lonePatient/BERT-NER-Pytorch

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2235 stars · 433 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A PyTorch library for Chinese Named Entity Recognition using BERT-based models with Softmax, CRF, and Span decoding heads. It also supports ALBERT variants and training tricks like focal loss, label smoothing, and adversarial training.

Use cases

• train a Chinese NER model with BERT and CRF
• run named entity recognition on Chinese text
• compare BERT softmax vs CRF vs span NER architectures
• fine-tune ALBERT for Chinese sequence labeling
• apply focal loss or label smoothing to NER training
• reproduce CLUENER benchmark results

When to choose

• you need Chinese NER with pretrained BERT/ALBERT models in PyTorch
• you want to experiment with different decoding heads and loss functions for sequence labeling
• you need benchmark baselines on CLUENER or CNER datasets

When to avoid

• you need NER for languages other than Chinese
• you want a maintained production library with recent PyTorch support (it targets PyTorch 1.1-1.5)
• you need multilingual or modern transformer NER pipelines

Facets

library · maturity maintenance

nlp machine-learning deep-learning machine-learning deep-learning python named-entity-recognition bert chinese-nlp crf pytorch albert sequence-labeling natural-language-processing linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="lonePatient/BERT-NER-Pytorch")