Gefu Tang

Results 3 issues of Gefu Tang

use highway network and employ decaying learning rate in the new code.

I think `tf.train.exponential_decay` is a more better and more standard method than `sess.run(tf.assign(model.lr, args.learning_rate * (args.decay_rate ** e)))` to employ decaying learning rate.

Here is my discuss of this probem: https://leetcode.com/discuss/35128/accepted-solution-use-backtracking-bitmask-easy-understand