Awesome Chinese LLM
Introduction
Awesome Chinese LLM is a comprehensive collection of resources related to Large Language Models (LLMs) in the Chinese language. As LLMs, exemplified by ChatGPT, have gained prominence due to their near-general artificial intelligence capabilities, this project aims to gather and systematically organize open-source Chinese LLMs, applications, datasets, and tutorials. It currently houses over 100 resources, catering to both industry practitioners and the wider community interested in Chinese language LLMs.
Objectives
The primary goal of the Awesome Chinese LLM project is to serve as a repository of knowledge and tools related to Chinese LLMs. The project encourages contributions from individuals who can add new models, applications, datasets, or other relevant resources via pull requests. Contributors should adhere to the project's format, including providing repository links, stars, and concise descriptions.
Supported Models
The project includes various foundational models, each with its own set of characteristics, such as model size, training tokens, and commercial availability. Some notable models include:
- ChatGLM: Known for its efficient performance in Chinese question-answering tasks, trained with approximately 1 trillion tokens.
- LLaMA: Offers multiple model configurations, supporting commercial use in certain cases.
- Baichuan: Developed by Baichuan Intelligence, supports bilingual capabilities in Chinese and English with models like Baichuan-7B and Baichuan-13B, optimized for commercial use.
- Qwen: From Alibaba Cloud, offers robust support for plugins and agent tasks along with multi-language capabilities.
- InternLM: Features a range of models, with key contributors including Hong Kong Chinese University and Shanghai Jiao Tong University, focusing on broader language capabilities.
- DeepSeek: An efficient expert-mixed language model designed for varied applications.
- XVERSE: Supports a broad context length, enhanced for multi-language tasks, offering models such as XVERSE-7B, XVERSE-13B, and XVERSE-65B.
Key Areas of Application
- Domain-Specific Fine-Tuning: Applications tailored to sectors like healthcare, legal, financial, education, technology, e-commerce, cybersecurity, and agriculture.
- LangChain Applications: Implements chain-of-thought reasoning in applications.
- Other Applications: Broad range of use cases.
Datasets and Frameworks
- Datasets: Includes pre-training datasets and specific fine-tuning datasets.
- Training and Deployment: Provides frameworks for LLM training and deployment.
- Evaluation and Tutorials: Offers resources for evaluating LLM performance and comprehensive tutorials for building and deploying LLM-based applications.
Contribution & Community
The project invites community contributions to further enrich the repository with diverse and innovative LLM resources pertinent to the Chinese language domain. This crowd-sourced approach ensures the collection stays updated and comprehensive, thereby supporting the ongoing evolution of LLM technologies.
Conclusion
Awesome Chinese LLM stands as a vital resource for anyone interested in the development and application of Chinese language models. It provides an organized repository of LLMs, applications, datasets, and educational materials, fostering both academic and practical advancements in the field. As the LLM landscape continues to evolve, this project represents a foundation for future innovations and applications.