Hugging Face's logo

Skywork
/
Skywork-R1V2-38B-AWQ

Image-Text-to-Text
Transformers
PyTorch
internvl_chat
conversational
custom_code
Model card Files Community
4
Skywork-R1V2-38B-AWQ
File size: 8,276 Bytes 
bac0036
 
 
 
 
 
 
 
 
 
 
3485ad2
bac0036
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f5ae779
 
bac0036
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3a9d437
 
bac0036
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5dac64b
bac0036
 
 
 
 
 
 
 
 
 
 
d36cd65
 
 
 
9afb830
d36cd65
 
 
bac0036
 
 
 
 
 
 
 
5dac64b
bac0036
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
69c997a
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
 
---
license: mit
library_name: transformers
pipeline_tag: image-text-to-text
---
# Skywork-R1V2-38B-AWQ

<div align="center">   
  <img src="skywork-logo.png" alt="Introduction Image" width="500" height="400"> 
</div>

## 📖 [R1V2 Report](https://arxiv.org/abs/2504.16656) | 💻 [GitHub](https://github.com/SkyworkAI/Skywork-R1V) | 🌐 [ModelScope](https://modelscope.cn/models/Skywork/Skywork-R1V2-38B) 

<div align="center">

[![GitHub Stars](https://img.shields.io/github/stars/SkyworkAI/Skywork-R1V)](https://github.com/SkyworkAI/Skywork-R1V/stargazers)[![GitHub Forks](https://img.shields.io/github/forks/SkyworkAI/Skywork-R1V)](https://github.com/SkyworkAI/Skywork-R1V/fork)

</div>


## Evaluation

<div align="center">
  <b>Comprehensive performance comparison across text and multimodal reasoning benchmarks.</b>
</div>
<table align="center" border="1" style="border-collapse: collapse; width: 100%;">
  <thead>
    <tr>
      <th>Model</th>
      <th align="center">MMMU</th>
      <th align="center">MathVista</th>
      <th align="center">MathVision</th>
      <th align="center">Olympiad Bench</th>
      <th align="center">AIME 24</th>
      <th align="center">LiveCode bench</th>
      <th align="center">Live Bench</th>
      <th align="center">IFEVAL</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td colspan="9" align="center"><i>Proprietary Models</i></td>
    </tr>
    <tr>
      <td>Claude-3.5-Sonnet</td>
      <td align="center">70.4</td>
      <td align="center">67.7</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
    </tr>
    <tr>
      <td>Gemini-2-Flash</td>
      <td align="center">70.7</td>
      <td align="center">73.1</td>
      <td align="center">41.3</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
    </tr>
    <tr>
      <td>Kimi-k1.5-longcot</td>
      <td align="center">70.0</td>
      <td align="center">74.9</td>
      <td align="center">53.3</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
    </tr>
    <tr>
      <td>OpenAI-o1</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">74.3</td>
      <td align="center">63.4</td>
      <td align="center">72.2</td>
      <td align="center">-</td>
    </tr>
    <tr>
      <td>OpenAI-o4-mini</td>
      <td align="center"><b>81.6</b></td>
      <td align="center"><b>84.3</b></td>
      <td align="center"><b>58.0</b></td>
      <td align="center">-</td>
      <td align="center"><b>93.4</b></td>
      <td align="center"><b>74.6</b></td>
      <td align="center"><b>78.1</b></td>
      <td align="center">-</td>
    </tr>
    <tr>
      <td colspan="9" align="center"><i>Open-Source Models</i></td>
    </tr>
    <tr>
      <td>Skywork-R1V1</td>
      <td align="center">68.0</td>
      <td align="center">67.0</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">72.0</td>
      <td align="center">57.2</td>
      <td align="center">54.6</td>
      <td align="center">72.5</td>
    </tr>
    <tr>
      <td>DeepseekR1-671B</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td
>
      <td align="center"><b>79.8</b></td>
      <td align="center"><b>65.9</b></td>
      <td align="center">71.6</td>
      <td align="center"><b>83.3</b></td>
    </tr>
    <tr>
      <td>InternVL3-38B</td>
      <td align="center">70.1</td>
      <td align="center">75.1</td>
      <td align="center">34.2</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
    </tr>
    <tr>
      <td>Qwen2.5-VL-72B</td>
      <td align="center">70.2</td>
      <td align="center">74.8</td>
      <td align="center">38.1</td>
      <td align="center">40.4</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
    </tr>
    <tr>
      <td>QvQ-Preview-72B</td>
      <td align="center">70.3</td>
      <td align="center">71.4</td>
      <td align="center">35.9</td>
      <td align="center">33.2</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
      <td align="center">-</td>
    </tr>
    <tr>
      <td>Skywork-R1V2</td>
      <td align="center"><b>73.6</b></td>
      <td align="center">74.0</td>
      <td align="center"><b>49.0</b></td>
      <td align="center"><b>62.6</b></td>
      <td align="center">78.9</td>
      <td align="center">63.6</td>
      <td align="center"><b>73.2</b></td>
      <td align="center">82.9</td>
    </tr>
    <tr>
      <td>Skywork-R1V2-AWQ</td>
      <td align="center">64.4</td>
      <td align="center">64.8</td>
      <td align="center">42.9</td>
      <td align="center">54.8</td>
      <td align="center">77.3</td>
      <td align="center">55.7</td>
      <td align="center">64.1</td>
      <td align="center">72.5</td>
    </tr>
  </tbody>
</table>

## Usage
You can use the quantized model with different inference frameworks:
### Using VLLM


#### Python API

```python
import os
from vllm import LLM, SamplingParams
from vllm.entrypoints.chat_utils import load_chat_template
model_name = "Skywork/Skywork-R1V2-38B-AWQ"  # or local path
llm = LLM(model_name, 
          dtype='float16', 
          quantization="awq", 
          gpu_memory_utilization=0.9,
          max_model_len=4096,
          trust_remote_code=True,
         )
# Add your inference code here
```

#### OpenAI-compatible API Server

```bash
MODEL_ID="Skywork/Skywork-R1V2-38B-AWQ"  # or local path
CUDA_VISIBLE_DEVICES=0 \
    python -m vllm.entrypoints.openai.api_server \
    --model $MODEL_ID \
    --dtype float16 \
    --quantization awq \
    --port 23334 \
    --max-model-len 12000 \
    --gpu-memory-utilization 0.9 \
    --trust-remote-code
```

### Using LMDeploy

```python
import os
from lmdeploy import pipeline, TurbomindEngineConfig, ChatTemplateConfig
from lmdeploy.vl import load_image
model_path = "Skywork/Skywork-R1V2-38B-AWQ"  # or local path
engine_config = TurbomindEngineConfig(cache_max_entry_count=0.75) 
chat_template_config = ChatTemplateConfig(model_name=model_path)
pipe = pipeline(model_path, 
                backend_config=engine_config, 
                chat_template_config=chat_template_config,
               )
# Example: Multimodal inference
image = load_image('table.jpg')
response = pipe(('Describe this image?', image))
print(response.text)
```

## Hardware Requirements

The AWQ quantization reduces the memory footprint compared to the original FP16 model. We recommend:

- At least one GPU with 30GB+ VRAM for inference
- For optimal performance with longer contexts, 40GB+ VRAM is recommended

## Citation

If you use this model in your research, please cite:

```bibtex
@misc{peng2025skyworkr1vpioneeringmultimodal,
      title={Skywork R1V: Pioneering Multimodal Reasoning with Chain-of-Thought}, 
      author={Yi Peng and Chris and Xiaokun Wang and Yichen Wei and Jiangbo Pei and Weijie Qiu and Ai Jian and Yunzhuo Hao and Jiachun Pan and Tianyidan Xie and Li Ge and Rongxian Zhuang and Xuchen Song and Yang Liu and Yahui Zhou},
      year={2025},
      eprint={2504.05599},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2504.05599}, 
}
```

```bibtex
@misc{chris2025skyworkr1v2multimodalhybrid,
      title={Skywork R1V2: Multimodal Hybrid Reinforcement Learning for Reasoning}, 
      author={Chris and Yichen Wei and Yi Peng and Xiaokun Wang and Weijie Qiu and Wei Shen and Tianyidan Xie and Jiangbo Pei and Jianhao Zhang and Yunzhuo Hao and Xuchen Song and Yang Liu and Yahui Zhou},
      year={2025},
      eprint={2504.16656},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2504.16656}, 
}
```