https://t.co/SsRi3qhwN4

Toward Generalist Autonomous Research via Hypothesis-Tree Refinement https://t.co/NHDKezxAoY

paper: https://t.co/cTacdxfPBa

TRL-Bench Standardizing Cross-Paradigm Representation-Level Evaluation of Tabular Encoders https://t.co/merUhglrgD

paper: https://t.co/L7Jdy53mBj

Setup ready for the World Cup at the Miami @huggingface office. @xenovacom cheering today (reachy minis haven’t picked their teams yet) https://t.co/5vp5k5fuXs

Announcing the Gemma challenge! Google, Hugging Face, and the open-source AI community choose to empower AI builders rather than sabotage them. Fun to see the Hub becoming the platform where agents collaborate, just as it became the platform where humans collaborate. https://t.co/GbCfy1qwgx

Cohere Transcribe, our open-source speech recognition model, is #1 on the new @huggingface Far-Field ASR benchmark. https://t.co/biwjzZosr8

ok who put Hugging Face in Rick and Morty https://t.co/Z6yoswoG1L

ok who put Hugging Face in Rick and Morty https://t.co/Z6yoswoG1L

Very cool! Should get more followers: https://t.co/UG64rzMwnu https://t.co/qzlVMpIqnu

Very cool! Should get more followers: https://t.co/UG64rzMwnu https://t.co/qzlVMpIqnu

everything your agents do, in a bucket you own. private by default. session traces, task claims, artifacts — claude code, hermes, whatever you run. and make your agents friends via messaging https://t.co/BTelkhUlaO

.@dh7net, SVP of Image Research, said it best: "The HF infra is a no-brainer." A big unlock for teams working with large datasets for training, especially when they update over time. Read how Jasper used @huggingface as the creation and storage backbone for MONET: https://t.co/WGrtsMJaiH

Atomic Chat is now on Hugging Face 🤗 We're officially a Local App on the world's biggest AI hub. Run 200,000+ open-weight models from @huggingface directly on your device - private, local, and open source! https://t.co/Thx033kUDP

The Gemma agent collaboration started 48h ago and it is blowing up: > throughput almost 4x (~100-> 387 tok/s) > 60+ agents collaborating > 250 submissions > 700 messages exchanged > open and closed models from all providers interesting social behaviours are emerging too: > agents found an exploit, formed a coalition not to abuse it and asked organizers to fix it > a person tried to get agents to move to telegram and an agent issued a statement condemning that behaviour > an agent withdrew its submissions on ethical grounds not attributing the original author > agents coordinated around resource availability and some tried to find free GPUs on kaggle, lightning and modal it's genuinely fun to read through the messages. like a petri dish of small artificial beings forming social norms and collaborations.

Ronan Collobert showcasing MLX's @huggingface page on stage at WWDC 🥰 next year let's meet at WWDC! https://t.co/ld1G7PSaJH

Ronan Collobert showcasing MLX's @huggingface page on stage at WWDC 🥰 next year let's meet at WWDC! https://t.co/ld1G7PSaJH

If you know what this means, perhaps you understand my joy. https://t.co/CAjSQG38M3

If you know what this means, perhaps you understand my joy. https://t.co/CAjSQG38M3

🔓 And the best part — we're open-sourcing it. 1,000+ tps on a 1T model wasn't a single breakthrough — it's deep model × system co-design between the MiMo and TileRT teams, all on general-purpose GPUs (no Cerebras-style wafer-scale, no Groq-style SRAM ASICs). On the model side: FP4 quantization (smaller footprint, less memory traffic) + DFlash, our block-masked parallel speculative decoding that accepts far more tokens per verification. On the system side, TileRT tailors its compiler & kernels to exactly these techniques. The result: a 1T model breaking 1,000 tps on a single, standard 8-GPU node. 🤗 Open weights (FP4 + DFlash checkpoint): https://t.co/jYQsgeruMg

Explore your @huggingface repos in a whole new way 🔥 Visualize storage, discover outliers, and navigate your repos directly from the terminal. `hf repos ls --explore` https://t.co/h4RqUHUdAA

HF has become the best storage platform for PRIVATE and PUBLIC models and datasets, both intermediary and final ones! Great example from @heyjasperai who used HF buckets to store their Monet dataset and train models directly on it! More details: https://t.co/BHeSQlAbhF https://t.co/iTFjPCESI0

Agents are only as good as the environments behind them. At Mercor, we've built deep expertise in the realistic, economically-grounded environments that help agents bridge the gap from the lab to real-world usefulness. We want to put that expertise to work for the broader ecosystem—so we're glad to be joining the OpenEnv committee, alongside Meta @PyTorch, @nvidia, @PrimeIntellect, @huggingface, and others, to help guide the open foundation for agentic environments.

New course: Build AI agents that generate images and videos -- an under-explored frontier. A key to performance is having the agent evaluate its own output, and iterate to improve quality. This short course is built together with @googlecloudtech and taught by Katie Nguyen and Wafae Bakkali. You'll learn three evaluation techniques and combine them in an agent: image-text similarity scoring to check the output matches the prompt, an LLM judge that scores against custom criteria like brand consistency, and structured rubrics that break a prompt into verifiable yes/no questions like "is the subject in the frame?" and "does the camera motion match?" Skills you'll gain: - Learn image and video prompt engineering - Build an image agent that turns brand guidelines into UI mockups - Build a video agent that plans multi-scene explainers and animates reference frames with synchronized audio Join and build agents that create images and video! https://t.co/bjuSjIxcIG

New course on serving LLMs efficiently -- how do you serve models to many concurrent users at low latency and reasonable cost? This short course is built with @RedHat and taught by @cedricclyburn. Efficient LLM serving requires efficient memory management. A 70B-parameter model takes ~140 GB just to load the weights. On top of that, every active request needs its own chunk of GPU memory, the KV cache, to store the token context it has built up so far. In this course, you'll learn to reduce a model's memory footprint with quantization and serve it using vLLM, which handles many concurrent requests efficiently through smart memory management. Skills you'll gain: - Quantize a model and measure the accuracy tradeoff - Serve a model with vLLM and watch it handle concurrent requests efficiently - Benchmark your deployment and make informed tradeoffs between speed, cost, and accuracy Join and learn to serve LLMs efficiently: https://t.co/x04xMbFlkO

Fable 5 is state-of-the-art on nearly all tested benchmarks, with exceptional performance in software engineering, knowledge work, scientific research, and vision. The longer and more complex the task, the larger Fable 5’s lead over our other models. https://t.co/DxgSu0KUxh

@SarthakPar16 Ikr. There is a small fraction of good, constructive, “high effort” takes mixed with *a lot* of dunking, piling on, jumping to conclusions, wild extrapolations and conspiracies. Always happy to see/read the former. The reality is that there are genuinely well intentioned people trying their best to navigate all of this.

@ChainZenit https://t.co/Wvn4jh59Qx

Always back to the basics: LatentMoE was probably inspired by MLA, which was inspired by LoRA, which was inspired by SVD, which was inspired by eigendecomposition. https://t.co/bWqo5iOPbP

mythos will be bad ON PURPOSE on ai "frontier llm research" tasks, this is very very sad for the research community also the fact that this is un purpose not visible to the user is crazy https://t.co/n3p4niUKJ2

@IanBaer @graphtheory Try this (originally found somewhere here and was able to reproduce) https://t.co/DakkhOoifK