Distributed Data Parallel #68
Closed
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
It is often useful to access convolutional layer attributes, e.g. for output shapes precalculation.
attribute readers for ConvNd
…t_mask Fixed generation of square subsequent mask
…ad:map_locaion, torchrun, Tests are provided
Owner
|
Hi @orlando-labs, thanks for the PR. I think most of this would be better as a separate gem for now, as I'm not in a position to support this functionality. (also, there should already be a |
Contributor
Author
|
Hi, @ankane. It's a really good idea to move this to a separate gem. However, some core functionality changes are needed to run DDP, such as improved device handling and mapping the location in |
Owner
|
Feel free to create individual PRs for those specific changes (for device handling, there's already a |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.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.
The code and its behavior mostly mimic those of PyTorch. It differs only in the multiprocessing section, where the Ruby flow differs from the Python one.
All code has been tested in a multi-GPU environment, which is not currently reproducible on GitHub CI:
We tried to maximize test coverage for every aspect of DDP communication. A benchmark and an example are also included.