TylorShine/MNP-SVC

Error when attempting to train

Opened this issue · 3 comments

 [DDSP Model] Minimum-Phase harmonic Source Combtooth Subtractive Synthesiser
Traceback (most recent call last):
  File "C:\Users\cross\Downloads\MNP-SVC\train.py", line 48, in <module>
    model = CombSubMinimumNoisedPhase(
  File "C:\Users\cross\Downloads\MNP-SVC\modules\vocoder.py", line 191, in __init__
    self.register_buffer("use_add_noise_env", torch.tensor(use_add_noise_env))
RuntimeError: Could not infer dtype of NoneType

It fixes it if i comment out this line

self.register_buffer("use_add_noise_env", torch.tensor(use_add_noise_env))

But I have no idea if that will break the entire result or not

Yeah that breaks the inference. I'm lost as to what to do.

Thanks for your report and So sorry for late reply. I didn't think the versioning, because I thought this repo is for researching...!
But I planning to versioning to this repo. I have already updated the repo to v2. You can update it and use it if you want!