Update README.md

README.md

@@ -20,6 +20,25 @@ Common-O is comprised of household objects:
 
 We have two subsets: Common-O (3 - 8 objects) and Common-O Complex (8 - 16 objects).
 
+## Evaluating a Multimodal LLM on Common-O
+
+```python
+from datasets import common_o
+
+
+
+question, image, image2, correct_response, _ = datasets.load("facebook/Common-O")[3]
+
+instructions = ""
+
+prompt = instructions + question
+
+
+output = model(prompt, image, image2)
+
+exact_match(output, correct_response)
+```
+
 
 ## Multimodal LLMs excel at single image perception, but struggle with multi-scene reasoning