1. Remove custom C++ ops from this repo.

   These are currently still included in base TensorFlow and C++ ABI issues
   are preventing us to conviently package them separately for the time being.
2. Update API to work with latest nightly Keras changes.

Author: Nick Johnston

README.md

@@ -5,20 +5,12 @@ This package contains data compression ops and layers for TensorFlow.
 For usage questions and discussions, please head over to our
 [Google group](https://groups.google.com/forum/#!forum/tensorflow-compression)!
 
-## Compiling
+## Prerequisite
 
 **Please note**: You need TensorFlow 1.9 (or the master branch as of May 2018)
 or later.
 
-First, compile the custom ops needed by TensorFlow.
-```shell
-cd compression
-chmod +x compile.sh
-./compile.sh
-cd ..
-```
-
-To make sure the compilation and library imports succeed, try running the two
+To make sure the library imports succeed, try running the two
 tests.
 ```
 python compression/python/ops/coder_ops_test.py

compression/__init__.py

@@ -21,9 +21,6 @@
 
 # pylint: disable=wildcard-import
 from compression.python.layers.entropybottleneck import *
-# TODO(jonycgn,nmjohn): revert to importing from here once we figure out why the
-# nightlies don't work.
-# from compression.python.ops.coder_ops import *
 from tensorflow.contrib.coder.python.ops.coder_ops import *
 # pylint: enable=wildcard-import