Add distributed training support with dask

Also fixes #87 (closed)

bob/learn/tensorflow/configs/MirroredStrategy.py

+import tensorflow as tf
+
+
+def strategy_fn():
+    print("Creating MirroredStrategy strategy.")
+    strategy = tf.distribute.MirroredStrategy()
+    print("MirroredStrategy strategy created.")
+    print("Number of devices: {}".format(strategy.num_replicas_in_sync))
+    return strategy