How sagorbrur/bangla-bert works
This repository distributes Bangla BERT Base, a pretrained Bengali masked-language model and vocabulary, with documented tokenizer and fill-mask usage plus downstream evaluation results and visualization notebooks. The principal user workflow is to load the model and tokenizer, provide Bengali text with a mask, and receive candidate replacements. The tree exposes no implementation excerpts, so relationships below rely on README statements; the bert entry is represented only at its listed path.
Main entry points: bert Bangla BERT · README.md Fill-mask inference · bangla-bert-evaluation-classification-task.ipynb Classification evaluation · wikiann WikiANN evaluation
Components 9
Model assets
- Bangla BERT
bert - BERT vocabulary
vocab.txt
Model usage
- BERT tokenizer
README.md - Fill-mask inference
README.md
Evaluations
- Classification evaluation
notebook/bangla-bert-evaluation-classification-task.ipynb - WikiANN evaluation
evaluations/wikiann - Attention visualization
notebook
Other components
- NLP user
- Hugging Face Hub
Connections 8
- NLP user → BERT tokenizer: tokenizes text
README.md - NLP user → Fill-mask inference: submits masked text
README.md - Fill-mask inference → Bangla BERT: predicts masks
README.md - Fill-mask inference → Hugging Face Hub: loads pretrained model
README.md - BERT tokenizer → Hugging Face Hub: loads tokenizer
README.md - Classification evaluation → Bangla BERT: evaluates
README.md - WikiANN evaluation → Bangla BERT: evaluates
README.md - Attention visualization → Bangla BERT: visualizes attention
README.md