correct usage example

README.md

@@ -26,6 +26,7 @@ Note: This model does not have a tokenizer as it was pretrained on audio alone.
 # Usage
 ```python
 from transformers import Wav2Vec2Model
+from transformers import Wav2Vec2FeatureExtractor
 
 feature_extractor = Wav2Vec2FeatureExtractor.from_pretrained('amsterdamNLP/Wav2Vec2-NL')
 model = Wav2Vec2Model.from_pretrained('amsterdamNLP/Wav2Vec2-NL')