QATQuantizationSettings

CLASS - QATQuantizationSettings(

• weights_num_bits: int =  8
• activations_num_bits: int =  8
• skip_tail_quantization: Optional[bool] =  None
• automatic_skip_quantization: bool =  True
• quantization_sensitivity_threshold: Union[int, float, None] =  None

)

Ancestors - (BaseQATQuantizationSettings, BaseQuantizationSettings)

Use this if you wish to do QAT Quantization.