README.md

metadata

library_name: transformers
license: mit
pipeline_tag: image-text-to-text

Skywork-R1V-38B-AWQ

📖 Technical Report | 💻 GitHub | 🌐 Wisemodel

Evaluation

Comparison with Larger-Scale Open-Source and Closed-Source Models

	Benchmark	LLM	VLM
		QwQ-32B-Preview	InternVL-2.5-38B	VILA 1.5-40B	InternVL2-40B	Skywork-R1V-38B	Skywork-R1V-AWQ
Reasoning	MATH-500	90.6	-	-	-	94.0	86.0
	AIME 2024	50.0	-	-	-	72.0	61.0
	GPQA	54.5	-	-	-	61.6	56.5
Vision	MathVista(mini)	-	71.9	49.5	63.7	67.5	59.9
	MMMU(Val)	-	63.9	55.1	55.2	69.0	60.1

Usage

You can use the quantized model with different inference frameworks:

Using VLLM

Python API

import os
from vllm import LLM, SamplingParams
from vllm.entrypoints.chat_utils import load_chat_template

model_name = "Skywork/Skywork-R1V-38B-AWQ"  # or local path
llm = LLM(model_name, 
          dtype='float16', 
          quantization="awq", 
          gpu_memory_utilization=0.85,
          max_model_len=4096,
          trust_remote_code=True,
         )

# Add your inference code here

OpenAI-compatible API Server

MODEL_ID="Skywork/Skywork-R1V-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

import os
from lmdeploy import pipeline, TurbomindEngineConfig, ChatTemplateConfig
from lmdeploy.vl import load_image

model_path = "Skywork/Skywork-R1V-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:

@article{skywork2025r1v,
  title     = {Skywork-R1V3 Technical Report},
  author    = {Yi Peng, Chris, Xiaokun Wang, Yichen Wei, Jiangbo Pei, Weijie Qiu, Ai Jian, Yunzhuo Hao, Jiachun Pan, Tianyidan Xie, Li Ge, Rongxian Zhuang, Xuchen Song, Yang Liu, Yahui Zhou},
  year      = {2025},
  journal   = {https://huggingface.co/papers/2507.06167},
  url       = {https://huggingface.co/papers/2507.06167}
}

Skywork-R1V-38B-AWQ (中文说明)

使用方法

您可以使用不同的推理框架来使用这个量化模型：

使用 VLLM

Python API

import os
from vllm import LLM, SamplingParams
from vllm.entrypoints.chat_utils import load_chat_template

model_name = "Skywork/Skywork-R1V-38B-AWQ"  # 或本地路径
llm = LLM(model_name, 
          dtype='float16', 
          quantization="awq", 
          gpu_memory_utilization=0.85,
          max_model_len=4096,
          trust_remote_code=True,
         )

# 在此添加您的推理代码

OpenAI 兼容的 API 服务器

MODEL_ID="Skywork/Skywork-R1V-38B-AWQ"  # 或本地路径

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

使用 LMDeploy

import os
from lmdeploy import pipeline, TurbomindEngineConfig, ChatTemplateConfig
from lmdeploy.vl import load_image

model_path = "Skywork/Skywork-R1V-38B-AWQ"  # 或本地路径

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,
               )

# 示例：多模态推理
image = load_image('table.jpg')
response = pipe(('描述这个图片？', image))
print(response.text)

硬件要求

与原始 FP16 模型相比，AWQ 量化减少了内存占用。我们建议：

• 至少一块 30GB+ 显存的 GPU 用于推理
• 对于更长上下文的最佳性能，建议使用 40GB+ 显存

引用

如果您在研究中使用此模型，请引用：

@misc{peng2025skyworkr1vpioneeringmultimodal,
      title={Skywork-R1V3 Technical Report}, 
      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},
      url={https://huggingface.co/papers/2507.06167}, 
}