fix: apply temperature scaling to logits before softmax in sampler#161
Open
zhuxiaoxuhit wants to merge 1 commit intoMoonshotAI:masterfrom
Open
fix: apply temperature scaling to logits before softmax in sampler#161zhuxiaoxuhit wants to merge 1 commit intoMoonshotAI:masterfrom
zhuxiaoxuhit wants to merge 1 commit intoMoonshotAI:masterfrom
Conversation
Temperature sampling requires dividing raw logits by T before softmax, i.e. softmax(logits / T). The previous code applied temperature after log_softmax (logprobs / T), which computes softmax(logits)^(1/T) — a mathematically different distribution whose denominator is (Σ exp(logit_j))^(1/T) instead of Σ exp(logit_j / T). This produces incorrect sampling distributions whenever temperature > 0 is used. Fix applied to both sample_audio_logits and sample_text_logits.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Temperature sampling requires dividing raw logits by T before softmax, i.e. softmax(logits / T). The previous code applied temperature after log_softmax (logprobs / T), which computes softmax(logits)^(1/T) — a mathematically different distribution whose denominator is (Σ exp(logit_j))^(1/T) instead of Σ exp(logit_j / T). This produces incorrect sampling distributions whenever temperature > 0 is used.
Fix applied to both sample_audio_logits and sample_text_logits.