{"id":15253161,"url":"https://github.com/OpenBuddy/OpenBuddy","last_synced_at":"2025-10-05T18:31:13.668Z","repository":{"id":156121129,"uuid":"632837174","full_name":"OpenBuddy/OpenBuddy","owner":"OpenBuddy","description":"Open Multilingual Chatbot for Everyone","archived":false,"fork":false,"pushed_at":"2024-05-04T03:07:37.000Z","size":2117,"stargazers_count":1239,"open_issues_count":15,"forks_count":70,"subscribers_count":23,"default_branch":"main","last_synced_at":"2024-12-07T09:43:40.104Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"https://openbuddy.ai","language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/OpenBuddy.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":"CODE_OF_CONDUCT.md","threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2023-04-26T08:29:49.000Z","updated_at":"2024-12-05T04:21:20.000Z","dependencies_parsed_at":null,"dependency_job_id":"a7d5826b-43d4-464f-88ba-4904125df925","html_url":"https://github.com/OpenBuddy/OpenBuddy","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OpenBuddy%2FOpenBuddy","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OpenBuddy%2FOpenBuddy/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OpenBuddy%2FOpenBuddy/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/OpenBuddy%2FOpenBuddy/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/OpenBuddy","download_url":"https://codeload.github.com/OpenBuddy/OpenBuddy/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":235432179,"owners_count":18989472,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-09-29T21:00:53.479Z","updated_at":"2025-10-05T18:31:13.329Z","avatar_url":"https://github.com/OpenBuddy.png","language":null,"funding_links":[],"categories":["A01_文本生成_文本对话","Others","App"],"sub_categories":["大语言对话模型及数据"],"readme":"# OpenBuddy - Open Multilingual Chatbot for Everyone\r\n\r\n\u003cdiv align=\"center\"\u003e\r\n  \u003cimg src=\"media/logo.png\" width=\"300px\"\u003e\r\n\u003c/div\u003e\r\n\r\n[中文](README.zh.md) | [English](README.md)\r\n\r\n微信公众号、微信群：\r\n\r\n\u003cimg src=\"media/mp.jpg\" width=\"285\"\u003e\r\n\r\n搜索“**开源智友**”，关注公众号并发送“加群”即可获得微信群邀请、参与新模型内测等活动\r\n\r\n欢迎关注我们的 ModelScope 社区，体验精选模型的高速下载和一键部署：\r\n\r\nhttps://modelscope.cn/organization/OpenBuddy\r\n\r\nWebsite: [https://openbuddy.ai](https://openbuddy.ai)\r\n\r\nGitHub: [https://github.com/OpenBuddy/OpenBuddy](https://github.com/OpenBuddy/OpenBuddy)\r\n\r\nHuggingface: https://huggingface.co/OpenBuddy\r\n\r\n![Demo](media/demo.png)\r\n\r\nOpenBuddy is a powerful open multilingual chatbot model aimed at global users, emphasizing conversational AI and seamless multilingual support for English, Chinese, and other languages.\r\n\r\nBuilt upon Tii's Falcon model and Facebook's LLaMA model, OpenBuddy is fine-tuned to include an extended vocabulary, additional common characters, and enhanced token embeddings. By leveraging these improvements and multi-turn dialogue datasets, OpenBuddy offers a robust model capable of answering questions and performing translation tasks across various languages.\r\n\r\nOur mission with OpenBuddy is to provide a free, open, and offline-capable AI model that operates on users' devices, irrespective of their language or cultural background. We strive to empower individuals worldwide to access and benefit from AI technology.\r\n\r\n## Online Demo\r\n\r\nCurrently, the OpenBuddy demo is available on our Discord server. Please join our Discord server to try it out!\r\n\r\nDiscord: [![Discord](https://img.shields.io/discord/1100710961549168640?color=blueviolet\u0026label=Discord)](https://discord.gg/6fU2s9cGjA)\r\n\r\n## Key Features\r\n\r\n- **Multilingual** conversational AI, Chinese, English, Japanese, Korean, French, Germany and more!\r\n- Enhanced vocabulary and support for common CJK characters\r\n- Fine-tuned with multi-turn dialogue datasets for improved performance\r\n- Various model sizes to suit your needs: 3B, 7B, 13B, 30B, 40B, 65B, 70B and more!\r\n- 3/4/5-bit quantization for CPU deployment via llama.cpp (with slightly reduced output quality)\r\n- Active development plans for future features and improvements\r\n\r\n## Future Plans\r\n\r\n- Enhancing multilingual performance\r\n- Optimizing model quality post-quantization\r\n- Developing a mechanism to assess content quality, safety, and inference capabilities\r\n- Investigating Reinforcement Learning with Human Feedback (RLHF)\r\n- Exploring the addition of multimodal capabilities for dialogues with image context\r\n\r\n## Model Download\r\n\r\nOpenBuddy currently offers model downloads on HuggingFace and ModelScope.\r\n\r\nMore information about downloading the models can be found in the [Models](models.md) page.\r\n\r\n## Prompt Format\r\n\r\nFor models with versions \u003e= 21.1, the prompt format is defined in the model card.\r\n\r\nFor models \u003c 21.1: refer to: [Legacy Prompt Format](legacy-prompt-format.md)\r\n\r\n## Inference with Ollama (recommended for personal users)\r\n\r\nOllama is a platform for locally deploying large models on consumer-grade hardware. It supports various inference methods such as CPU, CUDA, ROCm, and automatically selects the best hardware accelerator based on the actual situation. Ollama supports model quantization deployment, which means that large models can also run on devices with small memory.\r\n\r\nOllama implements one-stop model download, local deployment, and running. After [installing Ollama](https://github.com/ollama/ollama), you can deploy the 4-bit quantized version of the 8B model with just one command:\r\n\r\n```sh\r\nollama run openbuddy/openbuddy-llama3-8b-v21.1-8k\r\n```\r\n\r\nMore of our models can be found at: https://ollama.com/openbuddy\r\n\r\n## High-concurrency Inference using `vllm` in Linux + CUDA GPU environment\r\n\r\nStarting form v21, OpenBuddy models have their prompt formats defined in the `tokenizer_config.json` file, allowing for direct deployment using `vllm` to provide an OpenAI-compatible API service.\r\n\r\nFor more information, please refer to the [vllm documentation](https://docs.vllm.ai/en/latest/serving/openai_compatible_server.html).\r\n\r\n`vllm` is more suitable for high concurrency, multiple users, long context and other scenarios. Through technologies such as `FP8 KV Cache`, the concurrency and long text performance can be further improved. `vllm` currently only supports the Linux operating system and usually requires a CUDA GPU.\r\n\r\n## Disclaimer\r\n\r\nAll OpenBuddy models have inherent limitations and may potentially produce outputs that are erroneous, harmful, offensive, or otherwise undesirable. Users should not use these models in critical or high-stakes situations that may lead to personal injury, property damage, or significant losses. Examples of such scenarios include, but are not limited to, the medical field, controlling software and hardware systems that may cause harm, and making important financial or legal decisions.\r\n\r\nOpenBuddy is provided \"as-is\" without any warranty of any kind, either express or implied, including, but not limited to, the implied warranties of merchantability, fitness for a particular purpose, and non-infringement. In no event shall the authors, contributors, or copyright holders be liable for any claim, damages, or other liabilities, whether in an action of contract, tort, or otherwise, arising from, out of, or in connection with the software or the use or other dealings in the software.\r\n\r\nBy using OpenBuddy, you agree to these terms and conditions, and acknowledge that you understand the potential risks associated with its use. You also agree to indemnify and hold harmless the authors, contributors, and copyright holders from any claims, damages, or liabilities arising from your use of OpenBuddy.\r\n\r\n## License Restrictions\r\n\r\nOpenBuddy-LLaMA series models are subject to Meta's licensing agreement. These models are intended for use only by individuals who have obtained approval from Meta and are eligible to download LLaMA. If you have not obtained approval from Meta, you must visit the https://ai.meta.com/llama/ page, read and agree to the model's licensing agreement, submit an application, and wait for approval from Meta before downloading the model from the model page.\r\n\r\nFor the OpenBuddy-Falcon-7B/40B, OpenBuddy-Mistral, OpenBuddy-Zephyr and OpenBuddy-OpenLLaMA series models, they are released under the Apache 2.0 License. Please refer to the Apache 2.0 License for applicable scope and restrictions.\r\n\r\nFor other models, they are usually released under the same license as the base model. Please refer to the Model Card for more information.\r\n\r\nRegarding the source code related to the OpenBuddy open-source project (including, but not limited to, example code), they are released under the Apache 2.0 License.\r\n\r\n## Acknowledgements\r\n\r\nWe extend our deepest gratitude to the open-source community for their selfless assistance and contributions to the OpenBuddy project.\r\n\r\nFirstly, we would like to specifically thank WeiKe Software for their robust support and help in the aspect of model training. We also want to thank [AIOS.club](https://github.com/aios-club) for their invaluable support.\r\n\r\nWe thank [Mr. Su Jianlin](https://kexue.fm/) for his valuable advice during the model training process. Not only did he provide professional advice, but he also proposed several methods for context expansion, enabling open models to support inference with long context, which has had a profound impact on our work.\r\n\r\nOur appreciation goes to [flysnow](https://www.flysnow.org/about/) and [jstzwj](https://github.com/jstzwj). They provided valuable advice during the early stages of model development and extended substantial support and assistance in model inference.\r\n\r\nAt the same time, we also wish to express our gratitude to camera and other enthusiasts of open language models. Their suggestions played a pivotal role in improving the model.\r\n\r\nOnce again, we thank everyone who has contributed to the OpenBuddy project. Our success is inseparable from your support and encouragement.","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FOpenBuddy%2FOpenBuddy","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FOpenBuddy%2FOpenBuddy","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FOpenBuddy%2FOpenBuddy/lists"}