You loaded an image, got a tensor shaped (batch, height, width, channels), and your convolution wants (batch, channels, height, width). Stack Overflow says permute. Someone else says transpose. And re
pytorchfromgroundup.hashnode.dev8 min read
No responses yet.