Skip to content

Conversation

@nchudleigh
Copy link
Contributor

I added logic to dynamically adjust audio_ctx based on the input audio length.
Gives better performance with no degradation of WER in my testing.

File Duration Opt Encode (ms) Base Encode (ms) Improvement
jfk.wav ~11s ~13.5 ~19.7 ~31%
en-1-16khz.wav ~20s ~148.6 ~155.4 ~4.4%
en-2-16khz.wav ~30s ~21.0 ~20.2 ~0% (Noise)
en-2.wav ~30s ~21.1 ~21.1 ~0% (Noise)

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