akreal's picture
Specify VoxPopuli part of dataset
0c09f8c
---
language:
- en
tags:
- mbart-50
license: apache-2.0
datasets:
- SLUE-VoxPopuli
metrics:
- f1
- label-f1
---
This model is `mbart-large-50-many-to-many-mmt` model fine-tuned on the text part of [SLUE-VoxPopuli](https://github.com/asappresearch/slue-toolkit) spoken language understanding dataset.
The scores on the dev set are 83.25% and 87.76% for F1 and label-F1 respectively.