Be able to use imatrix computed with merged ffn_gate_up_exps#1419
Merged
Be able to use imatrix computed with merged ffn_gate_up_exps#1419
Conversation
Contributor
|
Thanks for holding the world together a little longer!! |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 PR is a sibling of #1418.
If one has an imatrix available that has been computed using a model with merged
ffn_gate_up_expstensors, but one has a model whereffn_up_expsandffn_gate_expsare separate, the PR allow the imatrix to be still used to quantize this model. Basically, theffn_up_exps,ffn_gate_exps, andffn_gate_up_expstensors "see" exactly the same activations, so one can use the imatrix data forffn_gate_up_expsalso forffn_up_expsandffn_gate_exps. Correspondingly, also the reverse case is now supported. I.e., one has an imatrix computed with separateffn_up_expsandffn_gate_expstensors, but now one wants to use it to quantize a model with mergedffn_gate_up_expstensors. This will also work with this PR.