NVlabs /Sana

SANA & Sol: Efficient Generative Models and Full-Stack Training & Inference Engines for Image, Video, and World Modeling

★ 9,837⑂ 807👁 96⑃ 273● 309▤ 4
PythonApache-2.0Created 2024-10-11nvlabs.github.io/Sana/docs/
下载:SANA-Video && SANA-WMSANA-Video-2.0.0 ↓2.0.02026-06-09
+11
日增长
9,837
stars
807
forks
96
watchers
273
prs
309
issues
4
releases

README

📚 Docs | SANA | SANA-1.5 | SANA-Sprint | SANA-Video | SANA-Video 2.0 | SANA-WM | SANA-Streaming | Sol-RL

Demo | 🎬 SANA-Video 2.0 Demo | 🤗 HuggingFace | ComfyUI | SGLang | Cosmos-RL




    
    
    
    
    
    



    
    
    
    



  
    
  


ICLR 2025 Oral | ICML 2025 | ICCV 2025 Highlight | ICLR 2026 Oral 

SANA is an efficiency-oriented codebase for high-resolution image and video generation, providing complete training and inference pipelines. This repository contains code for SANA, SANA-1.5, SANA-Sprint, SANA-Video, SANA-Video 2.0, SANA-WM, SANA-Streaming, and Sol-RL. More details can be found in our 📚 documentation.

Join our Discord to engage in discussions with the community! If you have any questions, run into issues, or are interested in contributing, don't hesitate to reach out!


  


 News

- 🔥 [2026/10] 🎮 StreamRender-H3 is released! A code-driven streaming render runtime connecting a controllable Three.js game engine to a multi-GPU H3 streaming renderer with browser video feedback. See Project | Code.
- 🔥 [2026/10] ⚡ Sol-H3 brings MiniMax-H3 to a single RTX 5090 or DGX Spark, now also in ComfyUI: one 1344×768 clip with audio in 35.5 s on one RTX 5090 (989.0 s for the original 50-step model). See Project | Code.
- 🔥 [2026/09] 🔍 SoL-Refiner is released! One-step refinement turns low-resolution video drafts into high-resolution video, with code and models for LTX-2.3 and MiniMax-H3. See Project | Code.
- 🔥 [2026/09] ⚡ SANA-Video 2.0 5B 4-Step Preview is released! The DMD preview generates 720p videos in four denoising steps and supports 5-second and 8-second outputs. See Online Demo | 4-Step Weights | Doc | Project.
- 🔥 [2026/08] 🎬 SANA-Video 2.0 training, inference, model architecture, and 5B 720p checkpoint are released! The 8-second model supports both text-to-video and text-image-to-video generation, with hybrid linear/softmax attention and Attention Residuals. See Online Demo | Project | Doc | Model Zoo | Weights.
- 🔥 [2026/08] ⚡ Sol Engine: Day-One MiniMax-H3 Acceleration is available! The 33B omni-modal audio+video DiT runs 3.95× faster on GB200, reached in 4.5 hours of optimization, and up to 4.52× on hardware that sits on a desk — 3.92× on DGX Spark, 4.52× on GeForce RTX 5090 — with no distillation, no LoRA, and no calibration pass. See GB200 Blog | On-Device Blog.
- 🔥 [2026/07] 🌍 SANA-Streaming training is released! Includes bidirectional and distillation training. See Doc.
- 🔥 [2026/07] 🌍 SANA-WM Stage-1 training is released! Includes bidirectional, chunk-causal, and distillation training. See Doc.
- 🔥 [2026/06] 🎬 SANA-Streaming: 2B Model for Real-time Streaming Editing is released! Supports 720p, 1-min video editing. A pioneer work for streaming editing. See Project | Doc | Paper | Reactor Demo.
- 🔥 [2026/05] 🌍 SANA-WM: 2.6B Controllable World Model is released! Supports 720p, 1-min video generation with 6-DoF camera control. A new baseline for World Modeling an

Top issues

Latest release

SANA-Video && SANA-WM v2.0.0 2026-06-09
What's Changed
 fix bugs and update inference examples; by @lawrence-cj in https://github.com/NVlabs/Sana/pull/217
 [fix bug] debug mode during training by @lawrence-cj in https://github.com/NVlabs/Sana/pull/220
 Adding svdquant required model type conversion script by @lawrence-cj in https://github.com/NVlabs/Sana/pull/222
 Fix args.image_size by @yujincheng08 in https://github.com/NVlabs/Sana/pull/223
 fix dead lock during training by @lawrence-cj in https://github.com/NVlabs/Sana/pull/224
 fix dataloader deadlock bug by @lawrence-cj in https://github.com/NVlabs/Sana/pull/225
 [ctrlne

Snapshot: 2026-11-16T09:23:14Z · Data license Apache-2.0 · JSON