--- layout: single title: "Large Language Models（大语言模型）" date: 2023-04-22 08:00:00 +0800 categories: [AI 与大模型, 编程开发] tags: [ChatGPT, ChatGLM, LLaMA, Alpaca, Vicuna, LLM] --- ## LLMS * [大语言模型综述 - A Survey of Large Language Models](https://arxiv.org/abs/2303.18223) ### ChatGPT * [document.ai](https://github.com/GanymedeNil/document.ai) ### LLaMA * [LLaMA](https://github.com/facebookresearch/llama) * [LLaMA: Open and Efficient Foundation Language Models](https://arxiv.org/abs/2302.13971) * [Llama Hub](https://llamahub.ai) * [llama.cpp](https://github.com/ggerganov/llama.cpp) ### ChatLLaMA * [ChatLLaMA](https://github.com/nebuly-ai/nebullvm/tree/main/apps/accelerate/chatllama) * [nebullvm](https://github.com/nebuly-ai/nebullvm) ### ChatGLM * [GLM](https://github.com/THUDM/GLM) * [ChatGLM-6B](https://github.com/THUDM/ChatGLM-6B) * [Hugging Face chatglm-6b](https://huggingface.co/THUDM/chatglm-6b) * [ChatGLM-webui](https://github.com/Akegarasu/ChatGLM-webui) * [langchain-ChatGLM](https://github.com/imClumsyPanda/langchain-ChatGLM) * [LangChain-ChatGLM-Webui](https://github.com/thomas-yanxin/LangChain-ChatGLM-Webui) * [闻达：一个大规模语言模型调用平台](https://github.com/l15y/wenda) * [InstructGLM:基于ChatGLM-6B在指令数据集上进行微调](https://github.com/yanqiangmiffy/InstructGLM) * [ChatGLM-6B 结合 langchain 实现本地知识库 QA Bot](https://www.heywhale.com/mw/project/643977aa446c45f4592a1e59) * [Hugging Face chatglm-6b](https://huggingface.co/THUDM/chatglm-6b/tree/main) ### Alpaca * [Stanford Alpaca: An Instruction-following LLaMA Mode](https://github.com/tatsu-lab/stanford_alpaca) * [alpaca_data.json](https://github.com/tatsu-lab/stanford_alpaca/blob/main/alpaca_data.json) contains 52K instruction-following data we used for fine-tuning the Alpaca model. This JSON file is a list of dictionaries, each dictionary contains the following fields: - instruction: str, describes the task the model should perform. Each of the 52K instructions is unique. - input: str, optional context or input for the task. For example, when the instruction is "Summarize the following article", the input is the article. Around 40% of the examples have an input. - output: str, the answer to the instruction as generated by text-davinci-003. * [Alpaca.cpp](https://github.com/antimatter15/alpaca.cpp) ### Chinese-Vicuna * [Chinese-Vicuna: A Chinese Instruction-following LLaMA-based Model](https://github.com/Facico/Chinese-Vicuna) ### Guanaco * [Guanaco: A Multilingual Instruction-Following Language Model Based on LLaMA 7B](https://github.com/Guanaco-Model/Guanaco-Model.github.io) ### Colossal-AI * [ColossalChat](https://chat.colossalai.org) * [Colossal-AI](https://github.com/hpcaitech/ColossalAI/blob/main/docs/README-zh-Hans.md) * [ColossalChat](https://github.com/hpcaitech/ColossalAI/tree/main/applications/Chat) * [ColossalChat: An Open-Source Solution for Cloning ChatGPT With a Complete RLHF Pipeline](https://medium.com/@yangyou_berkeley/colossalchat-an-open-source-solution-for-cloning-chatgpt-with-a-complete-rlhf-pipeline-5edf08fb538b) ### Open-Assistant * [Open Assistant](https://open-assistant.io/zh/) * [Open-Assistant](https://github.com/LAION-AI/Open-Assistant) ### ChatRWKV * [ChatRWKV](https://github.com/BlinkDL/ChatRWKV) * [Hugging Face ChatRWKV Demo](https://huggingface.co/spaces/BlinkDL/ChatRWKV-gradio) * [Model Scope ChatRWKV Demo](https://modelscope.cn/studios/BlinkDL/RWKV-CHN-2/summary) * [发布几个RWKV的Chat模型（包括英文和中文）7B/14B](https://zhuanlan.zhihu.com/p/618011122) ### BELLE * [BELLE: Be Everyone's Large Language model Engine](https://github.com/LianjiaTech/BELLE) ### Dolly * [Dolly](https://github.com/databrickslabs/dolly) * [dolly-v2-12b](https://huggingface.co/databricks/dolly-v2-12b) * [Free Dolly: Introducing the World's First Truly Open Instruction-Tuned LLM](https://www.databricks.com/blog/2023/04/12/dolly-first-open-commercially-viable-instruction-tuned-llm) ### Koala * [Koala: A Dialogue Model for Academic Research](https://bair.berkeley.edu/blog/2023/04/03/koala/) * [Koala Demo](https://chat.lmsys.org/?model=koala-13b) ### Baize * [Baize](https://github.com/project-baize/baize-chatbot) ### LMFlow * [LMFlow](https://github.com/OptimalScale/LMFlow) ### FreedomGPT * [FreedomGPT](https://github.com/ohmplatform/FreedomGPT) ### MiniGPT-4 * [MiniGPT-4](https://minigpt-4.github.io) * [Vision-CAIR/minigpt4](https://huggingface.co/spaces/Vision-CAIR/minigpt4) * [Vision-CAIR/MiniGPT-4](https://github.com/Vision-CAIR/MiniGPT-4) ### AutoGPT * [Auto-GPT: An Autonomous GPT-4 Experiment](https://github.com/Torantulino/Auto-GPT.git) * [AutoGPT](https://autogpt.net) * [AUTOGPT INSTALLATION AND FEATURES](https://autogpt.net/autogpt-installation-and-features/) ### JARVIS * [JARVIS](https://github.com/microsoft/JARVIS) ### FastChat * [Chat with Open Large Language Models](https://chat.lmsys.org) An open platform for training, serving, and evaluating large language model based chatbots. ### 通用工具 * [Text generation web UI](https://github.com/oobabooga/text-generation-webui) * [EasyLM](https://github.com/young-geng/EasyLM) * [Auto-evaluator](https://github.com/PineappleExpress808/auto-evaluator) ## 文生图 ### Midjourney * [Use Midjourney to generate and publish coloring books](https://aituts.com/ai-generated-coloring-books/) ## 数据集 * [nlp_chinese_corpus](https://github.com/brightmart/nlp_chinese_corpus) * [ChineseNlpCorpus](https://github.com/SophonPlus/ChineseNlpCorpus/) ## 参考资料 * [全球首个完全开源的指令跟随大模型；T5 到 GPT-4 最全盘点](https://xie.infoq.cn/article/df717392db619c37079f17cba) * [北大综述论文：In-context Learning进展、挑战和方向](https://hub.baai.ac.cn/view/23037) * [大型语言模型综述全新出炉：从T5到GPT-4最全盘点，国内20余位研究者联合撰写](https://mp.weixin.qq.com/s?__biz=MzA3MzI4MjgzMw==&mid=2650872546&idx=1&sn=068d372498ca57cadc6b43a6d6c162ae&scene=21#wechat_redirect) * [開放中文轉換](https://github.com/yichen0831/opencc-python) * [ChatGPT 火了，一大批开源平替也来了](https://my.oschina.net/oscpyaqxylk/blog/8671170) * [国产类ChatGPT密集开测，单卡能跑大模型，生成式AI热潮爆发](https://36kr.com/p/2172029387371011) * [如何（不靠运气）变得富有 — — 来自AngelList创办人的经验](https://xuwenhao.com/如果-不靠运气-变得富有-来自angellist创办人的经验-aa9f64f5883e) * [LLaMa 惊天泄露引爆 ChatGPT 平替狂潮，开源 LLM 领域变天](https://www.ithome.com/0/685/656.htm) * [寻找那些ChatGPT/GPT4开源“平替”们](https://github.com/chenking2020/FindTheChatGPTer) * [开源ChatGPT替代模型项目整理](https://zhuanlan.zhihu.com/p/618790279)