Adding distributed pipeline parallelism example #749
Merged
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
mrshenli
changed the title
osalpekar
approved these changes
Apr 14, 2020
| labels = torch.zeros(batch_size, num_classes) \ | ||
| .scatter_(1, one_hot_indices, 1) | ||
|
|
||
| with dist_autograd.context() as context_id: |
There was a problem hiding this comment.
Should we add a short comment about what dist_autograd/dist_optimizer is doing here?
|
|
||
| return nn.Sequential(*layers) | ||
|
|
||
|
|
There was a problem hiding this comment.
Maybe we should include some comments about what we're doing here at a high level (defining resnet with 2 partitions so we can place them on separate machines.) Also, should we call these Partitions or Shards instead of parts?
| Distributed Pipeline Parallel Example | ||
|
|
||
| This example shows how to distribute a ResNet50 model on two RPC workers and | ||
| then implement distributed pipeline parallelism using RPC. |
There was a problem hiding this comment.
Should we include a quick description of the pipelining strategy (pipelining micro-batches within a batch and then synchronously running the optimizer step)? Since this is like GPipe, should we also link the paper here?
YinZhengxun
pushed a commit
to YinZhengxun/mt-exercise-02
that referenced
this pull request
Mar 30, 2025
Co-authored-by: Shen Li <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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Outdated suggestions cannot be applied.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.
This example shows how to use RPC to implement pipeline parallelism. This can be viewed as a distributed version of single machine multiple GPU pipeline parallelism.
The numbers below show how the total execution time decreases with the increase of
num_split.