how to run this for videos with high resolution? When I test on Vid4 dataset it works fine (180x144 resolution) but when I test it on my own dataset (640x360)) resolution, it shows RuntimeError: CUDA error: an illegal memory access was encountered which is actually due to out of GPU memory (16 GB GPU memory). Is there a way to fix this? I am also pasting the stack trace:
Traceback (most recent call last):
File "test.py", line 139, in
main(dataset)
File "test.py", line 129, in main
Klens_demo_test(net, test_loader, opt.scale_factor, dataset_name)
File "test.py", line 68, in Klens_demo_test
SR = net(LR)
File "/data3/conda/envs/mehran/pythonlatest/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, kwargs)
File "/data2/superresolution/video_sr/D3Dnet/code/model.py", line 47, in forward
out = self.residual_layer(out)
File "/data3/conda/envs/mehran/pythonlatest/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, *kwargs)
File "/data3/conda/envs/mehran/pythonlatest/lib/python3.7/site-packages/torch/nn/modules/container.py", line 117, in forward
input = module(input)
File "/data3/conda/envs/mehran/pythonlatest/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(input, kwargs)
File "/data2/superresolution/video_sr/D3Dnet/code/model.py", line 64, in forward
return self.dcn1(self.lrelu(self.dcn0(x))) + x
File "/data3/conda/envs/mehran/pythonlatest/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/data2/superresolution/video_sr/D3Dnet/code/dcn/modules/deform_conv.py", line 287, in forward
self.im2col_step)
File "/data2/superresolution/video_sr/D3Dnet/code/dcn/functions/deform_conv_func.py", line 34, in forward
ctx.im2col_step)
RuntimeError: CUDA error: an illegal memory access was encountered
how to run this for videos with high resolution? When I test on Vid4 dataset it works fine (180x144 resolution) but when I test it on my own dataset (640x360)) resolution, it shows RuntimeError: CUDA error: an illegal memory access was encountered which is actually due to out of GPU memory (16 GB GPU memory). Is there a way to fix this? I am also pasting the stack trace: Traceback (most recent call last): File "test.py", line 139, in
main(dataset)
File "test.py", line 129, in main
Klens_demo_test(net, test_loader, opt.scale_factor, dataset_name)
File "test.py", line 68, in Klens_demo_test
SR = net(LR)
File "/data3/conda/envs/mehran/pythonlatest/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, kwargs)
File "/data2/superresolution/video_sr/D3Dnet/code/model.py", line 47, in forward
out = self.residual_layer(out)
File "/data3/conda/envs/mehran/pythonlatest/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, *kwargs)
File "/data3/conda/envs/mehran/pythonlatest/lib/python3.7/site-packages/torch/nn/modules/container.py", line 117, in forward
input = module(input)
File "/data3/conda/envs/mehran/pythonlatest/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(input, kwargs)
File "/data2/superresolution/video_sr/D3Dnet/code/model.py", line 64, in forward
return self.dcn1(self.lrelu(self.dcn0(x))) + x
File "/data3/conda/envs/mehran/pythonlatest/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/data2/superresolution/video_sr/D3Dnet/code/dcn/modules/deform_conv.py", line 287, in forward
self.im2col_step)
File "/data2/superresolution/video_sr/D3Dnet/code/dcn/functions/deform_conv_func.py", line 34, in forward
ctx.im2col_step)
RuntimeError: CUDA error: an illegal memory access was encountered