Details on OpenAI's new assistants RAG

*Hard* creep into vectorstore territory

Thoughts:
* Default chunk overlap of 50%, super interesting
* Metadata filtering, super interesting how this dips into vectorstore territory
* Unsure about what chunking method they use - 800 tokens

some vectorstore analysis - what do folks like when building for production?

Wrote my first custom chain in LangChain as part of improving our AI search in stockinsights.ai

ModifyDocumentsChain adds additional context to docs retrieved from vectorstore based on the metadata and runs before StuffDocumentChain. This helps us #buildinpublic

All the code that is required for a simple, yet effective RAG pipeline.

VectorStore on Supabase + Langchain is incredibly powerful.

Run as a build script on your documentation markdown.

And watch what happens.

Just added long-term memory to my GymBro AI Buddy.

Now it remembers snippets of my conversation across sessions.

How I did it:
* Used @langchain's VectorStore-Backed Memory
* Added AI Adventure by Steamship Vector DB

Pushing the source code in my next Tweet!

Confused about using LlamaIndex 🦙 indices or configuring parameters for building robust QA Systems?

See my new video on:

1️⃣ VectorStore Index
2️⃣ Storage Context
3️⃣ Service Context
4️⃣ List Index
5️⃣ KeyWordTable Index

📺 Video: rb.gy/ofmvo
📔 Notebook:

📢 New FREE Bubble template built with @finetuner_ai 🔥 100% plug and play, ZERO coding req.🙌

ChatGPT plus:
✅ Tools incl. Browsing
✅ Vectorstore (PDF, CSV, PPTX, DOCX, Scraper)
✅ Conversational Memory
✅ Full system prompt exposed
✅ GPT 4 and Claude 2

♻️ RT & DM -> I'll

⚡️ Mindwave Progress:
Got the architecture sketched out, 'Concept' docs written, interfaces for the building blocks defined.

✅ OpenaAI Chat and Complete API support added,
✅ InMemory Array Vectorstore is working (for testing)
✅ text-embedding-ada-002 Embedding API working

Previously I've talked about the amazing Ingestion Pipeline from LlamaIndex 🦙.

Here's how to use Redis (@Redisinc) as the docstore, vectorstore and cache for the pipeline.

LlamaIndex abstractions make it really easy to just use Redis for the entire pipeline 🔥👇

5/13: Private RAG
How to do RAG w/ local LLM/embeddings/vectorstore
python.langchain.com/docs/use_cases…

🧠 Mindwave update:

✅ Re-worked the vectorstore interface, its not much simpler. (we can expand it later if needed YAGNIKISS etc)
✅ Weaviate Driver now works.
✅ Added Gmail OAuth to the Mindwave demo (chat with your emails) etc, will be available as a Document loader in core

LangChain 🪢 Weaviate

Weaviate • vector database's open source vectorstore has features ranging from native multi-tenancy to advanced filtering, and it's now accessible through a standalone integration package: `langchain-weaviate`!

Its wide range of features even power the retrieval behind

5人目は森 一弥氏「AI全盛時代に備えるナレッジデータ管理〜VectorStoreの選び方〜」のトークです。
RAG（検索拡張生成）環境の基本概念とそのビジネスへの応用可能性を解説。実際に触ってみた感触も踏まえてトーク！
#生成AI協会 #gais #ChatGPT #ジェネレーティブＡＩ勉強会

The Generative AI ecosystem is very young (9 months)

- LangChain & LlamaIndex 🦙: best frameworks
- Chroma: OSS Vectorstore
- PrivateGPT: awesome Private-based chat from Iván Martínez
- GPT4ALL from Nomic AI: OSS LLM
- Quivr (YC W24) 🧠: doc retrieval app

What a year!

Master Transformer in 18 Hours with PyTorch🚀

Check out on my YouTube Channel

🟠 youtube.com/playlist?list=…

#MachineLearning #opensource #LLM #vectorstore #NLP #ArtificialIntelligence #datascience #textprocessing #deeplearning #deeplearning ai

🧠 Mindwave Update:

✅ Pinecone Vectorstore Driver implemented.
✅ The docs made an appearance in the AIWithLaravel.com course (buy it, its great)

🤔 Also figured out that the Pinecone API uses duped-params like '?id=1&id=2' instead of this '?id[]=1&id[]=2', which took

🐘🦜 Announcing a new `langchain-postgres` integration package, implementing:

1️⃣ Vectorstore based on pgvector
2️⃣ Chat Message History
3️⃣ Checkpoint Saver for LangGraph

Vector store docs: python.langchain.com/docs/integrati…

Chat message history docs: github.com/langchain-ai/l…

Checkpoint

LlamaParse with LangChain 🔥

Uploaded a video on how I used Llamaparse from LlamaIndex 🦙 with LangChain to create a super simple RAG app using Chainlit

- Qdrant as vectorstore
- mixtral from Mistral AI via Groq Inc
- Langsmith for traces

Video:

Want to create AI apps based on your own documents and knowledge bases?

Here's how to easily create your own free vectorstore using LangChain and FAISS.

First things first, if you want the LLM to have access to your documents, you need to:

1. Load them

2. Preprocess them

New Steamship LangChain release

🚢🚢 VectorStore and Loader Support 🚢🚢

Deploy your LangChains to scalable APIs with zero setup.

Check it out!

github.com/steamship-core…