Leona/fix dstraining (#2453)

* IntelGPU backend: Separate backend and executable classes (#2447)

* Add notice to main ToC page about distributed training

doc/sphinx/source/distr/index.rst

@@ -4,6 +4,12 @@
 Distributed Training in nGraph
 ##############################
 
+
+.. important:: Distributed training is not officially supported in version |version|;
+   however, some configuration options have worked for nGraph devices with mixed or 
+   limited success in testing environments.
+
+
 Why distributed training?
 =========================
 

doc/sphinx/source/howto/distribute-train.rst

@@ -4,11 +4,16 @@
 Distribute training across multiple nGraph backends 
 ===================================================
 
+.. important:: Distributed training is not officially supported in version |version|;
+    however, the following configuration options have worked for nGraph devices 
+    with mixed or limited success in testing.
+
 In the :doc:`previous section <../howto/derive-for-training>`, we described the 
 steps needed to create a "trainable" nGraph model. Here we demonstrate how to 
 train a data parallel model by distributing the graph to more than one device.
 
-These options are currently supported for available backends:
+Frameworks can implement distributed training with nGraph versions prior to 
+`0.13`:
 
 * Use ``-DNGRAPH_DISTRIBUTED_OMPI_ENABLE=TRUE`` to enable distributed training 
   with OpenMPI. Use of this flag requires that OpenMPI be a pre-existing library 
@@ -18,7 +23,7 @@ These options are currently supported for available backends:
 * Use ``-DNGRAPH_DISTRIBUTED_MLSL_ENABLE=TRUE`` to enable the option for 
   :abbr:`Intel® Machine Learning Scaling Library (MLSL)` for Linux* OS:
 
-  .. important:: The Intel® MLSL option applies to Intel® Architecture CPUs 
+  .. note:: The Intel® MLSL option applies to Intel® Architecture CPUs 
      (``CPU``) and ``Interpreter`` backends only. For all other backends, 
      ``OpenMPI`` is presently the only supported option. We recommend the 
      use of `Intel MLSL` for CPU backends to avoid an extra download step.