Latte
Latent diffusion transformer for video generation with spatial-temporal attention.
About
Latte is a latent diffusion transformer for video generation that factorizes spatial and temporal modeling into dedicated attention blocks, reaching competitive quality with efficient training. The repository provides PyTorch model definitions, pretrained weights, and training, sampling, and evaluation code from the paper. Developed by Monash University and Shanghai AI Lab. Released under the Apache 2.0 license.
Reviews (0)
Leave a Review
No reviews yet. Be the first to review!
Details
- Category
- Video Generation
- Price
- Free
- Platform
- Local/Desktop
- Difficulty
- Advanced (4/5)
- License
- Apache-2.0
- Minimum VRAM
- 16 GB
- Added
- Apr 3, 2026
Related Tools
Open-source video generation model by Tencent with text and image conditioning.
Image-to-video generation model by Alibaba DAMO Academy.
Updated CogVideo model by Zhipu AI with improved video quality.
Infinite-length music-driven video generation with visual conditioning.
Text-to-video generation framework with cascaded latent diffusion.
Open-source text-to-video model by Zhipu AI/Tsinghua with 2B and 5B variants.