Soraberta: Unsupervised Language Model Pre-training for Wolof

Soraberta is pretrained roberta-base model on wolof language . Roberta was introduced in this paper

Soraberta models

Model name Number of layers Attention Heads Embedding Dimension Total Parameters
soraberta-base 6 12 514 83 M

Using Soraberta with Hugging Face's Transformers

>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='abdouaziiz/soraberta')
>>> unmasker("juroom naari jullit man nanoo boole jend aw nag walla <mask>.")

[{'sequence': 'juroom naari jullit man nanoo boole jend aw nag walla gileem.',
  'score': 0.9783930778503418,
  'token': 4621,
  'token_str': ' gileem'},
 {'sequence': 'juroom naari jullit man nanoo boole jend aw nag walla jend.',
  'score': 0.009271537885069847,
  'token': 2155,
  'token_str': ' jend'},
 {'sequence': 'juroom naari jullit man nanoo boole jend aw nag walla aw.',
  'score': 0.0027585660573095083,
  'token': 704,
  'token_str': ' aw'},
 {'sequence': 'juroom naari jullit man nanoo boole jend aw nag walla pel.',
  'score': 0.001120452769100666,
  'token': 1171,
  'token_str': ' pel'},
 {'sequence': 'juroom naari jullit man nanoo boole jend aw nag walla juum.',
  'score': 0.0005133090307936072,
  'token': 5820,
  'token_str': ' juum'}]

Training data

The data sources are Bible OT , WOLOF-ONLINE

Contact

Please contact abdouaziz@gmail.com for any question, feedback or request.

New

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

Downloads last month
121
Hosted inference API
Fill-Mask
Mask token: <mask>
This model can be loaded on the Inference API on-demand.