Note: use nvidia-docker to run all GPU builds.
To get the latest source, rerun the docker builds using the Dockerfiles.
Docker images at https://hub.docker.com/r/caffe2ai/caffe2/ are a few months old, but will be refreshed soon.
Build like: docker build -t caffe2:cuda8-cudnn6-all-options .
Run like: nvidia-docker run --rm -it caffe2:cuda8-cudnn6-all-options python -m caffe2.python.operator_test.relu_op_test
For Docker on USB related instructions you can find some help on the gh-pages branch here