afe.ir.transform_ir_generator
=============================

.. py:module:: afe.ir.transform_ir_generator

.. autoapi-nested-parse::

   Generation of AwesomeNodes from a tensor transform.



Classes
-------

.. autoapisummary::

   afe.ir.transform_ir_generator.TransformIRGenerationContext
   afe.ir.transform_ir_generator.TransformIRGenerator


Module Contents
---------------

.. py:class:: TransformIRGenerationContext

   Context for generating AwesomeNodes for a tensor transform.


.. py:class:: TransformIRGenerator

   Generator of AwesomeNodes for a tensor transform.  Class instances
   are produced from tensor transforms.


   .. py:method:: generate(generation_info: TransformIRGenerationContext, input_type: afe.ir.tensor_type.TensorType, input_name: Optional[afe.ir.defines.NodeName], output_name: Optional[afe.ir.defines.NodeName]) -> Tuple[afe.ir.defines.NodeName, afe.ir.defines.NodeName, afe.ir.tensor_type.TensorType, List[afe.ir.node.AwesomeNode]]
      :abstractmethod:


      Generate AwesomeNodes for the transform.

      A caller must pass None for at least one of input_name and output_name.

      :param generation_info: The context where generation occurs
      :param input_type: Type of the input node.
      :param input_name: Name of the input node.  If not given, and a name is needed,
         a name will be created.
      :param output_name:  Name of the output node.  If not given, and a name is needed,
         a name will be created.
      :return: Tuple of input node name, output node name, output type, and
         newly created nodes.  The nodes should be inserted into an AwesomeNet.