Add benchmark for training training #6933
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/6933
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ❌ 3 New FailuresAs of commit 547af2d with merge base ce77995 ( NEW FAILURES - The following jobs have failed:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D65905482 |
davidlin54
force-pushed
the
export-D65905482
branch
from
5029d74 to
6b9e7b0
Compare
Summary: This adds benchmarking for training. It loads the 3 methods required for training (param names, param index, gradient index) and runs the exported forward+backwards, then runs the optimizer. Reviewed By: JacobSzwejbka Differential Revision: D65905482
|
This pull request was exported from Phabricator. Differential Revision: D65905482 |
|
@pytorchbot label "topic: not user facing" |
Summary: This adds benchmarking for training. It loads the 3 methods required for training (param names, param index, gradient index) and runs the exported forward+backwards, then runs the optimizer. Reviewed By: JacobSzwejbka Differential Revision: D65905482
davidlin54
force-pushed
the
export-D65905482
branch
from
6b9e7b0 to
547af2d
Compare
|
This pull request was exported from Phabricator. Differential Revision: D65905482 |
| MmapUseMlockIgnoreErrors, | ||
| }; | ||
|
|
||
| struct MethodHolder { |
There was a problem hiding this comment.
can you leave it private and just make a friend class for benchmarking?
There was a problem hiding this comment.
hmm I don't want to access any data in Module (and don't even have a reference to Module, since benchmarking uses program/method), just want to use MethodHolder as a helper structure. I can also just copy and paste it into the benchmark runner code if you prefer this to be private
|
Hi @davidlin54! Thank you for your pull request. We require contributors to sign our Contributor License Agreement, and yours needs attention. You currently have a record in our system, but the CLA is no longer valid, and will need to be resubmitted. ProcessIn order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA. Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with If you have received this in error or have any questions, please contact us at cla@meta.com. Thanks! |
|
@JacobSzwejbka, @tarun292, @Gasoonjia can you please take a look? |
4 participants
Move MethodHolder to public to be reused in internal benchmarking
Reviewed By: JacobSzwejbka
Differential Revision: D65905482