Update gemma_backbone.py for sharding config. #1491
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
|
PTAL again. |
mattdangerw
approved these changes
Mar 12, 2024
|
Ack, I will leave the PR here and feel free to merge it when ready. |
abuelnasr0
pushed a commit
to abuelnasr0/keras-nlp
that referenced
this pull request
Apr 2, 2024
* Update gemma_backbone.py for sharding config. * Update unit test and fix format. * Update sharding spec for gemma based on gemma training.
This was referenced Jan 3, 2025
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Outdated suggestions cannot be applied.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.
This is trying to address the #1464.
The new setting is based on the Gemma training script internally.
Here is some perf benchmark on TPU v3-8:
(Smaller value are better)
===================
base line (current setting):
generate: 1342 ms per 100 token
finetune with lora: 125ms/step
=====================
This PR setting
generate: 1245 ms per 100 token
finetune with lora: 64ms/step