M1 fixes cherry #6167
Merged
M1 fixes cherry #6167
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This would package libpng and libjpeg.dylib into wheel files Add a very simple test step, copied from https://github.com/pytorch/pytorch.github.io/blob/1eaa33a3d3f1b83b64c5031c6dd04dbb238d6105/scripts/test_install.py#L78 Cherry-picked from https://github.com/pytorch/builder/blob/d0bc74cc363a9da5a8b6a40e883d40d25d050036/build_m1_domains.sh#L22
* [BE] Unify version computation Instead of hardcoding dev version in various script, use one from `version.txt` if `setup_build_version` is called without arguments Also, pass `--pre` option to M1 build/test pip install commands to build TorchVision against nightly pytorch * Pin torchvision dependency to a specific pytorch version
Explicitly set PATH to point to `conda` binary, otherwise libjpeg detection logic does not work Pin libjpeg to the same version on x86 and m1 Add simple tests that jpeg can be decoded by a generated wheel
* Add M1 testing job * libjpeg -> jpeg<=9b in test-m1.yml * Added export PATH=~/miniconda3/bin... from 6122 * Tests were OK, let's see if we can remove the pinning
Clean up Conda build folder before every run
Enable artifact upload to GitHub for every workflow run, but upload to Conda/S3 only on nightly pushes
Test plan: `conda install -c pytorch-nightly torchvision; python -c "import torchvision;print(torchvision.io.read_image('hummingbird.jpg').shape)"`
By passing `--pre` option to `pip install`, otherwise torchvision were always tested against last PyTorch release
* Adding tagged builds * Testing * Testing * Testing * Testing * Adding conda builds
Replace `steps.extract_branch.outputs.branch` (which were probably taken from https://stackoverflow.com/questions/58033366/how-to-get-the-current-branch-within-github-actions ) with straightforward `github.event.ref` Tested in https://github.com/malfet/deleteme/runs/6822647720?check_suite_focus=true and https://github.com/malfet/deleteme/runs/6822691158?check_suite_focus=true
s#ref/heads/#refs/heads/# I should have noticed it while copy-n-pasting the condition. Unfortunately there are no way to test is other than in prod, but nightly builds are still not getting pushed, see https://github.com/pytorch/vision/runs/6860407007?check_suite_focus=true for example
|
As discussed offline, this is a potential duplicate of #6112 |
datumbox
approved these changes
Jun 15, 2022
|
Thanks @atalman |
NicolasHug
approved these changes
Jun 15, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
M1 cherry picks creating for CI signal