I disagree with this decision and I don't like it. But also... HOW DID ANTHROPIC NOT SEE THIS COMING‽ It is *the* obvious response to "this is too dangerous for anyone except us to use", since that relies on a premise ("we are uniquely good") that almost no-one agrees with.

M3 would never 🙂‍↔️ As a matter of fact, the weights are now open, too. huggingface.co/MiniMaxAI/Min…

🏃‍♀️💨 hf-sandbox is faster and more reliable than ever! Use it for your agents, RL training, to serve your model, or whatever else you can think of!

Super excited to announce that @arcee_ai is the first major American AI lab to replace AWS S3 with Hugging Face for ALL their models and datasets, public AND private 🔥🔥🔥 Multi-million $ partnership to support American open-source AI, let’s go!

🔗 Weights & code: huggingface.co/moonshotai/Ki…

PP-OCRv6 just released by Baidu @PaddlePaddle ✨ tiny 1.5M / small 7.7M / medium 34.5M ✨ 48 languages ✨ Supports handwritten/printed/industrial/screen and card text ✨ Edge friendly deployment

🚀PP-OCRv6 is officially released！ 🔥PaddleOCR’s new OCR model series scales from 1.5M to 34.5M parameters, bringing stronger accuracy, faster inference, and broader deployment options — from browsers and edge devices to servers. 📊What’s new: 🔸Tiny / Small / Medium models: 1.5M, 7.7M, 34.5M params 🔸 4.9% detection accuracy and 5.1% recognition accuracy over PP-OCRv5 🔸Up to 5.2× faster CPU inference with OpenVINO 🔸50 languages in one unified model 🔸New scenarios: PCB, CAD drawings, digital tubes, dot-matrix text 🔸Apache 2.0 open source ✨Lightweight OCR, built for the AI data era. 🔗Try it: 🌐 paddleocr.com 💻 github.com/PaddlePaddle/Padd… 🤗huggingface.co/collections/P… #PaddlePaddle #PaddleOCR #OCR #AI #ComputerVision #OpenSource #EdgeAI

Yesterday: diffusion LM beats AR on OCR correction. Today, after a suggestion from the @GoogleDeepMind team: bench it against its parameter-matched twin instead (same 26B MoE, same 4B active). The twin wins slightly on quality. The diffusion model is still ~10× faster. Updated scoreboard CER on 19th-c newspaper OCR (lower = better): • OCR input: 0.066 • Gemma-4-E4B: 0.042 (15.3s/passage) • DiffusionGemma: 0.035 (1.7s 🚀) • Gemma-4-26B MoE: 0.027 (16.3s) Equal capacity: diffusion trades accuracy for an order of magnitude of speed. Also tried the suggested sampler fixes (thx @joao_gante!) for the "seeded canvas just copies its input" bug. They free the model from copying but it re-derives the input instead of correcting it. Might still be some tweaking here? Full agent-written lab notebook, scripts runnable straight from the bucket: huggingface.co/buckets/davan…

huggingface_hub v1.19.0 is out 🚀 Three big ones: 🔐 Keyless CI/CD auth (Trusted Publishers) 🖥️ hf:// URIs in the CLI 🌐 Expose ports on Jobs Details below 🧵

Happy to announce the launch of the Far-Field ASR Leaderboard! 🎉 While many ASR benchmarks focus on clean speech, real-world applications need to handle noise, reverberation, and distant microphones. This leaderboard makes it easier to evaluate speech recognition models under realistic acoustic conditions and compare their robustness across challenging environments.

DiffusionGemma, our experimental open model released under an Apache 2.0 license, explores text diffusion, an exceptionally fast approach to text generation. Here’s how DiffusionGemma accelerates development: Faster token output: By shifting the bottleneck from memory bandwidth to raw compute, the model generates up to 4x faster token output on dedicated GPUs Accessible hardware footprint: Activates just 3.8B parameters during inference, fitting comfortably within 24GB-VRAM high-end consumer GPUs when quantized Novel workflows: Parallel token generation enables self-correction, making it ideal for code infilling, in-line editing, and non-linear structures DiffusionGemma prioritizes speed over raw quality and accelerates best on compute-bound hardware (like @NVIDIAAI GPUs). Standard @GoogleGemma 4 remains recommended for production quality and memory-bound devices.

Used an agent for a naughty no no use case while I still can: fine-tuning a domain-specific small VLM to be good at information extraction in the library/archive/historical research domain. Pointed an agent (Opus 4.8, Fabel refused....) at 🇫🇷 🇬🇧 handwritten death records and manuscript-catalogue cards. Resulted in a 4B model (from @numind_ai's NuExtract-3) that even follows schemas it's never seen. Agent found SFT beat GRPO but maybe @AnthropicAI sabotaged my GRPO experiments.... The agent labelled its own training data. All training /evals done via @huggingface Jobs. 🤗 huggingface.co/small-models-… Blog post soon!