iris-3b
A 3-billion-parameter diffusion model that paints images pixel by pixel instead of compressing through a VAE first, and the same base model also sharpens blurry photos and estimates depth.
View speridlabs/iris-3b on GitHub- GitHub stars
- 225
- Forks
- —
- Language
- Python
Project details
- Repository
- speridlabs/iris-3b
- Licence
- Apache-2.0
- Primary language
- Python
- Catalog status
- Available
- Catalog check
- 2026-11-13
Repository description
Iris-3B: pixel-space generation & general vision learner. A 3B diffusion transformer that generates every pixel directly (no VAE); the same prior, fine-tuned, does monocular depth estimation and 4x image restoration. Paper, weights, training code and demo.
Explore by topic
More tools in Images & design
- iris-3b-metal
tky-27 · Apache-2.0
A native macOS app and local engine that runs the Iris-3B pixel-space text-to-image model directly on Apple Silicon GPUs, so you can generate, restore and upscale…
- qwen-image-2.1-studio
janishar · MIT
A free local studio that lets you generate and edit images from text prompts on your own Mac using the Qwen-Image-2.1 model, with a prompt enhancer and a browser UI, no…
- comfyui
comfy-org · GPL-3.0
ComfyUI is a free, node-based visual interface for running diffusion models like Stable Diffusion locally, so you can build AI image and video generation workflows by…
- stable-diffusion.cpp
leejet · MIT
Lets you run popular image and video generation models like Stable Diffusion, Flux, and Qwen Image without Python or a GPU-heavy setup.