pytorch/.github/scripts
Andrey Talman 7275f28045 Fix cuda 12.9 aarch64 GPU builds. Update CUDA_STABLE variable. (#157630)
This contains 2 fixes that required in main and will need to be cherry-picked to Release 2.8 branch:
1. The PR https://github.com/pytorch/pytorch/pull/155819 missed to include triton change.
2. CUDA STABLE variable needs to be set to 12.8. Updating CUDA stable updates full static build

Pull Request resolved: https://github.com/pytorch/pytorch/pull/157630
Approved by: https://github.com/Skylion007, https://github.com/jeanschmidt
2025-07-04 18:08:31 +00:00
..
amd [BE] fix typos in .ci/, .circleci/, .github/ (#156069) 2025-06-17 09:43:14 +00:00
s390x-ci s390x: use qemu issue workaround for runner registration too (#154030) 2025-05-21 18:30:25 +00:00
windows Revert "Simplify nvtx3 CMake handling, always use nvtx3 (#153784)" 2025-06-24 20:02:07 +00:00
build_triton_wheel.py [CI] Fix triton version split issue (#155670) 2025-06-17 12:42:40 +00:00
check_labels.py [ez] Fix check labels error when deleting comment (#140578) 2024-11-13 23:00:58 +00:00
cherry_pick.py PEP585: .github (#145707) 2025-01-27 21:21:01 +00:00
close_nonexistent_disable_issues.py Fix workflow for closing nonexistent disable issues (#146447) 2025-02-05 22:29:05 +00:00
collect_ciflow_labels.py PEP585: .github (#145707) 2025-01-27 21:21:01 +00:00
comment_on_pr.py
convert_lintrunner_annotations_to_github.py [BE][Easy][2/19] enforce style for empty lines in import segments in .ci/ and .github/ (#129753) 2024-07-16 09:40:00 +00:00
delete_old_branches.py [ez][BE] Tag deletion script to delete any old ciflow + autorevert tags (#157468) 2025-07-02 20:42:32 +00:00
docathon-label-sync.py Update GH action to use the correct label (#154126) 2025-05-23 00:29:43 +00:00
drci_mocks.json.gz Skip signals from older runs of the same workflows (#129291) 2024-06-26 03:49:09 +00:00
ensure_actions_will_cancel.py Revert "Use absolute path path.resolve() -> path.absolute() (#129409)" 2025-01-04 14:17:20 +00:00
export_pytorch_labels.py [BE][Easy][2/19] enforce style for empty lines in import segments in .ci/ and .github/ (#129753) 2024-07-16 09:40:00 +00:00
file_io_utils.py PEP585: .github (#145707) 2025-01-27 21:21:01 +00:00
filter_test_configs.py [ez][CI] Print set output in CI (#157477) 2025-07-02 21:47:19 +00:00
generate_binary_build_matrix.py Fix cuda 12.9 aarch64 GPU builds. Update CUDA_STABLE variable. (#157630) 2025-07-04 18:08:31 +00:00
generate_ci_workflows.py Enable manywheel build and smoke test on main branch for ROCm (#153287) 2025-06-14 19:14:31 +00:00
generate_docker_release_matrix.py PEP585: .github release triggers (#145708) 2025-02-04 01:02:46 +00:00
generate_pytorch_version.py [BE][Easy][2/19] enforce style for empty lines in import segments in .ci/ and .github/ (#129753) 2024-07-16 09:40:00 +00:00
get_aws_session_tokens.py
get_ci_variable.py Automate stable CUDA update and linter using min Python verison (#148912) 2025-03-12 18:12:34 +00:00
get_workflow_job_id.py [ez][CI] Print set output in CI (#157477) 2025-07-02 21:47:19 +00:00
github_utils.py Fix broken URLs (#152237) 2025-04-27 09:56:42 +00:00
gitutils.py [BE] fix typos in .ci/, .circleci/, .github/ (#156069) 2025-06-17 09:43:14 +00:00
gql_mocks.json.gz [trymerge] Error on ghstack commit with multiple PRs (#154941) 2025-06-11 20:26:50 +00:00
kill_active_ssh_sessions.ps1
label_utils.py Fix more URLs (#153277) 2025-05-14 16:23:50 +00:00
lint_native_functions.py Revert "Use absolute path path.resolve() -> path.absolute() (#129409)" 2025-01-04 14:17:20 +00:00
lintrunner.sh [BE][CI][Easy] Run lintrunner on generated .pyi stub files (#150732) 2025-05-27 14:58:02 +00:00
parse_ref.py [ez][CI] Print set output in CI (#157477) 2025-07-02 21:47:19 +00:00
pr-sanity-check.sh [BE] fix typos in .ci/, .circleci/, .github/ (#156069) 2025-06-17 09:43:14 +00:00
pytest_cache.py [BE][Easy][2/19] enforce style for empty lines in import segments in .ci/ and .github/ (#129753) 2024-07-16 09:40:00 +00:00
pytest_caching_utils.py Enable ruff rule S324 (#147665) 2025-02-25 18:27:34 +00:00
README.md
report_git_status.sh Fix timeout check workflow lint job (#142476) 2024-12-10 20:47:22 +00:00
rockset_mocks.json.gz
runner_determinator.py Revert "[BE]: Enable RUFF TRY400 rule - log.exception (#153473)" 2025-05-16 08:29:26 +00:00
stop_runner_service.sh
td_llm_indexer.sh [TD] Pin numpy to 1.26.0 in indexer (#130442) 2024-07-10 20:52:24 +00:00
test_check_labels.py PEP585: .github (#145707) 2025-01-27 21:21:01 +00:00
test_delete_old_branches.py [ez][BE] Tag deletion script to delete any old ciflow + autorevert tags (#157468) 2025-07-02 20:42:32 +00:00
test_filter_test_configs.py [BE] fix typos in .ci/, .circleci/, .github/ (#156069) 2025-06-17 09:43:14 +00:00
test_gitutils.py Revert "Use absolute path path.resolve() -> path.absolute() (#129409)" 2025-01-04 14:17:20 +00:00
test_label_utils.py
test_pytest_caching_utils.py
test_runner_determinator.py Enable opting out of experiments even when they're being rolled out (#140433) 2024-11-14 19:18:24 +00:00
test_trymerge.py [BE] fix typos in .ci/, .circleci/, .github/ (#156069) 2025-06-17 09:43:14 +00:00
test_tryrebase.py
trymerge_explainer.py [BE][CI] bump ruff to 0.9.0: string quote styles (#144569) 2025-02-24 19:56:09 +00:00
trymerge.py [BE] fix typos in .ci/, .circleci/, .github/ (#156069) 2025-06-17 09:43:14 +00:00
tryrebase.py Fix more URLs (#153277) 2025-05-14 16:23:50 +00:00
update_runner_determinator.py [BE] Add script to keept the runner-determinator scripts in sync (#136794) 2024-10-01 22:26:28 +00:00
upload_aws_ossci.sh Build magma binary tarballs for various cuda (#139888) 2024-11-08 13:28:27 +00:00
wait_for_ssh_to_drain.ps1

pytorch/.github

NOTE: This README contains information for the .github directory but cannot be located there because it will overwrite the repo README.

This directory contains workflows and scripts to support our CI infrastructure that runs on GitHub Actions.

Workflows

  • Pull CI (pull.yml) is run on PRs and on main.
  • Trunk CI (trunk.yml) is run on trunk to validate incoming commits. Trunk jobs are usually more expensive to run so we do not run them on PRs unless specified.
  • Scheduled CI (periodic.yml) is a subset of trunk CI that is run every few hours on main.
  • Binary CI is run to package binaries for distribution for all platforms.

Templates

Templates written in Jinja are located in the .github/templates directory and used to generate workflow files for binary jobs found in the .github/workflows/ directory. These are also a couple of utility templates used to discern common utilities that can be used amongst different templates.

(Re)Generating workflow files

You will need jinja2 in order to regenerate the workflow files which can be installed using:

pip install -r .github/requirements/regenerate-requirements.txt

Workflows can be generated / regenerated using the following command:

.github/regenerate.sh

Adding a new generated binary workflow

New generated binary workflows can be added in the .github/scripts/generate_ci_workflows.py script. You can reference examples from that script in order to add the workflow to the stream that is relevant to what you particularly care about.

Different parameters can be used to achieve different goals, i.e. running jobs on a cron, running only on trunk, etc.

ciflow (trunk)

The label ciflow/trunk can be used to run trunk only workflows. This is especially useful if trying to re-land a PR that was reverted for failing a non-default workflow.

Infra

Currently most of our self hosted runners are hosted on AWS, for a comprehensive list of available runner types you can reference .github/scale-config.yml.

Exceptions to AWS for self hosted:

  • ROCM runners

Adding new runner types

New runner types can be added by committing changes to .github/scale-config.yml. Example: https://github.com/pytorch/pytorch/pull/70474

NOTE: New runner types can only be used once the changes to .github/scale-config.yml have made their way into the default branch

Testing pytorch/builder changes

In order to test changes to the builder scripts:

  1. Specify your builder PR's branch and repo as builder_repo and builder_branch in .github/templates/common.yml.j2.
  2. Regenerate workflow files with .github/regenerate.sh (see above).
  3. Submit fake PR to PyTorch. If changing binaries build, add an appropriate label like ciflow/binaries to trigger the builds.