Using torch.nn.BatchNorm2d , we can implement Batch ... x = torch.flatten(x, 1 ). x = self .fc1(x). x = F.relu( self .fc1_bn(x)).
確定! 回上一頁