QuantileStrategy.update

QuantileStrategy.update(samples: Dict[str, torch.Tensor | array], objectives: torch.Tensor | array, validity: torch.Tensor | array)

Updates the state of the sampling methods by providing the performance of the generated samples in the form of objectives and validity.

Parameters:

• samples (Dict[str, Union[torch.Tensor, np.array]]) – The samnples at which the objectives and conditions where evaluated.

• objectives (Union[torch.Tensor, np.array]) – The performance of the generated samples according to some objective, higher the better (of type float and shape (n, 1)).

• valid (Union[torch.Tensor, np.array]) – Whether the generated samples were valid or not (of type boolean and shape (n, 1))