This repository is a comprehensive, curated list of resources for Japanese Natural Language Processing (NLP), targeting researchers, developers, and power users. It provides categorized links to GitHub repositories, Hugging Face models and datasets, and tools covering a wide spectrum of NLP tasks, from morphological analysis and parsing to machine translation, OCR, and LLM evaluation.

How It Works

The project acts as a central index, meticulously gathering and organizing links to open-source projects and datasets relevant to Japanese NLP. It categorizes these resources by task (e.g., morphology, parsing, machine translation) and programming language (Python, C++, Rust, JavaScript, Go, Java), enabling users to quickly discover relevant tools and data. The inclusion of Hugging Face repositories further bridges the gap between research and practical application.

Quick Start & Requirements

This is a curated list, not a software package. No installation or execution is required. Users navigate the README to find links to external resources.

Highlighted Details

• Lists information on 697 GitHub repositories and 1837 Hugging Face repositories (models and datasets).
• Includes a dedicated search tool for navigating the extensive repository information.
• Categorizes resources by NLP task (morphology, parsing, conversion, preprocessor, sentence splitter, sentiment analysis, machine translation, NER, OCR) and programming language.
• Features sections on LLMs, dictionaries, and corpora, with specific sub-categories for pre-trained models (Word2Vec, Transformer-based, ChatGPT).

Maintenance & Community

The repository is maintained by taishi-i. Notable contributors are listed, with links to their websites or social media.

Licensing & Compatibility

The repository itself is a list and does not have a specific license. Individual linked resources will have their own licenses, which users must consult.

Limitations & Caveats

As a curated list, the quality and maintenance status of linked resources vary. Users are responsible for vetting the individual projects and datasets. The list is extensive but may not be exhaustive.