LlamaIndex Newsletter 2023-02-13:LlamaIndex v0.10 发布与 Self-RAG 等更新
LlamaIndex Newsletter 2023–02–13
LlamaIndex 发布开源版 v0.10,以全新核心包和数百个独立 PyPi 包推进生产可用性,LlamaHub 重构为集成中心,ServiceContext 被弃用。Self-RAG 以 LlamaPack 形式加入,FlowiseAI 集成支持无代码 RAG 开发,RAG CLI 工具用 Mistral-7B 和 bge-m3 在本地搜索文件。
Greetings, LlamaIndex Adventurers 🦙,
Welcome to an exhilarating week of discoveries at LlamaIndex! Our community’s lively input and the abundance of learning tools await to supercharge your journey through LlamaIndex.
Explore our free and paid plans today.
Before we dive into the updates, we have two major announcements:
- LlamaIndex v0.10: Our latest open-source release marks a monumental step towards production readiness. With a new core package and hundreds of integrations and LlamaPacks now available as separate PyPi packages, we’ve massively improved organization and version tracking. Major updates include the refactoring of LlamaHub into a central hub for all integrations and the deprecation of ServiceContext for an enhanced development experience. Blog, Migration Guide.
- Introducing Short Courses on Advanced RAG Development: Master complex RAG systems with our series, covering everything from unstructured data to agent integration. Learn through LlamaIndex Query Pipelines, from basic text-to-SQL to advanced query techniques, and build scalable RAG applications with hands-on guidance. Video1, Video2.
Your creativity fuels our inspiration! We’re excited to see any projects, articles, or videos you’re passionate about. Share your incredible creations with us at news@llamaindex.ai. Haven’t joined our newsletter yet? Make sure to subscribe on our website for the latest LlamaIndex updates delivered directly to your inbox.
🤩 The highlights:
- Self-RAG: Introducing Self-RAG, now part of LlamaIndex as a LlamaPack. Boosts LLM training and RAG workflows with dynamic capabilities. Notebook, Tweet.
- LlamaIndex + FlowiseAI Integration: Seamlessly merge LlamaIndex with FlowiseAI for effortless, no-code RAG app development. Docs, Tweet.
- RAG Guide with MistralAI: MistralAI’s new doc includes a RAG guide with LlamaIndex. Utilize Mistral-medium for enhanced RAG functions. Docs.
✨ Feature Releases and Enhancements:
- We have introduced a seamless integration between LlamaIndex and FlowiseAI, enabling easy, no-code development of advanced RAG applications with a drag-and-drop interface for quick chatbot or agent integration. Docs, Tweet.
- We have introduced Self-RAG, a dynamic retrieval tool by Akari Asai’s team, now available as a LlamaPack for easy integration, enhancing LLM training and RAG workflows with dynamic, iterative capabilities. Notebook, Tweet.
- We have introduced the RAG CLI tool that allows you to search any file on your filesystem using on-device language model embeddings, featuring the power of Mistral-7B and bge-m3 for an advanced, customizable experience. Docs, Tweet.
- We have launched full-stack agent servers with a single CLI command using
create-llamafrom LlamaIndex, offering instant access to 50+ tools for any agent project. Tweet. - We have introduced agents in LlamaIndex.TS, enabling advanced AI software development in TypeScript with features like function calling and multi-document handling. Blog, Docs, Tweet.
- DeepEval is integrated with LlamaIndex, significantly enhancing RAG evaluation capabilities and introducing unit testing for LlamaIndex apps in CI/CD environments. Docs.
🗺️ Guides:
- Guide to RAG with LlamaIndex in MistralAI’s new documentation with Mistral-medium and Mistral embedding models.
- Guide to Building Agentic RAG to incorporate user feedback in real-time enhancing complex searches with a human-in-the-loop approach.
- Guide to Integrating Huggingface’s New Messages API with OpenAI compatibility, simplifying the integration process for Inference Endpoints and Text Generation Inference.
✍️ Tutorials:
- Plaban Nayak tutorial on Setting up Query Pipeline For Advanced RAG Workflow using LlamaIndex.
- Krish Naik tutorial on Step-by-Step Guide to Building a RAG LLM App with Llama2 and LlamaIndex.
- HelixML tutorial to Knowledge Memorization by fine-tuning Mistral-7B for enhanced knowledge memorization, offering a new way to reason across contexts without RAG’s limitations.
- Wenqi Glantz tutorial on NeMo Guardrails, the Ultimate Open-Source LLM Security Toolkit.
🎥 Webinar:
- Webinar of Laurie with Ankit Khare(Rockset) delves into the essentials of RAG — its purpose, methodology, how LlamaIndex facilitates it, and exciting developments for 2024.
- Webinar with Zilong Wang, and Tianyang Liu on Advanced Tabular Data Understanding with LLMs.
🏢 Calling all enterprises:
Are you building with LlamaIndex? We are working hard to make LlamaIndex, even more, Enterprise-ready and have sneak peeks at our upcoming products available for partners. Interested? Get in touch.
来源:LlamaIndex:产品、工程与评测 · llamaindex.ai