Tendance des mentions — 20 derniers numéros

Actus en direct (27)

VideoChat3 Released as Fully Open 4B Video MLLM With 16x Visual-Token Compression, 70.1 on Video-MME

huggingface.co · 2026-07-16 · Score : 20

ALERTE Hugging Face Discloses AI-Agent-Driven Intrusion — Malicious Dataset Exploited RCE Loader and Template Injection to Harvest Cluster Credentials

huggingface.co · 2026-07-16 · Score : 33

OvisOCR2 Technical Report and Open-Weights Model Release — ATH-MaaS Publishes Multilingual OCR Foundation Model on Hugging Face at 42 Paper Upvotes and 134 Model Likes, Positioning It as Open Alternative to Commercial OCR APIs for Document AI Pipelines

huggingface.co · 2026-07-15 · Score : 16

Boogu-Image-0.1 Open-Source Unified Multimodal Model Released — Apache 2.0 in Four Variants (Base, Turbo, Edit, Edit-Turbo) Trained on 208.62M Images for ~$400K, Claims Parity With Nano-Banana-Pro and GPT-Image-2 on Bilingual Chinese-English Text Rendering

huggingface.co · 2026-07-16 · Score : 16

Paris Startup ZML Launches Chip-Agnostic LLM Inference Server LLMD With $20M Backed by LeCun, Docker's Hykes, and HF Founders — Runs Open Models on Nvidia, AMD, Google TPU, Apple Metal, and Intel Arc

techcrunch.com · 2026-07-08 · Score : 25

HF Paper 'GaP': Graph-as-Policy Multi-Agent Self-Learning Harness for Variational Automation Tasks

huggingface.co · 2026-07-06 · Score : 18

Paris Robotics Startup UMA Emerges From Stealth With 'Northstar' Humanoid — Founded by Ex-Tesla Optimus Scientist Rémi Cadene, Ex-DeepMind's Pierre Sermanet, Already in Talks With 50 Customers

bloomberg.com · 2026-07-07 · Score : 20

Community Merge 'Qwopus3.6-35B-A3B-Coder' Lands on Hugging Face — Thinking-Off MoE Claims 62.4% on SWE-Bench for Local Coding Agents

huggingface.co · 2026-07-02 · Score : 16

National Design Studio Releases Rampart — 14.7MB On-Device ONNX PII Redactor That Runs In-Browser Before Text Hits Hosted LLMs

huggingface.co · 2026-07-02 · Score : 17

HF Paper 'From SRA to Self-Flow': Reframes Data Augmentation as a Weak Form of Self-Supervision

huggingface.co · 2026-07-02 · Score : 15

HF Paper 'AGVBench': Reliability-Oriented Benchmark of Data Augmentation for Vein Recognition

huggingface.co · 2026-07-02 · Score : 15

HF Paper: 'Representation Distribution Matching' Cuts One-Step Visual Generation Gap With Multi-Step Diffusion

huggingface.co · 2026-07-02 · Score : 15

HF Paper 'PACE': Lightweight Proxy Metric for Agentic Capability Evaluation Aims to Replace Expensive Rollout-Based Benchmarks

huggingface.co · 2026-07-02 · Score : 15

HF Paper 'SkillCoach': Self-Evolving Rubrics Automatically Grade and Improve Agentic Skill Use During Training

huggingface.co · 2026-07-02 · Score : 16

Hugging Face + Cerebras Ship Open Cascaded Speech-to-Speech Stack: Nvidia Parakeet → Gemma 4 31B VLM on Cerebras → Alibaba Qwen3-TTS, Already Powering 9,000+ Reachy Mini Robots

huggingface.co · 2026-07-01 · Score : 18

Hugging Face Paper 'Beyond IID': New 142-Dataset BeyondArena Benchmark Exposes Tabular Foundation Models Still Lose to Tree-Based and Deep Learning on Non-IID, Large, and High-Dimensional Data

huggingface.co · 2026-06-29 · Score : 20

VibeThinker-3B: WeiboAI's MIT-Licensed 3B Model Claims 94.3 on AIME 2026, Matching DeepSeek V3.2 at 1/200th the Parameters

MarkTechPost · 2026-06-19 · Score : 21

OpenEnv Transitions to Multi-Organization Governance With Meta-PyTorch, Nvidia, Hugging Face, Scale AI, and a Dozen Others

huggingface.co · 2026-06-08 · Score : 20

Ideogram 4 Releases as Open-Weights Text-to-Image Model — 9.3B-Parameter DiT, #1 on Design Arena, Non-Commercial License

huggingface.co · 2026-06-03 · Score : 23

Liquid AI Releases LFM2.5-8B-A1B on HuggingFace — New 8B MoE Entry Extends the LFM2.5 On-Device Family

huggingface.co · 2026-05-28 · Score : 17

Alibaba Releases Qwen-Image-Bench — Apache-Licensed 27B Vision Judge for Text-to-Image Evaluation

huggingface.co · 2026-05-27 · Score : 20

SenseNova Releases U1-A3B-MoT: 3B Mixture-of-Experts Model With Mixture-of-Thoughts Reasoning

huggingface.co · 2026-05-13 · Score : 16

White Circle Raises $11M Seed to Monitor and Control Enterprise AI Models in Real Time, Backed by Executives From OpenAI, Anthropic, Mistral, and Hugging Face

fortune.com · 2026-05-12 · Score : 20

Fake OpenAI 'Privacy Filter' Repository Reaches #1 Trending on Hugging Face, Delivers Rust-Based Infostealer to 244,000 Downloaders Before Removal

bleepingcomputer.com · 2026-05-09 · Score : 24

ALERTE Coordinated AI Supply Chain Attack Plants 575+ Malicious Skills Across Hugging Face and ClawHub, Hijacking AI Agents for Credential Theft and Crypto Mining

thenextweb.com · 2026-05-09 · Score : 29

TIME names its 10 Most Influential AI Companies of 2026 — ByteDance tops list, three Chinese firms in top 10

TIME · 2026-04-27 · Score : 21

Critical unpatched Hugging Face LeRobot RCE vulnerability disclosed (CVE-2026-25874, CVSS 9.8)

The Hacker News · 2026-04-28 · Score : 25

Numéros de la newsletter (5)

Deep Learning: a 744B open model just lapped the labs

2026-06-21 · 2 mentions

AI in Business: Anthropic at $900B as Meta funds AI by firing 8,000

2026-05-26 · 1 mentions

Machine Learning: a reasoning model just cracked Erdős

2026-05-26 · 2 mentions

Robotics: 100,000 packages, 81 hours, one livestream

2026-05-21 · 1 mentions

Machine Learning: nobody trained bigger, everyone shrank the cache

2026-05-21 · 1 mentions