coreai_opt.quantization.spec.QParamsCalculatorBase

class coreai_opt.quantization.spec.QParamsCalculatorBase(dtype, qscheme, granularity, target_dtype, quant_min, quant_max, range_calculator, float_range, scale_dtype=None, **kwargs)

Bases: ClassRegistryMixin, Module

Abstract base for qparams calculators — common configuration and helpers.

Concrete subclasses inherit from either:

• StatefulQParamsCalculatorBase — has scale/zp/minval buffers; used by Static, MovingAverage, GlobalMinMax.

• StatelessQParamsCalculatorBase — no buffers, qparams recomputed per forward; used by Dynamic. FakeQuantizeImplBase and Quantizer detect this subclass to keep the observer always on and to reject export.

Subclasses must implement forward(tensor) -> (scale, zero_point, minval).

Parameters:

• dtype (torch.dtype)

• qscheme (QuantizationScheme)

• granularity (QuantizationGranularity)

• target_dtype (torch.dtype)

• quant_min (int)

• quant_max (int)

• range_calculator (RangeCalculatorBase)

• float_range (tuple[float | None, float | None])

• scale_dtype (torch.dtype | None)

__init__(dtype, qscheme, granularity, target_dtype, quant_min, quant_max, range_calculator, float_range, scale_dtype=None, **kwargs)

Parameters:

• dtype (dtype)

• qscheme (QuantizationScheme)

• granularity (QuantizationGranularity)

• target_dtype (dtype)

• quant_min (int)

• quant_max (int)

• range_calculator (RangeCalculatorBase)

• float_range (tuple[float | None, float | None])

• scale_dtype (dtype | None)

Methods

compute_qparams(tensor, min_val, max_val)	Given the observed min/max range, return (scale, zero_point, minval).
forward(tensor)	Compute and return (scale, zero_point, minval) for tensor.
get_class(key)
list_registry_keys()
list_registry_values()
register(key)	Register a virtual subclass of an ABC.
resolve(data)	Resolve a string key or class type against this registry.

compute_qparams(tensor, min_val, max_val)

Given the observed min/max range, return (scale, zero_point, minval).

Default implementation: pure function of the supplied range, no running state. RunningRangeMixin overrides this to apply a running-range smoothing rule before computing qparams.

Parameters:

• tensor (Tensor)

• min_val (Tensor)

• max_val (Tensor)

Return type:

tuple[Tensor, Tensor | None, Tensor | None]

abstract forward(tensor)

Compute and return (scale, zero_point, minval) for tensor.

Parameters:

tensor (Tensor)

Return type:

tuple[Tensor, Tensor | None, Tensor | None]

property granularity: QuantizationGranularity

Getter for granularity.