We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Implementing some experiments from "Qualitatively characterizing neural network optimization problems", ICLR 2015