use DataCollatorWithFlattening when not sample packing #2167
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
ran a quick experiment against l3-3B with microbatch size of 8. losses stack on top pretty closely, but flattening is good bit faster.
|
|
I should make this have an auto option that automatically enables this when not using packing with micro batch size > 1 |
winglian
force-pushed
the
flattening-collator
branch
from
503ead8 to
44f22d2
Compare
winglian
force-pushed
the
flattening-collator
branch
from
73a0e9d to
838aae1
Compare
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Outdated suggestions cannot be applied.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.
https://research.ibm.com/blog/hugging-face-training-flash-attention