78Alpha

I have a 3090 myself and run into this as well. Could be the latest driver which seems to prefer shared memory, enforcing 2 devices. Do you happen to be...

If setting index ratio to 1 doesn't add enough accent, then you'd need someone already singing with the accent. RVC layers a voice on top, it didn't usually do such...

getting it the right length would be a bit of an issue; Especially since it will drop sentences entirely, at random.

Adding on to this, I loaded 4909 small files and the same issue occurred. Now, this is a fresh install on a different drive because I didn't want my SSD...

That is a byproduct of the sentence splitter. It will just drop things here and there. The painful alternative is to do it sentence by sentence. An automated alternative would...

> > Is there any plan and if yes when about fine-tuning on custom data? > > @Abdulk084 > > https://github.com/bensonbs/VALL-E-X-Trainer Link now 404s

> @78Alpha did you even bother to look at that user's other repos? > > https://github.com/bensonbs/VALL-E-X-Trainer-by-CustomData Of course not. But regardless, the link is invalid

You could do with CPU memory fallback. For windows it is a setting in control panel and allows for a total of 24 GB VRAM + 24 GB Shared. If...

I've managed to get to: ``` for bib in range(len(output_lengths)): mel_length_pred = output_lengths[bib] mel_length_gt = int(mel_input_length[bib].item() / 2) if mel_length_gt

They're going to be zero for a while unless the conditions it's looking for are met. On about 1 epoch of training, my tensorboard only showed 60 steps worth of...