Rag over all physics-related Wikipedia pages
A few weeks ago, during the LK-99 hype, I made a demo app to explain superconductors in simple terms using RAG over relevant Wikipedia pages. I thought it would be cool to extend this to all physics-related Wikipedia pages (which turns out to be ~ 14K). After getting the text and splitting the pages, I ended up with around 100K chunks. I created all the embeddings using OpenAI’s embedding API, which cost around $7. I stored all the vectors in Pinecone.
Does the same job
all alternatives →- AHA HN/Reddit-style site for scientific pre-prints and publications2018 · upvote.pub · ▲275
- ETEasy to use Wikipedia API for Python2013 · github.com · ▲218
- ABA bookmarklet to remove clickgates on New York Times, Medium, etc.2019 · ▲71
Recently i stumbled on too many clickgates on the Medium blog Towards Data science. Considering that most people publish to share knowledge on Medium and are driven into putting their content behind a paywall, without actually getting paid for it, including myself. I felt like Medium is running the academic publishing scheme. Get free content and get paid for it. So I decided to create a small script to bypass the paywall on Medium, it turns out it also works on other newssites. Heres the website: https://sugoidesune.github.io/readium/ For the curious I will explain the…
- PFPlayground for comparing embedding models on Wikipedia+book retrieval2023 · embeds.ai · ▲5
Introducing embeds.ai: an embedding playground to compare how embedding models work on a real world use case (retrieval augmented generation for Wikipedia articles + Elad Gil's High growth handbook) A few weeks ago, Shreyan and I were looking for an embedding model to use for RAG. We eventually came across the MTEB leaderboard, but we struggled to understand the benchmark scores. We wanted a tool to test various embedding models with example queries on real-world datasets. After unsuccessfully looking for such a “playground”, we decided to just build one ourselves! We embedded HuggingFace’s…
- AEAn extensive set of RAG implementations+many different strategies2024 · github.com · ▲6
Hi all, Sharing a repo I was working on for a while. It’s open-source and includes many different strategies for RAG (currently 17), including tutorials, and visualizations. This is great learning and reference material. Open issues, suggest more strategies, and use as needed. Enjoy!
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
