herbert-large-cased / README.md
piotr-rybak's picture
Update README.md
8c3e1ae
metadata
language: pl
tags:
  - herbert
license: cc-by-4.0

HerBERT

HerBERT is a BERT-based Language Model trained on Polish corpora using Masked Language Modelling (MLM) and Sentence Structural Objective (SSO) with dynamic masking of whole words. For more details, please refer to: HerBERT: Efficiently Pretrained Transformer-based Language Model for Polish.

Model training and experiments were conducted with transformers in version 2.9.

Corpus

HerBERT was trained on six different corpora available for Polish language:

Corpus Tokens Documents
CCNet Middle 3243M 7.9M
CCNet Head 2641M 7.0M
National Corpus of Polish 1357M 3.9M
Open Subtitles 1056M 1.1M
Wikipedia 260M 1.4M
Wolne Lektury 41M 5.5k

Tokenizer

The training dataset was tokenized into subwords using a character level byte-pair encoding (CharBPETokenizer) with a vocabulary size of 50k tokens. The tokenizer itself was trained with a tokenizers library.

We kindly encourage you to use the Fast version of the tokenizer, namely HerbertTokenizerFast.

Usage

Example code:

from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("allegro/herbert-large-cased")
model = AutoModel.from_pretrained("allegro/herbert-large-cased")

output = model(
    **tokenizer.batch_encode_plus(
        [
            (
                "A potem szedł środkiem drogi w kurzawie, bo zamiatał nogami, ślepy dziad prowadzony przez tłustego kundla na sznurku.",
                "A potem leciał od lasu chłopak z butelką, ale ten ujrzawszy księdza przy drodze okrążył go z dala i biegł na przełaj pól do karczmy."
            )
        ],
    padding='longest',
    add_special_tokens=True,
    return_tensors='pt'
    )
)

License

CC BY 4.0

Citation

If you use this model, please cite the following paper:

@inproceedings{mroczkowski-etal-2021-herbert,
    title = "{H}er{BERT}: Efficiently Pretrained Transformer-based Language Model for {P}olish",
    author = "Mroczkowski, Robert  and
      Rybak, Piotr  and
      Wr{\'o}blewska, Alina  and
      Gawlik, Ireneusz",
    booktitle = "Proceedings of the 8th Workshop on Balto-Slavic Natural Language Processing",
    month = apr,
    year = "2021",
    address = "Kiyv, Ukraine",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/2021.bsnlp-1.1",
    pages = "1--10",
}

Authors

The model was trained by Machine Learning Research Team at Allegro and Linguistic Engineering Group at Institute of Computer Science, Polish Academy of Sciences.

You can contact us at: klejbenchmark@allegro.pl