Important
xla-das 是面向 DAS 架构的 xla 下游适配发行版，
基于 ROCm/xla 的 76282465a00b54ddcef59a2ee95412cd5ecf5551 基线构建并集成 DAS 支持。
本项目不是 xla 官方发行版，而是基于 ROCm/xla 二次开发；ROCm xla 源自 openxla/xla 二次开发。
- 目标架构： xla
- 上游分支： rocm-jaxlib-v0.10.0
- 上游版本： 76282465a00b54ddcef59a2ee95412cd5ecf5551
- 上游基线：
https://github.com/ROCm/xla/tree/76282465a00b54ddcef59a2ee95412cd5ecf5551 - 上游许可： Apache-2.0
- 源码编译： BUILDING.md
- 已知问题： BUILDING.md 中的已知问题
- Python 发布包名：
xla-das - Python 导入名：
xla
xla-das与上游xla会安装同名xla模块， 请勿在同一虚拟环境或容器中混装。
XLA (Accelerated Linear Algebra) is an open-source machine learning (ML) compiler for GPUs, CPUs, and ML accelerators.
The XLA compiler takes models from popular ML frameworks such as PyTorch, TensorFlow, and JAX, and optimizes them for high-performance execution across different hardware platforms including GPUs, CPUs, and ML accelerators.
openxla.org is the project's website.
If you want to use XLA to compile your ML project, refer to the corresponding documentation for your ML framework:
If you're not contributing code to the XLA compiler, you don't need to clone and build this repo. Everything here is intended for XLA contributors who want to develop the compiler and XLA integrators who want to debug or add support for ML frontends and hardware backends.
If you'd like to contribute to XLA, review How to Contribute and then see the developer guide.
- For questions, contact the maintainers - maintainers at openxla.org
While under TensorFlow governance, all community spaces for SIG OpenXLA are subject to the TensorFlow Code of Conduct.
This repository is licensed under the Apache License 2.0. See LICENSE. Some files are adapted from xla and other third-party sources; see Third-Party Notices for source and modification notices.
HCU adaptations, modifications, and original contributions by Hygon Information Technology Co., Ltd. are licensed under the Apache License, Version 2.0.
Modified by Hygon Information Technology Co., Ltd.