diff --git a/tensorflow/python/keras/engine/data_adapter.py b/tensorflow/python/keras/engine/data_adapter.py
index 3672ef64da3..33c868d02be 100644
--- a/tensorflow/python/keras/engine/data_adapter.py
+++ b/tensorflow/python/keras/engine/data_adapter.py
@@ -681,6 +681,7 @@ class DatasetAdapter(DataAdapter):
                y=None,
                sample_weights=None,
                steps=None,
+               batch_size=None,
                **kwargs):
     super(DatasetAdapter, self).__init__(x, y, **kwargs)
     # Note that the dataset instance is immutable, its fine to reuse the user
@@ -690,7 +691,7 @@ class DatasetAdapter(DataAdapter):
     # The user-provided steps.
     self._user_steps = steps
 
-    self._validate_args(y, sample_weights, steps)
+    self._validate_args(y, sample_weights, steps, batch_size)
 
   def get_dataset(self):
     return self._dataset
@@ -719,7 +720,7 @@ class DatasetAdapter(DataAdapter):
     return (self._user_steps is None or
             cardinality.cardinality(self._dataset).numpy() == self._user_steps)
 
-  def _validate_args(self, y, sample_weights, steps):
+  def _validate_args(self, y, sample_weights, steps, batch_size):
     """Validates `__init__` arguments."""
     # Arguments that shouldn't be passed.
     if not is_none_or_empty(y):
@@ -729,6 +730,10 @@ class DatasetAdapter(DataAdapter):
       raise ValueError("`sample_weight` argument is not supported when using "
                        "dataset as input.")
 
+    if batch_size is not None:
+      raise ValueError("`batch_size` argument must not be specified when "
+                       "using dataset as input.")
+
     if steps is None:
       if _is_distributed_dataset(self._dataset):
         raise ValueError("When providing a distributed dataset, you must "