Closed alexdemartos closed 1 year ago
@alexdemartos oh gosh yes :man_facepalming: thank you for catching this
https://github.com/lucidrains/audiolm-pytorch/commit/e8ee51f1efb5bb1bba26ec691191763dbece0ecc
@alexdemartos does this mean you have already gotten to the stage of sampling from the coarse and fine transformers?
thanks for detecting this bug! It helps me a lot
ok, i'm guessing he's slinking back off into the darkness to pen his next paper :laughing:
i'll leave him be
In the
decode_from_codebook_indices
method there is a residual connection after the decoder attention:https://github.com/lucidrains/audiolm-pytorch/blob/d244a6c90b33627bad1e38ae2e20e821e6818253/audiolm_pytorch/soundstream.py#L494
Which is not present in the
forward
method:https://github.com/lucidrains/audiolm-pytorch/blob/d244a6c90b33627bad1e38ae2e20e821e6818253/audiolm_pytorch/soundstream.py#L567
Is this correct?
In my case, I found significantly higher audio quality during inference after removing this residual connection (matching training conditions).