Resize features in pytorch

vision
1

Hi, I would like to resize the features, which obtained from VGG, for example, from 512×64×64 to 512×128×128, on GPU. I am sorry, I can not find the function to do this. So, I want to know, how can I do this?

2

Use Upsample+Conv as below

>>> print(input.size())
512×64×64
>>> up = nn.Upsample(mode='nearest', scale_factor=2)
>>> up_input = m(input)
>>> net = nn.Conv2d(512, 512, kernal_size=3, padding=1)
>>> out = net(up_input)
>>> print(out.size())
512×128×128