[cherry-pick] fix dataloader single process return (#28259)

* fix DataLoader return same format between static & dynamic in single mode. test=develop

python/paddle/fluid/dataloader/dataloader_iter.py

@@ -341,6 +341,12 @@ class _DataLoaderIterSingleProcess(_DataLoaderIterBase):
                 return self._reader.read_next_var_list()
             else:
                 if self._return_list:
+                    # static graph organized data on multi-device with list, if
+                    # place number is 1, there is only 1 device, extra the data
+                    # from list for devices to be compatible with dygraph mode
+                    if len(self._places) == 1:
+                        return self._reader.read_next_list()[0]
+                    else:
                         return self._reader.read_next_list()
                 else:
                     return self._reader.read_next()

python/paddle/fluid/tests/unittests/test_multiprocess_dataloader_static.py

@@ -170,5 +170,50 @@ class TestStaticDataLoader(unittest.TestCase):
             self.assertLess(diff, 1e-2)
 
 
+class TestStaticDataLoaderReturnList(unittest.TestCase):
+    def test_single_place(self):
+        scope = fluid.Scope()
+        image = fluid.data(
+            name='image', shape=[None, IMAGE_SIZE], dtype='float32')
+        label = fluid.data(name='label', shape=[None, 1], dtype='int64')
+        with fluid.scope_guard(scope):
+            dataset = RandomDataset(SAMPLE_NUM, CLASS_NUM)
+            dataloader = DataLoader(
+                dataset,
+                feed_list=[image, label],
+                num_workers=0,
+                batch_size=BATCH_SIZE,
+                drop_last=True,
+                return_list=True)
+
+            for d in dataloader:
+                assert isinstance(d, list)
+                assert len(d) == 2
+                assert not isinstance(d[0], list)
+                assert not isinstance(d[1], list)
+
+    def test_multi_place(self):
+        scope = fluid.Scope()
+        image = fluid.data(
+            name='image', shape=[None, IMAGE_SIZE], dtype='float32')
+        label = fluid.data(name='label', shape=[None, 1], dtype='int64')
+        with fluid.scope_guard(scope):
+            dataset = RandomDataset(SAMPLE_NUM, CLASS_NUM)
+            dataloader = DataLoader(
+                dataset,
+                feed_list=[image, label],
+                num_workers=0,
+                batch_size=BATCH_SIZE,
+                places=[fluid.CPUPlace()] * 2,
+                drop_last=True,
+                return_list=True)
+
+            for d in dataloader:
+                assert isinstance(d, list)
+                assert len(d) == 2
+                assert isinstance(d[0], list)
+                assert isinstance(d[1], list)
+
+
 if __name__ == '__main__':
     unittest.main()