Might this limit be outlined somewhere? RuntimeError: input tensor must fit into 32-bit index math

#8
by PsiPi - opened

It wasn't immediately clear what the numerical upper limit might be for the following, expressed in seconds.

File "C:\Python38\lib\site-packages\audiocraft\modules\conv.py", line 200, in forward
x = pad1d(x, (padding_left, padding_right + extra_padding), mode=self.pad_mode)
File "C:\Python38\lib\site-packages\audiocraft\modules\conv.py", line 84, in pad1d
padded = F.pad(x, paddings, mode, value)
RuntimeError: input tensor must fit into 32-bit index math

Might it be be prudent to outline the bound?
Or are we thinking 'exercere ad lectorem'

Many thanks :)

duration=820

2498.04 seconds to process

When I use the musicgen-melody-large model, compression_model = load_compression_model(name, device=device) will access the encodec_32khz model, and return error No such file or directory: compression_state_dict.bin. How to solve this problem?

When I use the musicgen-melody-large model, compression_model = load_compression_model(name, device=device) will access the encodec_32khz model, and return error No such file or directory: compression_state_dict.bin. How to solve this problem?

image.png

@JustKowalskii https://pinokio.computer/ that will fix your problem. install, run, install the musicgen project, run it. done.

Sign up or log in to comment