Downloads · 30 days
1.4K
2% of all-time downloads
AraBART
moussaKam/AraBART
AraBART is a fill-mask model from moussaKam. Use it when you need the model to fill a missing word. It is set up for transformers. The card lists the license as apache-2.0.
AraBART is the first Arabic model in which the encoder and the decoder are pretrained end-to-end, based on BART. AraBART follows the architecture of BART-Base which has 6 encoder and 6 decoder layers and 768 hidden di…
All-time downloads
75.3K
Public
Repo size
1.1 GB
Likes
18
Public
Repo makeup
Click a slice to open those files.
.bin557 MB · 100%
At a glance
- Task
- Fill-Mask
- Library
- transformers
- License
- apache-2.0
- Model type
- mbart
- Access
- Public
- Created
- Mar 9, 2022
- Updated
- May 5, 2022
- SHA
- cb67d617
Try a prompt
Model card
From the Hugging Face model README
AraBART is the first Arabic model in which the encoder and the decoder are pretrained end-to-end, based on BART. AraBART follows the architecture of BART-Base which has 6 encoder and 6 decoder layers and 768 hidden dimensions. In total AraBART has 139M parameters.
AraBART achieves the best performance on multiple abstractive summarization datasets, outperforming strong baselines including a pretrained Arabic BERT-based models and multilingual mBART and mT5 models.
- Task
- Fill-Mask
- Library
- transformers
- Type
- mbart
- License
- apache-2.0
- Languages
- ar
- Hf Inference
- live
- Created
- Mar 9, 2022
- Updated
- May 5, 2022