quark.onnx.equalization

quark.onnx.equalization#

Module Contents#

Classes#

Functions#

class quark.onnx.equalization.CLE_PAIR_TYPE(*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.

class quark.onnx.equalization.Equalization(model: onnx.ModelProto, op_types_to_quantize: List[str], nodes_to_quantize: List[str] | None, nodes_to_exclude: List[str] | None)#

A class for layers equalization

Args:

model (onnx.ModelProto): The ONNX model to be optimized. op_types_to_quantize (list): A list of operation types to be quantized. nodes_to_quantize (list): A list of node names to be quantized. nodes_to_exclude (list): A list of node names to be excluded from quantization.

quark.onnx.equalization.cle_transforms(model: onnx.ModelProto, op_types_to_quantize: List[str], nodes_to_quantize: List[str], nodes_to_exclude: List[str], cle_steps: int = -1, cle_balance_method: str = 'max', cle_weight_threshold: float = 0.5, cle_scale_append_bias: bool = True, cle_scale_use_threshold: bool = True, cle_total_layer_diff_threshold: float = 1.9e-07) → Any#

Equanlization transform models.

Contents