Fix inaccurate linear learner hyper-parameter and bump version to 1.2…

….2 (aws#120)

* Fix inaccurate range and doc in linear learner hyper-parameters

* bump to verison 1.2.2

Login to ECR if needed for Local Mode (aws#121)

Instead of having docker-compose generate a random network each time use a constant network.

When running local mode the SDK will perform an ECR login if required.

Bump version to 1.2.1

Update default versions to TensorFlow 1.6 and MXNet 1.1 (aws#118)

* Update default TF and MXNet versions

* Update changelog

* Bump version to 1.2.0 due to TF / MXNet version changes

* Reorder changelog entries

* Update version info in README

Add new hyper-parameters for linear learner and bump version to v1.1.3 (

aws#112)

* add new hyper-parameters for linear learner

adding hyper-parameters for early stopping and new loss types.

* bump version to 1.1.3

bump version to 1.1.2 (aws#98)

Bump the version from 1.1.0 to 1.1.1 (aws#89)

Update changelog for 1.1.0 (aws#85)

Add ntm algorithm with doc, unit tests, integ tests (aws#73)

Add FM and LDA to the documentation. (aws#66)

Add wrapper for FactorizationMachiones algorithm. (aws#38)

* Add FactorizationMachines class with unit tests.
* Add integration test for FactorizationMachines.
* Add CHANGELOG file.