I'm getting ERROR: Unexpected bus error encountered in worker. This might be caused by insufficient shared memory (shm). when trying to launch this after following the steps on your README.
My setup is this docker image ran with docker run --gpus all -it --ipc=host nvidia/cuda:10.1-cudnn7-devel /bin/bash. Host operating system is Ubuntu 18.04. 128 GB of RAM and an NVIDIA RTX 2080 Ti.
Any thoughts?
Edit: This is the command: ./mmdetection/tools/dist_train.sh ./configs/reppoints_moment_r101_fpn_2x_mt.py 1 --validate
I'm closing this. The issue was brought on because I had attempted to modify a pre-existing container to support the --ipc=host flag. For some reason, this didn't work. Running a new image with this flag worked.
I'm getting
ERROR: Unexpected bus error encountered in worker. This might be caused by insufficient shared memory (shm).
when trying to launch this after following the steps on your README.My setup is this docker image ran with
docker run --gpus all -it --ipc=host nvidia/cuda:10.1-cudnn7-devel /bin/bash
. Host operating system is Ubuntu 18.04. 128 GB of RAM and an NVIDIA RTX 2080 Ti.Any thoughts?
Edit: This is the command:
./mmdetection/tools/dist_train.sh ./configs/reppoints_moment_r101_fpn_2x_mt.py 1 --validate