core.experiment

core.experiment

Classes

Name	Description
ExperimentControl	Controls the experiment configuration.

ExperimentControl

core.experiment.ExperimentControl(
    dataset,
    model_class,
    hyperparameters,
    seed=123,
    device='cpu',
    num_workers=0,
    epochs=None,
    batch_size=32,
    optimizer_class=None,
    loss_function=None,
    metrics=list(),
    n_initial=10,
    max_evals=50,
    experiment_name='default_experiment',
    verbosity=0,
)

Controls the experiment configuration.

This data class serves as the central configuration object for optimization experiments, holding the dataset, model configuration, hyperparameters, and other settings.

Attributes

Name	Type	Description
dataset	SpotDataSet	The dataset object containing the data.
model_class	Any	The class of the model to be instantiated.
hyperparameters	Any	The hyperparameters for the model.
seed	int	The random seed for reproducibility.
device	str	The device to run the model on (e.g. “cpu” or “cuda”).
num_workers	int	The number of workers to use for data loading.
epochs	Optional[int]	The number of epochs to train the model.
batch_size	int	The batch size for training.
optimizer_class	Optional[Any]	The optimizer class to use for training.
loss_function	Optional[Any]	The loss function to use for training.
metrics	List[str]	The metrics to track during training.
n_initial	int	The number of initial design points.
max_evals	int	The maximum number of evaluations.
experiment_name	str	The name of the experiment.
verbosity	int	The verbosity level.

Public Methods

to_dict(): Convert the object to a dictionary. torch_device(): Return the torch device object.

Examples

import numpy as np
import torch
from spotoptim.core.data import SpotDataFromArray
from spotoptim.core.experiment import ExperimentControl
from spotoptim.nn.mlp import MLP

# 1. Prepare Data
X = np.array([[0.1, 0.2], [0.3, 0.4]])
y = np.array([[1.0], [2.0]])
dataset = SpotDataFromArray(x_train=X, y_train=y)

# 2. Define Hyperparameters
params = {"l1": 16, "num_hidden_layers": 1, "lr": 1e-3}

# 3. Initialize Control with Real Model
exp = ExperimentControl(
    dataset=dataset,
    model_class=MLP,
    hyperparameters=params,
    experiment_name="real_model_run",
    seed=42
)
print(exp.to_dict())

{'dataset': <spotoptim.core.data.SpotDataFromArray object at 0x11eec8050>, 'model_class': <class 'spotoptim.nn.mlp.MLP'>, 'hyperparameters': {'l1': 16, 'num_hidden_layers': 1, 'lr': 0.001}, 'seed': 42, 'device': 'cpu', 'num_workers': 0, 'epochs': None, 'batch_size': 32, 'optimizer_class': None, 'loss_function': None, 'metrics': [], 'n_initial': 10, 'max_evals': 50, 'experiment_name': 'real_model_run', 'verbosity': 0}

Methods

Name	Description
to_dict	Convert to dictionary for partial backward compatibility or logging.

to_dict

core.experiment.ExperimentControl.to_dict()

Convert to dictionary for partial backward compatibility or logging.

Returns

Name	Type	Description
	Dict[str, Any]	Dict[str, Any]: Dictionary representation of the object.

Examples

import numpy as np
from spotoptim.core.data import SpotDataFromArray
from spotoptim.core.experiment import ExperimentControl

# Setup
dataset = SpotDataFromArray(np.zeros((5,2)), np.zeros((5,1)))
exp = ExperimentControl(dataset, model_class=None, hyperparameters={})

# Convert to dict
config = exp.to_dict()
print(config['seed'])

123