pytorch 1.0 doesn't work

vasgaowei / pytorch_MELM

The pytorch implementation of the Min-Entropy Latent Model for Weakly Supervised Object Detection

104 stars 19 forks source link

Open tankche1 opened 4 years ago

tankche1 commented 4 years ago

Good work! Can you add a readme for pytorch1.0? There are no layer_util in lib. so I can't bash make.sh.

vasgaowei commented 4 years ago

For pytorch 1.2 branch, you com compile by command: bash make_cuda.sh.

tankche1 commented 4 years ago

I succesfully compile. Then when I train I met this, any solution? Also check mmdetection issue and didn't find solution.

vasgaowei commented 4 years ago

You can try to check if the .so file for nms_cpu exists. Also you can use nms implementated in torchvision.

Sean-Huang65 commented 4 years ago

I met this problem too, and the .so file for nms_cpu exists....