PyTorch 基金会发布开源 AI 软件栈路线图
PyTorch 基金会想让你在任意芯片、任意云上都能跑模型，旗下六个项目贡献者一年涨到 2.4 倍，开源栈的野心不小。
PyTorch 基金会执行总监 Mark Collier 在 #PyTorchCon 主题演讲中宣布，正在构建一套开源软件，用于在任意芯片和任意云上训练、优化、部署和运行模型。基金会旗下六个项目（PyTorch、vLLM、Helion、Ray、Safetensors、ExecuTorch）的贡献者数量达到一年前的 2.4 倍。所有六个项目的社区规模均实现增长。
PyTorch Foundation is building an open source software to train, optimize, serve, and run any models on any chip in any cloud.
In his #PyTorchCon keynote, Mark Collier (@sparkycollier), Executive Director of PyTorch Foundation, shared key metrics reflecting the momentum across the PyTorch Foundation ecosystem.
Every project grew its community, with 2.4x as many contributors as a year ago that spans all six foundation-hosted projects: PyTorch, vLLM, Helion, Ray, Safetensors, and https://t.co/DaKyHYnfg7.
Thank you to every contributor, maintainer, and member organization driving open source AI forward together.
