seed

set_all_seeds(seed)

Set the seed for all relevant random number generators to ensure reproducibility. This function sets the seed for Python’s built-in random module, NumPy, and PyTorch’s CPU and GPU (CUDA) random number generators. It also configures PyTorch’s settings to improve the reproducibility of experiments, which is crucial when debugging or comparing model performances.

Parameters:

Name	Type	Description	Default
seed	int	The seed value to be set for all random number generators.	required

Example

from spotpython.utils.seed import set_all_seeds set_all_seeds(42)

Proceed with model initialization or data processing to ensure results can be reproduced

model = SomeModel() # Replace with actual model train_model(model) # Replace with actual training function

Notes

• Setting torch.backends.cudnn.deterministic to True can make computations more reproducible but at the potential cost of performance.
• Additional considerations may be necessary for complete reproducibility in distributed or multi-threaded setups.

Source code in spotpython/utils/seed.py

def set_all_seeds(seed: int):
    """Set the seed for all relevant random number generators to ensure reproducibility.
    This function sets the seed for Python's built-in `random` module, NumPy,
    and PyTorch's CPU and GPU (CUDA) random number generators. It also configures
    PyTorch's settings to improve the reproducibility of experiments, which is
    crucial when debugging or comparing model performances.

    Args:
        seed (int): The seed value to be set for all random number generators.

    Example:
        >>> from spotpython.utils.seed import set_all_seeds
        >>> set_all_seeds(42)
        >>> # Proceed with model initialization or data processing to ensure results can be reproduced
        >>> model = SomeModel()  # Replace with actual model
        >>> train_model(model)   # Replace with actual training function

    Notes:
        - Setting `torch.backends.cudnn.deterministic` to `True` can make computations
          more reproducible but at the potential cost of performance.
        - Additional considerations may be necessary for complete reproducibility
          in distributed or multi-threaded setups.
    """
    random.seed(seed)
    np.random.seed(seed)
    torch.manual_seed(seed)
    if torch.cuda.is_available():
        torch.cuda.manual_seed_all(seed)
        torch.backends.cudnn.deterministic = True  # Improvements for reproducibility
        torch.backends.cudnn.benchmark = False

    os.environ["PYTHONHASHSEED"] = str(seed)  # Ensuring hash-based functions use a consistent seed