This repository is a curated collection of open-source Chinese Large Language Models (LLMs), focusing on smaller, deployable, and cost-effective models. It serves researchers, developers, and power users by cataloging base models, domain-specific fine-tuned models, datasets, and tutorials, aiming to foster the development and application of Chinese LLMs.

How It Works

The project acts as a comprehensive directory, meticulously organizing over 100 resources related to Chinese LLMs. It categorizes information into models (text and multimodal), applications (domain-specific fine-tuning, LangChain, etc.), datasets (pre-training, SFT, preference), LLM training/fine-tuning frameworks, inference/deployment frameworks, evaluation benchmarks, and tutorials. The README includes a detailed table comparing key features of popular base models like ChatGLM, LLaMA, Qwen, and others.

Quick Start & Requirements

This repository is a collection of links and information, not a runnable application itself. Users are directed to individual GitHub repositories for specific models and tools, each with its own installation and usage instructions.

Highlighted Details

• Comprehensive catalog of over 100 Chinese LLM resources.
• Detailed comparison table of major Chinese LLM base models.
• Extensive categorization covering models, applications, datasets, frameworks, evaluation, and tutorials.
• Focus on models that are smaller, privately deployable, and have lower training costs.

Maintenance & Community

The project is actively maintained and welcomes community contributions via Pull Requests. Users can find links to related repositories and discussions.

Licensing & Compatibility

Licensing varies by the individual projects linked within the repository. Users must consult the specific licenses of each model, dataset, or tool for commercial use or closed-source linking compatibility.

Limitations & Caveats

As a curated list, this repository does not provide direct functionality. Users must navigate to individual project repositories to assess their specific features, requirements, and licenses. The quality and status of linked projects may vary.