a bug in SRModel.py
Closed this issue · 1 comments
ChassWind commented
i think the code
self.conv_out = nn.Conv2d(32, (2*sr_rate)**2 * sr_rate, kernel_size=3, padding=1, bias=False)
should be change as behind
self.conv_out = nn.Conv2d(32, (2*sr_rate)**2 * 3, kernel_size=3, padding=1, bias=False)
so that it not only can run in X3 but also can run in other upsample factor
aselsan-research-imaging-team commented
You're right, thanks, the file is updated accordingly.