Skip to content
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

Jump to bottom

Resolve Inference Selection Bug Affecting Transcription Quality #1377

Open
wants to merge 3 commits into
base: main
Choose a base branch
from
Open

Resolve Inference Selection Bug Affecting Transcription Quality #1377

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
a72a044
Return best option on fallback
TheoBoyer May 21, 2023
f677284
Return the best only if all fallbacks failed
TheoBoyer May 21, 2023
bf2612f
Merge branch 'main' into return_best_text
TheoBoyer Dec 18, 2023
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
6 changes: 6 additions & 0 deletions whisper/transcribe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -155,6 +155,7 @@ def decode_with_fallback(segment: torch.Tensor) -> DecodingResult:
[temperature] if isinstance(temperature, (int, float)) else temperature
)
decode_result = None
results = {}

for t in temperatures:
kwargs = {**decode_options}
Expand All @@ -169,6 +170,8 @@ def decode_with_fallback(segment: torch.Tensor) -> DecodingResult:
options = DecodingOptions(**kwargs, temperature=t)
decode_result = model.decode(segment, options)

results[t] = decode_result

needs_fallback = False
if (
compression_ratio_threshold is not None
Expand All @@ -187,6 +190,9 @@ def decode_with_fallback(segment: torch.Tensor) -> DecodingResult:
needs_fallback = False # silence
if not needs_fallback:
break
else:
# all failed
return max(results.values(), key=lambda r: r.avg_logprob)

return decode_result

Expand Down