Distilled layers: [r, r, -1] (0.75)

Trained with normal text replication (+ bart noise [Multilingual BERT])

Finetuned on simplification task

Trained with normal text replication (+ bart noise [ + CTRL Tokens])

Trained with normal text replication (+ bart noise [Multilingual BERT])

Trained with normal text replication

Update config.json

Initial commit