.role == "user" %}
40
+ {{- '<|im_start|>' + message.role + '\n' }}
41
+ {%- if message.content is string %}
42
+ {{- message.content }}
43
+ {%- else %}
44
+ {%- for content in message.content %}
45
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
46
+ {%- set image_count.value = image_count.value + 1 %}
47
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
48
+ <|vision_start|><|image_pad|><|vision_end|>
49
+ {%- elif content.type == 'video' or 'video' in content %}
50
+ {%- set video_count.value = video_count.value + 1 %}
51
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
52
+ <|vision_start|><|video_pad|><|vision_end|>
53
+ {%- elif 'text' in content %}
54
+ {{- content.text }}
55
+ {%- endif %}
56
+ {%- endfor %}
57
+ {%- endif %}
58
+ {{- '<|im_end|>\n' }}
59
+ {%- elif message.role == "assistant" %}
60
+ {{- '<|im_start|>' + message.role + '\n' }}
61
+ {%- if message.content is string %}
62
+ {{- message.content }}
63
+ {%- else %}
64
+ {%- for content_item in message.content %}
65
+ {%- if 'text' in content_item %}
66
+ {{- content_item.text }}
67
+ {%- endif %}
68
+ {%- endfor %}
69
+ {%- endif %}
70
+ {%- if message.tool_calls %}
71
+ {%- for tool_call in message.tool_calls %}
72
+ {%- if (loop.first and message.content) or (not loop.first) %}
73
+ {{- '\n' }}
74
+ {%- endif %}
75
+ {%- if tool_call.function %}
76
+ {%- set tool_call = tool_call.function %}
77
+ {%- endif %}
78
+ {{- '<tool_call>\n{"name": "' }}
79
+ {{- tool_call.name }}
80
+ {{- '", "arguments": ' }}
81
+ {%- if tool_call.arguments is string %}
82
+ {{- tool_call.arguments }}
83
+ {%- else %}
84
+ {{- tool_call.arguments | tojson }}
85
+ {%- endif %}
86
+ {{- '}\n</tool_call>' }}
87
+ {%- endfor %}
88
+ {%- endif %}
89
+ {{- '<|im_end|>\n' }}
90
+ {%- elif message.role == "tool" %}
91
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
92
+ {{- '<|im_start|>user' }}
93
+ {%- endif %}
94
+ {{- '\n<tool_response>\n' }}
95
+ {%- if message.content is string %}
96
+ {{- message.content }}
97
+ {%- else %}
98
+ {%- for content in message.content %}
99
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
100
+ {%- set image_count.value = image_count.value + 1 %}
101
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
102
+ <|vision_start|><|image_pad|><|vision_end|>
103
+ {%- elif content.type == 'video' or 'video' in content %}
104
+ {%- set video_count.value = video_count.value + 1 %}
105
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
106
+ <|vision_start|><|video_pad|><|vision_end|>
107
+ {%- elif 'text' in content %}
108
+ {{- content.text }}
109
+ {%- endif %}
110
+ {%- endfor %}
111
+ {%- endif %}
112
+ {{- '\n</tool_response>' }}
113
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
114
+ {{- '<|im_end|>\n' }}
115
+ {%- endif %}
116
+ {%- endif %}
117
+ {%- endfor %}
118
+ {%- if add_generation_prompt %}
119
+ {{- '<|im_start|>assistant\n' }}
120
+ {%- endif %}
checkpoint-1000/chat_template.jinja ADDED
@@ -0,0 +1,120 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0].role == 'system' %}
4
+ {%- if messages[0].content is string %}
5
+ {{- messages[0].content }}
6
+ {%- else %}
7
+ {%- for content in messages[0].content %}
8
+ {%- if 'text' in content %}
9
+ {{- content.text }}
10
+ {%- endif %}
11
+ {%- endfor %}
12
+ {%- endif %}
13
+ {{- '\n\n' }}
14
+ {%- endif %}
15
+ {{- "# 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>" }}
16
+ {%- for tool in tools %}
17
+ {{- "\n" }}
18
+ {{- tool | tojson }}
19
+ {%- endfor %}
20
+ {{- "\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" }}
21
+ {%- else %}
22
+ {%- if messages[0].role == 'system' %}
23
+ {{- '<|im_start|>system\n' }}
24
+ {%- if messages[0].content is string %}
25
+ {{- messages[0].content }}
26
+ {%- else %}
27
+ {%- for content in messages[0].content %}
28
+ {%- if 'text' in content %}
29
+ {{- content.text }}
30
+ {%- endif %}
31
+ {%- endfor %}
32
+ {%- endif %}
33
+ {{- '<|im_end|>\n' }}
34
+ {%- endif %}
35
+ {%- endif %}
36
+ {%- set image_count = namespace(value=0) %}
37
+ {%- set video_count = namespace(value=0) %}
38
+ {%- for message in messages %}
39
+ {%- if message.role == "user" %}
40
+ {{- '<|im_start|>' + message.role + '\n' }}
41
+ {%- if message.content is string %}
42
+ {{- message.content }}
43
+ {%- else %}
44
+ {%- for content in message.content %}
45
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
46
+ {%- set image_count.value = image_count.value + 1 %}
47
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
48
+ <|vision_start|><|image_pad|><|vision_end|>
49
+ {%- elif content.type == 'video' or 'video' in content %}
50
+ {%- set video_count.value = video_count.value + 1 %}
51
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
52
+ <|vision_start|><|video_pad|><|vision_end|>
53
+ {%- elif 'text' in content %}
54
+ {{- content.text }}
55
+ {%- endif %}
56
+ {%- endfor %}
57
+ {%- endif %}
58
+ {{- '<|im_end|>\n' }}
59
+ {%- elif message.role == "assistant" %}
60
+ {{- '<|im_start|>' + message.role + '\n' }}
61
+ {%- if message.content is string %}
62
+ {{- message.content }}
63
+ {%- else %}
64
+ {%- for content_item in message.content %}
65
+ {%- if 'text' in content_item %}
66
+ {{- content_item.text }}
67
+ {%- endif %}
68
+ {%- endfor %}
69
+ {%- endif %}
70
+ {%- if message.tool_calls %}
71
+ {%- for tool_call in message.tool_calls %}
72
+ {%- if (loop.first and message.content) or (not loop.first) %}
73
+ {{- '\n' }}
74
+ {%- endif %}
75
+ {%- if tool_call.function %}
76
+ {%- set tool_call = tool_call.function %}
77
+ {%- endif %}
78
+ {{- '<tool_call>\n{"name": "' }}
79
+ {{- tool_call.name }}
80
+ {{- '", "arguments": ' }}
81
+ {%- if tool_call.arguments is string %}
82
+ {{- tool_call.arguments }}
83
+ {%- else %}
84
+ {{- tool_call.arguments | tojson }}
85
+ {%- endif %}
86
+ {{- '}\n</tool_call>' }}
87
+ {%- endfor %}
88
+ {%- endif %}
89
+ {{- '<|im_end|>\n' }}
90
+ {%- elif message.role == "tool" %}
91
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
92
+ {{- '<|im_start|>user' }}
93
+ {%- endif %}
94
+ {{- '\n<tool_response>\n' }}
95
+ {%- if message.content is string %}
96
+ {{- message.content }}
97
+ {%- else %}
98
+ {%- for content in message.content %}
99
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
100
+ {%- set image_count.value = image_count.value + 1 %}
101
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
102
+ <|vision_start|><|image_pad|><|vision_end|>
103
+ {%- elif content.type == 'video' or 'video' in content %}
104
+ {%- set video_count.value = video_count.value + 1 %}
105
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
106
+ <|vision_start|><|video_pad|><|vision_end|>
107
+ {%- elif 'text' in content %}
108
+ {{- content.text }}
109
+ {%- endif %}
110
+ {%- endfor %}
111
+ {%- endif %}
112
+ {{- '\n</tool_response>' }}
113
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
114
+ {{- '<|im_end|>\n' }}
115
+ {%- endif %}
116
+ {%- endif %}
117
+ {%- endfor %}
118
+ {%- if add_generation_prompt %}
119
+ {{- '<|im_start|>assistant\n' }}
120
+ {%- endif %}
checkpoint-1000/config.json ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3VLForConditionalGeneration"
4
+ ],
5
+ "bos_token_id": null,
6
+ "dtype": "bfloat16",
7
+ "eos_token_id": 151645,
8
+ "image_token_id": 151655,
9
+ "model_type": "qwen3_vl",
10
+ "pad_token_id": 151643,
11
+ "text_config": {
12
+ "attention_bias": false,
13
+ "attention_dropout": 0.0,
14
+ "bos_token_id": 151643,
15
+ "dtype": "bfloat16",
16
+ "eos_token_id": 151645,
17
+ "head_dim": 128,
18
+ "hidden_act": "silu",
19
+ "hidden_size": 2560,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 9728,
22
+ "max_position_embeddings": 262144,
23
+ "model_type": "qwen3_vl_text",
24
+ "num_attention_heads": 32,
25
+ "num_hidden_layers": 36,
26
+ "num_key_value_heads": 8,
27
+ "pad_token_id": null,
28
+ "rms_norm_eps": 1e-06,
29
+ "rope_parameters": {
30
+ "mrope_interleaved": true,
31
+ "mrope_section": [
32
+ 24,
33
+ 20,
34
+ 20
35
+ ],
36
+ "rope_theta": 5000000,
37
+ "rope_type": "default"
38
+ },
39
+ "tie_word_embeddings": true,
40
+ "use_cache": true,
41
+ "vocab_size": 151936
42
+ },
43
+ "tie_word_embeddings": true,
44
+ "transformers_version": "5.0.0",
45
+ "use_cache": false,
46
+ "video_token_id": 151656,
47
+ "vision_config": {
48
+ "deepstack_visual_indexes": [
49
+ 5,
50
+ 11,
51
+ 17
52
+ ],
53
+ "depth": 24,
54
+ "dtype": "bfloat16",
55
+ "hidden_act": "gelu_pytorch_tanh",
56
+ "hidden_size": 1024,
57
+ "in_channels": 3,
58
+ "initializer_range": 0.02,
59
+ "intermediate_size": 4096,
60
+ "model_type": "qwen3_vl",
61
+ "num_heads": 16,
62
+ "num_position_embeddings": 2304,
63
+ "out_hidden_size": 2560,
64
+ "patch_size": 16,
65
+ "spatial_merge_size": 2,
66
+ "temporal_patch_size": 2
67
+ },
68
+ "vision_end_token_id": 151653,
69
+ "vision_start_token_id": 151652
70
+ }
checkpoint-1000/generation_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_sample": true,
3
+ "eos_token_id": [
4
+ 151645,
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.0,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "5.0.0"
14
+ }
checkpoint-1000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:acded431caf1974316b56adaf6e52007ce27a1207c5fc92818587938f3759bcc
3
+ size 8875719408
checkpoint-1000/processor_config.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "data_format": "channels_first",
4
+ "do_convert_rgb": true,
5
+ "do_normalize": true,
6
+ "do_rescale": true,
7
+ "do_resize": true,
8
+ "image_mean": [
9
+ 0.5,
10
+ 0.5,
11
+ 0.5
12
+ ],
13
+ "image_processor_type": "Qwen2VLImageProcessorFast",
14
+ "image_std": [
15
+ 0.5,
16
+ 0.5,
17
+ 0.5
18
+ ],
19
+ "merge_size": 2,
20
+ "patch_size": 16,
21
+ "resample": 3,
22
+ "rescale_factor": 0.00392156862745098,
23
+ "size": {
24
+ "longest_edge": 16777216,
25
+ "shortest_edge": 65536
26
+ },
27
+ "temporal_patch_size": 2
28
+ },
29
+ "processor_class": "Qwen3VLProcessor",
30
+ "video_processor": {
31
+ "data_format": "channels_first",
32
+ "default_to_square": true,
33
+ "do_convert_rgb": true,
34
+ "do_normalize": true,
35
+ "do_rescale": true,
36
+ "do_resize": true,
37
+ "do_sample_frames": true,
38
+ "fps": 2,
39
+ "image_mean": [
40
+ 0.5,
41
+ 0.5,
42
+ 0.5
43
+ ],
44
+ "image_std": [
45
+ 0.5,
46
+ 0.5,
47
+ 0.5
48
+ ],
49
+ "max_frames": 768,
50
+ "merge_size": 2,
51
+ "min_frames": 4,
52
+ "patch_size": 16,
53
+ "resample": 3,
54
+ "rescale_factor": 0.00392156862745098,
55
+ "return_metadata": false,
56
+ "size": {
57
+ "longest_edge": 25165824,
58
+ "shortest_edge": 4096
59
+ },
60
+ "temporal_patch_size": 2,
61
+ "video_processor_type": "Qwen3VLVideoProcessor"
62
+ }
63
+ }
checkpoint-1000/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
3
+ size 11422650
checkpoint-1000/tokenizer_config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": null,
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|im_end|>",
7
+ "errors": "replace",
8
+ "extra_special_tokens": [
9
+ "<|im_start|>",
10
+ "<|im_end|>",
11
+ "<|object_ref_start|>",
12
+ "<|object_ref_end|>",
13
+ "<|box_start|>",
14
+ "<|box_end|>",
15
+ "<|quad_start|>",
16
+ "<|quad_end|>",
17
+ "<|vision_start|>",
18
+ "<|vision_end|>",
19
+ "<|vision_pad|>",
20
+ "<|image_pad|>",
21
+ "<|video_pad|>"
22
+ ],
23
+ "is_local": false,
24
+ "model_max_length": 262144,
25
+ "pad_token": "<|endoftext|>",
26
+ "processor_class": "Qwen3VLProcessor",
27
+ "split_special_tokens": false,
28
+ "tokenizer_class": "Qwen2Tokenizer",
29
+ "unk_token": null
30
+ }
checkpoint-1000/trainer_state.json ADDED
@@ -0,0 +1,206 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.0341526520051747,
6
+ "eval_steps": 250,
7
+ "global_step": 1000,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0517464424320828,
14
+ "grad_norm": 17.125,
15
+ "learning_rate": 3.356164383561644e-06,
16
+ "loss": 3.4162884521484376,
17
+ "step": 50
18
+ },
19
+ {
20
+ "epoch": 0.1034928848641656,
21
+ "grad_norm": 17.25,
22
+ "learning_rate": 6.78082191780822e-06,
23
+ "loss": 1.7713389587402344,
24
+ "step": 100
25
+ },
26
+ {
27
+ "epoch": 0.15523932729624837,
28
+ "grad_norm": 9.8125,
29
+ "learning_rate": 9.999970742404346e-06,
30
+ "loss": 1.523115997314453,
31
+ "step": 150
32
+ },
33
+ {
34
+ "epoch": 0.2069857697283312,
35
+ "grad_norm": 11.1875,
36
+ "learning_rate": 9.990871149623236e-06,
37
+ "loss": 1.4373092651367188,
38
+ "step": 200
39
+ },
40
+ {
41
+ "epoch": 0.258732212160414,
42
+ "grad_norm": 6.9375,
43
+ "learning_rate": 9.965551392499958e-06,
44
+ "loss": 1.3974497985839844,
45
+ "step": 250
46
+ },
47
+ {
48
+ "epoch": 0.258732212160414,
49
+ "eval_loss": 1.3120511770248413,
50
+ "eval_runtime": 185.4563,
51
+ "eval_samples_per_second": 53.921,
52
+ "eval_steps_per_second": 0.75,
53
+ "step": 250
54
+ },
55
+ {
56
+ "epoch": 0.31047865459249674,
57
+ "grad_norm": 6.90625,
58
+ "learning_rate": 9.924093759398476e-06,
59
+ "loss": 1.3703665161132812,
60
+ "step": 300
61
+ },
62
+ {
63
+ "epoch": 0.36222509702457956,
64
+ "grad_norm": 8.125,
65
+ "learning_rate": 9.86663298624003e-06,
66
+ "loss": 1.35594482421875,
67
+ "step": 350
68
+ },
69
+ {
70
+ "epoch": 0.4139715394566624,
71
+ "grad_norm": 6.75,
72
+ "learning_rate": 9.793355818615905e-06,
73
+ "loss": 1.3408810424804687,
74
+ "step": 400
75
+ },
76
+ {
77
+ "epoch": 0.46571798188874514,
78
+ "grad_norm": 7.75,
79
+ "learning_rate": 9.7045004048705e-06,
80
+ "loss": 1.3330262756347657,
81
+ "step": 450
82
+ },
83
+ {
84
+ "epoch": 0.517464424320828,
85
+ "grad_norm": 10.0,
86
+ "learning_rate": 9.600355522127174e-06,
87
+ "loss": 1.3203207397460937,
88
+ "step": 500
89
+ },
90
+ {
91
+ "epoch": 0.517464424320828,
92
+ "eval_loss": 1.2533953189849854,
93
+ "eval_runtime": 178.6366,
94
+ "eval_samples_per_second": 55.98,
95
+ "eval_steps_per_second": 0.778,
96
+ "step": 500
97
+ },
98
+ {
99
+ "epoch": 0.5692108667529108,
100
+ "grad_norm": 5.9375,
101
+ "learning_rate": 9.481259637772251e-06,
102
+ "loss": 1.3150039672851563,
103
+ "step": 550
104
+ },
105
+ {
106
+ "epoch": 0.6209573091849935,
107
+ "grad_norm": 6.46875,
108
+ "learning_rate": 9.347599809447331e-06,
109
+ "loss": 1.3117788696289063,
110
+ "step": 600
111
+ },
112
+ {
113
+ "epoch": 0.6727037516170763,
114
+ "grad_norm": 6.9375,
115
+ "learning_rate": 9.199810427124904e-06,
116
+ "loss": 1.3062660217285156,
117
+ "step": 650
118
+ },
119
+ {
120
+ "epoch": 0.7244501940491591,
121
+ "grad_norm": 12.625,
122
+ "learning_rate": 9.03837180135546e-06,
123
+ "loss": 1.2984295654296876,
124
+ "step": 700
125
+ },
126
+ {
127
+ "epoch": 0.7761966364812419,
128
+ "grad_norm": 6.21875,
129
+ "learning_rate": 8.86380860227426e-06,
130
+ "loss": 1.287723846435547,
131
+ "step": 750
132
+ },
133
+ {
134
+ "epoch": 0.7761966364812419,
135
+ "eval_loss": 1.2287719249725342,
136
+ "eval_runtime": 177.5762,
137
+ "eval_samples_per_second": 56.314,
138
+ "eval_steps_per_second": 0.783,
139
+ "step": 750
140
+ },
141
+ {
142
+ "epoch": 0.8279430789133247,
143
+ "grad_norm": 8.3125,
144
+ "learning_rate": 8.676688154440947e-06,
145
+ "loss": 1.293363037109375,
146
+ "step": 800
147
+ },
148
+ {
149
+ "epoch": 0.8796895213454075,
150
+ "grad_norm": 7.90625,
151
+ "learning_rate": 8.477618593053693e-06,
152
+ "loss": 1.2859170532226563,
153
+ "step": 850
154
+ },
155
+ {
156
+ "epoch": 0.9314359637774903,
157
+ "grad_norm": 6.25,
158
+ "learning_rate": 8.26724688753015e-06,
159
+ "loss": 1.2893869018554687,
160
+ "step": 900
161
+ },
162
+ {
163
+ "epoch": 0.9831824062095731,
164
+ "grad_norm": 5.21875,
165
+ "learning_rate": 8.04625673887846e-06,
166
+ "loss": 1.2764347076416016,
167
+ "step": 950
168
+ },
169
+ {
170
+ "epoch": 1.0341526520051747,
171
+ "grad_norm": 4.90625,
172
+ "learning_rate": 7.815366357691863e-06,
173
+ "loss": 1.266595458984375,
174
+ "step": 1000
175
+ },
176
+ {
177
+ "epoch": 1.0341526520051747,
178
+ "eval_loss": 1.2155812978744507,
179
+ "eval_runtime": 177.1968,
180
+ "eval_samples_per_second": 56.434,
181
+ "eval_steps_per_second": 0.784,
182
+ "step": 1000
183
+ }
184
+ ],
185
+ "logging_steps": 50,
186
+ "max_steps": 2901,
187
+ "num_input_tokens_seen": 0,
188
+ "num_train_epochs": 3,
189
+ "save_steps": 1000,
190
+ "stateful_callbacks": {
191
+ "TrainerControl": {
192
+ "args": {
193
+ "should_epoch_stop": false,
194
+ "should_evaluate": false,
195
+ "should_log": false,
196
+ "should_save": true,
197
+ "should_training_stop": false
198
+ },
199
+ "attributes": {}
200
+ }
201
+ },
202
+ "total_flos": 5.250430427952316e+18,
203
+ "train_batch_size": 18,
204
+ "trial_name": null,
205
+ "trial_params": null
206
+ }
checkpoint-1000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25f31ca829e8d2c0fd93d4b4b550d0bb0e9c801e4b84c4542753a7beaf8dcf8c
3
+ size 5201
checkpoint-2000/chat_template.jinja ADDED
@@ -0,0 +1,120 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0].role == 'system' %}
4
+ {%- if messages[0].content is string %}
5
+ {{- messages[0].content }}
6
+ {%- else %}
7
+ {%- for content in messages[0].content %}
8
+ {%- if 'text' in content %}
9
+ {{- content.text }}
10
+ {%- endif %}
11
+ {%- endfor %}
12
+ {%- endif %}
13
+ {{- '\n\n' }}
14
+ {%- endif %}
15
+ {{- "# 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>" }}
16
+ {%- for tool in tools %}
17
+ {{- "\n" }}
18
+ {{- tool | tojson }}
19
+ {%- endfor %}
20
+ {{- "\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" }}
21
+ {%- else %}
22
+ {%- if messages[0].role == 'system' %}
23
+ {{- '<|im_start|>system\n' }}
24
+ {%- if messages[0].content is string %}
25
+ {{- messages[0].content }}
26
+ {%- else %}
27
+ {%- for content in messages[0].content %}
28
+ {%- if 'text' in content %}
29
+ {{- content.text }}
30
+ {%- endif %}
31
+ {%- endfor %}
32
+ {%- endif %}
33
+ {{- '<|im_end|>\n' }}
34
+ {%- endif %}
35
+ {%- endif %}
36
+ {%- set image_count = namespace(value=0) %}
37
+ {%- set video_count = namespace(value=0) %}
38
+ {%- for message in messages %}
39
+ {%- if message.role == "user" %}
40
+ {{- '<|im_start|>' + message.role + '\n' }}
41
+ {%- if message.content is string %}
42
+ {{- message.content }}
43
+ {%- else %}
44
+ {%- for content in message.content %}
45
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
46
+ {%- set image_count.value = image_count.value + 1 %}
47
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
48
+ <|vision_start|><|image_pad|><|vision_end|>
49
+ {%- elif content.type == 'video' or 'video' in content %}
50
+ {%- set video_count.value = video_count.value + 1 %}
51
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
52
+ <|vision_start|><|video_pad|><|vision_end|>
53
+ {%- elif 'text' in content %}
54
+ {{- content.text }}
55
+ {%- endif %}
56
+ {%- endfor %}
57
+ {%- endif %}
58
+ {{- '<|im_end|>\n' }}
59
+ {%- elif message.role == "assistant" %}
60
+ {{- '<|im_start|>' + message.role + '\n' }}
61
+ {%- if message.content is string %}
62
+ {{- message.content }}
63
+ {%- else %}
64
+ {%- for content_item in message.content %}
65
+ {%- if 'text' in content_item %}
66
+ {{- content_item.text }}
67
+ {%- endif %}
68
+ {%- endfor %}
69
+ {%- endif %}
70
+ {%- if message.tool_calls %}
71
+ {%- for tool_call in message.tool_calls %}
72
+ {%- if (loop.first and message.content) or (not loop.first) %}
73
+ {{- '\n' }}
74
+ {%- endif %}
75
+ {%- if tool_call.function %}
76
+ {%- set tool_call = tool_call.function %}
77
+ {%- endif %}
78
+ {{- '<tool_call>\n{"name": "' }}
79
+ {{- tool_call.name }}
80
+ {{- '", "arguments": ' }}
81
+ {%- if tool_call.arguments is string %}
82
+ {{- tool_call.arguments }}
83
+ {%- else %}
84
+ {{- tool_call.arguments | tojson }}
85
+ {%- endif %}
86
+ {{- '}\n</tool_call>' }}
87
+ {%- endfor %}
88
+ {%- endif %}
89
+ {{- '<|im_end|>\n' }}
90
+ {%- elif message.role == "tool" %}
91
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
92
+ {{- '<|im_start|>user' }}
93
+ {%- endif %}
94
+ {{- '\n<tool_response>\n' }}
95
+ {%- if message.content is string %}
96
+ {{- message.content }}
97
+ {%- else %}
98
+ {%- for content in message.content %}
99
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
100
+ {%- set image_count.value = image_count.value + 1 %}
101
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
102
+ <|vision_start|><|image_pad|><|vision_end|>
103
+ {%- elif content.type == 'video' or 'video' in content %}
104
+ {%- set video_count.value = video_count.value + 1 %}
105
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
106
+ <|vision_start|><|video_pad|><|vision_end|>
107
+ {%- elif 'text' in content %}
108
+ {{- content.text }}
109
+ {%- endif %}
110
+ {%- endfor %}
111
+ {%- endif %}
112
+ {{- '\n</tool_response>' }}
113
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
114
+ {{- '<|im_end|>\n' }}
115
+ {%- endif %}
116
+ {%- endif %}
117
+ {%- endfor %}
118
+ {%- if add_generation_prompt %}
119
+ {{- '<|im_start|>assistant\n' }}
120
+ {%- endif %}
checkpoint-2000/config.json ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3VLForConditionalGeneration"
4
+ ],
5
+ "bos_token_id": null,
6
+ "dtype": "bfloat16",
7
+ "eos_token_id": 151645,
8
+ "image_token_id": 151655,
9
+ "model_type": "qwen3_vl",
10
+ "pad_token_id": 151643,
11
+ "text_config": {
12
+ "attention_bias": false,
13
+ "attention_dropout": 0.0,
14
+ "bos_token_id": 151643,
15
+ "dtype": "bfloat16",
16
+ "eos_token_id": 151645,
17
+ "head_dim": 128,
18
+ "hidden_act": "silu",
19
+ "hidden_size": 2560,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 9728,
22
+ "max_position_embeddings": 262144,
23
+ "model_type": "qwen3_vl_text",
24
+ "num_attention_heads": 32,
25
+ "num_hidden_layers": 36,
26
+ "num_key_value_heads": 8,
27
+ "pad_token_id": null,
28
+ "rms_norm_eps": 1e-06,
29
+ "rope_parameters": {
30
+ "mrope_interleaved": true,
31
+ "mrope_section": [
32
+ 24,
33
+ 20,
34
+ 20
35
+ ],
36
+ "rope_theta": 5000000,
37
+ "rope_type": "default"
38
+ },
39
+ "tie_word_embeddings": true,
40
+ "use_cache": true,
41
+ "vocab_size": 151936
42
+ },
43
+ "tie_word_embeddings": true,
44
+ "transformers_version": "5.0.0",
45
+ "use_cache": false,
46
+ "video_token_id": 151656,
47
+ "vision_config": {
48
+ "deepstack_visual_indexes": [
49
+ 5,
50
+ 11,
51
+ 17
52
+ ],
53
+ "depth": 24,
54
+ "dtype": "bfloat16",
55
+ "hidden_act": "gelu_pytorch_tanh",
56
+ "hidden_size": 1024,
57
+ "in_channels": 3,
58
+ "initializer_range": 0.02,
59
+ "intermediate_size": 4096,
60
+ "model_type": "qwen3_vl",
61
+ "num_heads": 16,
62
+ "num_position_embeddings": 2304,
63
+ "out_hidden_size": 2560,
64
+ "patch_size": 16,
65
+ "spatial_merge_size": 2,
66
+ "temporal_patch_size": 2
67
+ },
68
+ "vision_end_token_id": 151653,
69
+ "vision_start_token_id": 151652
70
+ }
checkpoint-2000/generation_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_sample": true,
3
+ "eos_token_id": [
4
+ 151645,
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.0,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "5.0.0"
14
+ }
checkpoint-2000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb322a82d96217d934db7593eef50fd724c93e02925ef8c9204c2f3906ed9e7f
3
+ size 8875719408
checkpoint-2000/processor_config.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "data_format": "channels_first",
4
+ "do_convert_rgb": true,
5
+ "do_normalize": true,
6
+ "do_rescale": true,
7
+ "do_resize": true,
8
+ "image_mean": [
9
+ 0.5,
10
+ 0.5,
11
+ 0.5
12
+ ],
13
+ "image_processor_type": "Qwen2VLImageProcessorFast",
14
+ "image_std": [
15
+ 0.5,
16
+ 0.5,
17
+ 0.5
18
+ ],
19
+ "merge_size": 2,
20
+ "patch_size": 16,
21
+ "resample": 3,
22
+ "rescale_factor": 0.00392156862745098,
23
+ "size": {
24
+ "longest_edge": 16777216,
25
+ "shortest_edge": 65536
26
+ },
27
+ "temporal_patch_size": 2
28
+ },
29
+ "processor_class": "Qwen3VLProcessor",
30
+ "video_processor": {
31
+ "data_format": "channels_first",
32
+ "default_to_square": true,
33
+ "do_convert_rgb": true,
34
+ "do_normalize": true,
35
+ "do_rescale": true,
36
+ "do_resize": true,
37
+ "do_sample_frames": true,
38
+ "fps": 2,
39
+ "image_mean": [
40
+ 0.5,
41
+ 0.5,
42
+ 0.5
43
+ ],
44
+ "image_std": [
45
+ 0.5,
46
+ 0.5,
47
+ 0.5
48
+ ],
49
+ "max_frames": 768,
50
+ "merge_size": 2,
51
+ "min_frames": 4,
52
+ "patch_size": 16,
53
+ "resample": 3,
54
+ "rescale_factor": 0.00392156862745098,
55
+ "return_metadata": false,
56
+ "size": {
57
+ "longest_edge": 25165824,
58
+ "shortest_edge": 4096
59
+ },
60
+ "temporal_patch_size": 2,
61
+ "video_processor_type": "Qwen3VLVideoProcessor"
62
+ }
63
+ }
checkpoint-2000/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
3
+ size 11422650
checkpoint-2000/tokenizer_config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": null,
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|im_end|>",
7
+ "errors": "replace",
8
+ "extra_special_tokens": [
9
+ "<|im_start|>",
10
+ "<|im_end|>",
11
+ "<|object_ref_start|>",
12
+ "<|object_ref_end|>",
13
+ "<|box_start|>",
14
+ "<|box_end|>",
15
+ "<|quad_start|>",
16
+ "<|quad_end|>",
17
+ "<|vision_start|>",
18
+ "<|vision_end|>",
19
+ "<|vision_pad|>",
20
+ "<|image_pad|>",
21
+ "<|video_pad|>"
22
+ ],
23
+ "is_local": false,
24
+ "model_max_length": 262144,
25
+ "pad_token": "<|endoftext|>",
26
+ "processor_class": "Qwen3VLProcessor",
27
+ "split_special_tokens": false,
28
+ "tokenizer_class": "Qwen2Tokenizer",
29
+ "unk_token": null
30
+ }
checkpoint-2000/trainer_state.json ADDED
@@ -0,0 +1,378 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 2.0683053040103494,
6
+ "eval_steps": 250,
7
+ "global_step": 2000,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0517464424320828,
14
+ "grad_norm": 17.125,
15
+ "learning_rate": 3.356164383561644e-06,
16
+ "loss": 3.4162884521484376,
17
+ "step": 50
18
+ },
19
+ {
20
+ "epoch": 0.1034928848641656,
21
+ "grad_norm": 17.25,
22
+ "learning_rate": 6.78082191780822e-06,
23
+ "loss": 1.7713389587402344,
24
+ "step": 100
25
+ },
26
+ {
27
+ "epoch": 0.15523932729624837,
28
+ "grad_norm": 9.8125,
29
+ "learning_rate": 9.999970742404346e-06,
30
+ "loss": 1.523115997314453,
31
+ "step": 150
32
+ },
33
+ {
34
+ "epoch": 0.2069857697283312,
35
+ "grad_norm": 11.1875,
36
+ "learning_rate": 9.990871149623236e-06,
37
+ "loss": 1.4373092651367188,
38
+ "step": 200
39
+ },
40
+ {
41
+ "epoch": 0.258732212160414,
42
+ "grad_norm": 6.9375,
43
+ "learning_rate": 9.965551392499958e-06,
44
+ "loss": 1.3974497985839844,
45
+ "step": 250
46
+ },
47
+ {
48
+ "epoch": 0.258732212160414,
49
+ "eval_loss": 1.3120511770248413,
50
+ "eval_runtime": 185.4563,
51
+ "eval_samples_per_second": 53.921,
52
+ "eval_steps_per_second": 0.75,
53
+ "step": 250
54
+ },
55
+ {
56
+ "epoch": 0.31047865459249674,
57
+ "grad_norm": 6.90625,
58
+ "learning_rate": 9.924093759398476e-06,
59
+ "loss": 1.3703665161132812,
60
+ "step": 300
61
+ },
62
+ {
63
+ "epoch": 0.36222509702457956,
64
+ "grad_norm": 8.125,
65
+ "learning_rate": 9.86663298624003e-06,
66
+ "loss": 1.35594482421875,
67
+ "step": 350
68
+ },
69
+ {
70
+ "epoch": 0.4139715394566624,
71
+ "grad_norm": 6.75,
72
+ "learning_rate": 9.793355818615905e-06,
73
+ "loss": 1.3408810424804687,
74
+ "step": 400
75
+ },
76
+ {
77
+ "epoch": 0.46571798188874514,
78
+ "grad_norm": 7.75,
79
+ "learning_rate": 9.7045004048705e-06,
80
+ "loss": 1.3330262756347657,
81
+ "step": 450
82
+ },
83
+ {
84
+ "epoch": 0.517464424320828,
85
+ "grad_norm": 10.0,
86
+ "learning_rate": 9.600355522127174e-06,
87
+ "loss": 1.3203207397460937,
88
+ "step": 500
89
+ },
90
+ {
91
+ "epoch": 0.517464424320828,
92
+ "eval_loss": 1.2533953189849854,
93
+ "eval_runtime": 178.6366,
94
+ "eval_samples_per_second": 55.98,
95
+ "eval_steps_per_second": 0.778,
96
+ "step": 500
97
+ },
98
+ {
99
+ "epoch": 0.5692108667529108,
100
+ "grad_norm": 5.9375,
101
+ "learning_rate": 9.481259637772251e-06,
102
+ "loss": 1.3150039672851563,
103
+ "step": 550
104
+ },
105
+ {
106
+ "epoch": 0.6209573091849935,
107
+ "grad_norm": 6.46875,
108
+ "learning_rate": 9.347599809447331e-06,
109
+ "loss": 1.3117788696289063,
110
+ "step": 600
111
+ },
112
+ {
113
+ "epoch": 0.6727037516170763,
114
+ "grad_norm": 6.9375,
115
+ "learning_rate": 9.199810427124904e-06,
116
+ "loss": 1.3062660217285156,
117
+ "step": 650
118
+ },
119
+ {
120
+ "epoch": 0.7244501940491591,
121
+ "grad_norm": 12.625,
122
+ "learning_rate": 9.03837180135546e-06,
123
+ "loss": 1.2984295654296876,
124
+ "step": 700
125
+ },
126
+ {
127
+ "epoch": 0.7761966364812419,
128
+ "grad_norm": 6.21875,
129
+ "learning_rate": 8.86380860227426e-06,
130
+ "loss": 1.287723846435547,
131
+ "step": 750
132
+ },
133
+ {
134
+ "epoch": 0.7761966364812419,
135
+ "eval_loss": 1.2287719249725342,
136
+ "eval_runtime": 177.5762,
137
+ "eval_samples_per_second": 56.314,
138
+ "eval_steps_per_second": 0.783,
139
+ "step": 750
140
+ },
141
+ {
142
+ "epoch": 0.8279430789133247,
143
+ "grad_norm": 8.3125,
144
+ "learning_rate": 8.676688154440947e-06,
145
+ "loss": 1.293363037109375,
146
+ "step": 800
147
+ },
148
+ {
149
+ "epoch": 0.8796895213454075,
150
+ "grad_norm": 7.90625,
151
+ "learning_rate": 8.477618593053693e-06,
152
+ "loss": 1.2859170532226563,
153
+ "step": 850
154
+ },
155
+ {
156
+ "epoch": 0.9314359637774903,
157
+ "grad_norm": 6.25,
158
+ "learning_rate": 8.26724688753015e-06,
159
+ "loss": 1.2893869018554687,
160
+ "step": 900
161
+ },
162
+ {
163
+ "epoch": 0.9831824062095731,
164
+ "grad_norm": 5.21875,
165
+ "learning_rate": 8.04625673887846e-06,
166
+ "loss": 1.2764347076416016,
167
+ "step": 950
168
+ },
169
+ {
170
+ "epoch": 1.0341526520051747,
171
+ "grad_norm": 4.90625,
172
+ "learning_rate": 7.815366357691863e-06,
173
+ "loss": 1.266595458984375,
174
+ "step": 1000
175
+ },
176
+ {
177
+ "epoch": 1.0341526520051747,
178
+ "eval_loss": 1.2155812978744507,
179
+ "eval_runtime": 177.1968,
180
+ "eval_samples_per_second": 56.434,
181
+ "eval_steps_per_second": 0.784,
182
+ "step": 1000
183
+ },
184
+ {
185
+ "epoch": 1.0858990944372575,
186
+ "grad_norm": 6.9375,
187
+ "learning_rate": 7.575326129988268e-06,
188
+ "loss": 1.2680277252197265,
189
+ "step": 1050
190
+ },
191
+ {
192
+ "epoch": 1.1376455368693401,
193
+ "grad_norm": 5.375,
194
+ "learning_rate": 7.326916178480796e-06,
195
+ "loss": 1.2629891204833985,
196
+ "step": 1100
197
+ },
198
+ {
199
+ "epoch": 1.189391979301423,
200
+ "grad_norm": 8.75,
201
+ "learning_rate": 7.070943827205048e-06,
202
+ "loss": 1.2643635559082032,
203
+ "step": 1150
204
+ },
205
+ {
206
+ "epoch": 1.2411384217335057,
207
+ "grad_norm": 11.9375,
208
+ "learning_rate": 6.808240977742987e-06,
209
+ "loss": 1.2576502227783204,
210
+ "step": 1200
211
+ },
212
+ {
213
+ "epoch": 1.2928848641655886,
214
+ "grad_norm": 5.25,
215
+ "learning_rate": 6.539661405570615e-06,
216
+ "loss": 1.2591765594482422,
217
+ "step": 1250
218
+ },
219
+ {
220
+ "epoch": 1.2928848641655886,
221
+ "eval_loss": 1.2061747312545776,
222
+ "eval_runtime": 179.247,
223
+ "eval_samples_per_second": 55.789,
224
+ "eval_steps_per_second": 0.775,
225
+ "step": 1250
226
+ },
227
+ {
228
+ "epoch": 1.3446313065976714,
229
+ "grad_norm": 5.03125,
230
+ "learning_rate": 6.266077985316224e-06,
231
+ "loss": 1.258944091796875,
232
+ "step": 1300
233
+ },
234
+ {
235
+ "epoch": 1.3963777490297542,
236
+ "grad_norm": 7.90625,
237
+ "learning_rate": 5.988379853947046e-06,
238
+ "loss": 1.2634922790527343,
239
+ "step": 1350
240
+ },
241
+ {
242
+ "epoch": 1.448124191461837,
243
+ "grad_norm": 7.59375,
244
+ "learning_rate": 5.7074695211038385e-06,
245
+ "loss": 1.2592449188232422,
246
+ "step": 1400
247
+ },
248
+ {
249
+ "epoch": 1.4998706338939198,
250
+ "grad_norm": 5.90625,
251
+ "learning_rate": 5.424259935974736e-06,
252
+ "loss": 1.2600762939453125,
253
+ "step": 1450
254
+ },
255
+ {
256
+ "epoch": 1.5516170763260027,
257
+ "grad_norm": 5.28125,
258
+ "learning_rate": 5.139671520240915e-06,
259
+ "loss": 1.2590911102294922,
260
+ "step": 1500
261
+ },
262
+ {
263
+ "epoch": 1.5516170763260027,
264
+ "eval_loss": 1.201247215270996,
265
+ "eval_runtime": 179.3983,
266
+ "eval_samples_per_second": 55.742,
267
+ "eval_steps_per_second": 0.775,
268
+ "step": 1500
269
+ },
270
+ {
271
+ "epoch": 1.6033635187580852,
272
+ "grad_norm": 5.875,
273
+ "learning_rate": 4.854629176736911e-06,
274
+ "loss": 1.25985595703125,
275
+ "step": 1550
276
+ },
277
+ {
278
+ "epoch": 1.6551099611901683,
279
+ "grad_norm": 6.09375,
280
+ "learning_rate": 4.570059283547328e-06,
281
+ "loss": 1.2547737121582032,
282
+ "step": 1600
283
+ },
284
+ {
285
+ "epoch": 1.7068564036222509,
286
+ "grad_norm": 6.15625,
287
+ "learning_rate": 4.286886683309053e-06,
288
+ "loss": 1.2532640075683594,
289
+ "step": 1650
290
+ },
291
+ {
292
+ "epoch": 1.758602846054334,
293
+ "grad_norm": 7.0625,
294
+ "learning_rate": 4.0060316775036025e-06,
295
+ "loss": 1.2515839385986327,
296
+ "step": 1700
297
+ },
298
+ {
299
+ "epoch": 1.8103492884864165,
300
+ "grad_norm": 6.8125,
301
+ "learning_rate": 3.728407035508089e-06,
302
+ "loss": 1.2524348449707032,
303
+ "step": 1750
304
+ },
305
+ {
306
+ "epoch": 1.8103492884864165,
307
+ "eval_loss": 1.198207974433899,
308
+ "eval_runtime": 178.7391,
309
+ "eval_samples_per_second": 55.947,
310
+ "eval_steps_per_second": 0.778,
311
+ "step": 1750
312
+ },
313
+ {
314
+ "epoch": 1.8620957309184993,
315
+ "grad_norm": 6.84375,
316
+ "learning_rate": 3.4549150281252635e-06,
317
+ "loss": 1.2492080688476563,
318
+ "step": 1800
319
+ },
320
+ {
321
+ "epoch": 1.9138421733505822,
322
+ "grad_norm": 6.59375,
323
+ "learning_rate": 3.1864444952335823e-06,
324
+ "loss": 1.2510016632080079,
325
+ "step": 1850
326
+ },
327
+ {
328
+ "epoch": 1.965588615782665,
329
+ "grad_norm": 9.6875,
330
+ "learning_rate": 2.9238679570873217e-06,
331
+ "loss": 1.2530317687988282,
332
+ "step": 1900
333
+ },
334
+ {
335
+ "epoch": 2.0165588615782664,
336
+ "grad_norm": 5.3125,
337
+ "learning_rate": 2.66803877865493e-06,
338
+ "loss": 1.254935760498047,
339
+ "step": 1950
340
+ },
341
+ {
342
+ "epoch": 2.0683053040103494,
343
+ "grad_norm": 6.875,
344
+ "learning_rate": 2.4197883962114164e-06,
345
+ "loss": 1.2447527313232423,
346
+ "step": 2000
347
+ },
348
+ {
349
+ "epoch": 2.0683053040103494,
350
+ "eval_loss": 1.1967895030975342,
351
+ "eval_runtime": 179.8167,
352
+ "eval_samples_per_second": 55.612,
353
+ "eval_steps_per_second": 0.773,
354
+ "step": 2000
355
+ }
356
+ ],
357
+ "logging_steps": 50,
358
+ "max_steps": 2901,
359
+ "num_input_tokens_seen": 0,
360
+ "num_train_epochs": 3,
361
+ "save_steps": 1000,
362
+ "stateful_callbacks": {
363
+ "TrainerControl": {
364
+ "args": {
365
+ "should_epoch_stop": false,
366
+ "should_evaluate": false,
367
+ "should_log": false,
368
+ "should_save": true,
369
+ "should_training_stop": false
370
+ },
371
+ "attributes": {}
372
+ }
373
+ },
374
+ "total_flos": 1.0500904956628828e+19,
375
+ "train_batch_size": 18,
376
+ "trial_name": null,
377
+ "trial_params": null
378
+ }
checkpoint-2000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25f31ca829e8d2c0fd93d4b4b550d0bb0e9c801e4b84c4542753a7beaf8dcf8c
3
+ size 5201
checkpoint-2901/chat_template.jinja ADDED
@@ -0,0 +1,120 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0].role == 'system' %}
4
+ {%- if messages[0].content is string %}
5
+ {{- messages[0].content }}
6
+ {%- else %}
7
+ {%- for content in messages[0].content %}
8
+ {%- if 'text' in content %}
9
+ {{- content.text }}
10
+ {%- endif %}
11
+ {%- endfor %}
12
+ {%- endif %}
13
+ {{- '\n\n' }}
14
+ {%- endif %}
15
+ {{- "# 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>" }}
16
+ {%- for tool in tools %}
17
+ {{- "\n" }}
18
+ {{- tool | tojson }}
19
+ {%- endfor %}
20
+ {{- "\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" }}
21
+ {%- else %}
22
+ {%- if messages[0].role == 'system' %}
23
+ {{- '<|im_start|>system\n' }}
24
+ {%- if messages[0].content is string %}
25
+ {{- messages[0].content }}
26
+ {%- else %}
27
+ {%- for content in messages[0].content %}
28
+ {%- if 'text' in content %}
29
+ {{- content.text }}
30
+ {%- endif %}
31
+ {%- endfor %}
32
+ {%- endif %}
33
+ {{- '<|im_end|>\n' }}
34
+ {%- endif %}
35
+ {%- endif %}
36
+ {%- set image_count = namespace(value=0) %}
37
+ {%- set video_count = namespace(value=0) %}
38
+ {%- for message in messages %}
39
+ {%- if message.role == "user" %}
40
+ {{- '<|im_start|>' + message.role + '\n' }}
41
+ {%- if message.content is string %}
42
+ {{- message.content }}
43
+ {%- else %}
44
+ {%- for content in message.content %}
45
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
46
+ {%- set image_count.value = image_count.value + 1 %}
47
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
48
+ <|vision_start|><|image_pad|><|vision_end|>
49
+ {%- elif content.type == 'video' or 'video' in content %}
50
+ {%- set video_count.value = video_count.value + 1 %}
51
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
52
+ <|vision_start|><|video_pad|><|vision_end|>
53
+ {%- elif 'text' in content %}
54
+ {{- content.text }}
55
+ {%- endif %}
56
+ {%- endfor %}
57
+ {%- endif %}
58
+ {{- '<|im_end|>\n' }}
59
+ {%- elif message.role == "assistant" %}
60
+ {{- '<|im_start|>' + message.role + '\n' }}
61
+ {%- if message.content is string %}
62
+ {{- message.content }}
63
+ {%- else %}
64
+ {%- for content_item in message.content %}
65
+ {%- if 'text' in content_item %}
66
+ {{- content_item.text }}
67
+ {%- endif %}
68
+ {%- endfor %}
69
+ {%- endif %}
70
+ {%- if message.tool_calls %}
71
+ {%- for tool_call in message.tool_calls %}
72
+ {%- if (loop.first and message.content) or (not loop.first) %}
73
+ {{- '\n' }}
74
+ {%- endif %}
75
+ {%- if tool_call.function %}
76
+ {%- set tool_call = tool_call.function %}
77
+ {%- endif %}
78
+ {{- '<tool_call>\n{"name": "' }}
79
+ {{- tool_call.name }}
80
+ {{- '", "arguments": ' }}
81
+ {%- if tool_call.arguments is string %}
82
+ {{- tool_call.arguments }}
83
+ {%- else %}
84
+ {{- tool_call.arguments | tojson }}
85
+ {%- endif %}
86
+ {{- '}\n</tool_call>' }}
87
+ {%- endfor %}
88
+ {%- endif %}
89
+ {{- '<|im_end|>\n' }}
90
+ {%- elif message.role == "tool" %}
91
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
92
+ {{- '<|im_start|>user' }}
93
+ {%- endif %}
94
+ {{- '\n<tool_response>\n' }}
95
+ {%- if message.content is string %}
96
+ {{- message.content }}
97
+ {%- else %}
98
+ {%- for content in message.content %}
99
+ {%- if content.type == 'image' or 'image' in content or 'image_url' in content %}
100
+ {%- set image_count.value = image_count.value + 1 %}
101
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
102
+ <|vision_start|><|image_pad|><|vision_end|>
103
+ {%- elif content.type == 'video' or 'video' in content %}
104
+ {%- set video_count.value = video_count.value + 1 %}
105
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
106
+ <|vision_start|><|video_pad|><|vision_end|>
107
+ {%- elif 'text' in content %}
108
+ {{- content.text }}
109
+ {%- endif %}
110
+ {%- endfor %}
111
+ {%- endif %}
112
+ {{- '\n</tool_response>' }}
113
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
114
+ {{- '<|im_end|>\n' }}
115
+ {%- endif %}
116
+ {%- endif %}
117
+ {%- endfor %}
118
+ {%- if add_generation_prompt %}
119
+ {{- '<|im_start|>assistant\n' }}
120
+ {%- endif %}
checkpoint-2901/config.json ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3VLForConditionalGeneration"
4
+ ],
5
+ "bos_token_id": null,
6
+ "dtype": "bfloat16",
7
+ "eos_token_id": 151645,
8
+ "image_token_id": 151655,
9
+ "model_type": "qwen3_vl",
10
+ "pad_token_id": 151643,
11
+ "text_config": {
12
+ "attention_bias": false,
13
+ "attention_dropout": 0.0,
14
+ "bos_token_id": 151643,
15
+ "dtype": "bfloat16",
16
+ "eos_token_id": 151645,
17
+ "head_dim": 128,
18
+ "hidden_act": "silu",
19
+ "hidden_size": 2560,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 9728,
22
+ "max_position_embeddings": 262144,
23
+ "model_type": "qwen3_vl_text",
24
+ "num_attention_heads": 32,
25
+ "num_hidden_layers": 36,
26
+ "num_key_value_heads": 8,
27
+ "pad_token_id": null,
28
+ "rms_norm_eps": 1e-06,
29
+ "rope_parameters": {
30
+ "mrope_interleaved": true,
31
+ "mrope_section": [
32
+ 24,
33
+ 20,
34
+ 20
35
+ ],
36
+ "rope_theta": 5000000,
37
+ "rope_type": "default"
38
+ },
39
+ "tie_word_embeddings": true,
40
+ "use_cache": true,
41
+ "vocab_size": 151936
42
+ },
43
+ "tie_word_embeddings": true,
44
+ "transformers_version": "5.0.0",
45
+ "use_cache": false,
46
+ "video_token_id": 151656,
47
+ "vision_config": {
48
+ "deepstack_visual_indexes": [
49
+ 5,
50
+ 11,
51
+ 17
52
+ ],
53
+ "depth": 24,
54
+ "dtype": "bfloat16",
55
+ "hidden_act": "gelu_pytorch_tanh",
56
+ "hidden_size": 1024,
57
+ "in_channels": 3,
58
+ "initializer_range": 0.02,
59
+ "intermediate_size": 4096,
60
+ "model_type": "qwen3_vl",
61
+ "num_heads": 16,
62
+ "num_position_embeddings": 2304,
63
+ "out_hidden_size": 2560,
64
+ "patch_size": 16,
65
+ "spatial_merge_size": 2,
66
+ "temporal_patch_size": 2
67
+ },
68
+ "vision_end_token_id": 151653,
69
+ "vision_start_token_id": 151652
70
+ }
checkpoint-2901/generation_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_sample": true,
3
+ "eos_token_id": [
4
+ 151645,
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.0,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "5.0.0"
14
+ }
checkpoint-2901/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:01454076c1f037c770800517410587300c263b5b3fbe6f382e3304592cfaf04a
3
+ size 8875719408
checkpoint-2901/processor_config.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "data_format": "channels_first",
4
+ "do_convert_rgb": true,
5
+ "do_normalize": true,
6
+ "do_rescale": true,
7
+ "do_resize": true,
8
+ "image_mean": [
9
+ 0.5,
10
+ 0.5,
11
+ 0.5
12
+ ],
13
+ "image_processor_type": "Qwen2VLImageProcessorFast",
14
+ "image_std": [
15
+ 0.5,
16
+ 0.5,
17
+ 0.5
18
+ ],
19
+ "merge_size": 2,
20
+ "patch_size": 16,
21
+ "resample": 3,
22
+ "rescale_factor": 0.00392156862745098,
23
+ "size": {
24
+ "longest_edge": 16777216,
25
+ "shortest_edge": 65536
26
+ },
27
+ "temporal_patch_size": 2
28
+ },
29
+ "processor_class": "Qwen3VLProcessor",
30
+ "video_processor": {
31
+ "data_format": "channels_first",
32
+ "default_to_square": true,
33
+ "do_convert_rgb": true,
34
+ "do_normalize": true,
35
+ "do_rescale": true,
36
+ "do_resize": true,
37
+ "do_sample_frames": true,
38
+ "fps": 2,
39
+ "image_mean": [
40
+ 0.5,
41
+ 0.5,
42
+ 0.5
43
+ ],
44
+ "image_std": [
45
+ 0.5,
46
+ 0.5,
47
+ 0.5
48
+ ],
49
+ "max_frames": 768,
50
+ "merge_size": 2,
51
+ "min_frames": 4,
52
+ "patch_size": 16,
53
+ "resample": 3,
54
+ "rescale_factor": 0.00392156862745098,
55
+ "return_metadata": false,
56
+ "size": {
57
+ "longest_edge": 25165824,
58
+ "shortest_edge": 4096
59
+ },
60
+ "temporal_patch_size": 2,
61
+ "video_processor_type": "Qwen3VLVideoProcessor"
62
+ }
63
+ }
checkpoint-2901/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
3
+ size 11422650
checkpoint-2901/tokenizer_config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": null,
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|im_end|>",
7
+ "errors": "replace",
8
+ "extra_special_tokens": [
9
+ "<|im_start|>",
10
+ "<|im_end|>",
11
+ "<|object_ref_start|>",
12
+ "<|object_ref_end|>",
13
+ "<|box_start|>",
14
+ "<|box_end|>",
15
+ "<|quad_start|>",
16
+ "<|quad_end|>",
17
+ "<|vision_start|>",
18
+ "<|vision_end|>",
19
+ "<|vision_pad|>",
20
+ "<|image_pad|>",
21
+ "<|video_pad|>"
22
+ ],
23
+ "is_local": false,
24
+ "model_max_length": 262144,
25
+ "pad_token": "<|endoftext|>",
26
+ "processor_class": "Qwen3VLProcessor",
27
+ "split_special_tokens": false,
28
+ "tokenizer_class": "Qwen2Tokenizer",
29
+ "unk_token": null
30
+ }
checkpoint-2901/trainer_state.json ADDED
@@ -0,0 +1,528 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 3.0,
6
+ "eval_steps": 250,
7
+ "global_step": 2901,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0517464424320828,
14
+ "grad_norm": 17.125,
15
+ "learning_rate": 3.356164383561644e-06,
16
+ "loss": 3.4162884521484376,
17
+ "step": 50
18
+ },
19
+ {
20
+ "epoch": 0.1034928848641656,
21
+ "grad_norm": 17.25,
22
+ "learning_rate": 6.78082191780822e-06,
23
+ "loss": 1.7713389587402344,
24
+ "step": 100
25
+ },
26
+ {
27
+ "epoch": 0.15523932729624837,
28
+ "grad_norm": 9.8125,
29
+ "learning_rate": 9.999970742404346e-06,
30
+ "loss": 1.523115997314453,
31
+ "step": 150
32
+ },
33
+ {
34
+ "epoch": 0.2069857697283312,
35
+ "grad_norm": 11.1875,
36
+ "learning_rate": 9.990871149623236e-06,
37
+ "loss": 1.4373092651367188,
38
+ "step": 200
39
+ },
40
+ {
41
+ "epoch": 0.258732212160414,
42
+ "grad_norm": 6.9375,
43
+ "learning_rate": 9.965551392499958e-06,
44
+ "loss": 1.3974497985839844,
45
+ "step": 250
46
+ },
47
+ {
48
+ "epoch": 0.258732212160414,
49
+ "eval_loss": 1.3120511770248413,
50
+ "eval_runtime": 185.4563,
51
+ "eval_samples_per_second": 53.921,
52
+ "eval_steps_per_second": 0.75,
53
+ "step": 250
54
+ },
55
+ {
56
+ "epoch": 0.31047865459249674,
57
+ "grad_norm": 6.90625,
58
+ "learning_rate": 9.924093759398476e-06,
59
+ "loss": 1.3703665161132812,
60
+ "step": 300
61
+ },
62
+ {
63
+ "epoch": 0.36222509702457956,
64
+ "grad_norm": 8.125,
65
+ "learning_rate": 9.86663298624003e-06,
66
+ "loss": 1.35594482421875,
67
+ "step": 350
68
+ },
69
+ {
70
+ "epoch": 0.4139715394566624,
71
+ "grad_norm": 6.75,
72
+ "learning_rate": 9.793355818615905e-06,
73
+ "loss": 1.3408810424804687,
74
+ "step": 400
75
+ },
76
+ {
77
+ "epoch": 0.46571798188874514,
78
+ "grad_norm": 7.75,
79
+ "learning_rate": 9.7045004048705e-06,
80
+ "loss": 1.3330262756347657,
81
+ "step": 450
82
+ },
83
+ {
84
+ "epoch": 0.517464424320828,
85
+ "grad_norm": 10.0,
86
+ "learning_rate": 9.600355522127174e-06,
87
+ "loss": 1.3203207397460937,
88
+ "step": 500
89
+ },
90
+ {
91
+ "epoch": 0.517464424320828,
92
+ "eval_loss": 1.2533953189849854,
93
+ "eval_runtime": 178.6366,
94
+ "eval_samples_per_second": 55.98,
95
+ "eval_steps_per_second": 0.778,
96
+ "step": 500
97
+ },
98
+ {
99
+ "epoch": 0.5692108667529108,
100
+ "grad_norm": 5.9375,
101
+ "learning_rate": 9.481259637772251e-06,
102
+ "loss": 1.3150039672851563,
103
+ "step": 550
104
+ },
105
+ {
106
+ "epoch": 0.6209573091849935,
107
+ "grad_norm": 6.46875,
108
+ "learning_rate": 9.347599809447331e-06,
109
+ "loss": 1.3117788696289063,
110
+ "step": 600
111
+ },
112
+ {
113
+ "epoch": 0.6727037516170763,
114
+ "grad_norm": 6.9375,
115
+ "learning_rate": 9.199810427124904e-06,
116
+ "loss": 1.3062660217285156,
117
+ "step": 650
118
+ },
119
+ {
120
+ "epoch": 0.7244501940491591,
121
+ "grad_norm": 12.625,
122
+ "learning_rate": 9.03837180135546e-06,
123
+ "loss": 1.2984295654296876,
124
+ "step": 700
125
+ },
126
+ {
127
+ "epoch": 0.7761966364812419,
128
+ "grad_norm": 6.21875,
129
+ "learning_rate": 8.86380860227426e-06,
130
+ "loss": 1.287723846435547,
131
+ "step": 750
132
+ },
133
+ {
134
+ "epoch": 0.7761966364812419,
135
+ "eval_loss": 1.2287719249725342,
136
+ "eval_runtime": 177.5762,
137
+ "eval_samples_per_second": 56.314,
138
+ "eval_steps_per_second": 0.783,
139
+ "step": 750
140
+ },
141
+ {
142
+ "epoch": 0.8279430789133247,
143
+ "grad_norm": 8.3125,
144
+ "learning_rate": 8.676688154440947e-06,
145
+ "loss": 1.293363037109375,
146
+ "step": 800
147
+ },
148
+ {
149
+ "epoch": 0.8796895213454075,
150
+ "grad_norm": 7.90625,
151
+ "learning_rate": 8.477618593053693e-06,
152
+ "loss": 1.2859170532226563,
153
+ "step": 850
154
+ },
155
+ {
156
+ "epoch": 0.9314359637774903,
157
+ "grad_norm": 6.25,
158
+ "learning_rate": 8.26724688753015e-06,
159
+ "loss": 1.2893869018554687,
160
+ "step": 900
161
+ },
162
+ {
163
+ "epoch": 0.9831824062095731,
164
+ "grad_norm": 5.21875,
165
+ "learning_rate": 8.04625673887846e-06,
166
+ "loss": 1.2764347076416016,
167
+ "step": 950
168
+ },
169
+ {
170
+ "epoch": 1.0341526520051747,
171
+ "grad_norm": 4.90625,
172
+ "learning_rate": 7.815366357691863e-06,
173
+ "loss": 1.266595458984375,
174
+ "step": 1000
175
+ },
176
+ {
177
+ "epoch": 1.0341526520051747,
178
+ "eval_loss": 1.2155812978744507,
179
+ "eval_runtime": 177.1968,
180
+ "eval_samples_per_second": 56.434,
181
+ "eval_steps_per_second": 0.784,
182
+ "step": 1000
183
+ },
184
+ {
185
+ "epoch": 1.0858990944372575,
186
+ "grad_norm": 6.9375,
187
+ "learning_rate": 7.575326129988268e-06,
188
+ "loss": 1.2680277252197265,
189
+ "step": 1050
190
+ },
191
+ {
192
+ "epoch": 1.1376455368693401,
193
+ "grad_norm": 5.375,
194
+ "learning_rate": 7.326916178480796e-06,
195
+ "loss": 1.2629891204833985,
196
+ "step": 1100
197
+ },
198
+ {
199
+ "epoch": 1.189391979301423,
200
+ "grad_norm": 8.75,
201
+ "learning_rate": 7.070943827205048e-06,
202
+ "loss": 1.2643635559082032,
203
+ "step": 1150
204
+ },
205
+ {
206
+ "epoch": 1.2411384217335057,
207
+ "grad_norm": 11.9375,
208
+ "learning_rate": 6.808240977742987e-06,
209
+ "loss": 1.2576502227783204,
210
+ "step": 1200
211
+ },
212
+ {
213
+ "epoch": 1.2928848641655886,
214
+ "grad_norm": 5.25,
215
+ "learning_rate": 6.539661405570615e-06,
216
+ "loss": 1.2591765594482422,
217
+ "step": 1250
218
+ },
219
+ {
220
+ "epoch": 1.2928848641655886,
221
+ "eval_loss": 1.2061747312545776,
222
+ "eval_runtime": 179.247,
223
+ "eval_samples_per_second": 55.789,
224
+ "eval_steps_per_second": 0.775,
225
+ "step": 1250
226
+ },
227
+ {
228
+ "epoch": 1.3446313065976714,
229
+ "grad_norm": 5.03125,
230
+ "learning_rate": 6.266077985316224e-06,
231
+ "loss": 1.258944091796875,
232
+ "step": 1300
233
+ },
234
+ {
235
+ "epoch": 1.3963777490297542,
236
+ "grad_norm": 7.90625,
237
+ "learning_rate": 5.988379853947046e-06,
238
+ "loss": 1.2634922790527343,
239
+ "step": 1350
240
+ },
241
+ {
242
+ "epoch": 1.448124191461837,
243
+ "grad_norm": 7.59375,
244
+ "learning_rate": 5.7074695211038385e-06,
245
+ "loss": 1.2592449188232422,
246
+ "step": 1400
247
+ },
248
+ {
249
+ "epoch": 1.4998706338939198,
250
+ "grad_norm": 5.90625,
251
+ "learning_rate": 5.424259935974736e-06,
252
+ "loss": 1.2600762939453125,
253
+ "step": 1450
254
+ },
255
+ {
256
+ "epoch": 1.5516170763260027,
257
+ "grad_norm": 5.28125,
258
+ "learning_rate": 5.139671520240915e-06,
259
+ "loss": 1.2590911102294922,
260
+ "step": 1500
261
+ },
262
+ {
263
+ "epoch": 1.5516170763260027,
264
+ "eval_loss": 1.201247215270996,
265
+ "eval_runtime": 179.3983,
266
+ "eval_samples_per_second": 55.742,
267
+ "eval_steps_per_second": 0.775,
268
+ "step": 1500
269
+ },
270
+ {
271
+ "epoch": 1.6033635187580852,
272
+ "grad_norm": 5.875,
273
+ "learning_rate": 4.854629176736911e-06,
274
+ "loss": 1.25985595703125,
275
+ "step": 1550
276
+ },
277
+ {
278
+ "epoch": 1.6551099611901683,
279
+ "grad_norm": 6.09375,
280
+ "learning_rate": 4.570059283547328e-06,
281
+ "loss": 1.2547737121582032,
282
+ "step": 1600
283
+ },
284
+ {
285
+ "epoch": 1.7068564036222509,
286
+ "grad_norm": 6.15625,
287
+ "learning_rate": 4.286886683309053e-06,
288
+ "loss": 1.2532640075683594,
289
+ "step": 1650
290
+ },
291
+ {
292
+ "epoch": 1.758602846054334,
293
+ "grad_norm": 7.0625,
294
+ "learning_rate": 4.0060316775036025e-06,
295
+ "loss": 1.2515839385986327,
296
+ "step": 1700
297
+ },
298
+ {
299
+ "epoch": 1.8103492884864165,
300
+ "grad_norm": 6.8125,
301
+ "learning_rate": 3.728407035508089e-06,
302
+ "loss": 1.2524348449707032,
303
+ "step": 1750
304
+ },
305
+ {
306
+ "epoch": 1.8103492884864165,
307
+ "eval_loss": 1.198207974433899,
308
+ "eval_runtime": 178.7391,
309
+ "eval_samples_per_second": 55.947,
310
+ "eval_steps_per_second": 0.778,
311
+ "step": 1750
312
+ },
313
+ {
314
+ "epoch": 1.8620957309184993,
315
+ "grad_norm": 6.84375,
316
+ "learning_rate": 3.4549150281252635e-06,
317
+ "loss": 1.2492080688476563,
318
+ "step": 1800
319
+ },
320
+ {
321
+ "epoch": 1.9138421733505822,
322
+ "grad_norm": 6.59375,
323
+ "learning_rate": 3.1864444952335823e-06,
324
+ "loss": 1.2510016632080079,
325
+ "step": 1850
326
+ },
327
+ {
328
+ "epoch": 1.965588615782665,
329
+ "grad_norm": 9.6875,
330
+ "learning_rate": 2.9238679570873217e-06,
331
+ "loss": 1.2530317687988282,
332
+ "step": 1900
333
+ },
334
+ {
335
+ "epoch": 2.0165588615782664,
336
+ "grad_norm": 5.3125,
337
+ "learning_rate": 2.66803877865493e-06,
338
+ "loss": 1.254935760498047,
339
+ "step": 1950
340
+ },
341
+ {
342
+ "epoch": 2.0683053040103494,
343
+ "grad_norm": 6.875,
344
+ "learning_rate": 2.4197883962114164e-06,
345
+ "loss": 1.2447527313232423,
346
+ "step": 2000
347
+ },
348
+ {
349
+ "epoch": 2.0683053040103494,
350
+ "eval_loss": 1.1967895030975342,
351
+ "eval_runtime": 179.8167,
352
+ "eval_samples_per_second": 55.612,
353
+ "eval_steps_per_second": 0.773,
354
+ "step": 2000
355
+ },
356
+ {
357
+ "epoch": 2.120051746442432,
358
+ "grad_norm": 5.40625,
359
+ "learning_rate": 2.1799236151982623e-06,
360
+ "loss": 1.2455268859863282,
361
+ "step": 2050
362
+ },
363
+ {
364
+ "epoch": 2.171798188874515,
365
+ "grad_norm": 5.21875,
366
+ "learning_rate": 1.9492239881327103e-06,
367
+ "loss": 1.2404389953613282,
368
+ "step": 2100
369
+ },
370
+ {
371
+ "epoch": 2.2235446313065976,
372
+ "grad_norm": 7.0,
373
+ "learning_rate": 1.7284392810881568e-06,
374
+ "loss": 1.2505016326904297,
375
+ "step": 2150
376
+ },
377
+ {
378
+ "epoch": 2.2752910737386802,
379
+ "grad_norm": 6.625,
380
+ "learning_rate": 1.5182870369794838e-06,
381
+ "loss": 1.2448814392089844,
382
+ "step": 2200
383
+ },
384
+ {
385
+ "epoch": 2.3270375161707633,
386
+ "grad_norm": 6.21875,
387
+ "learning_rate": 1.3194502435725787e-06,
388
+ "loss": 1.2437591552734375,
389
+ "step": 2250
390
+ },
391
+ {
392
+ "epoch": 2.3270375161707633,
393
+ "eval_loss": 1.1964588165283203,
394
+ "eval_runtime": 177.4301,
395
+ "eval_samples_per_second": 56.36,
396
+ "eval_steps_per_second": 0.783,
397
+ "step": 2250
398
+ },
399
+ {
400
+ "epoch": 2.378783958602846,
401
+ "grad_norm": 5.4375,
402
+ "learning_rate": 1.1325751137969381e-06,
403
+ "loss": 1.2385246276855468,
404
+ "step": 2300
405
+ },
406
+ {
407
+ "epoch": 2.430530401034929,
408
+ "grad_norm": 5.3125,
409
+ "learning_rate": 9.582689855752515e-07,
410
+ "loss": 1.2423560333251953,
411
+ "step": 2350
412
+ },
413
+ {
414
+ "epoch": 2.4822768434670115,
415
+ "grad_norm": 6.0625,
416
+ "learning_rate": 7.970983479954631e-07,
417
+ "loss": 1.2424100494384767,
418
+ "step": 2400
419
+ },
420
+ {
421
+ "epoch": 2.5340232858990945,
422
+ "grad_norm": 7.125,
423
+ "learning_rate": 6.495870002401822e-07,
424
+ "loss": 1.2452584838867187,
425
+ "step": 2450
426
+ },
427
+ {
428
+ "epoch": 2.585769728331177,
429
+ "grad_norm": 6.25,
430
+ "learning_rate": 5.162143492568511e-07,
431
+ "loss": 1.2483766174316406,
432
+ "step": 2500
433
+ },
434
+ {
435
+ "epoch": 2.585769728331177,
436
+ "eval_loss": 1.196244478225708,
437
+ "eval_runtime": 177.9314,
438
+ "eval_samples_per_second": 56.201,
439
+ "eval_steps_per_second": 0.781,
440
+ "step": 2500
441
+ },
442
+ {
443
+ "epoch": 2.63751617076326,
444
+ "grad_norm": 7.0625,
445
+ "learning_rate": 3.9741385170120273e-07,
446
+ "loss": 1.2461642456054687,
447
+ "step": 2550
448
+ },
449
+ {
450
+ "epoch": 2.6892626131953428,
451
+ "grad_norm": 5.5,
452
+ "learning_rate": 2.935716052176163e-07,
453
+ "loss": 1.2479273223876952,
454
+ "step": 2600
455
+ },
456
+ {
457
+ "epoch": 2.7410090556274254,
458
+ "grad_norm": 6.53125,
459
+ "learning_rate": 2.050250936346937e-07,
460
+ "loss": 1.2457350158691407,
461
+ "step": 2650
462
+ },
463
+ {
464
+ "epoch": 2.7927554980595084,
465
+ "grad_norm": 4.96875,
466
+ "learning_rate": 1.3206209015409532e-07,
467
+ "loss": 1.2492985534667969,
468
+ "step": 2700
469
+ },
470
+ {
471
+ "epoch": 2.8445019404915914,
472
+ "grad_norm": 5.03125,
473
+ "learning_rate": 7.491972209725807e-08,
474
+ "loss": 1.2479338836669922,
475
+ "step": 2750
476
+ },
477
+ {
478
+ "epoch": 2.8445019404915914,
479
+ "eval_loss": 1.196183681488037,
480
+ "eval_runtime": 177.5352,
481
+ "eval_samples_per_second": 56.327,
482
+ "eval_steps_per_second": 0.783,
483
+ "step": 2750
484
+ },
485
+ {
486
+ "epoch": 2.896248382923674,
487
+ "grad_norm": 5.71875,
488
+ "learning_rate": 3.3783700249533925e-08,
489
+ "loss": 1.2453683471679688,
490
+ "step": 2800
491
+ },
492
+ {
493
+ "epoch": 2.9479948253557566,
494
+ "grad_norm": 9.3125,
495
+ "learning_rate": 8.787715306344258e-09,
496
+ "loss": 1.2509371185302733,
497
+ "step": 2850
498
+ },
499
+ {
500
+ "epoch": 2.9997412677878397,
501
+ "grad_norm": 4.875,
502
+ "learning_rate": 1.3003382891563221e-11,
503
+ "loss": 1.2454129791259765,
504
+ "step": 2900
505
+ }
506
+ ],
507
+ "logging_steps": 50,
508
+ "max_steps": 2901,
509
+ "num_input_tokens_seen": 0,
510
+ "num_train_epochs": 3,
511
+ "save_steps": 1000,
512
+ "stateful_callbacks": {
513
+ "TrainerControl": {
514
+ "args": {
515
+ "should_epoch_stop": false,
516
+ "should_evaluate": false,
517
+ "should_log": false,
518
+ "should_save": true,
519
+ "should_training_stop": true
520
+ },
521
+ "attributes": {}
522
+ }
523
+ },
524
+ "total_flos": 1.523115236024071e+19,
525
+ "train_batch_size": 18,
526
+ "trial_name": null,
527
+ "trial_params": null
528
+ }
checkpoint-2901/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25f31ca829e8d2c0fd93d4b4b550d0bb0e9c801e4b84c4542753a7beaf8dcf8c
3
+ size 5201
config.json ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3VLForConditionalGeneration"
4
+ ],
5
+ "bos_token_id": null,
6
+ "dtype": "bfloat16",
7
+ "eos_token_id": 151645,
8
+ "image_token_id": 151655,
9
+ "model_type": "qwen3_vl",
10
+ "pad_token_id": 151643,
11
+ "text_config": {
12
+ "attention_bias": false,
13
+ "attention_dropout": 0.0,
14
+ "bos_token_id": 151643,
15
+ "dtype": "bfloat16",
16
+ "eos_token_id": 151645,
17
+ "head_dim": 128,
18
+ "hidden_act": "silu",
19
+ "hidden_size": 2560,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 9728,
22
+ "max_position_embeddings": 262144,
23
+ "model_type": "qwen3_vl_text",
24
+ "num_attention_heads": 32,
25
+ "num_hidden_layers": 36,
26
+ "num_key_value_heads": 8,
27
+ "pad_token_id": null,
28
+ "rms_norm_eps": 1e-06,
29
+ "rope_parameters": {
30
+ "mrope_interleaved": true,
31
+ "mrope_section": [
32
+ 24,
33
+ 20,
34
+ 20
35
+ ],
36
+ "rope_theta": 5000000,
37
+ "rope_type": "default"
38
+ },
39
+ "tie_word_embeddings": true,
40
+ "use_cache": true,
41
+ "vocab_size": 151936
42
+ },
43
+ "tie_word_embeddings": true,
44
+ "transformers_version": "5.0.0",
45
+ "use_cache": false,
46
+ "video_token_id": 151656,
47
+ "vision_config": {
48
+ "deepstack_visual_indexes": [
49
+ 5,
50
+ 11,
51
+ 17
52
+ ],
53
+ "depth": 24,
54
+ "dtype": "bfloat16",
55
+ "hidden_act": "gelu_pytorch_tanh",
56
+ "hidden_size": 1024,
57
+ "in_channels": 3,
58
+ "initializer_range": 0.02,
59
+ "intermediate_size": 4096,
60
+ "model_type": "qwen3_vl",
61
+ "num_heads": 16,
62
+ "num_position_embeddings": 2304,
63
+ "out_hidden_size": 2560,
64
+ "patch_size": 16,
65
+ "spatial_merge_size": 2,
66
+ "temporal_patch_size": 2
67
+ },
68
+ "vision_end_token_id": 151653,
69
+ "vision_start_token_id": 151652
70
+ }
generation_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_sample": true,
3
+ "eos_token_id": [
4
+ 151645,
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.0,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "5.0.0"
14
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:01454076c1f037c770800517410587300c263b5b3fbe6f382e3304592cfaf04a
3
+ size 8875719408
processor_config.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "data_format": "channels_first",
4
+ "do_convert_rgb": true,
5
+ "do_normalize": true,
6
+ "do_rescale": true,
7
+ "do_resize": true,
8
+ "image_mean": [
9
+ 0.5,
10
+ 0.5,
11
+ 0.5
12
+ ],
13
+ "image_processor_type": "Qwen2VLImageProcessorFast",
14
+ "image_std": [
15
+ 0.5,
16
+ 0.5,
17
+ 0.5
18
+ ],
19
+ "merge_size": 2,
20
+ "patch_size": 16,
21
+ "resample": 3,
22
+ "rescale_factor": 0.00392156862745098,
23
+ "size": {
24
+ "longest_edge": 16777216,
25
+ "shortest_edge": 65536
26
+ },
27
+ "temporal_patch_size": 2
28
+ },
29
+ "processor_class": "Qwen3VLProcessor",
30
+ "video_processor": {
31
+ "data_format": "channels_first",
32
+ "default_to_square": true,
33
+ "do_convert_rgb": true,
34
+ "do_normalize": true,
35
+ "do_rescale": true,
36
+ "do_resize": true,
37
+ "do_sample_frames": true,
38
+ "fps": 2,
39
+ "image_mean": [
40
+ 0.5,
41
+ 0.5,
42
+ 0.5
43
+ ],
44
+ "image_std": [
45
+ 0.5,
46
+ 0.5,
47
+ 0.5
48
+ ],
49
+ "max_frames": 768,
50
+ "merge_size": 2,
51
+ "min_frames": 4,
52
+ "patch_size": 16,
53
+ "resample": 3,
54
+ "rescale_factor": 0.00392156862745098,
55
+ "return_metadata": false,
56
+ "size": {
57
+ "longest_edge": 25165824,
58
+ "shortest_edge": 4096
59
+ },
60
+ "temporal_patch_size": 2,
61
+ "video_processor_type": "Qwen3VLVideoProcessor"
62
+ }
63
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be75606093db2094d7cd20f3c2f385c212750648bd6ea4fb2bf507a6a4c55506
3
+ size 11422650
tokenizer_config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": null,
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|im_end|>",
7
+ "errors": "replace",
8
+ "extra_special_tokens": [
9
+ "<|im_start|>",
10
+ "<|im_end|>",
11
+ "<|object_ref_start|>",
12
+ "<|object_ref_end|>",
13
+ "<|box_start|>",
14
+ "<|box_end|>",
15
+ "<|quad_start|>",
16
+ "<|quad_end|>",
17
+ "<|vision_start|>",
18
+ "<|vision_end|>",
19
+ "<|vision_pad|>",
20
+ "<|image_pad|>",
21
+ "<|video_pad|>"
22
+ ],
23
+ "is_local": false,
24
+ "model_max_length": 262144,
25
+ "pad_token": "<|endoftext|>",
26
+ "processor_class": "Qwen3VLProcessor",
27
+ "split_special_tokens": false,
28
+ "tokenizer_class": "Qwen2Tokenizer",
29
+ "unk_token": null
30
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:25f31ca829e8d2c0fd93d4b4b550d0bb0e9c801e4b84c4542753a7beaf8dcf8c
3
+ size 5201