mizoru

Ah, it is fastaudio that requires an older torchaudio version. So to recreate the issue you `pip install fastaudio` first.

Sorry, I'm a bit in over my head on how to contribute this properly right now, but: This should work ```python @staticmethod def _set_return_timestamps(return_timestamps, is_shortform, generation_config): if return_timestamps is None...

Really want for this to be created

The patch to fix this should be very easy to make. [This browser extension](https://chrome.google.com/webstore/detail/ruavatar-youtube-by-lazyh/lbmdmkocbkkedjcflldfbfnjkbdabmna) just readdresses from [yt3.ggpht.com](https://yt3.ggpht.com/) to [yt4.ggpht.com](https://yt4.ggpht.com/), and it works, e.g. from [yt3](https://yt3.ggpht.com//F02Qk_T2Hlip9e8MfUkuy_0yE2YYj_HK9O086UTnySiuUvM4uH8ZlEepgXvpsjY9NGjvvL-Pd_ID_Q=s640-nd-v1) to [yt4](https://yt4.ggpht.com//F02Qk_T2Hlip9e8MfUkuy_0yE2YYj_HK9O086UTnySiuUvM4uH8ZlEepgXvpsjY9NGjvvL-Pd_ID_Q=s640-nd-v1), the link...

This never materialized, right?

@tarun7r Did you have any success? Currently trying to do batch inference as well

I fine-tuned whisper using DoRa through huggingface peft, I was able to then just merge the DoRa weights into the original and treat it as any old regular fine-tune. "We...

"The library integrates the [Silero VAD](https://github.com/snakers4/silero-vad) model to filter out parts of the audio without speech" This does exactly what you're asking