Added support for MPS (Apple metal) GPU #839
Open
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
|
Thanks for submitting this. We're not going to merge the changes at this time. The changes are more involved than expected, and we don't have a way to test this ourselves. However, we'll leave this pull request up so that anyone else wanting to use MPS can see what they would need to tweak to get it working. If we get interest and feedback from more users, we may try to incorporate this in the future. |
Closed
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Outdated suggestions cannot be applied.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.
Previously, specifying "device = torch.device('mps')" would raise errors, as code handled any call to a device besides the cpu as a call to cuda. Pytorch also supports mps (metal's GPU API), although there are some limitations and the function calls are different. I coded everything so that where mps and cuda support differ, there are conditionals to catch which device is being used. This way I haven't changed any cuda-related code.