Polish version of BERT language model is here! It is now available in two variants: cased and uncased, both can be downloaded and used via HuggingFace transformers library. I recommend using the cased model, more info on the differences and benchmark results below.
PolBERT image
Cased and uncased variants
I initially trained the uncased model, the corpus and training details are referenced below. Here are some issues I found after I published the uncased model:
Some Polish characters and accents are not tokenized correctly through the BERT tokenizer when applying lowercase. This doesn't impact sequence classification much, but may influence token classfication tasks significantly.
I noticed a lot of duplicates in the Open Subtitles dataset, which dominates the training corpus.
I didn't use Whole Word Masking.
The cased model improves on the uncased model in the following ways:
All Polish characters and accents should now be tokenized correctly.
I removed duplicates from Open Subtitles dataset. The corpus is smaller, but more balanced now.
The model is trained with Whole Word Masking.
Pre-training corpora
Below is the list of corpora used along with the output of wc command (counting lines, words and characters). These corpora were divided into sentences with srxsegmenter (see references), concatenated and tokenized with HuggingFace BERT Tokenizer.
Polbert is released via HuggingFace Transformers library.
For an example use as language model, see this notebook file.
Uncased
python
1from transformers import*2model = BertForMaskedLM.from_pretrained("dkleczek/bert-base-polish-uncased-v1")3tokenizer = BertTokenizer.from_pretrained("dkleczek/bert-base-polish-uncased-v1")4nlp = pipeline('fill-mask', model=model, tokenizer=tokenizer)5for pred in nlp(f"Adam Mickiewicz wielkim polskim {nlp.tokenizer.mask_token} był."):6print(pred)7# Output:8# {'sequence': '[CLS] adam mickiewicz wielkim polskim poeta był. [SEP]', 'score': 0.47196975350379944, 'token': 26596}9# {'sequence': '[CLS] adam mickiewicz wielkim polskim bohaterem był. [SEP]', 'score': 0.09127858281135559, 'token': 10953}10# {'sequence': '[CLS] adam mickiewicz wielkim polskim człowiekiem był. [SEP]', 'score': 0.0647173821926117, 'token': 5182}11# {'sequence': '[CLS] adam mickiewicz wielkim polskim pisarzem był. [SEP]', 'score': 0.05232388526201248, 'token': 24293}12# {'sequence': '[CLS] adam mickiewicz wielkim polskim politykiem był. [SEP]', 'score': 0.04554257541894913, 'token': 44095}
Cased
python
1model = BertForMaskedLM.from_pretrained("dkleczek/bert-base-polish-cased-v1")2tokenizer = BertTokenizer.from_pretrained("dkleczek/bert-base-polish-cased-v1")3nlp = pipeline('fill-mask', model=model, tokenizer=tokenizer)4for pred in nlp(f"Adam Mickiewicz wielkim polskim {nlp.tokenizer.mask_token} był."):5print(pred)6# Output:7# {'sequence': '[CLS] Adam Mickiewicz wielkim polskim pisarzem był. [SEP]', 'score': 0.5391148328781128, 'token': 37120}8# {'sequence': '[CLS] Adam Mickiewicz wielkim polskim człowiekiem był. [SEP]', 'score': 0.11683262139558792, 'token': 6810}9# {'sequence': '[CLS] Adam Mickiewicz wielkim polskim bohaterem był. [SEP]', 'score': 0.06021466106176376, 'token': 17709}10# {'sequence': '[CLS] Adam Mickiewicz wielkim polskim mistrzem był. [SEP]', 'score': 0.051870670169591904, 'token': 14652}11# {'sequence': '[CLS] Adam Mickiewicz wielkim polskim artystą był. [SEP]', 'score': 0.031787533313035965, 'token': 35680}
See the next section for an example usage of Polbert in downstream tasks.
Evaluation
Thanks to Allegro, we now have the KLEJ benchmark, a set of nine evaluation tasks for the Polish language understanding. The following results are achieved by running standard set of evaluation scripts (no tricks!) utilizing both cased and uncased variants of Polbert.
Model
Average
NKJP-NER
CDSC-E
CDSC-R
CBD
PolEmo2.0-IN
PolEmo2.0-OUT
DYK
PSC
AR
Polbert cased
81.7
93.6
93.4
93.8
52.7
87.4
71.1
59.1
98.6
85.2
Polbert uncased
81.4
90.1
93.9
93.5
55.0
88.1
68.8
59.4
98.8
85.4
Note how the uncased model performs better than cased on some tasks? My guess this is because of the oversampling of Open Subtitles dataset and its similarity to data in some of these tasks. All these benchmark tasks are sequence classification, so the relative strength of the cased model is not so visible here.
Bias
The data used to train the model is biased. It may reflect stereotypes related to gender, ethnicity etc. Please be careful when using the model for downstream task to consider these biases and mitigate them.
Also appreciate the help from Timo Möller from deepset for sharing tips and scripts based on their experience training German BERT model.
Big thanks to Allegro for releasing KLEJ Benchmark and specifically to Piotr Rybak for help with the evaluation and pointing out some issues with the tokenization.
Finally, thanks to Rachel Thomas, Jeremy Howard and Sylvain Gugger from fastai for their NLP and Deep Learning courses!
