search

zylo117 / Yet-Another-EfficientDet-Pytorch

The pytorch re-implement of the official efficientdet with SOTA performance in real time and pretrained weights.
GNU Lesser General Public License v3.0
5.2k stars 1.27k forks source link

No such operator torchvision::nms #696

Closed Mankvis closed 2 years ago

Mankvis commented 2 years ago

If the CPU is used, use_cuda=False, then cuda will not be used, and why will it report an error? RuntimeError: No such operator torchvision::nms torch 1.4.0+cu92 torchvision 0.5.0

zylo117 commented 2 years ago

can you confirm that torchvision version matches torch version? AFAIK, they have strong version binding relationship. Also, maybe some cuda versions just can't work with cpu

Mankvis commented 2 years ago

can you confirm that torchvision version matches torch version? AFAIK, they have strong version binding relationship. Also, maybe some cuda versions just can't work with cpu

I haven't changed the pytorch version or cuda version I just changed the python version from 3.6 to 3.7 and it worked

zylo117 commented 2 years ago

actually, when you switch python version, the python deps will be different. For example, torch 1.4 for python3.6 != torch 1.4 for python3.7

Mankvis commented 2 years ago

actually, when you switch python version, the python deps will be different. For example, torch 1.4 for python3.6 != torch 1.4 for python3.7

That's it, thank you