TensorFlow是比较底层的深度学习API，TF-Slim、TFLearn、Keras和TensorLayer均尝试简化TensorFlow/Theano，它们对TensorFlow/Theano的进行高层封装，API实现更加的工程化。从1.0开始，TensorFlow官方支持Keras，所以Keras会成为主流。
TF-Slim: A high level library to define complex models in TensorFlow.
链接：https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/slim
TFLearn: A deep learning library featuring a higher-level API for TensorFlow.
链接：http://tflearn.org/
Keras: A deep learning library for TensorFlow/Theano in Python. Convnets, recurrent neural networks, and more.
链接：https://keras.io/
TensorLayer: A deep learning and reinforcement learning library for TensorFlow.
链接：http://tensorlayer.readthedocs.io/en/latest/index.html
PS: 一个对比的讨论：
https://www.zhihu.com/question/50030898/answer/119785042