wgpu-native-installer

Install wgpu-native

Introduction

Install the GitHub assets from gfx-rs/wgpu-native.

The installer has additional options to customize the installation.

Installation

Choose one of the following methods:

1. Install the package from PyPI:

   Website link

   pip install wgpu-native-installer
   

2. Install the package from GitHub:

   Website link

   curl -LO $ASSET_URL_ends_with_whl
   pip install *.whl
   

Usage

Choose one of the following methods. It depends on your requirements.

1. Download the artifacts of wgpu-native on your machine:

   python -m wgpu_native_installer -v
   ls -l
   # c++ -L. -lwgpu_native ....
   

2. Import the CLI as a package:

   from asyncio import run
   from wgpu_native_installer import main as install_wgpu_native
   
   if __name__ == "__main__":
       run(install_wgpu_native(["-vv", "--library", "dynamic", "static"]))
   

   from asyncio import run
   from wgpu_native_installer.github_release import parse_release_latest
   
   async def amain():
       release = await parse_release_latest("pytorch", "pytorch")
       asset = release.search_assets(r".*\.tar\.gz")[0]
       async for _ in asset.download("pytorch-latest.tar.gz", show_progress=True):
           pass
   
   if __name__ == "__main__":
       run(amain())
   

3. Clone the repository and reuse the codes you like:

   git clone https://github.com/AsherJingkongChen/wgpu-native-installer.git
   

Details

Goals

As one of my Python package templates, it should be TSPC-compliant.

• Tested
• Simple
• Performant
• Customizable

License

MIT