Fix num_examples for Vietnamese dataset

Change:
eval num_examples: 2912 -> 1137
train num_examples: 2773 -> 4548

for fixing Dataset.load_dataset does not work with Vietnamese dataset

README.md

@@ -98,16 +98,11 @@ dataset_info:
     dtype: float64
   splits:
   - name: train
-    num_bytes: 56584901.453
-    num_examples: 2773
+    num_examples: 4548
   - name: test
-    num_bytes: 69598082.31
     num_examples: 3437
   - name: dev
-    num_bytes: 57617298.896
-    num_examples: 2912
-  download_size: 181789393
-  dataset_size: 183800282.659
+    num_examples: 1137
 configs:
 - config_name: Chinese
   data_files: