Install the dependencies the first time run the benchmark. #694
Conversation
This file contains hidden or 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
|
@xuzhao9 has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
facebook-github-bot
pushed a commit
to pytorch/pytorch
that referenced
this pull request
Jan 17, 2022
Summary: This PR removes the PyTorch nightly dependencies of TorchBench CI. Instead, it relies on the bisection script to install TorchBench dependencies (pytorch/benchmark#694). This will unblock TorchBench CI users when the nightly build fails (e.g., #71260) RUN_TORCHBENCH: resnet18 TORCHBENCH_BRANCH: xz9/optimize-bisection Pull Request resolved: #71323 Reviewed By: wconstab Differential Revision: D33591713 Pulled By: xuzhao9 fbshipit-source-id: f1308ea33ece1f18196c993b40978351160ccc0c
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.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.
Strictly speaking, the bisection workflow shouldn't depend on PyTorch nightly: instead, it should install the TorchBench deps after the first time it compiles PyTorch and its domain packages.
This PR does this optimization and removes PyTorch nightly dependencies for bisection jobs. So the bisection jobs and OnDemand CI jobs can proceed even if the nightly packages are unavailable.
The OnDemand PR CI link is here: pytorch/pytorch#71323. It shows we can run PR CI even without installing nightly packages.