展示HN:Kiln – 用于AI产品开发的GitHub+Jira+Figma整合工具

1作者: scosman20 天前原帖
我一直在努力开发 Kiln(拥有 3000 个 GitHub 星标,约 1 万用户): - GitHub: [https://github.com/Kiln-AI/Kiln](https://github.com/Kiln-AI/Kiln) - 演示视频: [https://getkiln.ai#demo](https://getkiln.ai#demo) 它的目标是以多种方式改善 AI 模型和产品开发流程: - 真实直观的桌面应用程序,基于我们的开源 Python 库 - 快速原型制作:快速尝试一系列最先进的模型和技术(无需编码) - 任务特定评估:系统在您的目标上的表现如何(不仅仅是全球基准) - 自信迭代:更新更好更快的模型/技术,确保在所有过去的评估/测试中有效 - 团队协作:让整个团队(产品经理/质量保证/工程/设计/数据科学)早期并频繁地参与。可以与 Git 同步的共享数据集和评估。 功能包括: - 微调数十个模型 - 交互式合成数据生成以构建训练/评估数据集 - 强大的自定义评估:LLM 作为评审和 G-Eval - 协作:通过 Git 或共享驱动器进行分享 - 运行任何 LLM:本地或远程 - 私有:本地软件 - 开源 Python 库 - 完全免费 - 还有更多:数据集管理、评分、提示生成器、JSON 支持、LLM 修复、推理模型等... 欢迎随时提问!
查看原文
I&#x27;ve been hard at work on Kiln (3k Github stars, ~10k users):<p>- Gihub: <a href="https:&#x2F;&#x2F;github.com&#x2F;Kiln-AI&#x2F;Kiln" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;Kiln-AI&#x2F;Kiln</a><p>- Demo Video: <a href="https:&#x2F;&#x2F;getkiln.ai#demo" rel="nofollow">https:&#x2F;&#x2F;getkiln.ai#demo</a><p>It&#x27;s goal is to makes AI model+product development processes better in a bunch of ways:<p>- Truly intuitive desktop apps, on top of our open-source python library<p>- Rapid prototyping: Quickly try a range of state of the art models+techniques (no code needed)<p>- Task specific evals: How does the system perform on your goals (not just global benchmarks)<p>- Iterate with confidence: update to better+faster models&#x2F;techniques, knowing it works on all past evals&#x2F;tests<p>- Team collaboration: get the entire team (PM&#x2F;QA&#x2F;Eng&#x2F;Design&#x2F;DS) engaged early and often. Shared datasets+evals you can sync with Git.<p>Features include:<p>- Fine tune dozens of models<p>- Interactive synthetic data generation to build train&#x2F;eval datasets<p>- Powerful custom evals: LLM as Judge and G-Eval<p>- Collaboration: share with Git or shared drives<p>- Runs any LLM: locally or remote<p>- Private: it&#x27;s local software<p>- Open Source Python library<p>- Completely free<p>- Much more: dataset management, ratings, prompt generators, JSON support, LLM repairs, reasoning models...<p>Happy to answer any questions!