pymllm.quantize.kai.w4a32

Classes

W4A32KAIQuantizePass

Helper class that provides a standard way to create an ABC using

Module Contents

class pymllm.quantize.kai.w4a32.W4A32KAIQuantizePass

Bases: pymllm.quantize.quantize_pass.QuantizeBasePass

Helper class that provides a standard way to create an ABC using inheritance.

prepare(quantize_config, tensor_dict, **kwargs)

Parameters:

tensor_dict (Dict)

Return type:

pymllm.quantize.quantize_pass.QuantizePlanPayload

match(quantize_config, tensor_dict, **kwargs)

Parameters:

tensor_dict (Dict)

Return type:

bool

run(quantize_config, tensor_dict, **kwargs)

Parameters:

tensor_dict (Dict)

Return type:

Dict