output padding must be smaller than either stride or dilation

[jvel07]

Hi, congrats on the repo! I am trying to use it with the following dataset (as I don't have access to wsj):

train_set, val_set = LibriMix.mini_from_download(task='sep_clean')

I replace the dataset in DataLoader() with the train_setabove. When training the network I have the following error:

RuntimeError: output padding must be smaller than either stride or dilation, but got output_padding_height: 0 output_padding_width: 8000 stride_height: 1 stride_width: 10 dilation_height: 1 dilation_width: 1

Is this due to the librimix dataset?

[wangkenpu]

Sorry for the late reply. For this repo, the waveform sample rate should be 8k. Did you check your data's sample rate?