I don’t think Wav2Vec2 has beem updated to work with saved models, cc @Rocketknight1
hey the guy you tagged hasn’t been active since a while so do you have an estimate of how long it will take for this to get fixed?
I’m here! But I don’t have a timeline for fixing this one, unfortunately. I’m currently working on some other parts of the codebase, but in the meantime, you could try the PyTorch version of that model?
I originally used the PyTorch version but my boss wants me to use the TensorFlow versions so I converted the pytroch one to TensorFlow and exported that from hugging face but unfortunately I am having problems loading the h5 format inside TensorFlow. Well I will try getting help from the TensorFlow team , hope you get the saved model version working eventually