cross-validation comparison with ReLU and Swish

[bdecost]

Benchmark ReLU vs Swish performance and workload -- I ran some informal comparisons between ReLU networks and swish networks this summer, but didn't record computational workload or run the tests in cross-validation. In the one-off comparison study I had done, ReLU and Swish networks gave roughly equivalent performance on JARVIS e_form and e_gap tasks, and we just switched to Swish networks from that point on