| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
| Name | Name | Last commit date | ||
|---|---|---|---|---|
GridSearcher is a pure Python project designed to simplify the process of running grid searches for Machine Learning projects. It serves as a robust alternative to traditional bash scripts, providing a more flexible and user-friendly way to manage and execute multiple programs in parallel.
⚠️ It is designed for systems where users have direct SSH access to machines and can run their python scripts right away.
Install GridSearcher via pip:
pip install gridsearcherWe provide a minimal working example in the file example.py. Just set debug=True with debug=False in the run method call to run on GPUs. The output of example.py is the following:
GridSearcher PID: 8940
command 1: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-2_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=1_2024-06-19_23-04-23 --seed 1 --lr 1e-2 --wd 1e-2 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-2_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8/seed=1_2024-06-19_23-04-23
command 2: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-2_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=1_2024-06-19_23-04-23 --seed 1 --lr 1e-2 --wd 1e-3 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-2_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8/seed=1_2024-06-19_23-04-23
command 3: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-3_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=1_2024-06-19_23-04-23 --seed 1 --lr 1e-3 --wd 1e-2 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-3_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8/seed=1_2024-06-19_23-04-23
command 4: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-3_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=1_2024-06-19_23-04-23 --seed 1 --lr 1e-3 --wd 1e-3 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-3_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8/seed=1_2024-06-19_23-04-23
command 5: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-2_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=2_2024-06-19_23-04-23 --seed 2 --lr 1e-2 --wd 1e-2 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-2_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8/seed=2_2024-06-19_23-04-23
command 6: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-2_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=2_2024-06-19_23-04-23 --seed 2 --lr 1e-2 --wd 1e-3 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-2_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8/seed=2_2024-06-19_23-04-23
command 7: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-3_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=2_2024-06-19_23-04-23 --seed 2 --lr 1e-3 --wd 1e-2 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-3_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8/seed=2_2024-06-19_23-04-23
command 8: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-3_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=2_2024-06-19_23-04-23 --seed 2 --lr 1e-3 --wd 1e-3 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-3_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8/seed=2_2024-06-19_23-04-23
command 9: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-2_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=3_2024-06-19_23-04-23 --seed 3 --lr 1e-2 --wd 1e-2 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-2_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8/seed=3_2024-06-19_23-04-23
command 10: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-2_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=3_2024-06-19_23-04-23 --seed 3 --lr 1e-2 --wd 1e-3 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-2_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8/seed=3_2024-06-19_23-04-23
command 11: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-3_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=3_2024-06-19_23-04-23 --seed 3 --lr 1e-3 --wd 1e-2 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-3_wd=1e-2_beta1=0.9_beta2=0.999_eps=1e-8/seed=3_2024-06-19_23-04-23
command 12: python3 myscript.py --batch_size 128 --epochs 100 --lr_decay_at 82 123 --wandb_project cifar10-training --wandb_group cifar10_rn18_adamw_E=100_bs=128 --wandb_job_type lr=1e-3_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8 --wandb_name seed=3_2024-06-19_23-04-23 --seed 3 --lr 1e-3 --wd 1e-3 --beta1 0.9 --beta2 0.999 --eps 1e-8 --root_folder ./results/cifar10-training/cifar10_rn18_adamw_E=100_bs=128/lr=1e-3_wd=1e-3_beta1=0.9_beta2=0.999_eps=1e-8/seed=3_2024-06-19_23-04-23We also added a wrapper for SBATCH that allows running SLURM jobs directly from Python!
SBATCH(
script='script.sh', # the shell script that will be launched using sbatch
env_vars=dict( # these variables will be set in the --export argument of sbatch and will be available in the h100-eval.sh script
root=ROOT,
wandb_project=wandb_project,
wandb_group=wandb_group,
wandb_job_type=wandb_job_type,
wandb_name=wandb_name,
task='imagenet',
batch_size=128,
),
sbatch_args={ # the variables below will be added to the sbatch command (e.g. --job_name)
'exclude': 'machine-1,machine-2',
'job-name': f'job-name',
'error': 'slurm_output/%j-%x.err',
'output': 'slurm_output/%j-%x.out',
'ntasks': 1,
'cpus-per-task': 20,
'time': '1-00:00:00',
'mem': '200G',
'partition': 'gpu',
'gres': 'gpu:H100:1',
}
).run()We welcome contributions! If you have suggestions for new features or improvements, feel free to open an issue or submit a pull request.
| Back | FazBrowse Home | New Git URL |