HybridModel

class neuralhydrology.modelzoo.hybridmodel.HybridModel(cfg: Config)

Bases: BaseModel

Wrapper to combine a deep learning model with a conceptual hydrological models.

In the current implementation, the deep learning model is always an LSTM. The conceptual model is configurable using the config argument conceptual_model. Currently supported is [‘SHM’].

Parameters:

cfg (Config) – The run configuration.

forward(data: dict[str, Tensor | dict[str, Tensor]]) → Dict[str, Tensor]

Perform a forward pass on the model.

Parameters:

data (Dict[str, torch.Tensor | dict[str, torch.Tensor]]) – Dictionary, containing input features as key-value pairs.

Returns:

Model outputs, dynamic parameters and intermediate states coming from the conceptual model

Return type:

Dict[str, torch.Tensor]