Minkowski Engine is an auto-diff neural network library for high-dimensional sparse tensors
2.47k
stars
367
forks
source link
How could I know about the value of the padding in convolution and expanded resolution of the upsample operation? #430
Open
coallar opened 2 years ago
case 1: ME.MinkowskiConvolutionTranspose( in_channels=32, out_channels=16, kernel_size=3, stride=2, dimension=D), case 2: ME.MinkowskiConvolution( in_channels=16, out_channels=32, kernel_size=3, stride=2, dimension=D),
Dose case 1 mean resolution expansion by 2 times? how to calculate the value of padding in convoluton? thanks