L2L in TensorFlow

Learning to Learn (L2L) is a TensorFlow implementation of 'Learning to learn by gradient descent by gradient descent', a fully differentiable approach which learns optimization algorithms by optimizing them.

OpenSource

14 Jun 2016