Hello, thank you for such a wonderful job！I have a question that I hope you can answer.
I added some convolution layers before segnet's self.bev_compressor. Why did my modification result in 100% utilization on GPU0, and 0% utilization on GPU1.2.3 even though there is the same memory footprint as GPU0? And the program neither reports errors nor continues running,it just stuck there.
Hello, thank you for such a wonderful job！I have a question that I hope you can answer.
I added some convolution layers before segnet's self.bev_compressor. Why did my modification result in 100% utilization on GPU0, and 0% utilization on GPU1.2.3 even though there is the same memory footprint as GPU0? And the program neither reports errors nor continues running,it just stuck there.