GitHub radar
Tag multiple texts at a time Read more
convai Innovations shared an open source text classification model that accepts text and a set of typed questions and outputs answers to all of them alongside their confidence levels in approximately 33 ms.
Laya is an open source text classification model developed by convaiinnovations on Hugging Face. Enter your text and list of typed questions (of various question types like multiple choice, true/false, etc.) and get back all the answers along with their confidence scores. Base model size: 421M. Supports more than 100 languages. Open source alternative to proprietary TypeSafe Jev decision API. According to this benchmark, Laya is 7.8X faster for one question and 3x better in terms of calibration accuracy. Licensed under Apache license. Self host or use for free online via Hugging Face demo space (no registration needed).
Why a vibe-coder should care
If you have a text processing product which requires routing or filtering messages based on their contents, now you don’t need to make LLM calls for each message. You can ask up to 5 questions at a time in a single request. Save time and cost by avoiding multiple ChatGPT queries. Use online on Hugging Face (no registration needed).
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Open this demo space: https://huggingface.co/spaces/convaiinnovations/laya-demo — paste a text like 'My name is Alex, I am interested in your service, please call me.' and write five yes/no or choice questions about it, such as: is this a hot lead? is this spam? should I call them? should I delete this? is a sale possible? Show me the answers with confidence scores.
No installation needed — runs in the browser.
Open on Hugging Face▌ More finds