Initial upload of synthetic OKVQA dataset with key objects

README.md

@@ -360,6 +360,30 @@ dataset_info:
     num_examples: 50
   download_size: 906313
   dataset_size: 2506908.0
+- config_name: from2000_2100
+  features:
+  - name: idx
+    dtype: float64
+  - name: img
+    dtype: image
+  - name: original_question
+    dtype: string
+  - name: gt_ans
+    dtype: string
+  - name: key_object
+    dtype: string
+  - name: question
+    dtype: string
+  - name: answer
+    dtype: string
+  - name: vqa_sample
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 2876385.0
+    num_examples: 50
+  download_size: 970782
+  dataset_size: 2876385.0
 - config_name: from200_300
   features:
   - name: idx
@@ -725,6 +749,10 @@ configs:
   data_files:
   - split: train
     path: data/from1900_2000/train-*
+- config_name: from2000_2100
+  data_files:
+  - split: train
+    path: data/from2000_2100/train-*
 - config_name: from200_300
   data_files:
   - split: train

data/from2000_2100/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f60423da69209f2835fd92727294a641cd75f7b06f1d00e4c385699d02e8c555
+size 970782