Not Lain

If you're curious about how VLMs process inputs across modalities me and @EdDotDaniels released this blogpost diving deeper into their architecture and what makes them turn, using SmolVLM as a case study. 📖 huggingface.co/blog/not-lain/…

@RisingSayak my entire feed is just pure slope right now, i just wish for more technical stuff, or for posts to be presented in a way that is not engagement farming.

Speedmaxxing to some XXX K followers on socials without any GENUINE substance is 🤮

@KShivendu_ @qdrant_engine @canva @Flipkart @HubSpot congrats 🥳 can't wait to see more of your achievements 🚀

2.5 years ago, I joined @qdrant_engine at the seed stage to help build the future of search 🔍 Since then we’ve got: - 100K+ daily Docker pulls - 30K GitHub stars - Teams at @canva, @Flipkart, @HubSpot and many others running Qdrant in production for search, recommendations, and AI agents. Proud that we scaled while staying true to our open-source roots, even as enterprise adoption accelerated. And recently we announced our $50M Series B! That's one small step for Qdrant, one giant leap for the future of search 👨‍🚀 We’re hiring. DM me with proof of exceptional work :)

😋

Our OSS engineer @itsclelia recently built 𝗹𝗶𝘁𝗲𝘀𝗲𝗮𝗿𝗰𝗵, a fully local document ingestion and retrieval CLI/TUI application powered by LiteParse ⚡ litesearch demonstrates how developers can assemble a high-performance, local-first retrieval pipeline using open tools from across the ecosystem: • Parsing: LiteParse, the fast and accurate document parser we recently open sourced • Chunking: @ChonkieAI • Embeddings: A local @nomic_ai model via @huggingface transformers.js • Vector storage: A local @qdrant_engine edge shard (custom-built in Rust and compiled as a native add-on) • Retrieval: Query stored files with optional path-based filtering and configurable relevance thresholds • Runtime: @bunjavascript for speed and versatility 💻 Check out the repository and try it yourself: github.com/AstraBert/lite… 📚 LiteParse docs: developers.llamaindex.ai/liteparse?utm_…

Hey there 👋 , I built 𝗹𝗶𝘁𝗲𝘀𝗲𝗮𝗿𝗰𝗵, a fully local document ingestion and retrieval CLI and TUI app, powered by LiteParse⚡ - Parse your unstructured documents with LiteParse, the lightning fast parser that we just open sourced at @llama_index - Chunk with @ChonkieAI - Embed with a local @nomic_ai model through @huggingface transformers.js - Store embeddings in a local @qdrant_engine edge shard (custom-built in Rust and compiled as a native add-on🦀) - Retrieve from stored files with (optional) path-based filtering and a relevance threshold The app runs on @bunjavascript, so make sure you have it installed🥞 Find all the code on GitHub: github.com/AstraBert/lite…

are there any good extensions for open weights quantized models to run in vscode ?

Introducing Chroma Context-1, a 20B parameter search agent. > pushes the pareto frontier of agentic search > order of magnitude faster > order of magnitude cheaper > Apache 2.0, open-source

New Voxtral-4B-TTS is beautiful 💛 it sounds so good that it's whispering at closed models: "your time is over" - Demo available on Hugging Face ⬇️

@itsclelia @Vish_owl Great read, thanks for working on this🚀

So excited to see live the article @Vish_owl and I wrote on LlamaParse x Google Gemini! In this guide, we show you how you can leverage LlamaParse's advanced agentic OCR to extract text and tables from complex financial documents, and then use Gemini 3 state of the art context understanding capabilities to turn the parsed content into human-friendly insights. Read the blog: developers.googleblog.com/build-a-smart-… Get started with LlamaCloud: cloud.llamaindex.ai/signup PS: this was really a team effort, thanks to everyone involved!✨🥰

Map-Anything v1 (Universal Feed-Forward Metric 3D Reconstruction) demo is now available on Hugging Face Spaces. Built with @Gradio and integrated with @rerundotio , it performs multi-image and video-based 3D reconstruction, depth, normal map, and interactive measurements.

@PixelEyePicture @huggingface looks great, thanks a lot for bringing this to my radar will give it a spin and see how it goes

@not_so_lain @huggingface Corridor Key is way better

🔥🚀 You can now directly load and use MatAnyone2 from @huggingface 🤗 🔗 huggingface.co/PeiqingYang/Ma… Loading and using the model now is as easy as a couple of lines of code 🤓💻

good morning

@dictator_jii granny would be sad, and after she left therapist told me to achieve what would have made her proud (graduate, get a job, ... ), she always said she's waiting for me to get a job and get lots of money, which i achieved, god rest her soul and i hope she's proud of me.

html table chunking goes wooooosh 🔥🚀 you can now keep your table header in each chunk when working with both html and markdown tables. feel free to support our open source work by starring our repo ✨ github.com/chonkie-inc/ch…

@OrwellNGoode not to flex but i'm connected to kana-chan on linkedin

i just gave up on leaving notes for AI generated prs and decided to build upon them ages ago. kinda hate it when i give feedback only to be tossed off the window

🚀 𝙈𝙖𝙩𝘼𝙣𝙮𝙤𝙣𝙚 2 is accepted to #CVPR2026! 🔥A stronger version of 𝙈𝙖𝙩𝘼𝙣𝙮𝙤𝙣𝙚 with finer details and enhanced robustness🔥 -🏡: pq-yang.github.io/projects/MatAn… -📜: arxiv.org/abs/2512.11782 -👩🏻‍💻: github.com/pq-yang/MatAny…

@peiqing001 The new model looks amazing, congrats on the new release 🔥

We are coming for your top spot, @openclaw