Axel Chemla--Romeu-Santos
Axel Chemla--Romeu-Santos
`torch.Generator` are not yet compatible with torchscript unfortunately. Controllable seed with come with a scriptable generator from `torch`
Obviously you have a problem with your data paths, no influence of RAVE version here. Double check the paths of your input folder (dragging / dropping into the console to...
Federico, as written in the FAQ (https://github.com/acids-ircam/RAVE?tab=readme-ov-file#frequently-asked-question-faq), and as I remembered in other issues, this problem is due to the facts that your sounds are not long enough to feed...
There was actually a bug in 2.1 that prevented random cropping during training, pertaining overfitting problems. This problem was fixed. I redirect you to how to solve the problem in...
Are you resuming a checkpoint? I cannot reproduce this bug from a fresh training
Can you please give - your system, torch / torch audio version - full preprocess command - train preprocess command and, if a folder is created, the .gin inside summarizing...
Ok, got that ; it is a problem with the default amount of audio channels. Meanwhile, Please add --channels X (X being your number of channels: 1 for mono, 2...
What version was your RAVE? Did you try with 2.3?
What kind of data was this trained on? Did you provide additional configurations?
The cropping is only useful for the training, dropping signal with zero gradients. Cropping it in validation_step would not have that much sense, and would mess with the output dimensionality....