Sampling: Support logit_bias on the exllamav3 backend - #445
Draft
sashko-zakharchuk wants to merge 1 commit into
Draft
Sampling: Support logit_bias on the exllamav3 backend#445sashko-zakharchuk wants to merge 1 commit into
sashko-zakharchuk wants to merge 1 commit into
Conversation
Passes logit_bias through to the exllamav3 backend instead of dropping it with an unsupported-parameter warning. Adds a logit_bias() method to ExllamaV3SamplerBuilder that prepends SS_LogitBias (it must run before the logits are transformed), calls it from the exllamav3 model when a request sets a bias, and removes logit_bias from UNSUPPORTED_PARAMS. Depends on turboderp-org/exllamav3#274 (adds SS_LogitBias).
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Passes
logit_biasthrough to the exllamav3 backend, so requests that set it are honored instead of dropped with a warning (turboderp-org/exllamav3#270).Three changes: drop
logit_biasfromUNSUPPORTED_PARAMS; add alogit_bias()method toExllamaV3SamplerBuilderthat prependsSS_LogitBias(it must run before the logits are transformed); and call it from the exllamav3 model when a request sets a bias.Depends on turboderp-org/exllamav3#274, which adds
SS_LogitBias— this won't import until that PR lands or an exllamav3 release includes it, so opening as a draft.