Implement Cosine Addition and Smooth Add Diff #67
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
|
Currently waiting on the Tensorsum PR to go through before finalizing this PR. |
The implementation for Cosine Addition is not complete nor identical to the one from Supermerger. Currently waiting for clarification on why a certain change was made to the two new methods before replacing the current implementation with them. Smooth_Add is the same implementation as supermerger, with cupy optimizations built in if a user has that installed and working properly, otherwise it falls back to standard scipy, which is much slower due to being CPU based.
After getting an explanation behind the change, I have implemented it.
|
@mariaWitch tensor sum is in |
|
I currently have not had enough time to continue working on this, but I will work on moving the required changes over there shortly. I'll keep this change open for posterity until I have enough time to move the related changes over. |
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Outdated suggestions cannot be applied.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 Pull request implements the Smooth_Add difference and Cosine Similarity merging methods from SuperMerger with some additional changes for Smooth_add in order to provide significant speed improvements over the Supermerger implementation. There is fallback behavior coded in, in order to make sure that even if CuPy or its dependencies aren't properly installed, the merge can still continue even if it will take a significantly longer amount of time.
NOTE: The implementation for Cosine Addition is not complete nor identical to the one currently in Supermerger. Currently waiting for clarification on why a certain change was made to the two new methods before replacing the current implementation with them.