dmx.compressor.modeling.nn.core.DmxGraph

class dmx.compressor.modeling.nn.core.DmxGraph(owning_module=None, tracer_cls=None, tracer_extras=None)

__init__(owning_module=None, tracer_cls=None, tracer_extras=None): Construct an empty Graph.

Note

Backwards-compatibility for this API is guaranteed.

Methods

`__init__`([owning_module, tracer_cls, ...])	Construct an empty Graph.
`call_function`(the_function[, args, kwargs, ...])	Insert a `call_function` `Node` into the `Graph`.
`call_method`(method_name[, args, kwargs, ...])	Insert a `call_method` `Node` into the `Graph`.
`call_module`(module_name[, args, kwargs, ...])	Insert a `call_module` `Node` into the `Graph`.
`create_node`(op, target[, args, kwargs, ...])	Create a `Node` and add it to the `Graph` at the current insert-point.
`create_placeholders`(names[, cast_names, ...])
`eliminate_dead_code`([is_impure_node])	Remove all dead code from the graph, based on each node's number of users, and whether the nodes have any side effects.
`erase_node`(to_erase)	Erases a `Node` from the `Graph`.
`find_nodes`(*, op[, target, sort])	Allows for fast query of nodes
`get_attr`(qualified_name[, cast_name, ...])	optional_arg: controlling whether None will be returned instead of a Node. e.g. optional_arg = linear_mod.bias will return None when the module does not have a bias term.
`graph_copy`(g, val_map[, return_output_node])	Copy all nodes from a given graph into `self`.
`inserting_after`([n])	Set the point at which create_node and companion methods will insert into the graph.
`inserting_before`([n])	Set the point at which create_node and companion methods will insert into the graph.
`lint`()	Runs various checks on this Graph to make sure it is well-formed.
`node_copy`(node[, arg_transform])	Copy a node from one graph into another. `arg_transform` needs to transform arguments from the graph of node to the graph of self. Example::.
`on_generate_code`(make_transformer)	Register a transformer function when python code is generated
`output`(result[, type_expr])	Insert an `output` `Node` into the `Graph`.
`output_node`()
`placeholder`(name[, cast_name, cast_format, ...])	Insert a `placeholder` node into the Graph.
`print_tabular`()	Prints the intermediate representation of the graph in tabular format.
`process_inputs`(*args)	Processes args so that they can be passed to the FX graph.
`process_outputs`(out)
`python_code`(root_module, *[, verbose, ...])	Turn this `Graph` into valid Python code.
`qdq_node`(node, cast_name, cast_format)
`set_codegen`(codegen)

Attributes

`nodes`	Get the list of Nodes that constitute this Graph.
`owning_module`