fix: declare train+test splits in dataset config

README.md

@@ -61,13 +61,18 @@ dataset_info:
     dtype: string
   splits:
   - name: train
-    num_bytes: 76083574
-    num_examples: 64723
+    num_bytes: 64000000
+    num_examples: 55014
+  - name: test
+    num_bytes: 11000000
+    num_examples: 9709
   download_size: 32194993
-  dataset_size: 76083574
+  dataset_size: 75000000
 configs:
 - config_name: default
   data_files:
   - split: train
     path: data/train-*
+  - split: test
+    path: data/test-*
 ---