pyannote-audio icon indicating copy to clipboard operation
pyannote-audio copied to clipboard

Published 20 hours ago • verified publisher icon pyannote

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

Open hbredin opened this issue 2 years ago • 4 comments 

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?

hbredin avatar Dec 13 '22 22:12 hbredin

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] avatar Jun 12 '23 06:06 stale[bot]

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] avatar Dec 09 '23 22:12 stale[bot]

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

metalgearsloth avatar Dec 14 '23 05:12 metalgearsloth

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] avatar Jun 13 '24 23:06 stale[bot]