1 paper
Jerry Chee, Arturs Backurs, Rainie Heck +4
Quantizing the weights of a neural network has two steps: (1) Finding a good low bit-complexity representation for weights (which we call the quantization grid) and (2) Rounding th…