Constant deepviewrt_sys::NNQuantizationType_NNQuantizationType_Affine_PerTensor

pub const NNQuantizationType_NNQuantizationType_Affine_PerTensor: NNQuantizationType = 1;

Affine quantization with parameters applied globally across the tensor.

The scale term is queried from @ref nn_tensor_scales() while the zero term is queried from @ref nn_tensor_zeros().

Quantization: \f$ f(x) = \frac{x}{scale} + zero \f$

Dequantization: \f$ f(x) = (x - zero) * scale \f$