Results 20 comments of Rony Fadel

+1, I'm seeing this as well on node v10.18.0

I keep hitting this for my macOS builds: ``` platform :mac do desc "Push a new release build to the App Store" lane :release do build_app(workspace: "VidCap.xcworkspace", scheme: "VidCap", catalyst_platform:...

It doesn't crash if I run `./bin/main -m ../models/ggml-large-v2.bin -f ../samples/gb0.wav --no-gpu` but it does in Xcode. Hmmm 🤔.

Can you give more info please? As is, the info provided is pretty mediocre.

There's been discussions in openai/whisper where you could skew the model to output shorter segments by tweaking `max_text_token_logprob`: https://github.com/openai/whisper/discussions/435#discussioncomment-4010615 Is something similar with the codebase in faster-whisper?

I just saw the addition of `length_penalty` today. How should it be used? Its default value is set to 1.

@guillaumekln from my testing, I've also had great results using the `token_timestamps` flag [here](https://github.com/ggerganov/whisper.cpp/blob/0d229163bbea769c7a3e0e500e45850c9a6e2e42/whisper.h#L276) Tbh, I don't know what CTranslate2 does to the underlying model, and if such capabilities are...

Please provide a code example or a working project