# CLUEbenchmark/CLUECorpus2020

Large-scale Pre-training Corpus for Chinese 100G 中文预训练语料

Repository: https://github.com/CLUEbenchmark/CLUECorpus2020
Canonical: https://ross.abutalabs.com/products/cluecorpus2020
Homepage: https://arxiv.org/abs/2003.01355
License: MIT
License Family: permissive
Topics: chinese, chinese-corpus, datasets, pretrain, corpus, nlp, bert, roberta, albert
Last push: 2026-02-06T11:43:34+00:00

## Health v2 (maintenance only)
Score: 62/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 66, release rhythm 35, longevity 100
- inputs: {"age_days": 2412, "days_push": 208, "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 1018, forks 83 (observed 2026-08-28T04:03:14.830178+00:00)

## What it is
CLUECorpus2020 is a 100GB cleaned Chinese text corpus (35 billion characters) derived from Common Crawl for pre-training language models, plus a smaller 14GB CLUECorpusSmall and a compact 8K-token Chinese vocabulary. It is distributed by the CLUE benchmark organization with an accompanying technical report and pre-trained models.

## Use cases
- pretrain a Chinese language model like BERT or RoBERTa
- find a large Chinese text corpus for language modeling
- train a Chinese text generation model
- reduce tokenizer vocabulary size for Chinese NLP
- get cleaned Chinese web text for self-supervised learning
- fine-tune Chinese NLP models on a quality corpus

## When to choose
- you need large-scale Chinese text for pretraining or language generation
- you want a smaller Chinese vocabulary to cut compute and memory costs
- you need pre-formatted training data with one sentence per line

## When to avoid
- you need multilingual corpora beyond Chinese
- you cannot or will not apply via email and agree to the no-redistribution terms
- you need an actively updated corpus rather than a fixed 2020 snapshot

## Facets
- artifact type: dataset
- maturity: stable
- function: machine-learning, nlp, data-generation
- domain: large-language-models, machine-learning
- platform: python, cross-platform
- tags: chinese-corpus, pretraining, language-modeling, common-crawl, vocabulary, bert, natural-language-processing

## Member repositories
- CLUEbenchmark/CLUECorpus2020 (main) score 62

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:14.830178+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 2026-08-30T07:09:25.492882+00:00, confidence not recorded.
  - readme: https://github.com/CLUEbenchmark/CLUECorpus2020 (fetched 2026-08-28T04:03:14.830178+00:00, sha 852fa296ea2b)
  - homepage: https://arxiv.org/abs/2003.01355 (fetched 2026-08-29T13:09:58.579799+00:00, sha fc36d967248d)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T13:09:58.627186+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T13:09:58.645583+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T13:09:58.647658+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T13:09:58.643687+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.