I tried implementing a DCGAN generator using the following code
class Generator(nn.Module):
"""
Input : `(batch_size, 100)`
Output: `(batch_size, 3, 64, 64)`
A generator that takes noise as input and generates images.
loss: Binary cross entropy loss
"""
def __init__(self):
super().__init__()
# layer_dims = [1024, 512, 256, 128, 64]
layer_dims = [ngf*8, ngf*4, ngf*2, ngf]
self.layers = nn.ModuleList([self.upscale_block(in_f, out_f, kernel_size=4, stride=2, padding=1)
for in_f, out_f in zip(layer_dims[:-1], layer_dims[1:])])
self.gen = nn.Sequential(
self.input_layer(nz, ngf*8, use_bias=False),
*self.layers,
self.output_layer(ngf, 3, use_bias=False),
)
def forward(self, x):
print("Generator input: ", x.shape)
# x = torch.tensor(self.dense(x, 4*4*1024, use_bias=False), dtype=torch.float32)
# x = x.view(-1, 1024, 4, 4)
# x = self.gen(x)
return self.gen(x)
def upscale_block(self, in_channels, out_channels, kernel_size, stride, padding):
return nn.Sequential(
nn.ConvTranspose2d(in_channels, out_channels, kernel_size, stride, padding, bias=False),
nn.BatchNorm2d(out_channels),
nn.ReLU(True)
)
def input_layer(self, in_f, out_f, kernel_size=4, stride=1, padding=0, use_bias=False):
return nn.Sequential(
nn.ConvTranspose2d(in_f, out_f, kernel_size, stride, padding, bias=use_bias),
nn.BatchNorm2d(out_f),
nn.ReLU(True),
)
def output_layer(self, in_f, out_f, use_bias=True):
return nn.Sequential(
nn.ConvTranspose2d(in_f, out_f, 4, 2, 1, use_bias),
nn.Tanh()
)
following code from the official pytorch DCGAN tutorial
# Generator Code
class Generator(nn.Module):
def __init__(self, ngpu):
super(Generator, self).__init__()
self.ngpu = ngpu
self.main = nn.Sequential(
# input is Z, going into a convolution
nn.ConvTranspose2d( nz, ngf * 8, 4, 1, 0, bias=False),
nn.BatchNorm2d(ngf * 8),
nn.ReLU(True),
# state size. (ngf*8) x 4 x 4
nn.ConvTranspose2d(ngf * 8, ngf * 4, 4, 2, 1, bias=False),
nn.BatchNorm2d(ngf * 4),
nn.ReLU(True),
# state size. (ngf*4) x 8 x 8
nn.ConvTranspose2d( ngf * 4, ngf * 2, 4, 2, 1, bias=False),
nn.BatchNorm2d(ngf * 2),
nn.ReLU(True),
# state size. (ngf*2) x 16 x 16
nn.ConvTranspose2d( ngf * 2, ngf, 4, 2, 1, bias=False),
nn.BatchNorm2d(ngf),
nn.ReLU(True),
# state size. (ngf) x 32 x 32
nn.ConvTranspose2d( ngf, nc, 4, 2, 1, bias=False),
nn.Tanh()
# state size. (nc) x 64 x 64
)
def forward(self, input):
return self.main(input)
but when i am running the code
z = torch.randn(64, nz, 1, 1)
gen_model = Generator()
gen_model_pytorch = Generator_pytorch()
print(gen_model_pytorch)
print(gen_model_pytorch)
i get the following as output
torch.Size([64, 3, 64, 64])
Generator input: torch.Size([64, 100, 1, 1])
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
/tmp/ipykernel_939/3945186195.py in <module>
3 gen_model_pytorch = Generator_pytorch()
4 print(gen_model_pytorch(z).shape)
----> 5 print(gen_model(z).shape)
6
7 # print(gen_model)
/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
1108 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
1109 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1110 return forward_call(*input, **kwargs)
1111 # Do not call functions when jit is used
1112 full_backward_hooks, non_full_backward_hooks = [], []
/tmp/ipykernel_939/2092446074.py in forward(self, x)
28 # x = x.view(-1, 1024, 4, 4)
29 # x = self.gen(x)
---> 30 return self.gen(x)
31
32
/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
1108 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
1109 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1110 return forward_call(*input, **kwargs)
1111 # Do not call functions when jit is used
1112 full_backward_hooks, non_full_backward_hooks = [], []
/opt/conda/lib/python3.8/site-packages/torch/nn/modules/container.py in forward(self, input)
139 def forward(self, input):
140 for module in self:
--> 141 input = module(input)
142 return input
143
/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
1108 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
1109 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1110 return forward_call(*input, **kwargs)
1111 # Do not call functions when jit is used
1112 full_backward_hooks, non_full_backward_hooks = [], []
/opt/conda/lib/python3.8/site-packages/torch/nn/modules/container.py in forward(self, input)
139 def forward(self, input):
140 for module in self:
--> 141 input = module(input)
142 return input
143
/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py in _call_impl(self, *input, **kwargs)
1108 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
1109 or _global_forward_hooks or _global_forward_pre_hooks):
-> 1110 return forward_call(*input, **kwargs)
1111 # Do not call functions when jit is used
1112 full_backward_hooks, non_full_backward_hooks = [], []
/opt/conda/lib/python3.8/site-packages/torch/nn/modules/conv.py in forward(self, input, output_size)
923 input, output_size, self.stride, self.padding, self.kernel_size, self.dilation) # type: ignore[arg-type]
924
--> 925 return F.conv_transpose2d(
926 input, self.weight, self.bias, self.stride, self.padding,
927 output_padding, self.groups, self.dilation)
TypeError: conv_transpose2d(): argument 'output_padding' (position 6) must be tuple of ints, not tuple
It's working for the tutorial code(obvio) but not for the same custom network in my Generator class, albeit I followed the architecture as the tutorial, where I cross checked the file both models layers by print() they were same, ... what layer causes this issue?