[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-1423":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":23,"hasPages":25,"topics":26,"createdAt":10,"pushedAt":10,"updatedAt":36,"readmeContent":37,"aiSummary":38,"trendingCount":16,"starSnapshotCount":16,"syncStatus":18,"lastSyncTime":39,"discoverSource":40},1423,"Open-Assistant","LAION-AI\u002FOpen-Assistant","LAION-AI","OpenAssistant is a chat-based assistant that understands tasks, can interact with third-party systems, and retrieve information dynamically to do so.","https:\u002F\u002Fopen-assistant.io",null,"Python",37399,3280,431,229,0,1,2,16,3,45,"Apache License 2.0",false,"main",true,[27,28,29,30,31,32,33,34,35],"ai","assistant","chatgpt","discord-bot","language-model","machine-learning","nextjs","python","rlhf","2026-06-12 02:00:27","\u003Ch1 align=\"center\">\n    \u003Cspan>Open-Assistant\u003C\u002Fspan>\n  \u003Cimg width=\"auto\" height=\"50px\" src=\"https:\u002F\u002Fgithub.com\u002FLAION-AI\u002FOpen-Assistant\u002Fblob\u002Fmain\u002Fassets\u002Flogo_crop.png\"\u002F>\n\u003C\u002Fh1>\n\n\u003Cblockquote>\n\u003Cp>:memo: \u003Cstrong>NOTE\u003C\u002Fstrong>: OpenAssistant is completed, and the project is now finished. Thank you to everyone who contributed! Check out our \u003Ca href=\"https:\u002F\u002Fprojects.laion.ai\u002FOpen-Assistant\u002Fblog\u002F2023\u002F10\u002F25\u002Fopen-assistant-is-completed\">blog post\u003C\u002Fa> for more information. The final published oasst2 dataset can be found on HuggingFace at \u003Ca href=\"https:\u002F\u002Fhuggingface.co\u002Fdatasets\u002FOpenAssistant\u002Foasst2\">OpenAssistant\u002Foasst2\u003C\u002Fa>\u003C\u002Fp>\n\u003C\u002Fblockquote>\n\n\u003Cdiv align=\"center\">\n\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FLAION-AI\u002FOpen-Assistant\u002Fstargazers\">![GitHub Repo stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FLAION-AI\u002FOpen-Assistant?style=social)\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Flaion-ai.github.io\u002FOpen-Assistant\u002F\">![Docs](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdocs-laion--ai.github.io%2FOpen--Assistant%2F-green)\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FLAION-AI\u002FOpen-Assistant\u002Factions\u002Fworkflows\u002Fbuild-frontend.yaml\">![GitHub Workflow Status](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002FLAION-AI\u002FOpen-Assistant\u002Fbuild-frontend.yaml?label=build-frontend)\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FLAION-AI\u002FOpen-Assistant\u002Factions\u002Fworkflows\u002Fbuild-postgres.yaml\">![GitHub Workflow Status](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002FLAION-AI\u002FOpen-Assistant\u002Fbuild-postgres.yaml?label=build-postgres)\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FLAION-AI\u002FOpen-Assistant\u002Factions\u002Fworkflows\u002Fpre-commit.yaml\">![GitHub Workflow Status](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002FLAION-AI\u002FOpen-Assistant\u002Fpre-commit.yaml?label=pre-commit)\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FLAION-AI\u002FOpen-Assistant\u002Factions\u002Fworkflows\u002Ftest-api-contract.yaml\">![GitHub Workflow Status](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002FLAION-AI\u002FOpen-Assistant\u002Ftest-api-contract.yaml?label=tests-api)\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FLAION-AI\u002FOpen-Assistant\u002Factions\u002Fworkflows\u002Ftest-e2e.yaml\">![GitHub Workflow Status](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002FLAION-AI\u002FOpen-Assistant\u002Ftest-e2e.yaml?label=tests-web)\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FLAION-AI\u002FOpen-Assistant\u002Factions\u002Fworkflows\u002Fdeploy-docs-site.yaml\">![GitHub Workflow Status](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002FLAION-AI\u002FOpen-Assistant\u002Fdeploy-docs-site.yaml?label=deploy-docs)\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FLAION-AI\u002FOpen-Assistant\u002Factions\u002Fworkflows\u002Fproduction-deploy.yaml\">![GitHub Workflow Status](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002FLAION-AI\u002FOpen-Assistant\u002Fproduction-deploy.yaml?label=deploy-production)\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FLAION-AI\u002FOpen-Assistant\u002Factions\u002Fworkflows\u002Frelease.yaml\">![GitHub Workflow Status](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002FLAION-AI\u002FOpen-Assistant\u002Frelease.yaml?label=deploy-release)\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FLAION-AI\u002FOpen-Assistant\u002Freleases\">![GitHub release (latest by date)](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fv\u002Frelease\u002FLAION-AI\u002FOpen-Assistant)\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub-com.translate.goog\u002FLAION-AI\u002FOpen-Assistant\u002Fblob\u002Fmain\u002FREADME.md?_x_tr_sl=auto&_x_tr_tl=en&_x_tr_hl=en&_x_tr_pto=wapp\">![Translate](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FTranslate-blue)\u003C\u002Fa>\n\n\u003C\u002Fdiv>\n\n# Table of Contents\n\n- [What is Open Assistant?](#what-is-open-assistant)\n- [Useful Links](#useful-links)\n- [How To Try It Out](#how-to-try-it-out)\n- [The Vision](#the-vision)\n- [The Plan](#the-plan)\n- [How You Can Help](#how-you-can-help)\n\n---\n\n## What is Open Assistant?\n\n\u003Cp align=\"center\">\nOpen Assistant is a project meant to give everyone access to a great chat based\nlarge language model.\n\u003C\u002Fp>\n\nWe believe that by doing this we will create a revolution in innovation in\nlanguage. In the same way that stable-diffusion helped the world make art and\nimages in new ways we hope Open Assistant can help improve the world by\nimproving language itself.\n\n# Useful Links\n\n- [Data Collection](https:\u002F\u002Fopen-assistant.io)\n\n- [Chat](https:\u002F\u002Fopen-assistant.io\u002Fchat)\n\n- [Project Documentation](https:\u002F\u002Fprojects.laion.ai\u002FOpen-Assistant\u002F)\n\n## How To Try It Out\n\n### Chatting with the AI\n\nThe chat frontend is now live [here](https:\u002F\u002Fopen-assistant.io\u002Fchat). Log in and\nstart chatting! Please try to react with a thumbs up or down for the assistant's\nresponses when chatting.\n\n### Contributing to Data Collection\n\nThe data collection frontend is now live [here](https:\u002F\u002Fopen-assistant.io\u002F). Log\nin and start taking on tasks! We want to collect a high volume of quality data.\nBy submitting, ranking, and labelling model prompts and responses you will be\ndirectly helping to improve the capabilities of Open Assistant.\n\n### Running the Development Setup Locally (without chat)\n\n**You do not need to run the project locally unless you are contributing to the\ndevelopment process. The website link above will take you to the public website\nwhere you can use the data collection app and the chat.**\n\nIf you would like to run the data collection app locally for development, you\ncan set up an entire stack needed to run **Open-Assistant**, including the\nwebsite, backend, and associated dependent services, with Docker.\n\nTo start the demo, run this in the root directory of the repository (check\n[this FAQ](https:\u002F\u002Fprojects.laion.ai\u002FOpen-Assistant\u002Fdocs\u002Ffaq#docker-compose-instead-of-docker-compose)\nif you have problems):\n\n```sh\ndocker compose --profile ci up --build --attach-dependencies\n```\n\n> **Note:** when running on MacOS with an M1 chip you have to use:\n> `DB_PLATFORM=linux\u002Fx86_64 docker compose ...`\n\nThen, navigate to `http:\u002F\u002Flocalhost:3000` (It may take some time to boot up) and\ninteract with the website.\n\n> **Note:** If an issue occurs with the build, please head to the\n> [FAQ](https:\u002F\u002Fprojects.laion.ai\u002FOpen-Assistant\u002Fdocs\u002Ffaq) and check out the\n> entries about Docker.\n\n> **Note:** When logging in via email, navigate to `http:\u002F\u002Flocalhost:1080` to\n> get the magic email login link.\n\n> **Note:** If you would like to run this in a standardized development\n> environment (a\n> [\"devcontainer\"](https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fdevcontainers\u002Fcontainers))\n> using\n> [vscode locally](https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fdevcontainers\u002Fcreate-dev-container#_create-a-devcontainerjson-file)\n> or in a web browser using\n> [GitHub Codespaces](https:\u002F\u002Fgithub.com\u002Ffeatures\u002Fcodespaces), you can use the\n> provided [`.devcontainer`](.devcontainer\u002F) folder.\n\n### Running the Development Setup Locally for Chat\n\n**You do not need to run the project locally unless you are contributing to the\ndevelopment process. The website link above will take you to the public website\nwhere you can use the data collection app and the chat.**\n\n**Also note that the local setup is only for development and is not meant to be\nused as a local chatbot, unless you know what you are doing.**\n\nIf you _do_ know what you are doing, then see the `inference` folder for getting\nthe inference system up and running, or have a look at `--profile inference` in\naddition to `--profile ci` in the above command.\n\n## The Vision\n\nWe are not going to stop at replicating ChatGPT. We want to build the assistant\nof the future, able to not only write email and cover letters, but do meaningful\nwork, use APIs, dynamically research information, and much more, with the\nability to be personalized and extended by anyone. And we want to do this in a\nway that is open and accessible, which means we must not only build a great\nassistant, but also make it small and efficient enough to run on consumer\nhardware.\n\n## The Plan\n\n##### We want to get to an initial MVP as fast as possible, by following the 3-steps outlined in the [InstructGPT paper](https:\u002F\u002Farxiv.org\u002Fabs\u002F2203.02155)\n\n1. Collect high-quality human generated Instruction-Fulfillment samples\n   (prompt + response), goal >50k. We design a crowdsourced process to collect\n   and reviewed prompts. We do not want to train on\n   flooding\u002Ftoxic\u002Fspam\u002Fjunk\u002Fpersonal information data. We will have a\n   leaderboard to motivate the community that shows progress and the most active\n   users. Swag will be given to the top-contributors.\n2. For each of the collected prompts we will sample multiple completions.\n   Completions of one prompt will then be shown randomly to users to rank them\n   from best to worst. Again this should happen crowd-sourced, e.g. we need to\n   deal with unreliable potentially malicious users. At least multiple votes by\n   independent users have to be collected to measure the overall agreement. The\n   gathered ranking-data will be used to train a reward model.\n3. Now follows the RLHF training phase based on the prompts and the reward\n   model.\n\nWe can then take the resulting model and continue with completion sampling step\n2 for a next iteration.\n\n### Slide Decks\n\n[Vision & Roadmap](https:\u002F\u002Fdocs.google.com\u002Fpresentation\u002Fd\u002F1n7IrAOVOqwdYgiYrXc8Sj0He8krn5MVZO_iLkCjTtu0\u002Fedit?usp=sharing)\n\n[Important Data Structures](https:\u002F\u002Fdocs.google.com\u002Fpresentation\u002Fd\u002F1iaX_nxasVWlvPiSNs0cllR9L_1neZq0RJxd6MFEalUY\u002Fedit?usp=sharing)\n\n## How You Can Help\n\nAll open source projects begin with people like you. Open source is the belief\nthat if we collaborate we can together gift our knowledge and technology to the\nworld for the benefit of humanity.\n\nCheck out our [contributing guide](CONTRIBUTING.md) to get started.\n","Open-Assistant 是一个基于聊天的助手，能够理解任务、与第三方系统交互并动态检索信息以完成任务。该项目使用 Python 语言开发，具备强大的自然语言处理能力，并采用了机器学习技术来提升其对话质量。它支持通过强化学习人类反馈（RLHF）进行优化，确保了模型的响应更加符合用户需求。适用于需要自动化客服、智能助手等场景，可以帮助企业或个人提高工作效率和服务质量。尽管项目已经完成，但其成果包括最终发布的 oasst2 数据集仍可在 HuggingFace 上获取，为后续研究和应用提供了宝贵资源。","2026-06-11 02:43:42","top_all"]