Add option to use hardware friendly 'softermax' #1
Conversation
This file contains hidden or 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 implements softermax and provides it as an option within the config. Link for softermax paper: https://arxiv.org/abs/2103.09301 Setting the default to still use conventional softmax, since softermax does is not yet incorporated into faster attention like flashattention.
msaligane
approved these changes
Oct 4, 2023
gkielian
added a commit
that referenced
this pull request
Jan 26, 2024
Add Randomness Seed Parameter PR #1
gkielian
pushed a commit
that referenced
this pull request
Mar 4, 2024
Patch: gated attn and sliding window
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.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 implements softermax and provides it as an option within the model.py config.
Link for softermax paper:
https://arxiv.org/abs/2103.09301
Setting the default to still use conventional softmax which trains a little faster, since softermax does is not yet incorporated into the faster attention libraries.