Post-match replays of Chinese NLP contests

What it does

This repository is a curated knowledge base that collects post-competition retrospectives—what the maintainer calls “reviews”—of top solutions from NLP contests. It catalogs entries by platform, task type, data format, model architecture, and evaluation metrics, with a pronounced focus on Chinese-language competitions and biomedical NLP challenges. It is essentially a reading list for practitioners who want to know what actually worked, not a library of reusable code.

The interesting bit

The project treats competitions as primary research artifacts. Instead of dumping notebooks, it breaks down the strategic thinking behind winning submissions—covering everything from ancient Chinese translation to liver-disease QA systems—offering a rare longitudinal view of how Chinese NLP benchmarks have evolved.

Key highlights

• Heavy medical-NLP focus: nearly thirty competitions covering entity extraction, dialogue generation, knowledge-graph construction, and clinical triage.
• Includes esoteric tasks like cuneiform machine translation (Sumerian to modern languages) and classical Chinese analysis.
• Tracks completion status openly, admitting some entries are “in progress” or merely planned.
• Links out to related meta-resources, including a separate repo for Chinese sentence-similarity tasks and a roundup of 55 large-model competitions.

Caveats

• Updates are explicitly ad hoc (“when I have time”), so coverage is uneven and some links may rot.
• Most content is in Chinese, which limits accessibility for non-Chinese readers.
• The repo is documentation-only; there is no runnable code or unified dataset.

Verdict

Great for NLP researchers and competition practitioners looking for battle-tested strategies in Chinese-language and medical domains. Skip it if you need copy-paste code or are only interested in general English-language benchmarks.