Padding question

[ValdemarQ]

I noticed you use padding for convolutions as:

padding=kernel/2

But it seems that your kernel size are not even numbers: n_filtersizes = [65, 33, 17, 9, 9, 9, 9, 9, 9]

So I get following errors, how did you guys overcame this? Since I am getting following error:

conv1d(): argument 'padding' must be tuple of ints, but found element of type float at pos 1