tabensemb.model.AbstractNN.forward

method

AbstractNN.forward(*tensors: Tensor, data_required_models: Dict[str, DataFrame] | None = None) → Tensor

A wrapper of the original forward of nn.Module for compatibility concerns.

Parameters:

tensors:

Input tensors to the torch model. They have the same structures as the tensors stored in tabensemb.data.datamodule.DataModule()

data_required_models:

The datasets for required models processed by their own AbstractModel._train_data_preprocess() or AbstractModel._data_preprocess() methods. See TorchModel._generate_dataset_for_required_models()

Returns:

torch.Tensor

The output from _forward().