Encounter segfault when running inference.py

HowieMa / NSRMhand

[WACV 2020] "Nonparametric Structure Regularization Machine for 2D Hand Pose Estimation"

104 stars 16 forks source link

Encounter segfault when running inference.py #15

Closed HollyDQWang closed 3 years ago

HollyDQWang commented 3 years ago

Running inference.py on Ubuntu and encountered segfault on line 117 state_dict = torch.load(args.resume) Would you mind suggesting any possible cause that this might be happening?

HollyDQWang commented 3 years ago

Okay this is resolved by specifying torch.load(args.resume, map_location=torch.device('cuda')) and commenting out line 115 for some reason. Referenced this post here https://discuss.pytorch.org/t/segmentation-fault-when-loading-weight/1381

HowieMa commented 3 years ago

Sorry for the inconvenience. The model I uploaded has the device CUDA, and you may directly load it on the CPU. I will update the version of the model later and make it feasible for the CPU.