search

pyannote / pyannote-audio

Neural building blocks for speaker diarization: speech activity detection, speaker change detection, overlapped speech detection, speaker embedding
http://pyannote.github.io
MIT License
5.57k stars 733 forks source link

Investigate why using large embedding batch size makes things slower... #1195

Open hbredin opened 1 year ago

hbredin commented 1 year ago 
from pyannote.audio.pipelines.utils.hook import ProgressHook
pipeline = Pipeline.from_pretrained("pyannote/speaker-diarization")

On my Mac, default pipeline.embedding_batch_size = 32 is much slower than pipeline.embedding_batch_size = 1.

Why is that?

stale[bot] commented 1 year ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

stale[bot] commented 7 months ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

metalgearsloth commented 7 months ago

See https://github.com/pyannote/pyannote-audio/issues/1580 for a potentially similar issue I had.

stale[bot] commented 1 month ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.