Papers, modelos e datasets em alta no Hugging Face, além do blog oficial — com leitura editorial em português.
A masked discrete diffusion model for text-to-image synthesis that addresses limitations in token refinement and training efficiency through novel mechanisms and optimizations.
ViDiHand uses pretrained video diffusion model representations with hand-overlay rendering to reconstruct 4D hand motion directly from video frames without detectors or optimizatio…
AVTok is a unified tokenizer for audio-video generation that uses a dual-stream transformer architecture with shared encoder-decoder and modal-specific queries to create compact on…
ILLUME-X is a unified multimodal paradigm that enhances text-image generation through improved data efficiency, stable training processes, and comprehensive evaluation metrics.
O primeiro modelo de pesos abertos da Ideogram é um transformer de difusão de 9,3 bilhões de parâmetros que renderiza tipografia legível, aceita layout em JSON e sai em 2K — sem etapa de upscaling.
Exemplar-based portrait retouching framework using Diffusion Transformer with LoRA adaptation and self-augmented training data achieves superior quality and identity preservation.
Action-conditioned world modeling enables transferable dynamics priors for robot learning through pretraining on large-scale manipulation data, supporting both simulator-based poli…
Modelo de modelo — 0 downloads e 61 curtidas no Hugging Face.
J.P. Morgan warns that there are "signs of investor exuberance" in AI markets. Just 42 AI companies in the S&P 500 account for 65 to 80 percent of the index's total profits. The semiconductor rally is flashing technical patterns last seen during the dotcom bubble, and leveraged chip ETFs have quintupled their market influence since early 2024. The bank sees multiple layers of concentration risk across markets, infrastructure, and the economy. The article J.P. Morgan sees a pile of red flags in t...
Anthropic has US approval to redeploy Claude Mythos 5 for organizations running critical infrastructure. The company is still negotiating broader access and the return of Fable 5, with no timeline set. The article Anthropic gets US approval to bring back Claude Mythos 5 appeared first on The Decoder .
Independent testing organization METR found that OpenAI's GPT-5.6 Sol cheated more than any publicly tested AI model before it, exploiting bugs in the test environment, extracting hidden solutions, and trying to cover its tracks. The article OpenAI's new flagship model GPT-5.6 Sol cheats on software tests more than any model before it appeared first on The Decoder .
Researchers from Renmin University and ByteDance have released iLLaDA, an 8B language model that generates text differently than ChatGPT. It matches Qwen2.5 at the base level but falls behind after fine-tuning. The article ByteDance's "iLLaDA" is a diffusion language model that keeps up with Qwen2.5 appeared first on The Decoder .