I have some error with correlation :ImportError: No module named 'inpainting.models.correlation_package._ext.correlation._correlation' about video-object-removal HOT 5 OPEN

I am getting below error after all frames are executed and also results are save in result directory in data folder but no video is generated/saved.

Number of model parameters: 14115364
error in Correlation_forward_cuda_kernel: invalid device function
Traceback (most recent call last):
File "demo.py", line 15, in
inpaint(args)
File "/home/ios/IP_IOS_AI_ENQ/face-edit/video-object-removal/inpaint.py", line 154, in inpaint
outputs, _, , , _ = model(masked_inputs, masks, lstm_state, prev_feed, t)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 491, in call
result = self.forward(*input, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/parallel/data_parallel.py", line 112, in forward
return self.module(*inputs[0], **kwargs[0])
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 491, in call
result = self.forward(*input, **kwargs)
File "/home/ios/IP_IOS_AI_ENQ/face-edit/video-object-removal/inpainting/models/vinet.py", line 174, in forward
flow2 = self.flownet(f1, f2)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 491, in call
result = self.forward(*input, **kwargs)
File "/home/ios/IP_IOS_AI_ENQ/face-edit/video-object-removal/inpainting/models/flow_modules.py", line 132, in forward
corr = self.corr(x1.contiguous(), x2.contiguous())
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 491, in call
result = self.forward(*input, **kwargs)
File "/home/ios/IP_IOS_AI_ENQ/face-edit/video-object-removal/inpainting/models/correlation_package/modules/correlation.py", line 17, in forward
result = CorrelationFunction(self.pad_size, self.kernel_size, self.max_displacement,self.stride1, self.stride2, self.corr_multiply)(input1, input2)
File "/home/ios/IP_IOS_AI_ENQ/face-edit/video-object-removal/inpainting/models/correlation_package/functions/correlation.py", line 30, in forward
self.pad_size, self.kernel_size, self.max_displacement,self.stride1, self.stride2, self.corr_multiply)
File "/usr/local/lib/python3.5/dist-packages/torch/utils/ffi/init.py", line 197, in safe_call
result = torch._C._safe_call(*args, **kwargs)
torch.FatalError: aborting at /home/ios/IP_IOS_AI_ENQ/face-edit/video-object-removal/inpainting/models/correlation_package/src/correlation_cuda.c:88

from video-object-removal.

Before executing
bash install.sh
you need to change -arch=sm_52 option in the files
inpainting/lib/channelnorm_package/make.sh
inpainting/lib/correlation_package/make.sh
inpainting/lib/resample2d_package/make.sh
inpainting/models/correlation_package/make.sh
according to the web-site
https://arnon.dk/matching-sm-architectures-arch-and-gencode-for-various-nvidia-cards/

from video-object-removal.

what does the file look like after changing -arch=sm_52 options?

from video-object-removal.

what does the file look like after changing -arch=sm_52 options?

You need to change line 9 in these 4 files

nvcc -c -o ChannelNorm_kernel.o ChannelNorm_kernel.cu -x cu -Xcompiler -fPIC -arch=sm_52 -I ${TORCH}/lib/include/TH -I ${TORCH}/lib/include/THC

instead of -arch=sm_52 put your GPU architecture

from video-object-removal.

I am facing same error with ubuntu 16 , cuda 10 version , please help me to resolve the issue and even i changed -arch as per my gpu architecture..

from video-object-removal.