Closed destinyzs closed 6 years ago
i also tried the solution https://github.com/jwyang/faster-rcnn.pytorch/issues/150, and his python is 3.x version, in 2.7 there is no set_start_method in multiprocessing.
you can try other methods mentioned in https://github.com/pytorch/pytorch/issues/1355 like:
The first method had solved my ploblem too. But it make code run slower.
thanks! actually, i have solved it by set cv2.setNumThreads(0)
.
anyway, thanks for your response. @neulrl
I am training my own dataset using fpn, i set the
num_workers > 0
like =4, and then it collapses after the first training iter, but it will be ok ifnum_workers = 0,
my device is following: OS: Ubuntu 16.04 PyTorch version: pytorch 0.3.1 Python version: python 2.7 CUDA version: 8.0 GPU models : Tesla P40i saw the same problem https://github.com/pytorch/pytorch/issues/1355, but his python verison is 3.x, and i tried that solution, it does not works. @jwyang