quantize

burn_cubecl::kernel::quantization

Function quantize 

Source 
pub fn quantize<R>(
    tensor: CubeTensor<R>,
    scheme: &QuantScheme,
    scale: CubeTensor<R>,
) -> CubeTensor<R>
where
    R: CubeRuntime,
Expand description

Convert the tensor to a lower precision data type based on the quantization scheme and parameters.