I'm using R3GAN on my datasets an I encountered an error involving matrix size:
File "/home/aike/Desktop/R3GAN/R3GAN/Networks.py", line 101, in forward
return self.LinearLayer(self.Basis(x).view(x.shape[0], -1))
RuntimeError: mat1 and mat2 shapes cannot be multiplied (1x64896 and 384x1)
could you provide some suggestions regarding the configuration of discriminator?my image size is 256x256,and I'm currently using configurations below for the discriminator:
Discriminator(WidthPerStage=[3 * x // 4 for x in [1024, 1024, 1024, 1024, 512]],CardinalityPerStage=[3 * x for x in [32, 32, 32, 32, 16]],BlocksPerStage=[2 * x for x in [1, 1, 1, 1, 1]],ExpansionFactor=2)
I'm using R3GAN on my datasets an I encountered an error involving matrix size:
File "/home/aike/Desktop/R3GAN/R3GAN/Networks.py", line 101, in forward
return self.LinearLayer(self.Basis(x).view(x.shape[0], -1))
RuntimeError: mat1 and mat2 shapes cannot be multiplied (1x64896 and 384x1)
could you provide some suggestions regarding the configuration of discriminator?my image size is 256x256,and I'm currently using configurations below for the discriminator:
Discriminator(WidthPerStage=[3 * x // 4 for x in [1024, 1024, 1024, 1024, 512]],CardinalityPerStage=[3 * x for x in [32, 32, 32, 32, 16]],BlocksPerStage=[2 * x for x in [1, 1, 1, 1, 1]],ExpansionFactor=2)