Training 4 channel images

[vnyk-dgrid]

I want to train 4 channel images bellow is my train command

python train.py --name train_w_orient_alpha --dataroot ../pix2pixHD/train_data_new/ --label_nc 0 --loadSize 512 --fineSize 512 --no_instance --gpu_ids 0,1,2,3 --input_nc 4 --output_nc 4 --batchSize 12

I am getting the following error

File "train.py", line 71, in <module>
    Variable(data['image']), Variable(data['feat']), infer=save_fake)
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 123, in forward
    outputs = self.parallel_apply(replicas, inputs, kwargs)
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 133, in parallel_apply
    return parallel_apply(replicas, inputs, kwargs, self.device_ids[:len(replicas)])
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 77, in parallel_apply
    raise output
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 53, in _worker
    output = module(*input, **kwargs)
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/hairgen/hairgen/pix2pixHDnew/models/pix2pixHD_model.py", line 221, in forward
    loss_G_VGG = self.criterionVGG(fake_image, real_image) * self.opt.lambda_feat
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/hairgen/hairgen/pix2pixHDnew/models/networks.py", line 120, in forward
    x_vgg, y_vgg = self.vgg(x), self.vgg(y)
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/hairgen/hairgen/pix2pixHDnew/models/networks.py", line 410, in forward
    h_relu1 = self.slice1(X)
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/container.py", line 91, in forward
    input = module(input)
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in __call__
    result = self.forward(*input, **kwargs)
  File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/conv.py", line 301, in forward
    self.padding, self.dilation, self.groups)
RuntimeError: Given groups=1, weight of size [64, 3, 3, 3], expected input[3, 4, 512, 512] to have 3 channels, but got 4 channels instead

I have also tried using the solution in #119 https://github.com/NVIDIA/pix2pixHD/issues/119#issuecomment-485658463 But it doesnt worked

[vishwas1234567]

just go to the kernel and check for the version and also it would be the error from the malfunctioning of the drivers.

On Fri, 10 Jan 2020 at 16:17, vnyk-dgrid notifications@github.com wrote:

I want to train 4 channel images bellow is my train command

python train.py --name train_w_orient_alpha --dataroot ../pix2pixHD/train_data_new/ --label_nc 0 --loadSize 512 --fineSize 512 --no_instance --gpu_ids 0,1,2,3 --input_nc 4 --output_nc 4 --batchSize 12

I am getting the following error

File "train.py", line 71, in Variable(data['image']), Variable(data['feat']), infer=save_fake) File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in call result = self.forward(*input, kwargs) File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 123, in forward outputs = self.parallel_apply(replicas, inputs, kwargs) File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 133, in parallel_apply return parallel_apply(replicas, inputs, kwargs, self.device_ids[:len(replicas)]) File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 77, in parallel_apply raise output File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 53, in _worker output = module(*input, *kwargs) File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in call result = self.forward(input, kwargs) File "/home/hairgen/hairgen/pix2pixHDnew/models/pix2pixHD_model.py", line 221, in forward loss_G_VGG = self.criterionVGG(fake_image, real_image) self.opt.lambda_feat File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in call result = self.forward(input, kwargs) File "/home/hairgen/hairgen/pix2pixHDnew/models/networks.py", line 120, in forward x_vgg, y_vgg = self.vgg(x), self.vgg(y) File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in call result = self.forward(*input, *kwargs) File "/home/hairgen/hairgen/pix2pixHDnew/models/networks.py", line 410, in forward h_relu1 = self.slice1(X) File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in call result = self.forward(input, kwargs) File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/container.py", line 91, in forward input = module(input) File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in call result = self.forward(*input, **kwargs) File "/opt/anaconda3/lib/python3.6/site-packages/torch/nn/modules/conv.py", line 301, in forward self.padding, self.dilation, self.groups) RuntimeError: Given groups=1, weight of size [64, 3, 3, 3], expected input[3, 4, 512, 512] to have 3 channels, but got 4 channels instead

I have also tried using the solution in #119 https://github.com/NVIDIA/pix2pixHD/issues/119

119 (comment)

https://github.com/NVIDIA/pix2pixHD/issues/119#issuecomment-485658463 But it doesnt worked

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/NVIDIA/pix2pixHD/issues/172?email_source=notifications&email_token=ALSXLORTHYBS66M7MYYFVJ3Q5BG3HA5CNFSM4KFF4KNKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4IFJ5NWQ, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALSXLOVEQE36RZIUQPWFXZDQ5BG3HANCNFSM4KFF4KNA .

[vnyk-dgrid]

@vishwas1234567 No , the problem is not with the drivers or dependencies

[WayneCho]

I think this error is raise because the VGG loss can only compute among three-channel images, in my application I can convert the four channel image to three channel. If cannot convert, can also simply set the option '--no_vgg_loss' as True.

[AndyRylynn]

Could I ask how do you solve this problem that using this code to train 4channel DNG? It really confuses me when I want to train my datasets.

[zhangxiaojuan66]

你好，邮件已收到，祝你万事如意，生活愉快！