There is a line surface error during yolov5 training in the image. The reason is that there is not enough SHM allocated by the docker container, so you need to manually set the SHM size to be small.

RuntimeError: DataLoader worker (pid 463) is killed by signal: Bus error. It is possible that dataloq

Rerun a container

sudo docker run --name '' -p 5000:5000 -v /home:/home --runtime=nvidia --shm-size 8g -itd torch:v1 bash