Quantizer

Quantize your network

Quantizer

 Quantizer (backend:str='x86', method:str='static',
            qconfig_mapping:Optional[Dict]=None,
            custom_configs:Optional[Dict]=None, use_per_tensor:bool=False,
            verbose:bool=False)

Initialize a quantizer with specified backend and options.

	Type	Default	Details
backend	str	x86	Target backend for quantization
method	str	static	Quantization method: ‘static’, ‘dynamic’, or ‘qat’
qconfig_mapping	Optional	None	Optional custom quantization config
custom_configs	Optional	None	Custom module-specific configurations
use_per_tensor	bool	False	Force per-tensor quantization
verbose	bool	False	Enable verbose output

source

Quantizer.quantize

 Quantizer.quantize (model:torch.nn.modules.module.Module,
                     calibration_dl:Any, max_calibration_samples:int=100,
                     device:Union[str,torch.device]='cpu')

Quantize a model using the specified method and settings.

	Type	Default	Details
model	Module		Model to quantize
calibration_dl	Any		Dataloader for calibration
max_calibration_samples	int	100	Maximum number of samples to use for calibration
device	Union	cpu	Device to use for calibration
Returns	Module