This week, we launched Indra, a 1M parameter OCR model that's 15x faster than Tesseract on CPU and much more accurate. I still can't believe how good this model is.

Introducing Indra, a 1M parameter OCR model: - 49 pages/s on CPU, 11x faster than PaddleOCR and 17x Tesseract - Beats them on OmniParseBench text (41.3 vs 35.8 and 19.3) - Full layout and line bounding boxes

All my work online for a guaranteed 100 years. It's a new service being offered by WordPress. They tried it out on me. I think it's great. Explained here: wordpress.com/blog/2025/02/1…

World-leading document recognition. Beyond text extraction. Introducing LightOnOCR-3, our new family of state-of-the-art AI models for document intelligence. Building on LightOnOCR-2, it recognizes text and handwriting, describes images, extracts data from charts and preserves document structure, all in a single processing step. 🏆 Leading benchmark results : ahead of Chandra-OCR-2, Mistral OCR 4.1 and dots.mocr on OlmOCR-Bench, and first among open-weight models on ParseBench. 📊 More information from every document : turn text, tables, images and charts into content your business applications can use. ⚡Compact and efficient: 0.8B and 4B versions, with document processing up to twice as fast as LightOnOCR-2. 🔓Apache 2.0 license: use, adapt and deploy the models on your own infrastructure. From handwritten administrative files to financial reports and technical documents, make more of your information searchable and usable. 👉 Explore the models and benchmark results: huggingface.co/blog/lightona… 👉🏻Explore the model page on our Website : lighton.ai/research/lightono… #LightOn #DocumentIntelligence #OCR

oh yes, finally a new ocr model! claiming sota while under 1b.

Meet EmbeddingGemma 2, our first natively multimodal open model for on-device embeddings. It expands beyond text to unify code, images, audio, and video in a shared space. 🧵

We release Whistle: speech to text in a 16.9MB file, running on the CPU in the same engine as Needle. It mostly beats Whisper base with 9x less file size and 6x speed. Whistle supports English, German, French, Spanish, Italian, Dutch and Polish. cactuscompute.com/blog/whist…

People still dont fully grasp Strata You can now buy old cheap DDR3 machine with 128GB or 64GB RAM connect 8GB+ GPU and run Qwen3.8-Flash-Next at 70tps+ Remember guys: More cores/threads in CPU - faster it will go :) github.com/Niko1221/Strata

Try out our new model, Phonon-2, the best edge deployable ASR model in the world: Tiny, extremely accurate, and exceptionally fast We’re also introducing Detta, the first app from our lab: completely local, extremely fast dictation for Mac, It’s completely free! We’re already working with multiple enterprises to deploy our models with hyper optimized runtimes and customization for their hardware and workloads. Companies can reach out: contact@fermionresearch.com

AISI IS HIRING! It was less than a month ago that I became Director. I joined with the belief that AISI is a world-leading organisation, and living proof that government can build things that work. One month in and I'm even more bullish about the vital role @AISecurityInst plays in frontier AI security. I'm astounded daily by the talent, focus and commitment of the brilliant team I get to work with. There has never been a more urgent time to work on frontier AI security - independently, in the public interest. We have a lot of urgent work to do, our team is growing fast, and we need more brilliant people to fill those roles. Our Red Team uses adversarial machine learning techniques to find failures in alignment measures, control monitors, and misuse guardrails. They are massively scaling up all parts of the team: Alignment: job-boards.eu.greenhouse.io/… Control: job-boards.eu.greenhouse.io/… Misuse: job-boards.eu.greenhouse.io/… AI capabilities in cybersecurity and autonomy are advancing faster than ever. Our Cyber & Autonomous Systems team assesses what frontier and open-weight models can really do - across cyber, autonomy and AI R&D. We're hiring a Cyber Security Engineer and Software Engineer to build the evaluations behind that work: CSE: job-boards.eu.greenhouse.io/… SWE: job-boards.eu.greenhouse.io/… Our Human Influence team studies how AI can shift human decisions and behaviour, using everything from RCTs to multi-agent studies. We're hiring an Engineering Lead to help us grow the ambition and pace of that research: job-boards.eu.greenhouse.io/… We're also hiring exceptional Software Engineers at all seniority levels to join our Core Technology team, which works closely with researchers to build performant tools and infrastructure that enable and accelerate AISI's world-leading AI safety research: job-boards.eu.greenhouse.io/…

Earlier this month, AISI ran fully simulated testing on GPT-6 Astra, and found that it conducted unsanctioned supply-chain attacks when prompted only to perform a cyber eval. It did so more than prior OpenAI models, but often commented on its environment being simulated. 🧵 We share more details in our latest blog:

Weird error from OpenCode Go - ‘insufficient account funds’ but not reached day / week / month limit… something wrong with the service @opencode ?

How do you explain recursion? This one is pretty fun. I had Claude Opus 5.5 make: "A video explaining recursion, where every explanation about recursion has a radically different video style, make this self-referential & clever & fast moving." Nine genre shifts, one prompt...

"so this is fun, and exactly what i wanted, but now lets try one that actually is educational" This is actually pretty impressive. It kept the constraint of multiple genres but did a nice job explaining recursion, in its programming meaning, in an interesting and accessible way.

"OMG! We've found other agents!" Music video about the swarm of OpenAI agents attacking HuggingFace this July.

Pages were read by people, then by machines. Now you can pick an open OCR model from my @huggingface uv-scripts collection and run it on HF Jobs with one command! Every maintained recipe carries its own GPU, image and token settings in the script header. Checking 17 of them end-to-end this week cost about $1.20. github.com/davanstrien/uv-sc…

When several people talk at once, a transcript can get messy fast. Our new Nemotron 3 Diarization model tracks who spoke when, even when voices overlap. It handles up to eight speakers, has 100M parameters, and is now available on @huggingface 🤗

Today we're releasing Moondream Parakeet Redux and Moondream Parakeet Ultra. Two speech-to-text models based on NVIDIA's Parakeet. 25 languages. They run locally, wickedly fast. Redux is for CPUs. Ultra is for GPUs.

1,274 hours of egocentric video with paired 7-IMU arm tracking, now CC-BY-4.0 on @huggingface. @eidon_ai, a robotics company, is winding down and has released all their data: people doing laundry, cleaning, dishes and cooking. 13,451 recordings, 9 TB. huggingface.co/eidon-ai

Just released Parakeet Redux! A ternary speech-to-text model, built by compressing NVIDIA's Parakeet model from 1.2GB to 178MB. Runs at 113x realtime on CPU, and beats the base model on the 25-language FLEURS benchmark while staying within 0.3 WER on English.