{"id":27884304,"url":"https://github.com/skyworkai/skywork-r1v","last_synced_at":"2025-05-14T11:10:37.050Z","repository":{"id":282925455,"uuid":"948904124","full_name":"SkyworkAI/Skywork-R1V","owner":"SkyworkAI","description":"Skywork-R1V2 : Multimodal Hybrid Reinforcement Learning for Reasoning(最好的多模态推理)","archived":false,"fork":false,"pushed_at":"2025-04-28T09:46:51.000Z","size":38086,"stargazers_count":2405,"open_issues_count":20,"forks_count":236,"subscribers_count":142,"default_branch":"main","last_synced_at":"2025-05-05T06:41:10.902Z","etag":null,"topics":["deepseek-r1","grpo","llm","multimodal-r1","multimodal-understanding","r1v","reasoning","reinforcement-learning","skywork-r1v","vlm","vlm-r1"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/SkyworkAI.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null}},"created_at":"2025-03-15T08:11:44.000Z","updated_at":"2025-05-05T06:34:09.000Z","dependencies_parsed_at":"2025-05-14T11:09:47.518Z","dependency_job_id":null,"html_url":"https://github.com/SkyworkAI/Skywork-R1V","commit_stats":null,"previous_names":["skyworkai/skywork-r1v"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SkyworkAI%2FSkywork-R1V","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SkyworkAI%2FSkywork-R1V/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SkyworkAI%2FSkywork-R1V/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SkyworkAI%2FSkywork-R1V/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/SkyworkAI","download_url":"https://codeload.github.com/SkyworkAI/Skywork-R1V/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":254129490,"owners_count":22019628,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["deepseek-r1","grpo","llm","multimodal-r1","multimodal-understanding","r1v","reasoning","reinforcement-learning","skywork-r1v","vlm","vlm-r1"],"created_at":"2025-05-05T06:36:03.372Z","updated_at":"2025-05-14T11:10:31.977Z","avatar_url":"https://github.com/SkyworkAI.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003c!-- markdownlint-disable first-line-h1 --\u003e\n\u003c!-- markdownlint-disable html --\u003e\n\u003c!-- markdownlint-disable no-duplicate-header --\u003e\n\n# Skywork-R1V: Pioneering Multimodal Reasoning with CoT\n\u003cfont size=7\u003e\u003cdiv align='center' \u003e  [[🤗 Skywork-R1V2-38B](https://huggingface.co/Skywork/Skywork-R1V2-38B)] [[🤖 R1V2 ModelScope](https://modelscope.cn/models/Skywork/Skywork-R1V2-38B)] [[📖 R1V2 Report](https://arxiv.org/abs/2504.16656)] \u003cbr\u003e\u003c/br\u003e[[🤗 Skywork-R1V-38B](https://huggingface.co/Skywork/Skywork-R1V-38B)] [[📖 R1V1 Report](https://arxiv.org/abs/2504.05599)] \u003c/div\u003e\u003c/font\u003e\n\n\n\u003cdiv align=\"center\"\u003e\n  \u003ctable\u003e\n    \u003ctr\u003e\n      \u003ctd\u003e\n        \u003cimg src=\"https://github.com/SkyworkAI/Skywork-R1V/blob/main/imgs/math_r1v.gif\" width=\"450\" height=\"400\" alt=\"math_r1v\" /\u003e\n      \u003c/td\u003e\n      \u003ctd\u003e\n        \u003cimg src=\"https://github.com/SkyworkAI/Skywork-R1V/blob/main/imgs/Chemistry_cn.gif\" width=\"450\" height=\"400\" alt=\"chemistry_1\" /\u003e\n      \u003c/td\u003e\n    \u003c/tr\u003e\n  \u003c/table\u003e\n\u003c/div\u003e\n\n\u003cbr\u003e\u003c/br\u003e\nWelcome to the Skywork-R1V repository! Here, you'll find the model weights and inference code for our state-of-the-art open-sourced multimodal reasoning model, enabling advanced visual and text thinking.\n\n## 🔥News\n\n**April 28, 2025**: We released awq quantized version of Skywork R1V2[[🤗 Skywork-R1V2-38B-AWQ](https://huggingface.co/Skywork/Skywork-R1V2-38B-AWQ)], supporting single-card (above 30GB) inference.\n\n **April 24, 2025**: We released **Skywork-R1V2**, a state-of-the-art, open-source multimodal reasoning model that achieves leading performance across multiple vision-language benchmarks.[[🤗 Skywork-R1V2-38B](https://huggingface.co/Skywork/Skywork-R1V2-38B)][[📖R1V2 Report](https://arxiv.org/abs/2504.16656)] \n \n**April 9, 2025**: Our technical report is currently available on arxiv: [[Skywork-R1V: Pioneering Multimodal Reasoning with CoT](https://arxiv.org/abs/2504.05599)].\n\n**April 1, 2025**: Skywork-R1V supports inference with [[vLLM](https://github.com/vllm-project/vllm)], On 4×L20Y GPUs, vLLM generates 1k tokens in ~12.3s, at least 5× faster than transformers.\n\n**Mar 26, 2025**: We released awq quantized version of Skywork R1V[[🤗 Skywork-R1V-38B-AWQ](https://huggingface.co/Skywork/Skywork-R1V-38B-AWQ)], supporting single-card (above 30GB) inference.\n\n**Mar 18, 2025**: We are thrilled to introduce Skywork R1V, the first industry open-sourced multimodal reasoning model with advanced visual chain-of-thought capabilities, pushing the boundaries of AI-driven vision and logical inference! 🚀\n\n\n\n## R1V2-38B Evaluation\n Skywork-R1V2-38B demonstrates state-of-the-art performance on both text and multimodal reasoning tasks.\n \u003cdiv align=\"center\"\u003e\n   \u003cb\u003eComparison of Skywork-R1V2 with Multimodal Open-Source and Proprietary Models\u003c/b\u003e\n \u003c/div\u003e\n \n \u003ctable align=\"center\"\u003e\n   \u003cthead\u003e\n     \u003ctr\u003e\n       \u003cth rowspan=\"2\"\u003eModel\u003c/th\u003e\n       \u003cth colspan=\"5\" align=\"center\"\u003e\u003cstrong\u003eText Reasoning (pass@1 or %)\u003c/strong\u003e\u003c/th\u003e\n       \u003cth colspan=\"5\" align=\"center\"\u003e\u003cstrong\u003eMultimodal Reasoning (%)\u003c/strong\u003e\u003c/th\u003e\n     \u003c/tr\u003e\n     \u003ctr\u003e\n       \u003cth\u003eAIME24\u003c/th\u003e\n       \u003cth\u003eLiveCodebench\u003c/th\u003e\n       \u003cth\u003eliveBench\u003c/th\u003e\n       \u003cth\u003eIFEVAL\u003c/th\u003e\n       \u003cth\u003eBFCL\u003c/th\u003e\n       \u003cth\u003eMMMU(val)\u003c/th\u003e\n       \u003cth\u003eMathVista(mini)\u003c/th\u003e\n       \u003cth\u003eMathVision(mini)\u003c/th\u003e\n       \u003cth\u003eOlympiadBench\u003c/th\u003e\n       \u003cth\u003emmmu-pro\u003c/th\u003e\n     \u003c/tr\u003e\n   \u003c/thead\u003e\n   \u003ctbody\u003e\n     \u003ctr\u003e\n       \u003ctd\u003e\u003cstrong\u003eSkywork-R1V2-38B\u003c/strong\u003e\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e\u003cstrong\u003e78.9\u003c/strong\u003e\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e\u003cstrong\u003e63.6\u003c/strong\u003e\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e\u003cstrong\u003e73.2\u003c/strong\u003e\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e\u003cstrong\u003e82.9\u003c/strong\u003e\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e\u003cstrong\u003e66.3\u003c/strong\u003e\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e\u003cstrong\u003e73.6\u003c/strong\u003e\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e74.0\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e49.0\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e\u003cstrong\u003e62.6\u003c/strong\u003e\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e\u003cstrong\u003e52.0\u003c/strong\u003e\u003c/td\u003e\n     \u003c/tr\u003e\n     \u003ctr\u003e\n       \u003ctd\u003eOpenAI-4o\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e74.6\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e9.3\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e49.9\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e69.1\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e63.8\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e\u003cstrong\u003e58.0\u003c/strong\u003e\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n     \u003c/tr\u003e\n     \u003ctr\u003e\n       \u003ctd\u003eClaude 3.5 Sonnet\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e16.0\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e65.0\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e66.4\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e65.3\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n     \u003c/tr\u003e\n     \u003ctr\u003e\n       \u003ctd\u003eKimi k1.5\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e77.5\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e70.0\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e74.9\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n     \u003c/tr\u003e\n     \u003ctr\u003e\n       \u003ctd\u003eQwen2.5-VL-72B\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e70.2\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e74.8\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e38.1\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e40.4\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n     \u003c/tr\u003e\n     \u003ctr\u003e\n       \u003ctd\u003eInternVL3-38B\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e70.1\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e\u003cstrong\u003e75.1\u003c/strong\u003e\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e34.2\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e-\u003c/td\u003e\n       \u003ctd align=\"center\"\u003e—\u003c/td\u003e\n     \u003c/tr\u003e\n   \u003c/tbody\u003e\n \u003c/table\u003e\n \n \n \u003cbr\u003e\u003c/br\u003e\n \u003cdiv align=\"center\"\u003e\n   \u003cb\u003eText Reasoning Performance\u003c/b\u003e\n \u003c/div\u003e\n \u003cdiv align=\"center\"\u003e\n   \u003cimg src=\"https://github.com/SkyworkAI/Skywork-R1V/blob/main/imgs/text_reasoning.png?raw=true\" width=\"100%\" alt=\"text_reasoning\" /\u003e\n \u003c/div\u003e\n \n \u003cbr\u003e\u003c/br\u003e\n \u003cdiv align=\"center\"\u003e\n   \u003cb\u003eMultimodal Reasoning vs Proprietary Models\u003c/b\u003e\n \u003c/div\u003e\n \u003cdiv align=\"center\"\u003e\n   \u003cimg src=\"https://github.com/SkyworkAI/Skywork-R1V/blob/main/imgs/multi_reasoning_pm.png?raw=true\" width=\"100%\" alt=\"multi_reasoning_pm\" /\u003e\n \u003c/div\u003e\n \n \u003cbr\u003e\u003c/br\u003e\n \u003cdiv align=\"center\"\u003e\n   \u003cb\u003eMultimodal Reasoning vs Open-Source Models\u003c/b\u003e\n \u003c/div\u003e\n \u003cdiv align=\"center\"\u003e\n   \u003cimg src=\"https://github.com/SkyworkAI/Skywork-R1V/blob/main/imgs/multi_reasoning_osm.png?raw=true\" width=\"100%\" alt=\"multi_reasoning_osm\" /\u003e\n \u003c/div\u003e\n \n## How to Run Locally\n\n### 1. Clone the Repository\n\n```shell\ngit clone https://github.com/SkyworkAI/Skywork-R1V.git\ncd skywork-r1v/inference\n```\n\n### 2. Set Up the Environment\n\n```shell\n# For Transformers  \nconda create -n r1-v python=3.10 \u0026\u0026 conda activate r1-v  \nbash setup.sh  \n# For vLLM  \nconda create -n r1v-vllm python=3.10 \u0026\u0026 conda activate r1v-vllm  \npip install -U vllm  \n```\n\n### 3. Run the Inference Script\n\n#### Using Transformers\n```shell\nCUDA_VISIBLE_DEVICES=\"0,1\" python inference_with_transformers.py \\\n    --model_path path \\\n    --image_paths image1_path \\\n    --question \"your question\"\n```\n#### Using vLLM\n```shell\npython inference_with_vllm.py \\\n    --model_path path \\\n    --image_paths image1_path image2_path \\\n    --question \"your question\" \\\n    --tensor_parallel_size 4\n```\n\n## License\nThis code repository is licensed under [the MIT License](https://github.com/SkyworkAI/Skywork-R1V/blob/main/LICENSE). \n✅ Commercial use permitted\n\n✅ Modification allowed\n\n✅ Distribution allowed\n\n❌ No liability\n\n\n## Citation\nIf you use Skywork-R1V in your research, please cite:\n```\n@misc{chris2025skyworkr1v2multimodalhybrid,\n      title={Skywork R1V2: Multimodal Hybrid Reinforcement Learning for Reasoning}, \n      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},\n      year={2025},\n      eprint={2504.16656},\n      archivePrefix={arXiv},\n      primaryClass={cs.CV},\n      url={https://arxiv.org/abs/2504.16656}, \n}\n```\n\n```\n@misc{peng2025skyworkr1vpioneeringmultimodal,\n      title={Skywork R1V: Pioneering Multimodal Reasoning with Chain-of-Thought}, \n      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},\n      year={2025},\n      eprint={2504.05599},\n      archivePrefix={arXiv},\n      primaryClass={cs.CV},\n      url={https://arxiv.org/abs/2504.05599}, \n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fskyworkai%2Fskywork-r1v","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fskyworkai%2Fskywork-r1v","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fskyworkai%2Fskywork-r1v/lists"}