metadata
language:
- fr
tags:
- mbart-50
license: apache-2.0
datasets:
- MEDIA
metrics:
- cer
- cver
This model is mbart-large-50-many-to-many-mmt model fine-tuned on the text part of MEDIA spoken language understanding dataset.
The scores on the test set are 16.50% and 19.09% for CER and CVER respectively.