Yoach Lacombe

Hey @ravenouse, of course!

Hey @ravenouse, thanks for updating the code and for the Colab, I'll take a look ASAP!

Hey @hungdinhxuan, note that you need to use the wav2vec2 feature extractor before passing your input to the model. Find an example on the docs [here](https://huggingface.co/docs/transformers/model_doc/wav2vec2#transformers.Wav2Vec2ForPreTraining.forward.example)! You can notice [here](https://github.com/huggingface/transformers/blob/539ed75d501f9af87b28dc0325b8d2f62b791cd9/src/transformers/models/wav2vec2/feature_extraction_wav2vec2.py#L227-L235)...

Hey @sourabharsh, I'd be happy to help you understand how to do that. Have you checked the difference between the input `history_prompt` that you get with the processor and the...

Thanks for opening the issue! Opened a PR to fix this!

I want to add a test, but I realized most of the slow tests of Whisper were already failing, independently of this PR. cc @sanchit-gandhi

I do, thanks !

Hey @monirome, thanks for opening this issue! seems like it might be related with your checkpoint, do you think you could send a code snippet and a checkpoint that would...

Hey @sanchit-gandhi, thanks for opening the issue! It's working on my environment, but it might be explained by the torch version I'm using (2.2). Nonetheless, before I dive deeper, could...

hey @cjekel, not that I'm aware of! The current issue is without GPU and with fp32! Feel free to open an issue for the other models with a reproducing script...