TY - RPRT TI - Practical Quasi-Newton Methods for Training Deep Neural Networks AU - Donald Goldfarb AU - Yi Ren AU - Achraf Bahamou PY - 2021 UR - https://arxiv.org/abs/2006.08877 ID - 2006.08877 ER -