[PP] Fix PP meta init #582
Open
+9
−8
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
wconstab
added a commit
that referenced
this pull request
Sep 18, 2024
Uses meta device for tensors/model used before pipeline splitting. *Important:* Relies on pytorch/pytorch#136243 to make PipelineStage avoid materializing the model and the input/output buffers eagerly. Relies on existing .to(device) in train.py to finally materialize the model. ghstack-source-id: 66fa9f1f78dff0b1af753dc4b2afcc09d897751d Pull Request resolved: #582
facebook-github-bot
added
the
CLA Signed
tianyu-l
reviewed
Sep 18, 2024
|
|
||
| model.to_empty(device=device) | ||
| stage = PipelineStage( |
There was a problem hiding this comment.
just to understand: the device arg for PipelineStage still needs to be the actual device, e.g. cuda, correct?
There was a problem hiding this comment.
Correct. And I want to remove that too in PipelineStage but I didn't do it yet.
Curious -- why would train.py make a .to call? Should init_weight create tensor on the right device directly? Or, if we are loading from DCP, would DCP return a state dict with DTensors on target device or just a state dict with DTensors on CPU? |
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Outdated suggestions cannot be applied.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.
Stack from ghstack (oldest at bottom):
Uses meta device for tensors/model used before pipeline splitting.
Important:
Relies on pytorch/pytorch#136243 to make PipelineStage avoid
materializing the model and the input/output buffers eagerly.
Relies on existing .to(device) in train.py to finally materialize the
model.