MA YUNHAN

Low precision and coarse-to-fine dynamic fixed-point quantization design in convolution neural network