About Panda

Welcome to our overseas Chinese large language model open source project – Panda! Launched in May 2023, the Panda project aims to help the entire community explore the entire technology stack of large models in the era of large models. Recently, we have upgraded the Panda project. Currently the Panda project is divided into three sub-projects: 1. PandaLLM 2. PandaLLMOps 3. PandaCommunity. The following is the specific introduction of each sub-project:

  1. PandaLLM open source large model. The Panda series of open source large models are currently conducting continuous pre-training in the Chinese field based on LLaMA1 and LLaMA2. We hope to provide a general-purpose basic tool for the field of Chinese natural language processing for exploration. The PandaLLM model and the Chinese dataset involved in training will be released as open source, and anyone can use it and participate in the development for free.

  2. PandaLLMOps Open source large model training, inference and deployment tools. PandaLLMOps is a set of tools that integrates large models from model training, inference, and deployment. We hope to provide a set of useful tools for large-scale model developers and enthusiasts, reduce the difficulty of learning, and improve everyone's efficiency in all aspects of large-scale model development, reasoning, and deployment. The following scenarios are currently supported: 1. Pre-training from scratch 2. Incremental pre-training based on the existing base, fine-tuning of all parameters, Lora, QLora 3. Rapid deployment, integrated with vllm, lightllm, etc., to maximize and optimize the inference speed

  3. PandaCommunity Large model Chinese community. PandaCommunity aims to build a learning community for large-scale model developers in the Chinese world, so that everyone can have an in-depth understanding of the large-scale model technology stack and communicate with each other, exert collective wisdom, and overcome technical difficulties on the road to large-scale model exploration. At the same time, for students who want to learn large-scale model technology, we will also regularly launch free technical tutorials, technical seminars, interpretation of papers, etc.

    We welcome developers from all over the world to participate in this project and jointly promote the development of natural language processing technology.

Visit Official Website


Community Posts
no data
Nothing to display