Split val into 22 parts of 5000 (last is remainder)

README.md

@@ -8,24 +8,14 @@ dataset_info:
   - name: short_captions
     dtype: string
   splits:
-  - name: train
-    num_bytes: 508311861
-    num_examples: 217965
-  - name: test
-    num_bytes: 14742278
-    num_examples: 5709
-  - name: val
-    num_bytes: 280378795
-    num_examples: 108456
-  download_size: 151441323
-  dataset_size: 295121073
+  - name: val_00
+    num_bytes: 12636847
+    num_examples: 5000
+  download_size: 6478535
+  dataset_size: 12636847
 configs:
 - config_name: default
   data_files:
-  - split: test
-    path: data/test-*
-  - split: val
-    path: data/val-*
-  - split: train
-    path: data/train-*
+  - split: val_00
+    path: data/val_00-*
 ---

data/val_00-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0702cb01f590edaf72319119abfd6c352f00bdc6276509ec36bd2125113d549f
+size 6478535