choose_qparams_affine_floatx

torchao.quantization.choose_qparams_affine_floatx(tensor: 张量, ebits: int, mbits: int) → 张量