rmihaylov/bert-base-pos-theseus-bg

BERT BASE (cased) finetuned on Bulgarian part-of-speech data

Pretrained model on Bulgarian language using a masked language modeling (MLM) objective. It was introduced in this paper and first released in this repository. This model is cased: it does make a difference between bulgarian and Bulgarian. The training data is Bulgarian text from OSCAR, Chitanka and Wikipedia.

It was finetuned on public part-of-speech Bulgarian data.

Then, it was compressed via progressive module replacing.

How to use

Here is how to use this model in PyTorch:

>>> from transformers import pipeline
>>> 
>>> model = pipeline(
>>>     'token-classification',
>>>     model='rmihaylov/bert-base-pos-theseus-bg',
>>>     tokenizer='rmihaylov/bert-base-pos-theseus-bg',
>>>     device=0,
>>>     revision=None)
>>> output = model('Здравей, аз се казвам Иван.')
>>> print(output)

[{'end': 7,
  'entity': 'INTJ',
  'index': 1,
  'score': 0.9640711,
  'start': 0,
  'word': '▁Здравей'},
 {'end': 8,
  'entity': 'PUNCT',
  'index': 2,
  'score': 0.9998927,
  'start': 7,
  'word': ','},
 {'end': 11,
  'entity': 'PRON',
  'index': 3,
  'score': 0.9998872,
  'start': 8,
  'word': '▁аз'},
 {'end': 14,
  'entity': 'PRON',
  'index': 4,
  'score': 0.99990034,
  'start': 11,
  'word': '▁се'},
 {'end': 21,
  'entity': 'VERB',
  'index': 5,
  'score': 0.99989736,
  'start': 14,
  'word': '▁казвам'},
 {'end': 26,
  'entity': 'PROPN',
  'index': 6,
  'score': 0.99990785,
  'start': 21,
  'word': '▁Иван'},
 {'end': 27,
  'entity': 'PUNCT',
  'index': 7,
  'score': 0.9999685,
  'start': 26,
  'word': '.'}]