Many layer transformation functions at torch.nn have parameters like N, C_out, L_out (see http://pytorch.org/docs/master/nn.html#torch.nn.Conv1d) or N, C, H_out,W_out (see torch.nn.MaxPool2d). What are they?
N
: batch size
C
: channels of conv
H
: img size height(for conv2d)
W
: img size width(for conv2d)
L
: seq length (for conv1d)