Reducing Communication for Split Learning by Randomized Top-k Sparsification

Open in new window