Just downloaded the zip and ran it in python. I picked 2 images for style transfer and I have put them separately in trainA and trainB in another folder called style. I did not make valA and valB folders. I called:
python train.py --dataroot ./datasets/style
everything runs until it hits x + self.conv_block(x) which is part of the residual layer then I get cuda error on running out of memory? GPUs = 2 GTX 980 with 3GB in each so 6GB in total. In the printed info below it shows that I have 11.378 million params. What did I do wrong? Do I need to lower my residual blocks and change the default epoch size?
Edited: Btw how do you calculate param size anways? How big is 11M param in memory?
CustomDatasetDataLoader
dataset [UnalignedDataset] was created
#training images = 1
C:\Users\name\Desktop\Painter\models\networks.py:45: UserWarning: nn.init.normal is now
deprecated in favor of nn.init.normal_.
init.normal(m.weight.data, 0.0, 0.02)
---------- Networks initialized -------------
[Network G_A] Total number of parameters : 11.378 M
[Network G_B] Total number of parameters : 11.378 M
[Network D_A] Total number of parameters : 2.765 M
[Network D_B] Total number of parameters : 2.765 M
-----------------------------------------------
Actual Error
model [CycleGANModel] was created
create web directory ./checkpoints\experiment_name\web...
THCudaCheck FAIL file=c:\programdata\miniconda3\conda-bld\pytorch_1524549877902\work\aten\src\thc\generic/THCStorage.cu line=58 error=2 : out of memory
Traceback (most recent call last):
File "train.py", line 31, in
model.optimize_parameters()
File "C:\Users\Luke Chen\Desktop\Painter\models\cycle_gan_model.py", line 157, in optimize_parameters
self.backward_G()
File "C:\Users\Luke Chen\Desktop\Painter\models\cycle_gan_model.py", line 146, in backward_G
self.rec_B = self.netG_A(self.fake_A)
File "C:\Users\Luke Chen\Anaconda3\lib\site-packages\torch\nn\modules\module.py", line 491, in call
result = self.forward(*input, kwargs)
File "C:\Users\Luke Chen\Anaconda3\lib\site-packages\torch\nn\parallel\data_parallel.py", line 112, in forward
return self.module(*inputs[0], *kwargs[0])
File "C:\Users\Luke Chen\Anaconda3\lib\site-packages\torch\nn\modules\module.py", line 491, in call
result = self.forward(input, kwargs)
File "C:\Users\Luke Chen\Desktop\Painter\models\networks.py", line 199, in forward
return self.model(input)
File "C:\Users\Luke Chen\Anaconda3\lib\site-packages\torch\nn\modules\module.py", line 491, in call
result = self.forward(*input, *kwargs)
File "C:\Users\Luke Chen\Anaconda3\lib\site-packages\torch\nn\modules\container.py", line 91, in forward
input = module(input)
File "C:\Users\Luke Chen\Anaconda3\lib\site-packages\torch\nn\modules\module.py", line 491, in call
result = self.forward(input, **kwargs)
File "C:\Users\Luke Chen\Desktop\Painter\models\networks.py", line 241, in forward
out = x + self.conv_block(x)
RuntimeError: cuda runtime error (2) : out of memory at c:\programdata\miniconda3\conda-bld\pytorch_1524549877902\work\aten\src\thc\generic/THCStorage.cu:58
Just downloaded the zip and ran it in python. I picked 2 images for style transfer and I have put them separately in trainA and trainB in another folder called style. I did not make valA and valB folders. I called:
python train.py --dataroot ./datasets/style
everything runs until it hits x + self.conv_block(x) which is part of the residual layer then I get cuda error on running out of memory? GPUs = 2 GTX 980 with 3GB in each so 6GB in total. In the printed info below it shows that I have 11.378 million params. What did I do wrong? Do I need to lower my residual blocks and change the default epoch size?
Edited: Btw how do you calculate param size anways? How big is 11M param in memory?
Actual Error