diff --git a/flytekit/extras/tensorflow/model.py b/flytekit/extras/tensorflow/model.py
index b9fbf24d4b..ee0c663f00 100644
--- a/flytekit/extras/tensorflow/model.py
+++ b/flytekit/extras/tensorflow/model.py
@@ -20,7 +20,7 @@ def get_literal_type(self, t: Type[tf.keras.Model]) -> LiteralType:
         return LiteralType(
             blob=_core_types.BlobType(
                 format=self.TENSORFLOW_FORMAT,
-                dimensionality=_core_types.BlobType.BlobDimensionality.MULTIPART,
+                dimensionality=_core_types.BlobType.BlobDimensionality.SINGLE,
             )
         )
 
@@ -34,14 +34,14 @@ async def async_to_literal(
         meta = BlobMetadata(
             type=_core_types.BlobType(
                 format=self.TENSORFLOW_FORMAT,
-                dimensionality=_core_types.BlobType.BlobDimensionality.MULTIPART,
+                dimensionality=_core_types.BlobType.BlobDimensionality.SINGLE,
             )
         )
 
-        local_path = ctx.file_access.get_random_local_path()
+        local_path = ctx.file_access.get_random_local_path() + ".keras"
         pathlib.Path(local_path).parent.mkdir(parents=True, exist_ok=True)
 
-        # save model in SavedModel format
+        # save model in Keras format
         tf.keras.models.save_model(python_val, local_path)
 
         remote_path = await ctx.file_access.async_put_raw_data(local_path)
@@ -55,8 +55,8 @@ async def async_to_python_value(
         except AttributeError:
             TypeTransformerFailedError(f"Cannot convert from {lv} to {expected_python_type}")
 
-        local_path = ctx.file_access.get_random_local_path()
-        await ctx.file_access.async_get_data(uri, local_path, is_multipart=True)
+        local_path = ctx.file_access.get_random_local_path() + ".keras"
+        await ctx.file_access.async_get_data(uri, local_path, is_multipart=False)
 
         # load model
         return tf.keras.models.load_model(local_path)
@@ -64,7 +64,7 @@ async def async_to_python_value(
     def guess_python_type(self, literal_type: LiteralType) -> Type[tf.keras.Model]:
         if (
             literal_type.blob is not None
-            and literal_type.blob.dimensionality == _core_types.BlobType.BlobDimensionality.MULTIPART
+            and literal_type.blob.dimensionality == _core_types.BlobType.BlobDimensionality.SINGLE
             and literal_type.blob.format == self.TENSORFLOW_FORMAT
         ):
             return tf.keras.Model
diff --git a/tests/flytekit/unit/extras/tensorflow/model/test_transformations.py b/tests/flytekit/unit/extras/tensorflow/model/test_transformations.py
index 9d7aa12737..7bc487b0a5 100644
--- a/tests/flytekit/unit/extras/tensorflow/model/test_transformations.py
+++ b/tests/flytekit/unit/extras/tensorflow/model/test_transformations.py
@@ -39,7 +39,7 @@ def get_tf_model() -> tf.keras.Model:
 )
 def test_get_literal_type(transformer, python_type, format):
     lt = transformer.get_literal_type(python_type)
-    assert lt == LiteralType(blob=BlobType(format=format, dimensionality=BlobType.BlobDimensionality.MULTIPART))
+    assert lt == LiteralType(blob=BlobType(format=format, dimensionality=BlobType.BlobDimensionality.SINGLE))
 
 
 @pytest.mark.parametrize(
@@ -58,7 +58,7 @@ def test_to_python_value_and_literal(transformer, python_type, format, python_va
     assert lv.scalar.blob.metadata == BlobMetadata(
         type=BlobType(
             format=format,
-            dimensionality=BlobType.BlobDimensionality.MULTIPART,
+            dimensionality=BlobType.BlobDimensionality.SINGLE,
         )
     )
     assert lv.scalar.blob.uri is not None