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.

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

Listen wip #10

Merged
merged 11 commits into from
Mar 9, 2024
Merged

Listen wip #10

merged 11 commits into from
Mar 9, 2024

Conversation

scosman
Copy link
Owner

@scosman scosman commented Dec 10, 2023

No description provided.

Upgrading whisper and breaking everything next!
- supress_non_speach_tokens && suppress_blank
- enable metal via -DGGML_USE_METAL. This was big speedup. 77% faster on my iPhone 14 Pro.
- Moved to small distil model as default. Really fast now. 22% CPU, offloading to GPU via metal.
- Release scheme for debugging. Needed for working on the whisper parts.
…s. No longer exponential increase to transcode time as recording gets longer!

Still need to
 - figure out how to make long running transciption. The transcoder only outputs last 15s, so we lose old content
 - move the float16 audio buffer into a circular buffer, never ending.
…transcribe buffer. This can now run forever with fixed memory usage.

Remove the time limit. Records until stopped.
@scosman scosman merged commit 06c6f43 into main Mar 9, 2024
2 of 3 checks passed
@scosman scosman deleted the listen-WIP branch March 9, 2024 19:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant