Like DARTS, except subtract the max weight gradients.
Max-W Weighting: \begin{equation} output_i = (1 - max(w) + w_i) * op_i(input_i) \label{eqn:max_w} \end{equation}
Source: sharpDARTS: Faster and More Accurate Differentiable Architecture SearchPaper | Code | Results | Date | Stars |
---|