universal-ckp: support llama model #287
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
Modify universal checkpoint parameter patterns based on the specific model configuration. This commit adds support for llama family of models. Signed-off-by: Moshe Island <[email protected]>
mosheisland
requested review from
jeffra,
tjruwase,
ShadenSmith,
conglongli,
awan-10,
eltonzheng,
minjiaz,
RezaYazdaniAminabadi,
duli2012,
mrwyattii,
arashb,
xiaoxiawu-microsoft and
GuanhuaWang
as code owners
tjruwase
reviewed
Nov 13, 2023
tjruwase
reviewed
Nov 13, 2023
|
@tjruwase , now that microsoft/DeepSpeed#4666 is merged, can you please merge this PR? |
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Outdated suggestions cannot be applied.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.
Modify universal checkpoint parameter patterns based on the specific model configuration.
This commit adds support for llama family of models.
This commit depends on PR microsoft/DeepSpeed#4666 that adds support for SWIGLU linear parameter.
Following tensorboard captures training of a tiny llama model with DP=TP=PP=2 for 200 steps.
In addition, a checkpoint is taken in step 100 and converted to universal checkpoint.
Few other 3D configurations are trained from this universal checkpoint for the remaining 100 steps (till step 200).
It can be observed that both training and validation curves match for original 3D (2/2/2) and the others that load from universal.