A study for Karpathy's The Unreasonable Effectiveness of Recurrent Neural Networks published on May 21, 2015.

I read this post in 2016 in the EZOI room, maybe 2017, I don't know. It was a translation posted on Zhihu. I was deeply surprised then, but I always had no enough bravery to try it myself. Only after the winter camp in SCUT had I got confident. It feels great to give out a try, and it always pays.