- Login to cluster
ssh rohitgarg617@perceptron.ri.cmu.edu
Request a node (use any one of these)
srun -w clamps -t 5-00:00 -p gpu --gres=gpu:1 --pty /bin/bash
srun -w roberto -t 5-00:00 -p gpu --gres=gpu:1 --pty /bin/bash
- to run your docker image:
nvidia-docker run -it --rm --ipc=host -e CUDA_VISIBLE_DEVICES=`echo $CUDA_VISIBLE_DEVICES` -v /data/datasets:/data/datasets -v /storage2/datasets:/storage2/datasets -v /local:/local -v /home/$USER:/home/$USER -v /storage1:/storage1 nvidia/cuda:8.0-cudnn5-devel-ubuntu16.04