Upload Eye-Q (checked locally first: embedded bytes + decode + unique ids)

README.md

@@ -1,22 +0,0 @@
----
-pretty_name: "Eye-Q"
-license: "unknown"
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*.parquet
----
-
-# Eye-Q
-
-Columns: `image`, `answer`, `lang`, `id` (sequential numeric ids across all rows).
-
-## Load
-```python
-from datasets import load_dataset
-ds = load_dataset('llm-lab/Eye-Q')
-print(ds['train'][0])
-```
-
-Total rows: 1342