crnn-audio-classification
crnn-audio-classification copied to clipboard
Error: Kernel size can't be greater than actual input size
in inference, if I choose audios with small size (e.g. fold10/7913-3-3-0.wav, fold10/7913-3-1-0.wav), the error will occur: Kernel size can't be greater than actual input size
How shall I fix the bug?
You'll need to either pad the input with silence, or modify the network (smaller kernel size, or stride, or padding).