Papers, modelos e datasets em alta no Hugging Face, além do blog oficial — com leitura editorial em português.
A large-scale multimodal tool-use dataset and agent are presented for complex image creation workflows that orchestrate multiple visual tools through multi-turn interactions and hy…
HunyuanOCR-1.5 is a lightweight end-to-end vision-language model that enhances OCR capabilities through improved efficiency via DFlash and enhanced capability via Agentic Data Flow…
Pesquisadores coreanos identificaram que a maior causa de erro em sistemas de resposta visual médica não é a falta de conhecimento — é um primeiro passo de raciocínio errado que contamina tudo depois. Um novo método de treino, o MRPO, penaliza esses erros iniciais com mais peso e reduziu a taxa de falhas precoces de 64% para 13%.
Bytedance's AI video tool Seedance is dividing Hollywood. A viral clip featuring AI-generated Brad Pitt and Tom Cruise prompted the Motion Picture Association's first-ever cease-and-desist against an AI company. But behind the scenes, studios are quietly using the tool on a "don't ask, don't tell" basis, says Simpsons animation producer Joel Kuwahara. The article Hollywood wants Seedance banned and reportedly also wants to keep using it appeared first on The Decoder .
A multimodal sexism identification system for memes uses hierarchical conditional soft-label prediction with vision-language embeddings and a lightweight Gated MLP trained via KL d…
ResearchStudio-Reel automates research dissemination by composing specialized skills around a shared paper extractor, generating consistent and editable artifacts including posters…
Modelo de visão e linguagem · 4 B de parâmetros — 753 downloads e 48 curtidas no Hugging Face.
Multimodal large language models exhibit distinct attention patterns during generation, with attention to visual and textual modalities shifting based on semantic requirements, and…
arXiv:2607.01378v1 Tipo de Anúncio: novo Resumo: Modelos de Visão-Linguagem-Ação (VLA) têm demonstrado capacidades promissoras de generalização em tarefas de manipulação robótica, mas sua implantação no mundo real permanece limitada pela falta de medidas de segurança eficazes. Especificamente, as medidas de segurança existentes apenas evitam colisões causadas pela próxima ação do robô. Neste artigo, propomos um mecanismo de orientação de segurança neuro-simbólico para VLAs baseados em correspondência de fluxo que possibilita a prevenção preditiva de colisões...
arXiv:2607.01883v1 Announce Type: new Abstract: Code is the medium through which large language models generate structured artifacts: charts, scientific figures, vector graphics, CAD models, 3D scenes, and hardware designs are all produced by writing programs. In this regime single pass inference is brittle, because the compiler, renderer, or simulator that decides whether the artifact exists is invisible to the model. We present PairCoder, which grounds review in the toolchain and realizes it a...
arXiv:2607.02092v1 Announce Type: new Abstract: Flow-matching vision-language-action policies generate robot action chunks through an iterative transport process, creating an opportunity for test-time guidance without retraining the base policy. We study this opportunity in Guided Action Flow, an inference-time framework that keeps a pretrained SmolVLA policy frozen and uses a learned action-chunk critic to guide its reverse-time flow sampler. The critic is trained from real success and failure ...
arXiv:2607.02195v1 Announce Type: new Abstract: General-purpose vision-language-action models benefit from large vision-language priors, but effective manipulation also requires anticipating action-relevant scene changes. Existing world-action models often rely on large generative world models or dense future rollouts, which are expensive and spend capacity on visual details weakly coupled to control. We present Bridge-WA, a lightweight world-action framework that distills a frozen future-change...