Add train_003 from tar 2

README.md

@@ -19,11 +19,16 @@ dataset_info:
   - name: train
     num_bytes: 990604365.0
     num_examples: 10000
-  download_size: 945799736
-  dataset_size: 990604365.0
+  - name: train_003
+    num_bytes: 489097550.0
+    num_examples: 5000
+  download_size: 1425581291
+  dataset_size: 1479701915.0
 configs:
 - config_name: default
   data_files:
   - split: train
     path: data/train-*
+  - split: train_003
+    path: data/train_003-*
 ---

data/train_003-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01ceafadaae8fec8edf5dd3855312c06cd3bfb8948f1e6f9627a761b375bc52e
+size 479781555