Understanding the effects of data parallelism and sparsity on neural network training.

Namhoon Lee, Thalaiyasingam Ajanthan, Philip H. S. Torr, Martin Jaggi

Browse the full ICLR paper archive.