za chto

def base_execution1(model: object, X_train: t.Tensor, X_test: t.Tensor, y_train: t.Tensor, y_test: t.Tensor) -> list:

    optimizer = t.optim.Adam(model.parameters(), lr=1e-3)

    loss = nn.MSELoss(reduction='mean')

    num_epochs = 1000
    b_size = 30

    train_data = DataLoader(TensorDataset(X_train, y_train), batch_size=b_size, shuffle=True)

    train_losses = []
    test_losses = []

    for epoch in range(num_epochs):

        total_train_loss = 0
        total_test_loss = 0

        for X_train, y_train in train_data:

            optimizer.zero_grad()

            preds = model.forward(X_train).flatten()

            loss_train = loss(preds, y_train)
            total_train_loss += math.sqrt(loss_train.item())

            loss_train.backward()

            optimizer.step()

        if epoch % 5 == 0:

            model.eval()
            test_preds = model.forward(X_test)
            loss_test = loss(test_preds, y_test)
            total_test_loss += math.sqrt(loss_test.item())

        train_losses.append(total_train_loss)
        test_losses.append(total_test_loss)

    plt.plot(train_losses)
    plt.show()

    return train_losses, test_losses