5fa1a76
1
2
Our approach enables a single multilingual speech recognition model which is competitive to strong individual models.
