A long way passed. Here is our Bangla-Bert! It is now available in huggingface model hub.
Bangla-Bert-Base is a pretrained language model of Bengali language using mask language modeling described in BERT and it's github repository
NB: If you use this model for any nlp task please share evaluation results with us. We will add it here.
| TF Version | Pytorch Version | Vocab | |
|---|---|---|---|
| Bangla BERT Base | ----- | Huggingface Hub | Vocab |
Corpus was downloaded from two main sources:
- Bengali commoncrawl copurs downloaded from OSCAR
- Bengali Wikipedia Dump Dataset
After downloading these corpus, we preprocessed it as a Bert format. which is one sentence per line and an extra newline for new documents.
sentence 1
sentence 2
sentence 1
sentence 2
We used BNLP package for training bengali sentencepiece model with vocab size 102025. We preprocess the output vocab file as Bert format. Our final vocab file availabe at https://github.com/sagorbrur/bangla-bert and also at huggingface model hub.
- Bangla-Bert was trained with code provided in Google BERT's github repository (https://github.com/google-research/bert)
- Currently released model follows bert-base-uncased model architecture (12-layer, 768-hidden, 12-heads, 110M parameters)
- Total Training Steps: 1 Million
- The model was trained on a single Google Cloud GPU
After training 1 millions steps here is the evaluation resutls.
global_step = 1000000
loss = 2.2406516
masked_lm_accuracy = 0.60641736
masked_lm_loss = 2.201459
next_sentence_accuracy = 0.98625
next_sentence_loss = 0.040997364
perplexity = numpy.exp(2.2406516) = 9.393331287442784
Loss for final step: 2.426227
- Evaluation on Bengali Classification Benchmark Datasets
Huge Thanks to Nick Doiron for providing evalution results of classification task. He used Bengali Classification Benchmark datasets for classification task. Comparing to Nick's Bengali electra and multi-lingual BERT, Bangla BERT Base achieves state of the art result. Here is the evaluation script. Check comparison between Bangla-BERT with recent other Bengali BERT here
| Model | Sentiment Analysis | Hate Speech Task | News Topic Task | Average |
|---|---|---|---|---|
| mBERT | 68.15 | 52.32 | 72.27 | 64.25 |
| Bengali Electra | 69.19 | 44.84 | 82.33 | 65.45 |
| Bangla BERT Base | 70.37 | 71.83 | 89.19 | 77.13 |
- Evaluation on Wikiann Datasets
We evaluated Bangla-BERT-Base with Wikiann Bengali NER datasets along with another benchmark three models(mBERT, XLM-R, Indic-BERT).
Bangla-BERT-Base got a third-place where mBERT got first and XML-R got second place after training these models 5 epochs.
| Base Pre-trained Model | F1 Score | Accuracy |
|---|---|---|
| mBERT-uncased | 97.11 | 97.68 |
| XLM-R | 96.22 | 97.03 |
| Indic-BERT | 92.66 | 94.74 |
| Bangla-BERT-Base | 95.57 | 97.49 |
All four model trained with transformers-token-classification notebook. You can find all models evaluation results here
Also you can check these below paper list. They evaluated this model on their datasets.
- DeepHateExplainer: Explainable Hate Speech Detection in Under-resourced Bengali Language
- Emotion Classification in a Resource Constrained Language Using Transformer-based Approach
- A Review of Bangla Natural Language Processing Tasks and the Utility of Transformer Models
NB: If you use this model for any nlp task please share evaluation results with us. We will add it here.
Check Bangla BERT Visualize
Bangla BERT Tokenizer
from transformers import AutoTokenizer, AutoModel
bnbert_tokenizer = AutoTokenizer.from_pretrained("sagorsarker/bangla-bert-base")
text = "আমি বাংলায় গান গাই।"
bnbert_tokenizer.tokenize(text)
# ['আমি', 'বাংলা', '##য', 'গান', 'গাই', '।']MASK Generation
You can use this model directly with a pipeline for masked language modeling:
from transformers import BertForMaskedLM, BertTokenizer, pipeline
model = BertForMaskedLM.from_pretrained("sagorsarker/bangla-bert-base")
tokenizer = BertTokenizer.from_pretrained("sagorsarker/bangla-bert-base")
nlp = pipeline('fill-mask', model=model, tokenizer=tokenizer)
for pred in nlp(f"আমি বাংলায় {nlp.tokenizer.mask_token} গাই।"):
print(pred)
# {'sequence': '[CLS] আমি বাংলায গান গাই । [SEP]', 'score': 0.13404667377471924, 'token': 2552, 'token_str': 'গান'}- Thanks to Google TensorFlow Research Cloud (TFRC) for providing the free GPU credits - thank you!
- Thank to all the people around, who always helping us to build something for Bengali.
If you find this model helpful, please cite this.
@misc{Sagor_2020,
title = {BanglaBERT: Bengali Mask Language Model for Bengali Language Understanding},
author = {Sagor Sarker},
year = {2020},
url = {https://github.com/sagorbrur/bangla-bert}
}