diff --git a/mindspore/dataset/engine/datasets.py b/mindspore/dataset/engine/datasets.py
index 63cb7d2d481149bd3017086073295d4469680e76..0058eebc5835c244ba1722d1092935c17950737c 100644
--- a/mindspore/dataset/engine/datasets.py
+++ b/mindspore/dataset/engine/datasets.py
@@ -2284,6 +2284,7 @@ class ImageFolderDatasetV2(MappableDataset):
         self.decode = decode
         self.num_shards = num_shards
         self.shard_id = shard_id
+        self.cur_dataset_size = None
 
     def get_args(self):
         args = super().get_args()
@@ -2305,6 +2306,9 @@ class ImageFolderDatasetV2(MappableDataset):
         Return:
             Number, number of batches.
         """
+        if self.cur_dataset_size is not None:
+            return self.cur_dataset_size
+
         if self.num_samples is None:
             num_samples = 0
         else:
@@ -2314,9 +2318,11 @@ class ImageFolderDatasetV2(MappableDataset):
         rows_from_sampler = self._get_sampler_dataset_size()
 
         if rows_from_sampler is None:
+            self.cur_dataset_size = rows_per_shard
             return rows_per_shard
 
-        return min(rows_from_sampler, rows_per_shard)
+        self.cur_dataset_size = min(rows_from_sampler, rows_per_shard)
+        return self.cur_dataset_size
 
     def num_classes(self):
         """
@@ -2509,6 +2515,7 @@ class MindDataset(SourceDataset):
         self.shuffle_option = shuffle
         self.distribution = ""
         self.sampler = sampler
+        self.cur_dataset_size = None
 
         if num_shards is None or shard_id is None:
             self.partitions = None
@@ -2578,6 +2585,9 @@ class MindDataset(SourceDataset):
             Number, number of batches.
         """
         if self._dataset_size is None:
+            if self.cur_dataset_size is not None:
+                return self.cur_dataset_size
+
             if self.load_dataset:
                 dataset_file = [self.dataset_file]
             else:
@@ -2591,6 +2601,7 @@ class MindDataset(SourceDataset):
                         raise RuntimeError(
                             "Dataset size plus number of padded samples is not divisible by number of shards.")
                     num_rows = num_rows // self.partitions[0] + 1
+            self.cur_dataset_size = num_rows
             return num_rows
         return self._dataset_size