| #SBATCH --nodes=1 | |
| #SBATCH --ntasks-per-node=1 | |
| #SBATCH --gres=gpu:1 | |
| #SBATCH --constraint=A100 | |
| #SBATCH --mem=17000 | |
| {'train_runtime': 55.1071, 'train_samples_per_second': 0.181, 'train_steps_per_second': 0.181, 'train_loss': 0.7973592102527618, 'epoch': 2.0} | |
| Time: 55.11 | |
| Samples/second: 0.18 | |
| GPU memory occupied: 5011 MB. |