to_affine_quantized_floatx

torchao.dtypes.to_affine_quantized_floatx(input_float: Tensor, block_size: Tuple[int, ...], target_dtype: dtype, _layout: Layout, scale_dtype: Optional[dtype] = None)