Remove all .cuda() calls
#61
Merged
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
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Outdated suggestions cannot be applied.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.
As correctly pointed out in #58, there are some parts in the codebase where
.cuda()is called. This causes errors when users want to run the benchmark on machines with multiple GPUs, and the device passed as an argument tobenchmarkis not used as expected.This PR solves this issue by using some PyTorch APIs. In particular,
DMWideResNetandDMPreActResNetmodels I savemeanandstdwithregister_buffer, so that the tensors are automatically moved to the same device as the overall model.Chen2020AdversarialNet,Diffenderfer2021CARD_Deck,Diffenderfer2021CARD_Deck_Binary), it is enough to save the sub-models usingnn.ModuleList, so that the sub-models are moved to the same device as the overall model.The overall API remains unchanged, and users can continue specifying the device to use for the benchmark by passing the
deviceargument tobenchmark.I have run the exact clean accuracy tests on all the models, and they pass.