Dropout1d

Dropout1d(p=0.5, inplace=False)

Zero whole channels of a channels-last 1-d signal, ``(N, L, C)`` or
``(L, C)``, with probability ``p`` and scale the rest by ``1 / (1 - p)``
while training; the identity in evaluation mode.

forward

forward(self, input: 'Tensor') -> 'Tensor'

forward(input) -> Tensor