apple
Mlpackage of StyleGAN2 gets wrong results. The output of ToRGB module has large difference between mlmodel and mlpackage.
🐞Describe the bug
It gets wrong result when the output adds a skip connection value. I tested the difference by absolute mean error. The error is zero when I only test a single ToRGB module. But when I test on the StyleGAN2 model and comment some lines, the error of ToRGB module is around 4e-2, and the error is below 1e-4 before adding the skip connection.
The results generated by mlmodel and mlpackage.
To Reproduce
ToRGB module (error is around 4e-2):
class ToRGB(nn.Module):
def __init__(self, in_channel, style_dim, upsample=True, blur_kernel=[1, 3, 3, 1], version='v1', fuse=False):
super().__init__()
if upsample:
self.upsample = Upsample(blur_kernel)
self.conv = ModulatedConv2d(in_channel, 3, 1, style_dim, demodulate=False)
self.bias = nn.Parameter(torch.zeros(1, 3, 1, 1))
def forward(self, input, style, skip=None, upsample=True):
out = self.conv(input, style)
out = out + self.bias
if skip is not None:
if upsample:
skip = self.upsample(skip)
out = out + skip
return out
Return the skip connection and error is below 1e-4:
class ToRGB(nn.Module):
def __init__(self, in_channel, style_dim, upsample=True, blur_kernel=[1, 3, 3, 1], version='v1', fuse=False):
super().__init__()
if upsample:
self.upsample = Upsample(blur_kernel)
self.conv = ModulatedConv2d(in_channel, 3, 1, style_dim, demodulate=False)
self.bias = nn.Parameter(torch.zeros(1, 3, 1, 1))
def forward(self, input, style, skip=None, upsample=True):
out = self.conv(input, style)
out = out + self.bias
if skip is not None:
if upsample:
skip = self.upsample(skip)
return skip
# out = out + skip
# return out
Return the output berfore adding skip connection and error is below 1e-8:
class ToRGB(nn.Module):
def __init__(self, in_channel, style_dim, upsample=True, blur_kernel=[1, 3, 3, 1], version='v1', fuse=False):
super().__init__()
if upsample:
self.upsample = Upsample(blur_kernel)
self.conv = ModulatedConv2d(in_channel, 3, 1, style_dim, demodulate=False)
self.bias = nn.Parameter(torch.zeros(1, 3, 1, 1))
def forward(self, input, style, skip=None, upsample=True):
out = self.conv(input, style)
out = out + self.bias
return out
# if skip is not None:
# if upsample:
# skip = self.upsample(skip)
# return skip
# out = out + skip
# return out
System environment (please complete the following information):
- MacOS
- coremltools == 5.2
- Pytorch==1.9
@PigletOS - I need more information in order to be able to reproduce the problem. Please include all the code I need to reproduce this problem. At a minimum, it should include the following:
- How to create an instances of
ToRGB(i.e. values ofin_channel,style_dim). - Getting predictions from the torch models.
- Converting the torch models to Core ML.
- Comparing the results.
Also what version of macOS are you using?
