Henry

Thanks for your tips! But when I pretrained faster rcnn, torch said apex.amp has been integrated into torch. So if torch version >= 1.6, just replace `from apex import amp`...

https://github.com/yubaoliu/RDS-SLAM/issues/4#issuecomment-932777773

run `xhost +` in terminal before you run docker-compose up, it works for me. After this, a lxterminal comes up. The docker image is running.

My gpu is 3060, so it can't work in cuda 10.2. I change dockerfile `FROM nvidia/cudagl:10.2-devel-ubuntu18.04` to `FROM nvidia/cudagl:11.1.1-devel-ubuntu18.04`, it works to me. When compile segnet files, change gpu arch...

you can find the method in [Bridging Knowledge Graphs to Generate Scene Graphs: Supplementary Material](https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123680596-supp.pdf), there is no code.