amitaie

Why do you use Interpolating in pitch feature?

1

Hey, didn't find any answer to this question. i saw few implementations of pitch feature ( with/out quantisation, with/out un-voice predictor) but all of them used the interpolation in some...

trim_silence in processor should get sr. [Bug]

1

### Describe the bug Hello, I think there is a bug in the processor. when we load_wav we can get different sr than self.sample_rate in order to resample the audio...

Why is ASR model goes to train mode in the training loop

11

Hey, I saw that the ASR model that is under "model" also transfer to train mode in the beginning of the training loop, why is that? I tried to leave...

Question about 'attention bias not supported for flash attention'

2

https://github.com/lucidrains/audiolm-pytorch/blob/879f3bd2668b4fc61126b8405bbfdea8fa2c8778/audiolm_pytorch/attend.py#L112 Why not to add the bias and the mask and create an attn_mask of type float and supply it to the scaled_dot_product_attention as attn_mask? is that not the same...