quark.onnx.graph_transformations.model_transformer

Module Contents

Classes

class quark.onnx.graph_transformations.model_transformer.ModelTransformer(model: onnx.ModelProto, transforms: List[Any], candidate_nodes: Optional[Dict[str, Any]] = None, node_metadata: Optional[Dict[str, Any]] = None)

Matches patterns to apply transforms in a tf.keras model graph.

class NodeType(*args, **kwds)

Create a collection of name/value pairs.

Example enumeration:

>>> class Color(Enum):
...     RED = 1
...     BLUE = 2
...     GREEN = 3

Access them by:

• attribute access:

  >>> Color.RED
  <Color.RED: 1>
  

• value lookup:

  >>> Color(1)
  <Color.RED: 1>
  

• name lookup:

  >>> Color['RED']
  <Color.RED: 1>
  

Enumerations can be iterated over, and know how many members they have:

>>> len(Color)
3

>>> list(Color)
[<Color.RED: 1>, <Color.BLUE: 2>, <Color.GREEN: 3>]

Methods can be added to enumerations, and members can have their own attributes – see the documentation for details.

transform() → Tuple[onnx.ModelProto, Dict[str, Any]]

Transforms the Onnx model by applying all the specified transforms.

This is the main entry point function used to apply the transformations to the Onnx model.

Not suitable for multi-threaded use. Creates and manipulates internal state.

Returns:

(Onnx model after transformation, Updated node metadata map)