PyTorch: Difference between revisions

Correction: virtual env needs to be set up on each node
mNo edit summary
(Correction: virtual env needs to be set up on each node)
Line 592: Line 592:


module load python # Using Default Python version - Make sure to choose a version that suits your application
module load python # Using Default Python version - Make sure to choose a version that suits your application
srun --tasks-per-node=1 bash << EOF
virtualenv --no-download $SLURM_TMPDIR/env
virtualenv --no-download $SLURM_TMPDIR/env
source $SLURM_TMPDIR/env/bin/activate
source $SLURM_TMPDIR/env/bin/activate
pip install torchvision --no-index
pip install torchvision --no-index
EOF


export NCCL_BLOCKING_WAIT=1  #Set this environment variable if you wish to use the NCCL backend for inter-GPU communication.
export NCCL_BLOCKING_WAIT=1  #Set this environment variable if you wish to use the NCCL backend for inter-GPU communication.
1

edit