NodePiece Transforms

NodePieceMLPTransform

The NodePieceMLPTransform converts a batch of data from the NodePieceLoader into a format that can be used in a MLP implemented in PyTorch.

_init_()

init(label: str, features: list, target_type: str = None)

Instantiate a NodePieceMLPTransform.

Parameters:

label (str): The attribute name that corresponds with the label being predicted. Supports numerical datatypes (INT, FLOAT, DOUBLE). * features (list of str, optional): List of attributes to use as features into the model. Supports numerical datatypes (INT, FLOAT, DOUBLE). * target_type (str, optional): The type of vertex to perform predictions on.

call()

call(data)

Perform the transform.

Parameter:

  • data (pd.DataFrame): Batch of data produced by the NodePiece Loader.