Update README.md

README.md

@@ -8,6 +8,10 @@ multilinguality:
 - translation
 task_ids:
 - machine-translation
+size_categories:
+- 10K<n<100K
+source_datasets:
+- original
 ---
 
 # Dataset Summary: 
@@ -49,4 +53,12 @@ DatasetDict({
     'engyay': 'own-thray into-ay isorder-day ey-thay eturned-ray ith-way ore-may ecipitation-pray an-thay is-ay usual-ay'
   }
 }
-```
+```
+
+### Data Fields
+
+- `translation`: a dictionary containing two strings paired with a key indicating the corresponding language.
+
+### Data Splits
+- `train`: most of the datas, 14,778 samples. 
+- `dev`: 1k holdout samples, created with the datasets.train_test_split() function