ml/10_pytorch_hp_example.py

import torch
from hyperopt import hp
from sklearn.datasets import load_iris
from sklearn.metrics import accuracy_score, f1_score, log_loss
from sklearn.model_selection import train_test_split
from torch import nn, optim

from astrodata.ml.metrics import SklearnMetric
from astrodata.ml.model_selection import HyperOptSelector
from astrodata.ml.models import PytorchModel

if __name__ == "__main__":
    X, y = load_iris(return_X_y=True)
    X_train, X_val, y_train, y_val = train_test_split(
        X, y, test_size=0.2, random_state=42
    )

    dataset = torch.utils.data.TensorDataset(
        torch.tensor(X_train, dtype=torch.float32),
        torch.tensor(y_train, dtype=torch.long),
    )

    dataset_val = torch.utils.data.TensorDataset(
        torch.tensor(X_val, dtype=torch.float32), torch.tensor(y_val, dtype=torch.long)
    )

    class IrisNet(nn.Module):
        def __init__(self, input_layers, output_layers):
            super().__init__()
            self.layers = nn.Sequential(
                nn.Linear(input_layers, 16), nn.ReLU(), nn.Linear(16, output_layers)
            )

        def forward(self, x):
            return self.layers(x)

    model = PytorchModel(
        model_class=IrisNet,
        loss_fn=nn.CrossEntropyLoss,
        optimizer=optim.AdamW,
        device="cpu",
    )

    print(model)

    param_grid = {
        "model": hp.choice("model", [model]),
        "model_params": hp.choice(
            "model_params", [{"input_layers": X.shape[1], "output_layers": 3}]
        ),
        "optimizer_params": {"lr": hp.uniform("lr", 1e-4, 1e-2)},
        "batch_size": hp.choice("batch_size", [32, 64]),
        "epochs": hp.choice("epochs", [5, 10, 15]),
    }

    accuracy = SklearnMetric(accuracy_score, greater_is_better=True)
    f1 = SklearnMetric(f1_score, average="micro")
    logloss = SklearnMetric(log_loss)

    metrics = [accuracy, f1, logloss]

    hos = HyperOptSelector(
        param_space=param_grid,
        scorer=accuracy,
        use_cv=False,
        val_size=0.1,
        random_state=42,
        max_evals=100,
        metrics=metrics,
    )

    hos.fit(dataset_train=dataset, dataset_val=dataset_val)

    print(hos.get_best_params())
    print(hos.get_best_metrics())