Initial upload of synthetic OKVQA dataset with key objects

README.md

@@ -1152,6 +1152,30 @@ dataset_info:
     num_examples: 49
   download_size: 812609
   dataset_size: 2328040.0
+- config_name: from4800_4900
+  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: 2312402.0
+    num_examples: 47
+  download_size: 853929
+  dataset_size: 2312402.0
 - config_name: from500_600
   features:
   - name: idx
@@ -1449,4 +1473,8 @@ configs:
   data_files:
   - split: train
     path: data/from4700_4800/train-*
+- config_name: from4800_4900
+  data_files:
+  - split: train
+    path: data/from4800_4900/train-*
 ---

data/from4800_4900/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b2e26ffb7a540ab0868b8f00db8e2a6f1acb1eb93645cb7877d6ab71b5c55e0
+size 853929