search

argmaxinc / WhisperKit

On-device Speech Recognition for Apple Silicon
https://takeargmax.com/blog/whisperkit
MIT License
3.73k stars 315 forks source link

Output of JSON and SRT are located at both audio file path and --report-path #211

Closed maxlund closed 2 weeks ago

maxlund commented 2 weeks ago

Hi!

When transcribing with e.g. 

./whisperkit-cli transcribe --skip-special-tokens --report --report-path '/Users/maxlund/whisper-reports' --chunking-strategy 'vad' --model 'large-v3_turbo_954MB' --audio-path '/Users/maxlund/audio/arthur-fx2.wav' --language 'en'

We get a JSON and SRT file as output in both the specified --model-path /Users/maxlund/whisper-reports, as well as next to the source audio file, i.e. at --audio-path '/Users/maxlund/audio/arthur-fx2.wav'.

We would expect it to only output files at the specified --report-path right?

latenitefilms commented 2 weeks ago

Ummm... I can't reproduce:

image
maxlund commented 2 weeks ago

Yea I'm an idiot, closing this.