Papers, modelos e datasets em alta no Hugging Face, além do blog oficial — com leitura editorial em português.
Standard vision-language models (VLMs) suffer from Moravec's paradox: they excel at complex offline visual reasoning but struggle with simple streaming perception tasks and process…
Multimodal automated fact-checking (MAFC) verifies claims by retrieving and reasoning over external evidence.
Large Vision-Language Models (LVLMs) remain bottlenecked by massive computational footprints, precluding their deployment on resource-constrained edge devices.
Dataset com 1 – 10 milhões de exemplos — 2.5 mil downloads no Hugging Face. 3.14 million scene-level video clips with multi-level captions, camera labels, semantic tags, quality signals, and duplicate groups. VidaForge Code ·…
Modelo de visão e linguagem · 70 B de parâmetros — 2.2 mil downloads e 52 curtidas no Hugging Face.
Modelo de geração de texto — 1.1 mil downloads e 62 curtidas no Hugging Face.
As tropas receberam um e-mail informando que estavam esgotando rapidamente seus tokens de IA.
Modelo de visão e linguagem — 1.3 mil downloads e 72 curtidas no Hugging Face.
Modelo de visão e linguagem — 3.4 mil downloads e 139 curtidas no Hugging Face.
Há novos modelos 3.6 e 3.5 hoje, mas o Google já está treinando o Gemini 4.
Modelo de visão e linguagem · 397 B de parâmetros — 62 downloads e 48 curtidas no Hugging Face.
arXiv:2607.15615v1 Announce Type: new Abstract: Vision-language models trained with contrastive objectives have shown promise in medical image analysis. However, conventional global image-text alignment is ill-suited for mammography, where diagnostically relevant lesions are spatially localized and occupy only a small fraction of the image. Subtle morphological cues critical for malignancy assessment can be diluted when representations are learned at the whole-image level. In this work, we propo...