Convert dataset to Parquet

Convert dataset to Parquet.

README.md

@@ -19,13 +19,20 @@ dataset_info:
     dtype: string
   splits:
   - name: train
-    num_bytes: 6389954
+    num_bytes: 6327209
     num_examples: 33410
   - name: validation
-    num_bytes: 376508
+    num_bytes: 372815
     num_examples: 1954
-  download_size: 2198056
-  dataset_size: 6766462
+  download_size: 3678635
+  dataset_size: 6700024
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+  - split: validation
+    path: data/validation-*
 ---
 
 # Dataset Card for "social_i_qa"

data/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e3ec841d22f62a02a7a1c024bde076d735c5adc580a3455d70ae89d0144b2d0
+size 3475355

data/validation-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70b1ce33fe17e0b843761a5c34717717a30669957abd184fd94ed463f5922192
+size 203280