diff --git a/RELEASE.md b/RELEASE.md
index 9136fd3de7d..f1e4aeaab37 100644
--- a/RELEASE.md
+++ b/RELEASE.md
@@ -47,6 +47,10 @@
         `SidecarEvaluator` evaluator. The evaluator regularly evaluates the
         model and exports it if the user-defined comparison function determines
         that it is an improvement.
+    *   Added warmup capabilities to `tf.keras.optimizers.schedules.CosineDecay`
+        learning rate scheduler. You can now specify an initial and target
+        learning rate, and our scheduler will perform a linear interpolation
+        between the two after which it will begin a decay phase.
 
 *   `tf.function`:
     * ConcreteFunction (`tf.types.experimental.ConcreteFunction`) as generated
diff --git a/tensorflow/tools/api/golden/v2/tensorflow.optimizers.schedules.-cosine-decay.pbtxt b/tensorflow/tools/api/golden/v2/tensorflow.optimizers.schedules.-cosine-decay.pbtxt
index 004d74073c6..851e2059c8d 100644
--- a/tensorflow/tools/api/golden/v2/tensorflow.optimizers.schedules.-cosine-decay.pbtxt
+++ b/tensorflow/tools/api/golden/v2/tensorflow.optimizers.schedules.-cosine-decay.pbtxt
@@ -5,7 +5,7 @@ tf_class {
   is_instance: "<type \'object\'>"
   member_method {
     name: "__init__"
-    argspec: "args=[\'self\', \'initial_learning_rate\', \'decay_steps\', \'alpha\', \'name\'], varargs=None, keywords=None, defaults=[\'0.0\', \'None\'], "
+    argspec: "args=[\'self\', \'initial_learning_rate\', \'decay_steps\', \'alpha\', \'name\', \'warmup_target\', \'warmup_steps\'], varargs=None, keywords=None, defaults=[\'0.0\', \'None\', \'None\', \'0\'], "
   }
   member_method {
     name: "from_config"