[Disco] Add loader for presharded params. #15957
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
|
This PR was developed in collaboration with @csullivan, and is based on #15676. |
This was referenced Oct 20, 2023
Lunderberg
force-pushed
the
disco_load_presharded_params
branch
from
e68fb67 to
9f019ca
Compare
|
Rebased onto main to re-run CI, as 2-week-old CI results are a bit stale for my preferences. @junrushao Could I get a review on this PR? |
Prior to this commit, sharding of model weights was always performed when initializing the model. This could cause slow initialization, especially for larger numbers of GPUs, as all model weights are initially transferred to GPU-0, before being scattered to all workers. This commit updates the `tvm::runtime::ShardLoaderObj` to also allow loading of pre-sharded model weights. With pre-sharded model weights, the tensors are sharded while the model is being built, and each worker independently loads the specific model weights that it requires.
Lunderberg
force-pushed
the
disco_load_presharded_params
branch
from
9f019ca to
8a62451
Compare
masahi
approved these changes
Nov 7, 2023
There was a problem hiding this comment.
I confirmed that it works. @junrushao Any concerns in merging this?
junrushao
reviewed
Nov 8, 2023
junrushao
reviewed
Nov 8, 2023
junrushao
reviewed
Nov 8, 2023
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Outdated suggestions cannot be applied.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.
Prior to this commit, sharding of model weights was always performed when initializing the model. This could cause slow initialization, especially for larger numbers of GPUs, as all model weights are initially transferred to GPU-0, before being scattered to all workers.
This commit updates the
tvm::runtime::ShardLoaderObjto also allow loading of pre-sharded model weights. With pre-sharded model weights, the tensors are sharded while the model is being built, and each worker independently loads the specific model weights that it requires.