Hugging Face's logo

acyildirimer
/
Qwen3.6-27B-int4-AutoRound

Safetensors
qwen3_5
4-bit precision
auto-round
Model card Files
xet
 Community
1
acyildirimer commited on
Commit
4c99ee6
·
0 Parent(s):

Update model

Browse files
Files changed (17) hide show
  1. .gitattributes +36 -0
  2. README.md +75 -0
  3. chat_template.jinja +154 -0
  4. config.json +547 -0
  5. generation_config.json +13 -0
  6. model-00001-of-00005.safetensors +3 -0
  7. model-00002-of-00005.safetensors +3 -0
  8. model-00003-of-00005.safetensors +3 -0
  9. model-00004-of-00005.safetensors +3 -0
  10. model-00005-of-00005.safetensors +3 -0
  11. model.safetensors.index.json +0 -0
  12. model_extra_tensors.safetensors +3 -0
  13. preprocessor_config.json +26 -0
  14. processor_config.json +60 -0
  15. quantization_config.json +398 -0
  16. tokenizer.json +3 -0
  17. tokenizer_config.json +33 -0
.gitattributes ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.bin filter=lfs diff=lfs merge=lfs -text
4
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
5
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
6
+ *.ftz filter=lfs diff=lfs merge=lfs -text
7
+ *.gz filter=lfs diff=lfs merge=lfs -text
8
+ *.h5 filter=lfs diff=lfs merge=lfs -text
9
+ *.joblib filter=lfs diff=lfs merge=lfs -text
10
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
+ *.mlmodel filter=lfs diff=lfs merge=lfs -text
12
+ *.model filter=lfs diff=lfs merge=lfs -text
13
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
14
+ *.npy filter=lfs diff=lfs merge=lfs -text
15
+ *.npz filter=lfs diff=lfs merge=lfs -text
16
+ *.onnx filter=lfs diff=lfs merge=lfs -text
17
+ *.ot filter=lfs diff=lfs merge=lfs -text
18
+ *.parquet filter=lfs diff=lfs merge=lfs -text
19
+ *.pb filter=lfs diff=lfs merge=lfs -text
20
+ *.pickle filter=lfs diff=lfs merge=lfs -text
21
+ *.pkl filter=lfs diff=lfs merge=lfs -text
22
+ *.pt filter=lfs diff=lfs merge=lfs -text
23
+ *.pth filter=lfs diff=lfs merge=lfs -text
24
+ *.rar filter=lfs diff=lfs merge=lfs -text
25
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
26
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
28
+ *.tar filter=lfs diff=lfs merge=lfs -text
29
+ *.tflite filter=lfs diff=lfs merge=lfs -text
30
+ *.tgz filter=lfs diff=lfs merge=lfs -text
31
+ *.wasm filter=lfs diff=lfs merge=lfs -text
32
+ *.xz 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
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,75 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model:
4
+ - Qwen/Qwen3.6-27B
5
+ ---
6
+ ## Model details
7
+
8
+ This model is an INT4 quantized version of the original **Qwen3.6-27B** model.
9
+
10
+ Quantization was performed with **AutoRound** using a calibration dataset derived from Pile-10k. The calibration set consists of 256 randomly selected samples from diverse categories, with sequence length 2048 and 300 optimization iterations.
11
+
12
+ The launch parameters below were tested on a headless RTX 3090 using the latest vLLM nightly build. MTP and prefix caching are confirmed to work.
13
+
14
+ ## vLLM launch example
15
+
16
+ ```bash
17
+ vllm serve acyildirimer/Qwen3.6-27B-int4-AutoRound \
18
+ --host 0.0.0.0 \
19
+ --port 8000 \
20
+ --dtype auto \
21
+ --kv-cache-dtype auto \
22
+ --max-model-len auto \
23
+ --reasoning-parser qwen3 \
24
+ --enable-auto-tool-choice \
25
+ --tool-call-parser qwen3_coder \
26
+ --gpu-memory-utilization 0.96 \
27
+ --enable-prefix-caching \
28
+ --max-num-seqs 1 \
29
+ --served-model-name Qwen3.6-27B \
30
+ --language-model-only \
31
+ --performance-mode interactivity \
32
+ --attention-backend auto \
33
+ --max-num-batched-tokens 4096 \
34
+ --generation-config auto \
35
+ --override-generation-config '{"temperature":0.6,"top_p":0.95,"top_k":20,"min_p":0.0,"repetition_penalty":1.0}' \
36
+ --speculative-config '{"method": "mtp", "num_speculative_tokens": 2}'
37
+ ```
38
+
39
+ ## Compatibility
40
+
41
+ - The launch command was tested with the latest vLLM nightly build.
42
+ - It should also work with the stable vLLM release, provided that the selected KV cache type is supported. If needed, use a supported `--kv-cache-dtype` value or set it to `auto`.
43
+ - `turboquant` is not currently working on Ampere GPUs, so it is not enabled in the launch example above.
44
+
45
+ ## Troubleshooting
46
+
47
+ - If you encounter out-of-memory errors, reduce `--gpu-memory-utilization`. For example, try `0.94`, `0.92`, or lower depending on your available VRAM and workload.
48
+ - In some runs, vLLM may incorrectly calculate `Available KV cache memory` during startup. If this happens, terminate the running instance and launch it again with the same parameters.
49
+
50
+ ## Citation
51
+
52
+ Please cite both the original Qwen3.6-27B model and the AutoRound quantization method when using this quantized model.
53
+
54
+ ### Base model
55
+
56
+ ```bibtex
57
+ @misc{qwen36-27b,
58
+ title = {{Qwen3.6-27B}: Flagship-Level Coding in a {27B} Dense Model},
59
+ author = {{Qwen Team}},
60
+ year = {2026},
61
+ month = {April},
62
+ url = {https://qwen.ai/blog?id=qwen3.6-27b}
63
+ }
64
+ ```
65
+
66
+ ### Quantization method
67
+
68
+ ```bibtex
69
+ @article{cheng2023optimize,
70
+ title = {Optimize Weight Rounding via Signed Gradient Descent for the Quantization of LLMs},
71
+ author = {Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao and Liu, Yi},
72
+ journal = {arXiv preprint arXiv:2309.05516},
73
+ year = {2023}
74
+ }
75
+ ```
chat_template.jinja ADDED
@@ -0,0 +1,154 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- set image_count = namespace(value=0) %}
2
+ {%- set video_count = namespace(value=0) %}
3
+ {%- macro render_content(content, do_vision_count, is_system_content=false) %}
4
+ {%- if content is string %}
5
+ {{- content }}
6
+ {%- elif content is iterable and content is not mapping %}
7
+ {%- for item in content %}
8
+ {%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
9
+ {%- if is_system_content %}
10
+ {{- raise_exception('System message cannot contain images.') }}
11
+ {%- endif %}
12
+ {%- if do_vision_count %}
13
+ {%- set image_count.value = image_count.value + 1 %}
14
+ {%- endif %}
15
+ {%- if add_vision_id %}
16
+ {{- 'Picture ' ~ image_count.value ~ ': ' }}
17
+ {%- endif %}
18
+ {{- '<|vision_start|><|image_pad|><|vision_end|>' }}
19
+ {%- elif 'video' in item or item.type == 'video' %}
20
+ {%- if is_system_content %}
21
+ {{- raise_exception('System message cannot contain videos.') }}
22
+ {%- endif %}
23
+ {%- if do_vision_count %}
24
+ {%- set video_count.value = video_count.value + 1 %}
25
+ {%- endif %}
26
+ {%- if add_vision_id %}
27
+ {{- 'Video ' ~ video_count.value ~ ': ' }}
28
+ {%- endif %}
29
+ {{- '<|vision_start|><|video_pad|><|vision_end|>' }}
30
+ {%- elif 'text' in item %}
31
+ {{- item.text }}
32
+ {%- else %}
33
+ {{- raise_exception('Unexpected item type in content.') }}
34
+ {%- endif %}
35
+ {%- endfor %}
36
+ {%- elif content is none or content is undefined %}
37
+ {{- '' }}
38
+ {%- else %}
39
+ {{- raise_exception('Unexpected content type.') }}
40
+ {%- endif %}
41
+ {%- endmacro %}
42
+ {%- if not messages %}
43
+ {{- raise_exception('No messages provided.') }}
44
+ {%- endif %}
45
+ {%- if tools and tools is iterable and tools is not mapping %}
46
+ {{- '<|im_start|>system\n' }}
47
+ {{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
48
+ {%- for tool in tools %}
49
+ {{- "\n" }}
50
+ {{- tool | tojson }}
51
+ {%- endfor %}
52
+ {{- "\n</tools>" }}
53
+ {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
54
+ {%- if messages[0].role == 'system' %}
55
+ {%- set content = render_content(messages[0].content, false, true)|trim %}
56
+ {%- if content %}
57
+ {{- '\n\n' + content }}
58
+ {%- endif %}
59
+ {%- endif %}
60
+ {{- '<|im_end|>\n' }}
61
+ {%- else %}
62
+ {%- if messages[0].role == 'system' %}
63
+ {%- set content = render_content(messages[0].content, false, true)|trim %}
64
+ {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
65
+ {%- endif %}
66
+ {%- endif %}
67
+ {%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
68
+ {%- for message in messages[::-1] %}
69
+ {%- set index = (messages|length - 1) - loop.index0 %}
70
+ {%- if ns.multi_step_tool and message.role == "user" %}
71
+ {%- set content = render_content(message.content, false)|trim %}
72
+ {%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
73
+ {%- set ns.multi_step_tool = false %}
74
+ {%- set ns.last_query_index = index %}
75
+ {%- endif %}
76
+ {%- endif %}
77
+ {%- endfor %}
78
+ {%- if ns.multi_step_tool %}
79
+ {{- raise_exception('No user query found in messages.') }}
80
+ {%- endif %}
81
+ {%- for message in messages %}
82
+ {%- set content = render_content(message.content, true)|trim %}
83
+ {%- if message.role == "system" %}
84
+ {%- if not loop.first %}
85
+ {{- raise_exception('System message must be at the beginning.') }}
86
+ {%- endif %}
87
+ {%- elif message.role == "user" %}
88
+ {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
89
+ {%- elif message.role == "assistant" %}
90
+ {%- set reasoning_content = '' %}
91
+ {%- if message.reasoning_content is string %}
92
+ {%- set reasoning_content = message.reasoning_content %}
93
+ {%- else %}
94
+ {%- if '</think>' in content %}
95
+ {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
96
+ {%- set content = content.split('</think>')[-1].lstrip('\n') %}
97
+ {%- endif %}
98
+ {%- endif %}
99
+ {%- set reasoning_content = reasoning_content|trim %}
100
+ {%- if (preserve_thinking is defined and preserve_thinking is true) or (loop.index0 > ns.last_query_index) %}
101
+ {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
102
+ {%- else %}
103
+ {{- '<|im_start|>' + message.role + '\n' + content }}
104
+ {%- endif %}
105
+ {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
106
+ {%- for tool_call in message.tool_calls %}
107
+ {%- if tool_call.function is defined %}
108
+ {%- set tool_call = tool_call.function %}
109
+ {%- endif %}
110
+ {%- if loop.first %}
111
+ {%- if content|trim %}
112
+ {{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
113
+ {%- else %}
114
+ {{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
115
+ {%- endif %}
116
+ {%- else %}
117
+ {{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
118
+ {%- endif %}
119
+ {%- if tool_call.arguments is defined %}
120
+ {%- for args_name, args_value in tool_call.arguments|items %}
121
+ {{- '<parameter=' + args_name + '>\n' }}
122
+ {%- set args_value = args_value | string if args_value is string else args_value | tojson | safe %}
123
+ {{- args_value }}
124
+ {{- '\n</parameter>\n' }}
125
+ {%- endfor %}
126
+ {%- endif %}
127
+ {{- '</function>\n</tool_call>' }}
128
+ {%- endfor %}
129
+ {%- endif %}
130
+ {{- '<|im_end|>\n' }}
131
+ {%- elif message.role == "tool" %}
132
+ {%- if loop.previtem and loop.previtem.role != "tool" %}
133
+ {{- '<|im_start|>user' }}
134
+ {%- endif %}
135
+ {{- '\n<tool_response>\n' }}
136
+ {{- content }}
137
+ {{- '\n</tool_response>' }}
138
+ {%- if not loop.last and loop.nextitem.role != "tool" %}
139
+ {{- '<|im_end|>\n' }}
140
+ {%- elif loop.last %}
141
+ {{- '<|im_end|>\n' }}
142
+ {%- endif %}
143
+ {%- else %}
144
+ {{- raise_exception('Unexpected message role.') }}
145
+ {%- endif %}
146
+ {%- endfor %}
147
+ {%- if add_generation_prompt %}
148
+ {{- '<|im_start|>assistant\n' }}
149
+ {%- if enable_thinking is defined and enable_thinking is false %}
150
+ {{- '<think>\n\n</think>\n\n' }}
151
+ {%- else %}
152
+ {{- '<think>\n' }}
153
+ {%- endif %}
154
+ {%- endif %}
config.json ADDED
@@ -0,0 +1,547 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3_5ForConditionalGeneration"
4
+ ],
5
+ "dtype": "bfloat16",
6
+ "image_token_id": 248056,
7
+ "language_model_only": false,
8
+ "model_type": "qwen3_5",
9
+ "quantization_config": {
10
+ "autoround_version": "0.13.0",
11
+ "bits": 4,
12
+ "block_name_to_quantize": [
13
+ "model.language_model.layers",
14
+ "mtp.layers"
15
+ ],
16
+ "data_type": "int",
17
+ "extra_config": {
18
+ "model.language_model.layers.0.linear_attn.in_proj_a": {
19
+ "bits": 16,
20
+ "data_type": "fp"
21
+ },
22
+ "model.language_model.layers.0.linear_attn.in_proj_b": {
23
+ "bits": 16,
24
+ "data_type": "fp"
25
+ },
26
+ "model.language_model.layers.1.linear_attn.in_proj_a": {
27
+ "bits": 16,
28
+ "data_type": "fp"
29
+ },
30
+ "model.language_model.layers.1.linear_attn.in_proj_b": {
31
+ "bits": 16,
32
+ "data_type": "fp"
33
+ },
34
+ "model.language_model.layers.10.linear_attn.in_proj_a": {
35
+ "bits": 16,
36
+ "data_type": "fp"
37
+ },
38
+ "model.language_model.layers.10.linear_attn.in_proj_b": {
39
+ "bits": 16,
40
+ "data_type": "fp"
41
+ },
42
+ "model.language_model.layers.12.linear_attn.in_proj_a": {
43
+ "bits": 16,
44
+ "data_type": "fp"
45
+ },
46
+ "model.language_model.layers.12.linear_attn.in_proj_b": {
47
+ "bits": 16,
48
+ "data_type": "fp"
49
+ },
50
+ "model.language_model.layers.13.linear_attn.in_proj_a": {
51
+ "bits": 16,
52
+ "data_type": "fp"
53
+ },
54
+ "model.language_model.layers.13.linear_attn.in_proj_b": {
55
+ "bits": 16,
56
+ "data_type": "fp"
57
+ },
58
+ "model.language_model.layers.14.linear_attn.in_proj_a": {
59
+ "bits": 16,
60
+ "data_type": "fp"
61
+ },
62
+ "model.language_model.layers.14.linear_attn.in_proj_b": {
63
+ "bits": 16,
64
+ "data_type": "fp"
65
+ },
66
+ "model.language_model.layers.16.linear_attn.in_proj_a": {
67
+ "bits": 16,
68
+ "data_type": "fp"
69
+ },
70
+ "model.language_model.layers.16.linear_attn.in_proj_b": {
71
+ "bits": 16,
72
+ "data_type": "fp"
73
+ },
74
+ "model.language_model.layers.17.linear_attn.in_proj_a": {
75
+ "bits": 16,
76
+ "data_type": "fp"
77
+ },
78
+ "model.language_model.layers.17.linear_attn.in_proj_b": {
79
+ "bits": 16,
80
+ "data_type": "fp"
81
+ },
82
+ "model.language_model.layers.18.linear_attn.in_proj_a": {
83
+ "bits": 16,
84
+ "data_type": "fp"
85
+ },
86
+ "model.language_model.layers.18.linear_attn.in_proj_b": {
87
+ "bits": 16,
88
+ "data_type": "fp"
89
+ },
90
+ "model.language_model.layers.2.linear_attn.in_proj_a": {
91
+ "bits": 16,
92
+ "data_type": "fp"
93
+ },
94
+ "model.language_model.layers.2.linear_attn.in_proj_b": {
95
+ "bits": 16,
96
+ "data_type": "fp"
97
+ },
98
+ "model.language_model.layers.20.linear_attn.in_proj_a": {
99
+ "bits": 16,
100
+ "data_type": "fp"
101
+ },
102
+ "model.language_model.layers.20.linear_attn.in_proj_b": {
103
+ "bits": 16,
104
+ "data_type": "fp"
105
+ },
106
+ "model.language_model.layers.21.linear_attn.in_proj_a": {
107
+ "bits": 16,
108
+ "data_type": "fp"
109
+ },
110
+ "model.language_model.layers.21.linear_attn.in_proj_b": {
111
+ "bits": 16,
112
+ "data_type": "fp"
113
+ },
114
+ "model.language_model.layers.22.linear_attn.in_proj_a": {
115
+ "bits": 16,
116
+ "data_type": "fp"
117
+ },
118
+ "model.language_model.layers.22.linear_attn.in_proj_b": {
119
+ "bits": 16,
120
+ "data_type": "fp"
121
+ },
122
+ "model.language_model.layers.24.linear_attn.in_proj_a": {
123
+ "bits": 16,
124
+ "data_type": "fp"
125
+ },
126
+ "model.language_model.layers.24.linear_attn.in_proj_b": {
127
+ "bits": 16,
128
+ "data_type": "fp"
129
+ },
130
+ "model.language_model.layers.25.linear_attn.in_proj_a": {
131
+ "bits": 16,
132
+ "data_type": "fp"
133
+ },
134
+ "model.language_model.layers.25.linear_attn.in_proj_b": {
135
+ "bits": 16,
136
+ "data_type": "fp"
137
+ },
138
+ "model.language_model.layers.26.linear_attn.in_proj_a": {
139
+ "bits": 16,
140
+ "data_type": "fp"
141
+ },
142
+ "model.language_model.layers.26.linear_attn.in_proj_b": {
143
+ "bits": 16,
144
+ "data_type": "fp"
145
+ },
146
+ "model.language_model.layers.28.linear_attn.in_proj_a": {
147
+ "bits": 16,
148
+ "data_type": "fp"
149
+ },
150
+ "model.language_model.layers.28.linear_attn.in_proj_b": {
151
+ "bits": 16,
152
+ "data_type": "fp"
153
+ },
154
+ "model.language_model.layers.29.linear_attn.in_proj_a": {
155
+ "bits": 16,
156
+ "data_type": "fp"
157
+ },
158
+ "model.language_model.layers.29.linear_attn.in_proj_b": {
159
+ "bits": 16,
160
+ "data_type": "fp"
161
+ },
162
+ "model.language_model.layers.30.linear_attn.in_proj_a": {
163
+ "bits": 16,
164
+ "data_type": "fp"
165
+ },
166
+ "model.language_model.layers.30.linear_attn.in_proj_b": {
167
+ "bits": 16,
168
+ "data_type": "fp"
169
+ },
170
+ "model.language_model.layers.32.linear_attn.in_proj_a": {
171
+ "bits": 16,
172
+ "data_type": "fp"
173
+ },
174
+ "model.language_model.layers.32.linear_attn.in_proj_b": {
175
+ "bits": 16,
176
+ "data_type": "fp"
177
+ },
178
+ "model.language_model.layers.33.linear_attn.in_proj_a": {
179
+ "bits": 16,
180
+ "data_type": "fp"
181
+ },
182
+ "model.language_model.layers.33.linear_attn.in_proj_b": {
183
+ "bits": 16,
184
+ "data_type": "fp"
185
+ },
186
+ "model.language_model.layers.34.linear_attn.in_proj_a": {
187
+ "bits": 16,
188
+ "data_type": "fp"
189
+ },
190
+ "model.language_model.layers.34.linear_attn.in_proj_b": {
191
+ "bits": 16,
192
+ "data_type": "fp"
193
+ },
194
+ "model.language_model.layers.36.linear_attn.in_proj_a": {
195
+ "bits": 16,
196
+ "data_type": "fp"
197
+ },
198
+ "model.language_model.layers.36.linear_attn.in_proj_b": {
199
+ "bits": 16,
200
+ "data_type": "fp"
201
+ },
202
+ "model.language_model.layers.37.linear_attn.in_proj_a": {
203
+ "bits": 16,
204
+ "data_type": "fp"
205
+ },
206
+ "model.language_model.layers.37.linear_attn.in_proj_b": {
207
+ "bits": 16,
208
+ "data_type": "fp"
209
+ },
210
+ "model.language_model.layers.38.linear_attn.in_proj_a": {
211
+ "bits": 16,
212
+ "data_type": "fp"
213
+ },
214
+ "model.language_model.layers.38.linear_attn.in_proj_b": {
215
+ "bits": 16,
216
+ "data_type": "fp"
217
+ },
218
+ "model.language_model.layers.4.linear_attn.in_proj_a": {
219
+ "bits": 16,
220
+ "data_type": "fp"
221
+ },
222
+ "model.language_model.layers.4.linear_attn.in_proj_b": {
223
+ "bits": 16,
224
+ "data_type": "fp"
225
+ },
226
+ "model.language_model.layers.40.linear_attn.in_proj_a": {
227
+ "bits": 16,
228
+ "data_type": "fp"
229
+ },
230
+ "model.language_model.layers.40.linear_attn.in_proj_b": {
231
+ "bits": 16,
232
+ "data_type": "fp"
233
+ },
234
+ "model.language_model.layers.41.linear_attn.in_proj_a": {
235
+ "bits": 16,
236
+ "data_type": "fp"
237
+ },
238
+ "model.language_model.layers.41.linear_attn.in_proj_b": {
239
+ "bits": 16,
240
+ "data_type": "fp"
241
+ },
242
+ "model.language_model.layers.42.linear_attn.in_proj_a": {
243
+ "bits": 16,
244
+ "data_type": "fp"
245
+ },
246
+ "model.language_model.layers.42.linear_attn.in_proj_b": {
247
+ "bits": 16,
248
+ "data_type": "fp"
249
+ },
250
+ "model.language_model.layers.44.linear_attn.in_proj_a": {
251
+ "bits": 16,
252
+ "data_type": "fp"
253
+ },
254
+ "model.language_model.layers.44.linear_attn.in_proj_b": {
255
+ "bits": 16,
256
+ "data_type": "fp"
257
+ },
258
+ "model.language_model.layers.45.linear_attn.in_proj_a": {
259
+ "bits": 16,
260
+ "data_type": "fp"
261
+ },
262
+ "model.language_model.layers.45.linear_attn.in_proj_b": {
263
+ "bits": 16,
264
+ "data_type": "fp"
265
+ },
266
+ "model.language_model.layers.46.linear_attn.in_proj_a": {
267
+ "bits": 16,
268
+ "data_type": "fp"
269
+ },
270
+ "model.language_model.layers.46.linear_attn.in_proj_b": {
271
+ "bits": 16,
272
+ "data_type": "fp"
273
+ },
274
+ "model.language_model.layers.48.linear_attn.in_proj_a": {
275
+ "bits": 16,
276
+ "data_type": "fp"
277
+ },
278
+ "model.language_model.layers.48.linear_attn.in_proj_b": {
279
+ "bits": 16,
280
+ "data_type": "fp"
281
+ },
282
+ "model.language_model.layers.49.linear_attn.in_proj_a": {
283
+ "bits": 16,
284
+ "data_type": "fp"
285
+ },
286
+ "model.language_model.layers.49.linear_attn.in_proj_b": {
287
+ "bits": 16,
288
+ "data_type": "fp"
289
+ },
290
+ "model.language_model.layers.5.linear_attn.in_proj_a": {
291
+ "bits": 16,
292
+ "data_type": "fp"
293
+ },
294
+ "model.language_model.layers.5.linear_attn.in_proj_b": {
295
+ "bits": 16,
296
+ "data_type": "fp"
297
+ },
298
+ "model.language_model.layers.50.linear_attn.in_proj_a": {
299
+ "bits": 16,
300
+ "data_type": "fp"
301
+ },
302
+ "model.language_model.layers.50.linear_attn.in_proj_b": {
303
+ "bits": 16,
304
+ "data_type": "fp"
305
+ },
306
+ "model.language_model.layers.52.linear_attn.in_proj_a": {
307
+ "bits": 16,
308
+ "data_type": "fp"
309
+ },
310
+ "model.language_model.layers.52.linear_attn.in_proj_b": {
311
+ "bits": 16,
312
+ "data_type": "fp"
313
+ },
314
+ "model.language_model.layers.53.linear_attn.in_proj_a": {
315
+ "bits": 16,
316
+ "data_type": "fp"
317
+ },
318
+ "model.language_model.layers.53.linear_attn.in_proj_b": {
319
+ "bits": 16,
320
+ "data_type": "fp"
321
+ },
322
+ "model.language_model.layers.54.linear_attn.in_proj_a": {
323
+ "bits": 16,
324
+ "data_type": "fp"
325
+ },
326
+ "model.language_model.layers.54.linear_attn.in_proj_b": {
327
+ "bits": 16,
328
+ "data_type": "fp"
329
+ },
330
+ "model.language_model.layers.56.linear_attn.in_proj_a": {
331
+ "bits": 16,
332
+ "data_type": "fp"
333
+ },
334
+ "model.language_model.layers.56.linear_attn.in_proj_b": {
335
+ "bits": 16,
336
+ "data_type": "fp"
337
+ },
338
+ "model.language_model.layers.57.linear_attn.in_proj_a": {
339
+ "bits": 16,
340
+ "data_type": "fp"
341
+ },
342
+ "model.language_model.layers.57.linear_attn.in_proj_b": {
343
+ "bits": 16,
344
+ "data_type": "fp"
345
+ },
346
+ "model.language_model.layers.58.linear_attn.in_proj_a": {
347
+ "bits": 16,
348
+ "data_type": "fp"
349
+ },
350
+ "model.language_model.layers.58.linear_attn.in_proj_b": {
351
+ "bits": 16,
352
+ "data_type": "fp"
353
+ },
354
+ "model.language_model.layers.6.linear_attn.in_proj_a": {
355
+ "bits": 16,
356
+ "data_type": "fp"
357
+ },
358
+ "model.language_model.layers.6.linear_attn.in_proj_b": {
359
+ "bits": 16,
360
+ "data_type": "fp"
361
+ },
362
+ "model.language_model.layers.60.linear_attn.in_proj_a": {
363
+ "bits": 16,
364
+ "data_type": "fp"
365
+ },
366
+ "model.language_model.layers.60.linear_attn.in_proj_b": {
367
+ "bits": 16,
368
+ "data_type": "fp"
369
+ },
370
+ "model.language_model.layers.61.linear_attn.in_proj_a": {
371
+ "bits": 16,
372
+ "data_type": "fp"
373
+ },
374
+ "model.language_model.layers.61.linear_attn.in_proj_b": {
375
+ "bits": 16,
376
+ "data_type": "fp"
377
+ },
378
+ "model.language_model.layers.62.linear_attn.in_proj_a": {
379
+ "bits": 16,
380
+ "data_type": "fp"
381
+ },
382
+ "model.language_model.layers.62.linear_attn.in_proj_b": {
383
+ "bits": 16,
384
+ "data_type": "fp"
385
+ },
386
+ "model.language_model.layers.8.linear_attn.in_proj_a": {
387
+ "bits": 16,
388
+ "data_type": "fp"
389
+ },
390
+ "model.language_model.layers.8.linear_attn.in_proj_b": {
391
+ "bits": 16,
392
+ "data_type": "fp"
393
+ },
394
+ "model.language_model.layers.9.linear_attn.in_proj_a": {
395
+ "bits": 16,
396
+ "data_type": "fp"
397
+ },
398
+ "model.language_model.layers.9.linear_attn.in_proj_b": {
399
+ "bits": 16,
400
+ "data_type": "fp"
401
+ },
402
+ "mtp.fc": {
403
+ "bits": 16,
404
+ "data_type": "fp"
405
+ }
406
+ },
407
+ "group_size": 128,
408
+ "iters": 300,
409
+ "nsamples": 256,
410
+ "packing_format": "auto_round:auto_gptq",
411
+ "quant_method": "auto-round",
412
+ "sym": true
413
+ },
414
+ "text_config": {
415
+ "attention_bias": false,
416
+ "attention_dropout": 0.0,
417
+ "attn_output_gate": true,
418
+ "bos_token_id": 248044,
419
+ "dtype": "bfloat16",
420
+ "eos_token_id": 248044,
421
+ "full_attention_interval": 4,
422
+ "head_dim": 256,
423
+ "hidden_act": "silu",
424
+ "hidden_size": 5120,
425
+ "initializer_range": 0.02,
426
+ "intermediate_size": 17408,
427
+ "layer_types": [
428
+ "linear_attention",
429
+ "linear_attention",
430
+ "linear_attention",
431
+ "full_attention",
432
+ "linear_attention",
433
+ "linear_attention",
434
+ "linear_attention",
435
+ "full_attention",
436
+ "linear_attention",
437
+ "linear_attention",
438
+ "linear_attention",
439
+ "full_attention",
440
+ "linear_attention",
441
+ "linear_attention",
442
+ "linear_attention",
443
+ "full_attention",
444
+ "linear_attention",
445
+ "linear_attention",
446
+ "linear_attention",
447
+ "full_attention",
448
+ "linear_attention",
449
+ "linear_attention",
450
+ "linear_attention",
451
+ "full_attention",
452
+ "linear_attention",
453
+ "linear_attention",
454
+ "linear_attention",
455
+ "full_attention",
456
+ "linear_attention",
457
+ "linear_attention",
458
+ "linear_attention",
459
+ "full_attention",
460
+ "linear_attention",
461
+ "linear_attention",
462
+ "linear_attention",
463
+ "full_attention",
464
+ "linear_attention",
465
+ "linear_attention",
466
+ "linear_attention",
467
+ "full_attention",
468
+ "linear_attention",
469
+ "linear_attention",
470
+ "linear_attention",
471
+ "full_attention",
472
+ "linear_attention",
473
+ "linear_attention",
474
+ "linear_attention",
475
+ "full_attention",
476
+ "linear_attention",
477
+ "linear_attention",
478
+ "linear_attention",
479
+ "full_attention",
480
+ "linear_attention",
481
+ "linear_attention",
482
+ "linear_attention",
483
+ "full_attention",
484
+ "linear_attention",
485
+ "linear_attention",
486
+ "linear_attention",
487
+ "full_attention",
488
+ "linear_attention",
489
+ "linear_attention",
490
+ "linear_attention",
491
+ "full_attention"
492
+ ],
493
+ "linear_conv_kernel_dim": 4,
494
+ "linear_key_head_dim": 128,
495
+ "linear_num_key_heads": 16,
496
+ "linear_num_value_heads": 48,
497
+ "linear_value_head_dim": 128,
498
+ "mamba_ssm_dtype": "float32",
499
+ "max_position_embeddings": 262144,
500
+ "model_type": "qwen3_5_text",
501
+ "mtp_num_hidden_layers": 1,
502
+ "mtp_use_dedicated_embeddings": false,
503
+ "num_attention_heads": 24,
504
+ "num_hidden_layers": 64,
505
+ "num_key_value_heads": 4,
506
+ "output_gate_type": "swish",
507
+ "pad_token_id": null,
508
+ "partial_rotary_factor": 0.25,
509
+ "rms_norm_eps": 1e-06,
510
+ "rope_parameters": {
511
+ "mrope_interleaved": true,
512
+ "mrope_section": [
513
+ 11,
514
+ 11,
515
+ 10
516
+ ],
517
+ "partial_rotary_factor": 0.25,
518
+ "rope_theta": 10000000,
519
+ "rope_type": "default"
520
+ },
521
+ "tie_word_embeddings": false,
522
+ "use_cache": true,
523
+ "vocab_size": 248320
524
+ },
525
+ "tie_word_embeddings": false,
526
+ "transformers_version": "5.6.2",
527
+ "video_token_id": 248057,
528
+ "vision_config": {
529
+ "deepstack_visual_indexes": [],
530
+ "depth": 27,
531
+ "dtype": "bfloat16",
532
+ "hidden_act": "gelu_pytorch_tanh",
533
+ "hidden_size": 1152,
534
+ "in_channels": 3,
535
+ "initializer_range": 0.02,
536
+ "intermediate_size": 4304,
537
+ "model_type": "qwen3_5_vision",
538
+ "num_heads": 16,
539
+ "num_position_embeddings": 2304,
540
+ "out_hidden_size": 5120,
541
+ "patch_size": 16,
542
+ "spatial_merge_size": 2,
543
+ "temporal_patch_size": 2
544
+ },
545
+ "vision_end_token_id": 248054,
546
+ "vision_start_token_id": 248053
547
+ }
generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 248044,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 248046,
6
+ 248044
7
+ ],
8
+ "pad_token_id": 248044,
9
+ "temperature": 1.0,
10
+ "top_k": 20,
11
+ "top_p": 0.95,
12
+ "transformers_version": "5.6.2"
13
+ }
model-00001-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:994d372da0812952c05a3fb1093db5b1245e314385290892f0dce3883f4c2643
3
+ size 2542796928
model-00002-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:388d57d83f5ff822553a0e712bc091d4bc47a3db0fdc88e258b73ebe17bc0d5f
3
+ size 4983417520
model-00003-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c0238413f8f6505c27dd1c41480639a13261d0e8e4df141add9b9cfcd9f91686
3
+ size 4959166944
model-00004-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e7ba2bd56ce4a6fc642934dc84984183df016570e4c703324619917c7c71639
3
+ size 4959166944
model-00005-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2c6c0f9c4c84321000c6aab280419d6141f33e627f995fefe45469ec9b69d1c
3
+ size 1253852480
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
model_extra_tensors.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7cc0f34cc732905a487bb4e4b5704f34090b61019cf282750f093aeca01de612
3
+ size 298305576
preprocessor_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_convert_rgb": true,
3
+ "do_normalize": true,
4
+ "do_rescale": true,
5
+ "do_resize": true,
6
+ "image_mean": [
7
+ 0.5,
8
+ 0.5,
9
+ 0.5
10
+ ],
11
+ "image_processor_type": "Qwen2VLImageProcessor",
12
+ "image_std": [
13
+ 0.5,
14
+ 0.5,
15
+ 0.5
16
+ ],
17
+ "merge_size": 2,
18
+ "patch_size": 16,
19
+ "resample": 3,
20
+ "rescale_factor": 0.00392156862745098,
21
+ "size": {
22
+ "longest_edge": 16777216,
23
+ "shortest_edge": 65536
24
+ },
25
+ "temporal_patch_size": 2
26
+ }
processor_config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "do_convert_rgb": true,
4
+ "do_normalize": true,
5
+ "do_rescale": true,
6
+ "do_resize": true,
7
+ "image_mean": [
8
+ 0.5,
9
+ 0.5,
10
+ 0.5
11
+ ],
12
+ "image_processor_type": "Qwen2VLImageProcessor",
13
+ "image_std": [
14
+ 0.5,
15
+ 0.5,
16
+ 0.5
17
+ ],
18
+ "merge_size": 2,
19
+ "patch_size": 16,
20
+ "resample": 3,
21
+ "rescale_factor": 0.00392156862745098,
22
+ "size": {
23
+ "longest_edge": 16777216,
24
+ "shortest_edge": 65536
25
+ },
26
+ "temporal_patch_size": 2
27
+ },
28
+ "processor_class": "Qwen3VLProcessor",
29
+ "video_processor": {
30
+ "do_convert_rgb": true,
31
+ "do_normalize": true,
32
+ "do_rescale": true,
33
+ "do_resize": true,
34
+ "do_sample_frames": true,
35
+ "fps": 2,
36
+ "image_mean": [
37
+ 0.5,
38
+ 0.5,
39
+ 0.5
40
+ ],
41
+ "image_std": [
42
+ 0.5,
43
+ 0.5,
44
+ 0.5
45
+ ],
46
+ "max_frames": 768,
47
+ "merge_size": 2,
48
+ "min_frames": 4,
49
+ "patch_size": 16,
50
+ "resample": 3,
51
+ "rescale_factor": 0.00392156862745098,
52
+ "return_metadata": false,
53
+ "size": {
54
+ "longest_edge": 25165824,
55
+ "shortest_edge": 4096
56
+ },
57
+ "temporal_patch_size": 2,
58
+ "video_processor_type": "Qwen3VLVideoProcessor"
59
+ }
60
+ }
quantization_config.json ADDED
@@ -0,0 +1,398 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bits": 4,
3
+ "data_type": "int",
4
+ "group_size": 128,
5
+ "sym": true,
6
+ "iters": 300,
7
+ "nsamples": 256,
8
+ "autoround_version": "0.13.0",
9
+ "block_name_to_quantize": "model.language_model.layers",
10
+ "quant_method": "auto-round",
11
+ "packing_format": "auto_round:auto_gptq",
12
+ "extra_config": {
13
+ "model.language_model.layers.0.linear_attn.in_proj_b": {
14
+ "bits": 16,
15
+ "data_type": "fp"
16
+ },
17
+ "model.language_model.layers.0.linear_attn.in_proj_a": {
18
+ "bits": 16,
19
+ "data_type": "fp"
20
+ },
21
+ "model.language_model.layers.1.linear_attn.in_proj_b": {
22
+ "bits": 16,
23
+ "data_type": "fp"
24
+ },
25
+ "model.language_model.layers.1.linear_attn.in_proj_a": {
26
+ "bits": 16,
27
+ "data_type": "fp"
28
+ },
29
+ "model.language_model.layers.2.linear_attn.in_proj_b": {
30
+ "bits": 16,
31
+ "data_type": "fp"
32
+ },
33
+ "model.language_model.layers.2.linear_attn.in_proj_a": {
34
+ "bits": 16,
35
+ "data_type": "fp"
36
+ },
37
+ "model.language_model.layers.4.linear_attn.in_proj_b": {
38
+ "bits": 16,
39
+ "data_type": "fp"
40
+ },
41
+ "model.language_model.layers.4.linear_attn.in_proj_a": {
42
+ "bits": 16,
43
+ "data_type": "fp"
44
+ },
45
+ "model.language_model.layers.5.linear_attn.in_proj_b": {
46
+ "bits": 16,
47
+ "data_type": "fp"
48
+ },
49
+ "model.language_model.layers.5.linear_attn.in_proj_a": {
50
+ "bits": 16,
51
+ "data_type": "fp"
52
+ },
53
+ "model.language_model.layers.6.linear_attn.in_proj_b": {
54
+ "bits": 16,
55
+ "data_type": "fp"
56
+ },
57
+ "model.language_model.layers.6.linear_attn.in_proj_a": {
58
+ "bits": 16,
59
+ "data_type": "fp"
60
+ },
61
+ "model.language_model.layers.8.linear_attn.in_proj_b": {
62
+ "bits": 16,
63
+ "data_type": "fp"
64
+ },
65
+ "model.language_model.layers.8.linear_attn.in_proj_a": {
66
+ "bits": 16,
67
+ "data_type": "fp"
68
+ },
69
+ "model.language_model.layers.9.linear_attn.in_proj_b": {
70
+ "bits": 16,
71
+ "data_type": "fp"
72
+ },
73
+ "model.language_model.layers.9.linear_attn.in_proj_a": {
74
+ "bits": 16,
75
+ "data_type": "fp"
76
+ },
77
+ "model.language_model.layers.10.linear_attn.in_proj_b": {
78
+ "bits": 16,
79
+ "data_type": "fp"
80
+ },
81
+ "model.language_model.layers.10.linear_attn.in_proj_a": {
82
+ "bits": 16,
83
+ "data_type": "fp"
84
+ },
85
+ "model.language_model.layers.12.linear_attn.in_proj_b": {
86
+ "bits": 16,
87
+ "data_type": "fp"
88
+ },
89
+ "model.language_model.layers.12.linear_attn.in_proj_a": {
90
+ "bits": 16,
91
+ "data_type": "fp"
92
+ },
93
+ "model.language_model.layers.13.linear_attn.in_proj_b": {
94
+ "bits": 16,
95
+ "data_type": "fp"
96
+ },
97
+ "model.language_model.layers.13.linear_attn.in_proj_a": {
98
+ "bits": 16,
99
+ "data_type": "fp"
100
+ },
101
+ "model.language_model.layers.14.linear_attn.in_proj_b": {
102
+ "bits": 16,
103
+ "data_type": "fp"
104
+ },
105
+ "model.language_model.layers.14.linear_attn.in_proj_a": {
106
+ "bits": 16,
107
+ "data_type": "fp"
108
+ },
109
+ "model.language_model.layers.16.linear_attn.in_proj_b": {
110
+ "bits": 16,
111
+ "data_type": "fp"
112
+ },
113
+ "model.language_model.layers.16.linear_attn.in_proj_a": {
114
+ "bits": 16,
115
+ "data_type": "fp"
116
+ },
117
+ "model.language_model.layers.17.linear_attn.in_proj_b": {
118
+ "bits": 16,
119
+ "data_type": "fp"
120
+ },
121
+ "model.language_model.layers.17.linear_attn.in_proj_a": {
122
+ "bits": 16,
123
+ "data_type": "fp"
124
+ },
125
+ "model.language_model.layers.18.linear_attn.in_proj_b": {
126
+ "bits": 16,
127
+ "data_type": "fp"
128
+ },
129
+ "model.language_model.layers.18.linear_attn.in_proj_a": {
130
+ "bits": 16,
131
+ "data_type": "fp"
132
+ },
133
+ "model.language_model.layers.20.linear_attn.in_proj_b": {
134
+ "bits": 16,
135
+ "data_type": "fp"
136
+ },
137
+ "model.language_model.layers.20.linear_attn.in_proj_a": {
138
+ "bits": 16,
139
+ "data_type": "fp"
140
+ },
141
+ "model.language_model.layers.21.linear_attn.in_proj_b": {
142
+ "bits": 16,
143
+ "data_type": "fp"
144
+ },
145
+ "model.language_model.layers.21.linear_attn.in_proj_a": {
146
+ "bits": 16,
147
+ "data_type": "fp"
148
+ },
149
+ "model.language_model.layers.22.linear_attn.in_proj_b": {
150
+ "bits": 16,
151
+ "data_type": "fp"
152
+ },
153
+ "model.language_model.layers.22.linear_attn.in_proj_a": {
154
+ "bits": 16,
155
+ "data_type": "fp"
156
+ },
157
+ "model.language_model.layers.24.linear_attn.in_proj_b": {
158
+ "bits": 16,
159
+ "data_type": "fp"
160
+ },
161
+ "model.language_model.layers.24.linear_attn.in_proj_a": {
162
+ "bits": 16,
163
+ "data_type": "fp"
164
+ },
165
+ "model.language_model.layers.25.linear_attn.in_proj_b": {
166
+ "bits": 16,
167
+ "data_type": "fp"
168
+ },
169
+ "model.language_model.layers.25.linear_attn.in_proj_a": {
170
+ "bits": 16,
171
+ "data_type": "fp"
172
+ },
173
+ "model.language_model.layers.26.linear_attn.in_proj_b": {
174
+ "bits": 16,
175
+ "data_type": "fp"
176
+ },
177
+ "model.language_model.layers.26.linear_attn.in_proj_a": {
178
+ "bits": 16,
179
+ "data_type": "fp"
180
+ },
181
+ "model.language_model.layers.28.linear_attn.in_proj_b": {
182
+ "bits": 16,
183
+ "data_type": "fp"
184
+ },
185
+ "model.language_model.layers.28.linear_attn.in_proj_a": {
186
+ "bits": 16,
187
+ "data_type": "fp"
188
+ },
189
+ "model.language_model.layers.29.linear_attn.in_proj_b": {
190
+ "bits": 16,
191
+ "data_type": "fp"
192
+ },
193
+ "model.language_model.layers.29.linear_attn.in_proj_a": {
194
+ "bits": 16,
195
+ "data_type": "fp"
196
+ },
197
+ "model.language_model.layers.30.linear_attn.in_proj_b": {
198
+ "bits": 16,
199
+ "data_type": "fp"
200
+ },
201
+ "model.language_model.layers.30.linear_attn.in_proj_a": {
202
+ "bits": 16,
203
+ "data_type": "fp"
204
+ },
205
+ "model.language_model.layers.32.linear_attn.in_proj_b": {
206
+ "bits": 16,
207
+ "data_type": "fp"
208
+ },
209
+ "model.language_model.layers.32.linear_attn.in_proj_a": {
210
+ "bits": 16,
211
+ "data_type": "fp"
212
+ },
213
+ "model.language_model.layers.33.linear_attn.in_proj_b": {
214
+ "bits": 16,
215
+ "data_type": "fp"
216
+ },
217
+ "model.language_model.layers.33.linear_attn.in_proj_a": {
218
+ "bits": 16,
219
+ "data_type": "fp"
220
+ },
221
+ "model.language_model.layers.34.linear_attn.in_proj_b": {
222
+ "bits": 16,
223
+ "data_type": "fp"
224
+ },
225
+ "model.language_model.layers.34.linear_attn.in_proj_a": {
226
+ "bits": 16,
227
+ "data_type": "fp"
228
+ },
229
+ "model.language_model.layers.36.linear_attn.in_proj_b": {
230
+ "bits": 16,
231
+ "data_type": "fp"
232
+ },
233
+ "model.language_model.layers.36.linear_attn.in_proj_a": {
234
+ "bits": 16,
235
+ "data_type": "fp"
236
+ },
237
+ "model.language_model.layers.37.linear_attn.in_proj_b": {
238
+ "bits": 16,
239
+ "data_type": "fp"
240
+ },
241
+ "model.language_model.layers.37.linear_attn.in_proj_a": {
242
+ "bits": 16,
243
+ "data_type": "fp"
244
+ },
245
+ "model.language_model.layers.38.linear_attn.in_proj_b": {
246
+ "bits": 16,
247
+ "data_type": "fp"
248
+ },
249
+ "model.language_model.layers.38.linear_attn.in_proj_a": {
250
+ "bits": 16,
251
+ "data_type": "fp"
252
+ },
253
+ "model.language_model.layers.40.linear_attn.in_proj_b": {
254
+ "bits": 16,
255
+ "data_type": "fp"
256
+ },
257
+ "model.language_model.layers.40.linear_attn.in_proj_a": {
258
+ "bits": 16,
259
+ "data_type": "fp"
260
+ },
261
+ "model.language_model.layers.41.linear_attn.in_proj_b": {
262
+ "bits": 16,
263
+ "data_type": "fp"
264
+ },
265
+ "model.language_model.layers.41.linear_attn.in_proj_a": {
266
+ "bits": 16,
267
+ "data_type": "fp"
268
+ },
269
+ "model.language_model.layers.42.linear_attn.in_proj_b": {
270
+ "bits": 16,
271
+ "data_type": "fp"
272
+ },
273
+ "model.language_model.layers.42.linear_attn.in_proj_a": {
274
+ "bits": 16,
275
+ "data_type": "fp"
276
+ },
277
+ "model.language_model.layers.44.linear_attn.in_proj_b": {
278
+ "bits": 16,
279
+ "data_type": "fp"
280
+ },
281
+ "model.language_model.layers.44.linear_attn.in_proj_a": {
282
+ "bits": 16,
283
+ "data_type": "fp"
284
+ },
285
+ "model.language_model.layers.45.linear_attn.in_proj_b": {
286
+ "bits": 16,
287
+ "data_type": "fp"
288
+ },
289
+ "model.language_model.layers.45.linear_attn.in_proj_a": {
290
+ "bits": 16,
291
+ "data_type": "fp"
292
+ },
293
+ "model.language_model.layers.46.linear_attn.in_proj_b": {
294
+ "bits": 16,
295
+ "data_type": "fp"
296
+ },
297
+ "model.language_model.layers.46.linear_attn.in_proj_a": {
298
+ "bits": 16,
299
+ "data_type": "fp"
300
+ },
301
+ "model.language_model.layers.48.linear_attn.in_proj_b": {
302
+ "bits": 16,
303
+ "data_type": "fp"
304
+ },
305
+ "model.language_model.layers.48.linear_attn.in_proj_a": {
306
+ "bits": 16,
307
+ "data_type": "fp"
308
+ },
309
+ "model.language_model.layers.49.linear_attn.in_proj_b": {
310
+ "bits": 16,
311
+ "data_type": "fp"
312
+ },
313
+ "model.language_model.layers.49.linear_attn.in_proj_a": {
314
+ "bits": 16,
315
+ "data_type": "fp"
316
+ },
317
+ "model.language_model.layers.50.linear_attn.in_proj_b": {
318
+ "bits": 16,
319
+ "data_type": "fp"
320
+ },
321
+ "model.language_model.layers.50.linear_attn.in_proj_a": {
322
+ "bits": 16,
323
+ "data_type": "fp"
324
+ },
325
+ "model.language_model.layers.52.linear_attn.in_proj_b": {
326
+ "bits": 16,
327
+ "data_type": "fp"
328
+ },
329
+ "model.language_model.layers.52.linear_attn.in_proj_a": {
330
+ "bits": 16,
331
+ "data_type": "fp"
332
+ },
333
+ "model.language_model.layers.53.linear_attn.in_proj_b": {
334
+ "bits": 16,
335
+ "data_type": "fp"
336
+ },
337
+ "model.language_model.layers.53.linear_attn.in_proj_a": {
338
+ "bits": 16,
339
+ "data_type": "fp"
340
+ },
341
+ "model.language_model.layers.54.linear_attn.in_proj_b": {
342
+ "bits": 16,
343
+ "data_type": "fp"
344
+ },
345
+ "model.language_model.layers.54.linear_attn.in_proj_a": {
346
+ "bits": 16,
347
+ "data_type": "fp"
348
+ },
349
+ "model.language_model.layers.56.linear_attn.in_proj_b": {
350
+ "bits": 16,
351
+ "data_type": "fp"
352
+ },
353
+ "model.language_model.layers.56.linear_attn.in_proj_a": {
354
+ "bits": 16,
355
+ "data_type": "fp"
356
+ },
357
+ "model.language_model.layers.57.linear_attn.in_proj_b": {
358
+ "bits": 16,
359
+ "data_type": "fp"
360
+ },
361
+ "model.language_model.layers.57.linear_attn.in_proj_a": {
362
+ "bits": 16,
363
+ "data_type": "fp"
364
+ },
365
+ "model.language_model.layers.58.linear_attn.in_proj_b": {
366
+ "bits": 16,
367
+ "data_type": "fp"
368
+ },
369
+ "model.language_model.layers.58.linear_attn.in_proj_a": {
370
+ "bits": 16,
371
+ "data_type": "fp"
372
+ },
373
+ "model.language_model.layers.60.linear_attn.in_proj_b": {
374
+ "bits": 16,
375
+ "data_type": "fp"
376
+ },
377
+ "model.language_model.layers.60.linear_attn.in_proj_a": {
378
+ "bits": 16,
379
+ "data_type": "fp"
380
+ },
381
+ "model.language_model.layers.61.linear_attn.in_proj_b": {
382
+ "bits": 16,
383
+ "data_type": "fp"
384
+ },
385
+ "model.language_model.layers.61.linear_attn.in_proj_a": {
386
+ "bits": 16,
387
+ "data_type": "fp"
388
+ },
389
+ "model.language_model.layers.62.linear_attn.in_proj_b": {
390
+ "bits": 16,
391
+ "data_type": "fp"
392
+ },
393
+ "model.language_model.layers.62.linear_attn.in_proj_a": {
394
+ "bits": 16,
395
+ "data_type": "fp"
396
+ }
397
+ }
398
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523
3
+ size 19989325
tokenizer_config.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "audio_bos_token": "<|audio_start|>",
4
+ "audio_eos_token": "<|audio_end|>",
5
+ "audio_token": "<|audio_pad|>",
6
+ "backend": "tokenizers",
7
+ "bos_token": null,
8
+ "clean_up_tokenization_spaces": false,
9
+ "eos_token": "<|im_end|>",
10
+ "errors": "replace",
11
+ "image_token": "<|image_pad|>",
12
+ "is_local": false,
13
+ "local_files_only": false,
14
+ "model_max_length": 262144,
15
+ "model_specific_special_tokens": {
16
+ "audio_bos_token": "<|audio_start|>",
17
+ "audio_eos_token": "<|audio_end|>",
18
+ "audio_token": "<|audio_pad|>",
19
+ "image_token": "<|image_pad|>",
20
+ "video_token": "<|video_pad|>",
21
+ "vision_bos_token": "<|vision_start|>",
22
+ "vision_eos_token": "<|vision_end|>"
23
+ },
24
+ "pad_token": "<|endoftext|>",
25
+ "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
26
+ "processor_class": "Qwen3VLProcessor",
27
+ "split_special_tokens": false,
28
+ "tokenizer_class": "Qwen2Tokenizer",
29
+ "unk_token": null,
30
+ "video_token": "<|video_pad|>",
31
+ "vision_bos_token": "<|vision_start|>",
32
+ "vision_eos_token": "<|vision_end|>"
33
+ }