Skip to content
Agientry
ZH

BLOG

Blog

WaterfallPagination
Open-H3-Regenerate-2K: High-Resolution Video Refinement with Few-Step Diffusion
FEATURED

Open-H3-Regenerate-2K: High-Resolution Video Refinement with Few-Step Diffusion

Open-H3-Regenerate-2K freezes H3 and trains only a refiner LoRA, combining latent upscaling with three diffusion steps to redraw drafts natively from 768P to 4K.

Video DiffusionFew-Step GenerationVideo Super-Resolution
Sensen Gao, Haozhe Liu, Yitong Li, Haopeng Li, Tian Ye, Jia-Wang Bian, Song Han, Enze Xie· 2026-10-28T00:00:00
All
Embodied AI17Humanoid11VLA11AI Agent10Reinforcement Learning10World Models8Agent Harness7Claude Code7NVIDIA7LLM Agent5MCP5OpenAI5vision-language-action5Boston Dynamics4Codex4Dexterous Manipulation4Game Development4Google DeepMind4LLM4System One Models4
AW-0: Planning With Code Inside a Joint World Action Model

AW-0: Planning With Code Inside a Joint World Action Model

AW-0 jointly denoises a generated skill program, future video and action chunk. Code guides control without being executed by default, while a gated hybrid mode can run verified programs.

World-action modelsEmbodied AI
Zhuoyang Liu, Dantong Niu, Botong Zhang et al. · 358 views · Oct 26, 2026
Astranomy: An Empirical Study of GPT-6 Astra as a Native Robot Policy

Astranomy: An Empirical Study of GPT-6 Astra as a Native Robot Policy

NVIDIA studies GPT-6 Astra as a native robot policy: Agentic Policy tops RoboLab-120 at 74.8%, but each call averages 18.56 s and a successful episode costs 30.9M tokens ($36.40).

Robot policiesEmbodied AI
Yang, Khetan & Zhang (NVIDIA SRL) · 1.1k views · Oct 14, 2026
Inside NVIDIA Isaac Sim Agent Skills: 38 Robot-Simulation Workflows Encoded for AI Coding Agents

Inside NVIDIA Isaac Sim Agent Skills: 38 Robot-Simulation Workflows Encoded for AI Coding Agents

NVIDIA ships 38 Isaac Sim Agent Skills with the distribution: procedures, handoff contracts and acceptance thresholds in SKILL.md teach coding agents to finish and validate robot sims like engineers.

Isaac SimAgent Skills
具身Agent · 993 views · Oct 9, 2026
Fine-tune, Convert and Deploy Language Models on NVIDIA Jetson with Unsloth and llama.cpp

Fine-tune, Convert and Deploy Language Models on NVIDIA Jetson with Unsloth and llama.cpp

Official Jetson AI Lab tutorial: fine-tune models directly on Jetson with Unsloth and JetPack 7.2 using memory-efficient QLoRA, export to GGUF, and run locally with llama.cpp. Two hands-on examples: Qwen3.5-4B vision-language model on Jetson Orin Nano (LaTeX OCR fine-tuning), and NVIDIA Nemotron 3.5 Lightning 30B-A3B on Jetson AGX Thor (3 training steps in 44.8s, 66.1 tokens/sec with Q4_K_M quantization). No cloud required.

NVIDIAJetson
Jetson AI Lab · 942 views · Sep 1, 2026
How to Train a Cross-Embodiment Robot Navigation Policy with AI Agents

How to Train a Cross-Embodiment Robot Navigation Policy with AI Agents

NVIDIA COMPASS framework enables cross-embodiment navigation via residual RL and skill synthesis. A coding agent automates environment validation, scene preparation, smoke testing, residual specialist training, and checkpoint evaluation with human approval gates. Uses Spot quadruped as reference across built-in warehouse, SAGE-10K, and NuRec captured environments.

NVIDIACOMPASS
NVIDIA · 669 views · Aug 28, 2026
ADEPT: Pre-Train Dexterity Once, Post-Train Every Task

ADEPT: Pre-Train Dexterity Once, Post-Train Every Task

ADEPT pre-trains one dexterous policy on 16 primitives, post-trains specialists without forgetting, and deploys zero-shot on Kuka-Allegro and Flexiv-Sharpa, 2-14x faster than gripper pipelines.

Sim-to-RealDexterous Manipulation
Jayjun Lee 等（NVIDIA / 密歇根大学） · 601 views · Aug 19, 2026

As an Amazon Associate, we earn from qualifying purchases.