mBART (a pre-trained model by Facebook) is pre-trained to de-noise multiple languages simultaneously with BART objective.

Checkpoint available in this repository is obtained after fine-tuning facebook/mbart-large-cc25 on all samples (~260K) from Bhasha (pib_v1.3) Hindi-English parallel corpus. This checkpoint gives decent results for Hindi-english translation.


Select AutoNLP in the “Train” menu to fine-tune this model automatically.

Downloads last month
Hosted inference API
Text2Text Generation
This model can be loaded on the Inference API on-demand.