You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
There was a question in the Office Hours: How to get a working TensorFlow setup with GPU on the Mila cluster.
I wasn't able to give a good, definitive answer. I looked online, tried a myriad of different ways to install tensorflow, but didn't manage to get any variation of this kind of script to work, at all.
Perhaps this kind of thing is trivial to someone like @obilaniu . If so, please let me know, and I'll send the instructions along to the student in question.
#!/bin/bash#SBATCH --job-name=tensorflow_gpu_setup_test#SBATCH --cpus-per-task=1#SBATCH --mem=16G#SBATCH --gres=gpu:1#SBATCH --time=0-00:10:00 # should setup in under 10 minutes, right?
module load miniconda/3
conda create -p $SLURM_TMPDIR/env python=3.9
conda activate $SLURM_TMPDIR/env
# Now do whatever is necessary (module load, conda install, pip install, env variable magic, ...)# to make the last line work.
python -c "import tensorflow as tf; assert tf.config.list_physical_devices('GPU') != []"
The text was updated successfully, but these errors were encountered:
There was a question in the Office Hours: How to get a working TensorFlow setup with GPU on the Mila cluster.
I wasn't able to give a good, definitive answer. I looked online, tried a myriad of different ways to install tensorflow, but didn't manage to get any variation of this kind of script to work, at all.
Perhaps this kind of thing is trivial to someone like @obilaniu . If so, please let me know, and I'll send the instructions along to the student in question.
The text was updated successfully, but these errors were encountered: