Update README.md

README.md

@@ -35,6 +35,7 @@ BLEURT is intended to be used for evaluating text produced by language models.
 This metric takes as input lists of predicted sentences and reference sentences:
 
 ```python
+>>> from evaluate import load
 >>> predictions = ["hello there", "general kenobi"]
 >>> references = ["hello there", "general kenobi"]
 >>> bleurt = load("bleurt", module_type="metric")