Use the omniglot dataset. - View it on GitHub
Star
0
Rank
13033951