Skip to content

Latest commit

 

History

History
116 lines (69 loc) · 5.23 KB

shadernn.adoc

File metadata and controls

116 lines (69 loc) · 5.23 KB
Raw

Project Proposal Template

  • Name of project — or proposed name (must be unique within LF AI) :

    ShaderNN – A mobile inference framework for convolutional neural network

  • Requested project maturity level:

    Sandbox

  • Project description (what it does, why it is valuable, origin and history, ongoing development) :

    ShaderNN is a lightweight deep learning inference framework optimized for Convolutional Neural Networks. It provides high-performance inference for deep learning applications in image and graphics process on mobile devices. It uses texture-based input/output, which provides an efficient, zero-copy integration with real-time graphics pipeline or image processing applications, thereby saves expensive data transfers & format conversion between CPU and GPU which is critical for real time application on mobile platforms. It is built on native OpenGL ES (ShaderNN 1.0 ) and Vulkan (ShaderNN 2.0) graphics stack, which can be easily integrated with the graphics rendering pipeline to maximize the use of computing resources, suits for rendering, image/video and game AI applications.

  • Statement on alignment with LF AI’s mission :

    ShaderNN falls within the mission of Deep Learning. This project provides a light weight inference framework for mobile image and graphics processing applications.

  • Have you identified possible collaboration opportunities with current LF AI hosted projects (https://lfai.foundation/projects/)? Please explain :

    We are working on collaborating possible LF AI hosted projects and will update it once they are finalized.

  • License name, version, and URL to license text :

    Apache-2.0 License, https://github.com/inferenceengine/shadernn/blob/dev_opensource_release/LICENSE

  • Source control (GitHub, etc.) - Please confirm tools in use :

    We use Github as source control

  • Does the project sits in its own GH organization? :

    We are working on tranform it to a GH organization

  • Do you have the GH DCO app active in the repos? :

    No

  • Issue tracker (GitHub, JIRA, etc) - Please confirm tools in use :

    We use Github issue tracker.

  • Collaboration tools (mailing lists, wiki, IRC, Slack, Glitter, etc.) - Please confirm tools in use and state request for tools you’d like to use :

    None

  • External dependencies including licenses (name and version) of those dependencies:

    • ncnn (BSD 3-Clause License)

    • MNN (Apache 2.0)

    • opencv (Apache 2.0)

    • libyuv (BSD 3-Clause License)

    • libjpeg (Covered three compatible BSD-style open source licenses: The IJG (Independent JPEG Group) License, The modified BSD-3 Clause License, The zlib License)

    • cli11 (BSD 3-Clause License)

    • cmrc (MIT License)

    • eigen (MPL2 License)

    • glad (MIT License)

    • glfw (Zlib License)

    • glm (MIT License)

    • picojson (BSD 2-Clause License)

    • stb_image (MIT License)

    • readerwriterqueue (Simplified BSD License)

    • uVkCompute (Apache 2.0)

    • volk (MIT License)

    • Tensorflow (Apache 2.0)

    • PyTorch (BSD 3-Clause License)

    • ONNX (Apache 2.0)

    • Netron (MIT License)

  • Initial committers (name, email, organization) and how long have they been working on project?

  • Have the project defined the roles of contributor, committer, maintainer, etc.? Please document it in MAINTAINERS.md.

    Project roles have not been formally defined yet.  Those info will be updated according OPPO’s policy after the project is approved.

  • Total number of contributors to the project including their affiliations :

    We just start this open source project. Right now, we collaborate with Prof Xiaogang Jin from Zhejiang University to optimize the MNSS (Neural Supersampling Framework for Real-Time Rendering on Mobile Devices) model. His team will be one of contributors.

  • Does the project have a release methodology? Please document it in RELEASES.md :

    Not fully fledged yet. Currently it is released with a git tag. A RELEASES.md will be provided upon project approval and guidance.

  • Does the project have a code of conduct? If yes, please share the URL. If no, please created CODE_OF_CONDUCT.md and point to https://lfprojects.org/policies/code-of-conduct/. You can use [email protected] as email for contact on this topic.

    Yes, https://github.com/inferenceengine/shadernn/blob/dev_opensource_release/CODE_OF_CONDUCT.md

  • Did the project achieve any of the CII best practices badges? A different badge is required depending on the requested incubation level :

    No

  • Do you have any specific infrastructure requests needed as part of hosting the project in the LF AI? :

    None

  • Project website - Do you have a web site? If no, did you reserve a domain, and would like you to have a website created? :

    No

  • Project governance - Do you have a working governance model for the project? Please provide URL to where it is documented, typically GOVERNANCE.md:

    No

  • Social media accounts - Do you have any Twitter/LinkedIn/Facebook/etc. project accounts? Please provide pointers:

    No

  • Existing sponsorship (e.g., whether any organization has provided funding or other support to date, and a description of that support), if any :

    No