One of the hardest challenges in AI safety is finding the right balance: how do we protect people from harm without undermining their agency? This tension is especially visible in conversational systems, where safeguards can sometimes feel more paternalistic than supportive.

The @hf.co community is awesome. Real work that moves everyone forward: huggingface.co/blog/rteb

🐋🐦🐸 We just launched an interactive demo of NatureLM-audio on @hf.co! 👉 Try the demo with your audio or ours, share your feedback, and help us shape the future of decoding animal communication: huggingface.co/blog/EarthSp...

You only need one line of code to start exploring the new @OpenAI models! gr.load("models/openai/gpt-oss-120b", provider="fireworks-ai").launch()

Well, it took just 2 hours for GPT-OSS to hit #1 on @hf.co. Don’t remember seeing anything rise that fast!

OpenAI have released their new open source models! One thing I really like about this release is that while they are only open weight, the model is not gated in any way (anyone can download it) and it has a permissive OSS license (apache 2). Very refreshing. huggingface.co/openai/gpt-o...

You can now generate synthetic data using OpenAIs GPT OSS models on @hf.co Jobs! One command, no setup: hf jobs uv run --flavor l4x4 [script-url] \ --input-dataset your/dataset \ --output-dataset your/output Works on L4 GPUs ⚡ huggingface.co/datasets/uv-...

Many VLM-based OCR models have been released recently. Are they useful for libraries and archives? I made a quick Space to compare VLM OCR with "traditional" OCR using 11k Scottish exam papers from @natlibscot.bsky.social huggingface.co/spaces/davanstrien/ocr-time-capsule

We just crossed 1 million repositories backed by Xet storage on @hf.co I celebrated by reviving the early 2000s web design aesthetics that I love so much. Here's our dashboard showing our progress converting the Hub from Git LFS to Xet (and demonstrating my questionable design sensibilities).

Built my 1st app exclusively using @hf.co Hub features! It helps me keep track/summarize the latest HF News. Uses Datasets, Inference Endpoints, the newly announced `hf jobs`, and Spaces to visualize results. Check it out here: huggingface.co/spaces/rajat...

`hf jobs` looks super interesting. Send off workloads easily to remote infra. And with inline metadata uv scripts `hf jobs uv run my_script.py` all dependencies can be defined in a single file. So simple and so useful. huggingface.co/docs/hugging...

Working to port VLaMy to an entirely free mode where you can just cache all your data in the browser for a project. Slowly adding all the features from the full version to this user-free version. Available now on @hf.co @danielvanstrien.bsky.social Link: huggingface.co/spaces/wjbma...

465 people. 122 languages. 58,185 annotations! FineWeb-C v1 is complete! Communities worldwide have built their own educational quality datasets, proving that we don't need to wait for big tech to support languages. Huge thanks to all who contributed! huggingface.co/blog/davanst...

Last week, we concluded the @gradio-hf.bsky.social‬ MCP hackathon with @hf.co‬. The project that one the @llamaindex.bsky.social prize was the "Nasa Space Explorer" 🔭🪐 3 servers that provide live data on: ☄️ Asteroids 🤖 the Mars Rover 🌌 Astronomy Here's the space: huggingface.co/spaces/Agen...

i love the simplicity of LeRobot from @hf.co to interact with robots, especially for beginners like me there is one very huge problem: it's written in python, but i love js introducing: LeRobot.js interact with your robot directly in the browser www.youtube.com/watch?v=H1iU...

🧬 super psyched to announce a new collaboration between @hf.co and Ginkgo Datapoints to open up high-quality biological datasets for the machine learning community! Just dropped the GDPx and GDPa dataset series on the Hub (x1000 boost to AI for drug development) 🔗 huggingface.co/ginkgo-datap...

If you are interested in a unified collection of common misinformation detection benchmarks, check out our recent repo @hf.co

Have high quality data sitting on Transkribus? Want to make it available on @hf.co with a single command line? Introducing Transkribus-HF which allows you to take a Transkribus export zip and make it into a HF dataset! It can parse pages, regions, lines, or windows! github.com/wjbmattingly...

Everyone’s dropping VLM-based OCR models lately… But are they actually better than traditional OCR engines, which output XML for historical docs? I built OCR Time Machine to test it! 📄 Upload image + ALTO/PAGE XML ⚖️ Compare outputs side by side 🔗 huggingface.co/spaces/davan...

MiniMax-M1 🔥 The first reasoning model by MiniMax AI is now live on @hf.co huggingface.co/collections/... ✨ 40k/80k thinking budget ✨ Powered by Hybrid MoE + Lightning Attention 👀 ✨ 1M context length 🤯 ✨ Apache 2.0 ✨ RL-trained for math, coding & real-world software

Over the past few months, we have worked on the @hf.co Kernel Hub. Kernel Hub allows you to get cutting-edge compute kernels directly from the hub in a few lines of code. David Holz made a great writeup of how you can use kernels in your projects: huggingface.co/blog/hello-h...

“AI Scraping Bots Are Breaking Open Libraries, Archives, and Museums” – interesting piece via @404media.co Not a perfect fix, but making ML-ready datasets from collections can help. If you want help getting your data on @hf.co, I'd be happy to help.

Institutional Books: Massive Historical Text Corpus - 983K books, 242B tokens, 386M pages - 19th-20th century texts in 254 languages - Refined OCR with quality scores & metadata - Noncommercial early-access release huggingface.co/datasets/ins...

How to Make Gallery, Library, Archive & Museum Collections Ready for AI 📚✨ Join me this Tuesday – I'm hosting a hands-on session for the #ai4lam community, where I'll show cultural institutions how to share data on Hub. 🗓️ June 17, 16:00 UK 📍 Details: docs.google.com/document/d/1...

huggingface.co/Belykh-Lab I’m trying to start a trend in the applied math community of releasing data sets on @hf.co

We look forward to growing Institutional Books through community. We welcome collaboration from researchers and model makers as we: - Evaluate the dataset’s impact on model outputs - Continuing to refine our OCR pipelines View the dataset on Hugging Face: huggingface.co/datasets/ins...

Hugging Face released their MCP server. You can see it here configured and running in Claude Desktop. @hf.co

first steps with the @hf.co SO-101 AI robot arm

Inspired by @hf.co's official MCP server, I built my own to expose my semantic search API for the HF ecosystem! Features AI-powered search, parameter analysis via safetensors, and tools to find similar models/datasets. Try: "Find non maths reasoning datasets from 2025"!

Such an important project: @hf.co put up an interactive site to see the real time energy costs of chatting with genAI. "Calculate how much water it would take to cool the world's largest supercomputer" took 13% of a smartphone battery. Complete with hallucinations. 😆 huggingface.co/spaces/jdela...

🗞️ Just released a Parquet version of the Newspaper Navigator dataset on @hf.co! - 3M+ visual elements from historic US newspapers — photos, maps, cartoons, OCR + metadata. - Parquet = fast filters, easier analysis. - Great for ML + cultural research. 👉 huggingface.co/datasets/big...

Finally documented the Beyond Words dataset from the @librarycongress.bsky.social labs / @bcgl.bsky.social for the BigLAM @hf.co org! - 3.5K annotated historical newspaper pages - Bounding boxes + category labels - Photos, ads, headlines, cartoons & more

Not sure how long it has been around but it’s a really great ML resource huggingface.co/learn by @hf.co

🚨 Just one month left to submit your solutions for The Structured Semantic 3D Reconstruction (S23DR-2025) Challenge!!! It is not too late to join! Comp on @hf.co, part of the Workshop on Urban Scene Modeling at @cvprconference.bsky.social 2025 🔥$25,000 prize pool. Deadline: June 5, 2025. 🧵 (1/7)

Ever notice how some AI assistants feel like tools while others feel like companions? Turns out, it's not always about fancy tech upgrades, because sometimes it's just clever design. huggingface.co/blog/giadap/...

Introducing DeepSeek-R1T-Chimera: our new open weights model adds #R1 reasoning to #DeepSeek V3-0324. In benchmarks, the hybrid child model appears to be as smart as R1 but uses 40% fewer output tokens. Available on @hf.co & Open Router.

New free HTR app nearly ready to share! This is a simple local app that creates a local database that lets you create projects which have document images that you can then send off to Qwen 2.5 VL and GliNER on my Caracal app on @hf.co entirely for free. You just need a free HF account/token.

Our @cardiffnlp.bsky.social team and collaborators have been releasing amazing NLP models and datasets over the last few years. We've now organised the content as @hf.co collections 🤗 A short thread summarising some of these NLP resources, in particular related to social media! 🧵

Need blazing-fast classifier inference with minimal code? ModernBERT now runs on vLLM — fast enough to process 200K+ arXiv papers in minutes. It makes running any of the 100s of ModernBERT models on the @hf.co Hub even quicker. Guide 👇 danielvanstrien.xyz/posts/2025/v...

Starting a fun new side-quest to create an open-source dataset for Voynich that contains all the variant transcriptions and encodings in a single place with some added polygons for regions and lines. The idea is to put it on @hf.co. @lisafdavis.bsky.social would this be useful? #medieval

The discussion pages for Open-R1 on @hf.co are such a goldmine for actual practical information on how to train a reasoning model. Like look at this! If you're not reading those community tabs you're missing so much! huggingface.co/spaces/open-...

JetBrains has been quietly building something special for the open-source LLM community. More details will be posted soon on Hugging Face. Stay tuned! 🧑‍💻

📢 HUGE NEWS! Project Sidewalk just crossed 1.1 MILLION manual sidewalk accessibility annotations across 21,000+ km of streets! 🎉 (1,163,413 labels to be exact!). To celebrate and help others build on our work, we have posted a sidewalk feature/barrier dataset to @hf.co huggingface.co/projectsidew...

Academic Chains: Scientific Thinking for LLMs - Distils reasoning from papers in biology & economics - Captures researcher intuition & exploration - Fine-tuning shows 7.2% improvement on MMLU-Pro Economics huggingface.co/datasets/mar...

Happy to see that @hf.co's new ChatUI Energy tool shows real-time energy use of AI chats, promoting #sustainable interactions. 🙂 Direct link: huggingface.co/spaces/jdela... #AI #Sustainability #TechNews

I just released Sentence Transformers v4.1; featuring ONNX and OpenVINO backends for rerankers offering 2-3x speedups and improved hard negatives mining which helps prepare stronger training datasets. Details in 🧵

🔵 𝐂𝐎𝐏-𝐆𝐄𝐍-𝐁𝐞𝐭𝐚: 𝐔𝐧𝐢𝐟𝐢𝐞𝐝 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐯𝐞 𝐌𝐨𝐝𝐞𝐥𝐥𝐢𝐧𝐠 𝐨𝐟 𝐂𝐎𝐏𝐞𝐫𝐧𝐢𝐜𝐮𝐬 𝐈𝐦𝐚𝐠𝐞𝐫𝐲 𝐓𝐡𝐮𝐦𝐛𝐧𝐚𝐢𝐥𝐬 Now on @hf.co ⚖️ 𝐌𝐨𝐝𝐞𝐥 huggingface.co/mespinosami/... 📱 𝐃𝐞𝐦𝐨 huggingface.co/spaces/mikon... 📄 𝐏𝐚𝐩𝐞𝐫 arxiv.org/abs/2504.08548

Xet infra now backs 1000s of repos on @hf.co , which means we get to put on our researcher hats and peer into the bytes 👀 🤓 Xet clients chunk files (~64KB) and skip uploads of duplicate content, but what if those chunks are already in _another_ repo? We skip those too.

Reasoning datasets dominate @hf.co trending datasets, but focus mainly on code/math. We launched a competition with Bespoke Labs and & Together AI to diversify this. Create a PoC reasoning dataset and win prizes to help scale it!

Dolphin 🐬 an open ASR model released by DataOceanAI , one of the biggest AI data provider in China 🔥 Model: huggingface.co/DataoceanAI/... huggingface.co/DataoceanAI/... Paper: huggingface.co/papers/2503....