search

pytorch / opacus

Training PyTorch models with differential privacy
https://opacus.ai
Apache License 2.0
1.68k stars 334 forks source link

disable padding=same for experimental hook #498

Closed ffuuugor closed 2 years ago

ffuuugor commented 2 years ago

The tests are failing because our experimental convolution as a backward GSM can't handle padding="same"

Since this grad sampler is not used by default and mostly exists for aesthetic reasons, I think disabling the test is the best course of action.

facebook-github-bot commented 2 years ago

@ffuuugor has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

facebook-github-bot commented 2 years ago

@ffuuugor has updated the pull request. You must reimport the pull request before landing.

facebook-github-bot commented 2 years ago

@facebook-github-bot has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.