pyfemtet.opt.prediction package

Module contents

class pyfemtet.opt.prediction.AbstractModel[source]

Bases: object

fit(x, y, bounds=None, **kwargs)[source]

Parameters:

x (ndarray) –
y (ndarray) –
bounds (ndarray | None) –

predict(x)[source]

Parameters:: x (ndarray) –
Return type:: tuple[ndarray, ndarray]

class pyfemtet.opt.prediction.PyFemtetModel[source]

Bases: object

current_model: AbstractModel

current_trans: _SearchSpaceTransform

fit(history, df, **kwargs)[source]

Parameters:

history (History) –
df (DataFrame) –

history: History

predict(x)[source]

Parameters:: x (ndarray) –
Return type:: tuple[ndarray, ndarray]

update_model(model)[source]

Parameters:: model (AbstractModel) –

class pyfemtet.opt.prediction.SingleTaskGPModel[source]

Bases: AbstractModel

KWARGS = {'device': 'cpu', 'dtype': torch.float64}

fit(x, y, bounds=None, observation_noise=None, likelihood_class=None, covar_module_settings=None)[source]

Parameters:

x (ndarray) –
y (ndarray) –
bounds (ndarray | None) –
covar_module_settings (dict | None) –

gp: SingleTaskGP

predict(x)[source]

Parameters:: x (ndarray) –