Maintenance Release 1.3.8-9

Compatible with TF 1.0 and old versions.

• Update
  • tl.cost.cross_entropy : name is not required by TF0.11~0.12
  • tl.cost.cross_entropy_seq : automatically compute the batch_size and num_steps, and support TF1.0
  • tl.rein.cross_entropy_reward_loss for TF1.0
  • shorten the terminal logging e.g. tensorlayer --> [TL]
• Example
  • A simple web service - TensorFlask by JoelKronander
• Documentation
  • @zsdonghao
• TODO
  • debug tutorial_dynamic_rnn.py, tutorial_tfrecord3.py and tutorial_translation.py(the wrapper will be deprecated)

Thanks to @wagamamaz for the useful tricks to use TensorLayer

This version works well under TF 0.10, 0.11, 0.12 and 1.0 backend, feel free to report if you found a bug.
Many thanks!