Watch keynotes, product sessions, workshops, and more from Google I/O See playlist


View source on GitHub

Conceptually tracks indices of arguments of "OpHint functions".

The inputs and arguments of these functions both use an instance of the class so they can have independent numbering.

function_name Name of the function that this tracks arguments for.
unique_function_id UUID of function that this tracks arguments for.
node_name_prefix How identities that are created are named.
attr_name Name of attribute to use to store the index for this hint. i.e. FUNCTION_INPUT_INDEX or FUNCTION_OUTPUT_INDEX
level Hierarchical level of the Ophint node, a number.
children_inputs_mappings Inputs/Outputs mapping for children hints.



View source

Return a wrapped tensor of an input tensor as an argument.

arg A TensorFlow tensor that should be considered an argument.
tag String tag to identify arguments that should be packed.
name Name of argument. This is included in the Identity hint op names.
aggregate Strategy to aggregate. Acceptable values are OpHint.AGGREGATE_FIRST, OpHint.AGGREGATE_LAST, and OpHint.AGGREGATE_STACK. Note, aggregate is only valid if tag is specified.
index_override Specify what input/output index should this be in the final stub. i.e. add(arg0, index=1); add(arg1, index=0) will make the final stub be as stub_func(inputs[arg1, arg0], outputs=[]) rather than the default call order based ordering.

A tensor representing the wrapped argument.

ValueError When indices are not consistent.