A arxiv.org Revisiting Small Batch Training for Deep Neural Networks 3 points tempw 8 years ago 0 comments