Skip to content

perf(whisper): remove redundant timestamp normalization #2

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
unohee merged 1 commit into from
Aug 21, 2026
Merged

perf(whisper): remove redundant timestamp normalization #2

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 10 additions & 7 deletions whisper/mlx_whisper/decoding.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -392,16 +392,19 @@ def apply(self, logits: mx.array, tokens: mx.array) -> mx.array:
)
mask = mask | (vocab > last_allowed)

# if sum of probability over timestamps is above any other token, sample timestamp
logprobs = logits - mx.logsumexp(logits, axis=-1, keepdims=True)
timestamp_logprob = logprobs[:, self.tokenizer.timestamp_begin :].logsumexp(
axis=-1, keepdims=True
# The common normalization constant cancels in this comparison. Avoid a
# redundant full-vocabulary logsumexp and subtraction on every decode step.
timestamp_logit = mx.logsumexp(
logits[:, self.tokenizer.timestamp_begin :],
axis=-1,
keepdims=True,
)
max_text_token_logprob = logprobs[:, : self.tokenizer.timestamp_begin].max(
axis=-1, keepdims=True
max_text_token_logit = logits[:, : self.tokenizer.timestamp_begin].max(
axis=-1,
keepdims=True,
)
mask = mask | (
(timestamp_logprob > max_text_token_logprob)
(timestamp_logit > max_text_token_logit)
& (vocab < self.tokenizer.timestamp_begin)
)
return logits + mx.where(mask, -mx.inf, 0.0)
Expand Down