mirror of
https://github.com/zebrajr/pytorch.git
synced 2025-12-06 12:20:52 +01:00
651e6aacf9
Fixes #144203. We build a custom libdrm when preparing our docker image. We attempt to locate the amdgpu.ids file relative to the python binary, but this is not possible for venv installs of pytorch when the python binary is a symlink. Not finding amdgpu.ids causes `torch.cuda.get_device_name()` to return "AMD Radeon Graphics" as a generic name instead of something specific such as "AMD Instinct MI250X / MI250". The libdrm warning is noisy, so we are removing it. Pull Request resolved: https://github.com/pytorch/pytorch/pull/147791 Approved by: https://github.com/jeffdaily |
||
|---|---|---|
| .. | ||
| almalinux | ||
| centos-rocm | ||
| ci_commit_pins | ||
| common | ||
| java | ||
| libtorch | ||
| linter | ||
| linter-cuda | ||
| manywheel | ||
| ubuntu | ||
| ubuntu-cuda | ||
| ubuntu-rocm | ||
| ubuntu-xpu | ||
| build.sh | ||
| README.md | ||
| requirements-ci.txt | ||
| requirements-docs.txt | ||
| triton_version.txt | ||
Docker images for GitHub CI and CD
This directory contains everything needed to build the Docker images that are used in our CI.
The Dockerfiles located in subdirectories are parameterized to
conditionally run build stages depending on build arguments passed to
docker build. This lets us use only a few Dockerfiles for many
images. The different configurations are identified by a freeform
string that we call a build environment. This string is persisted in
each image as the BUILD_ENVIRONMENT environment variable.
See build.sh for valid build environments (it's the giant switch).
Docker CI builds
build.sh-- dispatch script to launch all buildscommon-- scripts used to execute individual Docker build stagesubuntu-- Dockerfile for Ubuntu image for CPU build and test jobsubuntu-cuda-- Dockerfile for Ubuntu image with CUDA support for nvidia-dockerubuntu-rocm-- Dockerfile for Ubuntu image with ROCm supportubuntu-xpu-- Dockerfile for Ubuntu image with XPU support
Docker CD builds
conda- Dockerfile and build.sh to build Docker images used in nightly conda buildsmanywheel- Dockerfile and build.sh to build Docker images used in nightly manywheel buildslibtorch- Dockerfile and build.sh to build Docker images used in nightly libtorch builds
Usage
# Build a specific image
./build.sh pytorch-linux-bionic-py3.8-gcc9 -t myimage:latest
# Set flags (see build.sh) and build image
sudo bash -c 'PROTOBUF=1 ./build.sh pytorch-linux-bionic-py3.8-gcc9 -t myimage:latest