NVIDIA Releases NLD-Uni: Unified Diffusion LLM for Visual Understanding and Generation
nvidia · hf · 2026-10-28
NVIDIA introduces Nemotron-Labs-Diffusion-Uni (NLD-Uni), a unified discrete diffusion LLM handling both visual understanding and generation, with two key innovations:
- Cross-paradigm on-policy distillation transferring autoregressive VLM understanding and continuous flow model generation abilities into a unified DLLM
- Progressive self-refining diffusion that gradually corrupts tokens across multiple steps during training, enabling intermediate token commitment and progressive refinement at inference
NLD-Uni achieves SOTA among unified DLLMs on VQA, multimodal reasoning, grounding, text-to-image generation and image editing. Model and code are publicly available.
More from Multimodal
- Hands-on with Baidu's Unlimited OCR 2: whole docs in one pass, no page limit — Baidu_Inc · 2026-10-28
- Agentic video editing goes pro: senior editors 5x output, costs to collapse — dotey · 2026-10-28
- Alibaba's OvisOCR3: a 0.8B model tops OmniDocBench at 96.74 via multi-teacher distillation + RL — ATH-MaaS · 2026-10-28
- CodeCap-Bench: executable code beats natural language as video captions — Jiafu Tang · 2026-10-28
- Moving Atoms Engine turns real-world spaces and cities into high-resolution 4D models for sim — Scobleizer · 2026-10-28
- QuiverAI Launches Arrow 2 Vector Model on FLORA with Layered SVG Editing — andpoul · 2026-10-28