MWMS tutorial: Replace tf.distribute.experimental.MultiWorkerMirrored…

…Strategy with tf.distribute.MultiWorkerMirroredStrategy. PiperOrigin-RevId: 348508697
edmontdants · Dec 21, 2020 · 187adca · 187adca
1 parent 2b72cbe
commit 187adca
Showing 1 changed file with 5 additions and 4 deletions.
diff --git a/site/en/tutorials/distribute/multi_worker_with_keras.ipynb b/site/en/tutorials/distribute/multi_worker_with_keras.ipynb
@@ -63,7 +63,7 @@
       "source": [
         "## Overview\n",
         "\n",
-        "This tutorial demonstrates multi-worker distributed training with Keras model using `tf.distribute.Strategy` API, specifically `tf.distribute.experimental.MultiWorkerMirroredStrategy`. With the help of this strategy, a Keras model that was designed to run on single-worker can seamlessly work on multiple workers with minimal code change.\n",
+        "This tutorial demonstrates multi-worker distributed training with Keras model using `tf.distribute.Strategy` API, specifically `tf.distribute.MultiWorkerMirroredStrategy`. With the help of this strategy, a Keras model that was designed to run on single-worker can seamlessly work on multiple workers with minimal code change.\n",
         "\n",
         "[Distributed Training in TensorFlow](../../guide/distributed_training.ipynb) guide is available for an overview of the distribution strategies TensorFlow supports for those interested in a deeper understanding of `tf.distribute.Strategy` APIs.\n"
       ]
@@ -422,7 +422,7 @@
         "* Asynchronous training, where the training steps are not strictly synced.\n",
         "\n",
         "`MultiWorkerMirroredStrategy`, which is the recommended strategy for synchronous multi-worker training, will be demonstrated in this guide.\n",
-        "To train the model, use an instance of `tf.distribute.experimental.MultiWorkerMirroredStrategy`.\n",
+        "To train the model, use an instance of `tf.distribute.MultiWorkerMirroredStrategy`.\n",
         "\n",
         "`MultiWorkerMirroredStrategy` creates copies of all variables in the model's layers on each device across all workers.  It uses `CollectiveOps`, a TensorFlow op for collective communication, to aggregate gradients and keep the variables in sync.  The [`tf.distribute.Strategy` guide](../../guide/distributed_training.ipynb) has more details about this strategy."
       ]
@@ -435,7 +435,7 @@
       },
       "outputs": [],
       "source": [
-        "strategy = tf.distribute.experimental.MultiWorkerMirroredStrategy()"
+        "strategy = tf.distribute.MultiWorkerMirroredStrategy()"
       ]
     },
     {
@@ -520,7 +520,7 @@
         "tf_config = json.loads(os.environ['TF_CONFIG'])\n",
         "num_workers = len(tf_config['cluster']['worker'])\n",
         "\n",
-        "strategy = tf.distribute.experimental.MultiWorkerMirroredStrategy()\n",
+        "strategy = tf.distribute.MultiWorkerMirroredStrategy()\n",
         "\n",
         "global_batch_size = per_worker_batch_size * num_workers\n",
         "multi_worker_dataset = mnist.mnist_dataset(global_batch_size)\n",
@@ -1126,6 +1126,7 @@
     "colab": {
       "collapsed_sections": [],
       "name": "multi_worker_with_keras.ipynb",
+      "provenance": [],
       "toc_visible": true
     },
     "kernelspec": {