Hello and thank you for an interesting paper!
I have a question concerning the optimization of the quantization step size. In section D.1 of the paper, you mentioned the usage of Adam optimizer for solving the AdaQuant optimization problem, but I have failed to find a description of the methodology used for backpropagating gradients through formally nondifferentiable rounding operation. Is it just Straight Through Estimator?
Thank you for your time and consideration.
Hello and thank you for an interesting paper!
I have a question concerning the optimization of the quantization step size. In section D.1 of the paper, you mentioned the usage of Adam optimizer for solving the AdaQuant optimization problem, but I have failed to find a description of the methodology used for backpropagating gradients through formally nondifferentiable rounding operation. Is it just Straight Through Estimator?
Thank you for your time and consideration.