# ymcui/Chinese-BERT-wwm

Pre-Training with Whole Word Masking for Chinese BERT（中文BERT-wwm系列模型）

Repository: https://github.com/ymcui/Chinese-BERT-wwm
Canonical: https://ross.abutalabs.com/products/chinese-bert-wwm
Homepage: https://ieeexplore.ieee.org/document/9599397
Language: Python
License: Apache-2.0
License Family: permissive
Topics: chinese-bert, tensorflow, pytorch, bert, nlp, roberta, bert-wwm, bert-wwm-ext, roberta-wwm, rbt
Last push: 2026-04-19T00:57:10+00:00

## Health v2 (maintenance only)
Score: 67/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 78, release rhythm 35, longevity 100
- inputs: {"age_days": 2632, "days_push": 137, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 10223, forks 1381 (observed 2026-08-28T04:10:39.949191+00:00)

## What it is
A collection of Chinese pre-trained language models (BERT-wwm, BERT-wwm-ext, RoBERTa-wwm-ext, RBT3, etc.) built with Whole Word Masking, released by HFL. Models are loadable via Hugging Face Transformers and TensorFlow 2.

## Use cases
- fine-tune a Chinese BERT model for text classification
- get Chinese sentence embeddings for semantic search
- named entity recognition on Chinese text
- compare Chinese pre-trained models on CLUE benchmarks
- download a small Chinese transformer model for production NLP

## When to choose
- you need well-established Chinese encoder models with published benchmark results
- you want whole word masking variants of BERT/RoBERTa for Chinese tasks
- you need small models like RBT3 for low-latency Chinese NLP

## When to avoid
- you need a generative large language model rather than an encoder
- you work with non-Chinese languages primarily
- you need the latest pre-training techniques like LERT or MacBERT (see sibling projects)

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, nlp, llm-training
- domain: machine-learning, large-language-models
- platform: python, cross-platform
- tags: bert, whole-word-masking, chinese-nlp, pretrained-language-models, roberta, huggingface-transformers, natural-language-processing, tensorflow

## Member repositories
- ymcui/Chinese-BERT-wwm (main) score 67

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:39.949191+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:19:30.034704+00:00, confidence not recorded.
  - readme: https://github.com/ymcui/Chinese-BERT-wwm (fetched 2026-08-28T04:10:39.949191+00:00, sha a1b7dc44b36a)
  - homepage: https://ieeexplore.ieee.org/document/9599397 (fetched 2026-08-29T08:18:56.552001+00:00, sha 44136fa355b3)
- Data as of 2026-08-30T08:39:29.467469+00:00.