Closed kaishijeng closed 7 years ago
I just ran the demo successfully on a laptop with no nvidia. Comment out lines like https://github.com/longcw/yolo2-pytorch/blob/master/demo.py#L34 and https://github.com/longcw/yolo2-pytorch/blob/master/demo.py#L47 which place tensors onto the GPU memory and you should be OK.
Thanks for the info. But the compilation make.sh can't be completed because it needs nvcc to compile roi_pooling. Also utils needs nvcc in order to build. How do you solve this?
@kaishijeng Can you tell me how you solved it ?
I have trained a network on GPU but I have to deploy my code on a CPU instance. The test.py script does not seem to work on CPU and gives 'ImportError: No module named _reorg_layer' error on from ._reorg_layer import lib as _lib, ffi as _ffi. This is probably because make.sh could not be compiled. Any idea how to achieve this? @kaishijeng @9thDimension @longcw
hi, how do you solve it?
I also want use it with pytorch-cpuonly,What should I do?
Plan to run this on Raspberry3 board. Does you code run on a CPU only platform?
Thanks,