Try creating cropped images from bounding boxes

app.py

@@ -57,12 +57,16 @@ def detect_objects(img: Image.Image):
     target_sizes = torch.tensor([tuple(reversed(img.size))])
     results = yolos_processor.post_process_object_detection(outputs, threshold=0.7, target_sizes=target_sizes)[0]
 
+    box_images = []
     for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
         box = [round(i, 2) for i in box.tolist()]
         print(
             f"Detected {yolos_model.config.id2label[label.item()]} with confidence "
             f"{round(score.item(), 3)} at location {box}"
         )
+        box_images.append(img[box[1]:box[3], box[0]:box[2]])
+
+    return box_images[0]
 
 
 if __name__ == "__main__":
@@ -90,6 +94,6 @@ if __name__ == "__main__":
                 moon_output = gr.TextArea(label="Output")
 
         moon_submit.click(answer_question, [moon_img, moon_prompt], moon_output)
-        yolos_button.click(detect_objects, [yolos_input])
+        yolos_button.click(detect_objects, [yolos_input], yolos_output)
 
     app.queue().launch()