forked from mrq/DL-Art-School
Fix scale torch warning
This commit is contained in:
parent
bcebed19b7
commit
8dd44182e6
|
@ -18,7 +18,7 @@ class MultiConvBlock(nn.Module):
|
|||
self.bnconvs = nn.ModuleList([ConvBnLelu(filters_in, filters_mid, kernel_size, norm=norm, bias=False, weight_init_factor=weight_init_factor)] +
|
||||
[ConvBnLelu(filters_mid, filters_mid, kernel_size, norm=norm, bias=False, weight_init_factor=weight_init_factor) for i in range(depth - 2)] +
|
||||
[ConvBnLelu(filters_mid, filters_out, kernel_size, activation=False, norm=False, bias=False, weight_init_factor=weight_init_factor)])
|
||||
self.scale = nn.Parameter(torch.full((1,), fill_value=scale_init))
|
||||
self.scale = nn.Parameter(torch.full((1,), fill_value=scale_init, dtype=torch.float))
|
||||
self.bias = nn.Parameter(torch.zeros(1))
|
||||
|
||||
def forward(self, x, noise=None):
|
||||
|
|
Loading…
Reference in New Issue
Block a user