Trait burn::tensor::quantization::Calibration

pub trait Calibration {
    // Required method
    fn compute_range<B, const D: usize>(
        &self,
        tensor: &Tensor<B, D>,
    ) -> CalibrationRange<B>
       where B: Backend;
}

Calibration method used to compute the quantization range mapping.

Required Methods

fn compute_range<B, const D: usize>( &self, tensor: &Tensor<B, D>, ) -> CalibrationRange<B>

where B: Backend,

Compute the input tensor range.

Object Safety

This trait is not object safe.

Implementors

impl Calibration for MinMaxCalibration