Part 2

1. Use ls /ceph/container/pytorch command to find an appropriate PyTorch container. Copy the path to the container.

2. Edit run.sh with nano (Nano guide) or vim (Vim guide) text editor.

3. Set the following parameters in the script (Guide to run a bash script)

   • Set #SBATCH --cpus-per-task= to 15
   • Set #SBATCH --mem= to 24G
   • Set #SBATCH --gres=gpu:1
   • Add the container path you copied to PATHTOCONTAINER=
   • Add name_generator.py to SCRIPT=

4. Save and exit the file.

5. Start run.sh using sbatch command

6. Use watch -n 1 squeue --me to check your job process (approx. 3 minutes to processs)

7. When the job done, press CTRL+C to exit watch-mode and check the content of result_x.out using cat command. The x is the job id.

Solution

1. Use ls /ceph/container to locate a PyTorch container image in the /ceph/container directory. Copy the path, e.g. /ceph/container/pytorch/pytorch_24.07.sif.
2. Make sure your current directory is user@student.aau.dk@ailab-fe01:~/generating-names-with-pytorch$.
   • If not, go into the directory by using cd generating-names-with-pytorch.
3. Use nano run.sh to open the bash script with nano text editor
   • Set #SBATCH --cpus-per-task= to 15
   • Set #SBATCH --mem= to 24G
   • Set #SBATCH --gres=gpu:1
   • Add the container path you copied to PATHTOCONTAINER=
   • Add name_generator.py to SCRIPT=
4. Press CTRL+O followed by ENTER to save the file, then press CTRL+X to exit the file.
5. Use sbatch run.sh to start the job.
6. Use watch -n 1 squeue --me to check your job process
7. When the job done, press CTRL+C to exit watch-mode and check the content of result_x.out using cat command