Revisiting Small Batch Training for Deep NNs; batch_size less than 64 is best (arxiv.org) 1 points by evc123 8y ago ↗ HN
0 comments
[ 2.0 ms ] story [ 7.7 ms ] threadNo comments yet.