Bug fixes and other changes
* typo in buildspec-release file
* add sagemaker-tensorflow-extensions to CodePipeline
* upgrade to TF 1.14
* Upgrade to TF 1.13
* add codebuild ci
* Add support for multipart recordio records
* Update README.rst
* Update README.rst
* Update README.rst
* Update README.rst
* add new line after hyperlink
* Upgrade to TF 1.12
* Validate channel name
* Update README.rst
* Update README.rst
* Update README.rst
* Update README.rst
* Update README.rst
* Update README.rst
* Update README.rst
* TF 1.11
* Update README.rst
* Update README.rst
* Update README.rst
* Set tox dependency on tensorflow to 1.10
* Support for TensorFlow 1.10
* Update readme.
* Add flake8 dependnecy on tensorflow. Update readme
* Remove explicit TensorFlow dependency.
* Update PipeModeDataset to open FIFO immediately on Iterator creation
* Update README to describe Python 3 support
* Remove flaky assertion
* Initial Python3 support
* Update integ tests to include 'sagemaker' in generated s3 keys
* Build correct version of package for inclusion in integ test docker image
* Upgrade Master to TF 1.9
* Update README.rst
* Update README.rst
* Update README.rst
* Update README.rst
* Update README.rst
* Update README.rst
* Update README.rst
* Add a check to fail fast if invoked in file mode
* Remove debug statement from RecordReader destructor
* Add extra logging in benchmark
* Integrate benchmark tc
* Add region to benchmark script creation
* Set region in benchmarking bucket helper
* Improve style. Add dependency on tensorflow to benchmark package.
* Add a local benchmark application to benchmarking. Report iterator metrics when benchmarking.
* Remove from_image build arg. Remove dummy.txt testing file
* Do a docker pull of base image prior to building integration test docker image
* Add region to create_integ_test_docker_images
* Remove erroneous comment from benchmarking Dockerfile
* Add region name to boto3 clients
* Add boto3 install_requires to pipemode benchmark
* Set benchmarking as a separate python package
* Add documentation to benchmark components
* Test TC / GH integration
* Remove setup_requires from setup.py setup function call
* Add tensorflow as a setup_requires
* Move cmake to a setup_requires dependency
* Add cmake as a flake8 dep
* Exclude benchmark from flake8
* Add tox dependency on cmake
* Add Header for TextLineRecordReaderTest
* Run python tests and flake8 with tox. Refactor RecordReader hierarchy to remove read-ahead buffer from base class.
* Use old-style super constructor call on CMakeExtension.
* Add a PipeMode Dataset implementation for TF 1.8
* Creating initial file from template
* Creating initial file from template
* Updating initial README.md from template
* Creating initial file from template
* Creating initial file from template
* Creating initial file from template
* Initial commit