small typo in readme

README.md

@@ -37,7 +37,7 @@ inputs = processor(images=image, return_tensors="pt")
 outputs = model(**inputs)
 
 # convert outputs (bounding boxes and class logits) to COCO API
-# let's only keep detections with score > 0.9
+# let's only keep detections with score > 0.8
 target_sizes = torch.tensor([image.size[::-1]])
 results = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.8)[0]