{"data":{"slug":"oneflow-inc-oneflow","name":"oneflow","tagline":"OneFlow is a deep learning framework designed to be user-friendly, scalable and efficient.","github_url":"https://github.com/Oneflow-Inc/oneflow","owner":"Oneflow-Inc","repo":"oneflow","owner_avatar_url":"https://avatars.githubusercontent.com/u/24632470?v=4","primary_language":"C++","stars":9409,"forks":1013,"topics":["cuda","deep-learning","deep-neural-networks","distributed","machine-learning","ml","neural-network"],"archived":false,"github_pushed_at":"2025-12-04T01:29:28+00:00","maintenance_label":"Slowing","url":"https://www.graphcanon.com/tools/oneflow-inc-oneflow","markdown_url":"https://www.graphcanon.com/tools/oneflow-inc-oneflow.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/oneflow-inc-oneflow","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=oneflow-inc-oneflow","description":"OneFlow is a deep learning framework designed to be user-friendly, scalable and efficient.","homepage_url":"http://www.oneflow.org","license":"Apache-2.0","open_issues":645,"watchers":361,"ai_summary":"A deep learning framework focusing on user-friendliness, scalability, and efficiency, supporting both CPU and CUDA installations.","readme_excerpt":"### Preinstall docker image\n\n```\ndocker pull oneflowinc/oneflow:nightly-cuda11.8\n```\n\n---\n\n### Pip Install\n\n- (**Highly recommended**) Upgrade pip\n\n  ```\n  python3 -m pip install --upgrade pip #--user\n  ```\n\n- To install latest stable release of OneFlow with CUDA support:\n\n  ```bash\n  python3 -m pip install oneflow\n  ```\n\n- To install nightly release of OneFlow with CPU-only support:\n\n  ```bash\n  python3 -m pip install --pre oneflow -f https://oneflow-staging.oss-cn-beijing.aliyuncs.com/branch/master/cpu\n  ```\n\n- To install nightly release of OneFlow with CUDA support:\n\n  ```bash\n  python3 -m pip install --pre oneflow -f https://oneflow-staging.oss-cn-beijing.aliyuncs.com/branch/master/cu118\n  ```\n\n  If you are in China, you could run this to have pip download packages from domestic mirror of pypi:\n  ```\n  python3 -m pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple\n  ```\n  For more information on this, please refer to [pypi 镜像使用帮助](https://mirror.tuna.tsinghua.edu.cn/help/pypi/)\n\n---\n\n### Install from Source\n\n<details>\n<summary>Clone Source Code</summary>\n\n- #### Option 1: Clone source code from GitHub\n\n  ```bash\n  git clone https://github.com/Oneflow-Inc/oneflow.git\n  ```\n\n- #### Option 2: Download from Aliyun(Only available in China)\n\n  ```bash\n  curl https://oneflow-public.oss-cn-beijing.aliyuncs.com/oneflow-src.zip -o oneflow-src.zip\n  unzip oneflow-src.zip\n  ```\n\n  </details>\n\n<details>\n<summary>Build OneFlow</summary>\n\n- Install dependencies\n  ```\n  apt install -y libopenblas-dev nasm g++ gcc python3-pip cmake autoconf libtool\n  ```\n  These dependencies are preinstalled in offical conda environment and docker image, you can use the offical conda environment [here](https://github.com/Oneflow-Inc/conda-env) or use the docker image by:\n  ```bash\n  docker pull oneflowinc/manylinux2014_x86_64_cuda11.2\n  ```\n- In the root directory of OneFlow source code, run:\n\n  ```\n  mkdir build\n  cd build\n  ```\n\n- Config the project, inside `build` directory:\n\n  - If you are in China\n\n    config for CPU-only like this:\n\n    ```\n    cmake .. -C ../cmake/caches/cn/cpu.cmake\n    ```\n\n    config for CUDA like this:\n\n    ```\n    cmake .. -C ../cmake/caches/cn/cuda.cmake -DCMAKE_CUDA_ARCHITECTURES=80 -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda -DCUDNN_ROOT_DIR=/usr/local/cudnn\n    ```\n\n  - If you are not in China\n\n    config for CPU-only like this:\n\n    ```\n    cmake .. -C ../cmake/caches/international/cpu.cmake\n    ```\n\n    config for CUDA like this:\n\n    ```\n    cmake .. -C ../cmake/caches/international/cuda.cmake -DCMAKE_CUDA_ARCHITECTURES=80 -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda -DCUDNN_ROOT_DIR=/usr/local/cudnn\n    ```\n    Here the DCMAKE\\_CUDA\\_ARCHITECTURES macro is used to specify the CUDA architecture, and the DCUDA\\_TOOLKIT\\_ROOT\\_DIR and DCUDNN\\_ROOT\\_DIR macros are used to specify the root path of the CUDA Toolkit and CUDNN.\n\n- Build the project, inside `build` directory, run:\n\n  ```\n  make -j$(nproc)\n  ```\n\n- Add oneflow to your PYTHONPATH, inside `build` directory, run:\n\n  ```\n  source source.sh\n  ```\n\n  Please note that this change is not permanent.\n\n- Simple validation\n\n  ```\n  python3 -m oneflow --doctor\n  ```\n\n  </details>\n\n---\n\n## Getting Started\n\n- Please refer to [QUICKSTART](https://docs.oneflow.org/en/master/basics/01_quickstart.html)\n- 中文版请参见 [快速上手](https://docs.oneflow.org/master/basics/01_quickstart.html)\n\n---\n\n## License\n\n[Apache License 2.0](LICENSE)","github_created_at":"2017-02-11T06:09:53+00:00","created_at":"2026-07-11T23:24:14.675556+00:00","updated_at":"2026-07-12T07:43:49.666706+00:00","categories":[{"slug":"model-training","name":"Model Training","url":"https://www.graphcanon.com/categories/model-training","markdown_url":"https://www.graphcanon.com/categories/model-training.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/model-training"}],"tags":[{"slug":"cuda","name":"cuda"},{"slug":"deep-learning","name":"deep-learning"},{"slug":"distributed","name":"distributed"},{"slug":"machine-learning","name":"machine-learning"},{"slug":"neural-networks","name":"neural-networks"}],"trust":{"provenance":{"is_fork":false,"github_id":81634683,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-07-11T23:24:19.592Z","maintenance":{"label":"Slowing","score":36,"methodology":"github_public_v1","releases_90d":0,"days_since_push":219,"last_release_at":"2024-03-11T03:18:18Z"},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-11T23:24:20.083Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-07-12T07:43:22.569Z"},"languages":{"value":["c++"],"source":"github.language","observed_at":"2026-07-12T07:43:22.569Z"},"license_spdx":{"value":"Apache-2.0","source":"github.license","observed_at":"2026-07-12T07:43:22.569Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["OneFlow is preferable when you need a user-friendly framework for both CPU and CUDA installations, aiming to streamline the deep learning workflow.","OneFlow should be selected when targeting applications requiring high scalability and efficiency without compromising ease of use.","OneFlow is beneficial in environments where a C++-based backend provides specific advantages over other languages."],"when_not_to_use":["Avoid OneFlow if your project requires extensive customization features not natively supported, as switching to another framework might offer better flexibility.","If the development environment lacks support for CUDA or Python3-based installation methods, consider an alternative framework that suits your hardware and software environment more closely.","OneFlow may not be ideal when working in regions with difficulty accessing external libraries due to dependency management tailored towards certain geographic locations."],"source":"enrich:decision_facts","observed_at":"2026-07-12T07:43:49.419Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"OneFlow is a deep learning framework built for user-friendly, scalable, and efficient performance in model training, with support via CUDA installations."}]}}