Struct F32ToInt8 

pub struct F32ToInt8 {
    pub params: QuantizationParams,
    pub calibration: Option<CalibrationData>,
    pub tensor_params: Vec<(u32, u32, QuantizationParams)>,
    pub calibration_result: Option<CalibrationResult>,
}

Converts array<f32> global variables to array<i8>.

When calibration data is provided, computes proper scale/zero_point per tensor from observed min/max ranges. When a CalibrationResult is attached, calibrated parameters are available for downstream backends to emit quantization metadata (QDQ nodes, TFLite quant params).

Fields

params: QuantizationParams

Per-tensor quantization parameters.

calibration: Option<CalibrationData>

Optional calibration data for computing scale/zero_point.

tensor_params: Vec<(u32, u32, QuantizationParams)>

Per-tensor computed parameters (populated after run).

calibration_result: Option<CalibrationResult>

Optional calibration result from the calibration pipeline.

Implementations

impl F32ToInt8

pub fn with_calibration(calibration: CalibrationData) -> Self

Create with calibration data.

pub fn with_calibration_result(result: CalibrationResult) -> Self

Create with a calibration result from the calibration pipeline.

Trait Implementations

impl Debug for F32ToInt8

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for F32ToInt8

fn default() -> Self

Returns the “default value” for a type.

impl Pass for F32ToInt8

fn name(&self) -> &str

Human-readable name of the pass.

fn run(&self, module: &mut Module) -> bool

Run the pass on a module. Returns true if anything was modified.