3 ms·
There's no easy way to turn a set of 10-layer Neural Network weights into faster optimized assembly language. You can optimize the Neural Network so that maybe
by dragontamer 9y ago
There's no easy way to turn a set of 10-layer Neural Network weights into faster optimized assembly language.
You can optimize the Neural Network so that maybe it runs faster (using NN specific hardware like maybe TensorCores or whatever), but I don't think anyone has any way to "generalize" a Neural Network into faster code.