Expected a value of type 'Tensor (inferred)' for argument 'input' but instead found type 'Optional[Tensor]'.

[gray5wolf]

[rank0]: forward(torch.torch.nn.modules.container.___torch_mangle_16.Sequential self, Tensor input) -> Tensor:
[rank0]: Expected a value of type 'Tensor (inferred)' for argument 'input' but instead found type 'Optional[Tensor]'.
[rank0]: Inferred 'input' to be of type 'Tensor' because it was not annotated with an explicit type.
[rank0]: :
[rank0]: File "/home/admin1/wespeaker/wespeaker/models/redimnet.py", line 623
[rank0]: skip = x
[rank0]: x = self.red_dim_conv(x)
[rank0]: x = self.tcm(x)
[rank0]: ~~~~~~~~ <--- HERE
[rank0]: x = self.exp_dim_conv(x)
[rank0]: return skip + x
I reported this error using redimnet.yaml. How to solve it?

[JiJiJiang]

It seems a error caused by torch.jit.script. If so, comment these lines 153~155 (

wespeaker/wespeaker/bin/train.py

Line 153 in 1d4164b

if frontend_type == 'fbank':

) and try again.