iris-3b
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.
- depth-estimation
- diffusion
- pixel-space
- pytorch
- super-resolution
- text-to-image
- flow-matching
- generative-prior
- image-restoration
- vision-foundation-model
- Stars
- 32
- Forks
- 2
- + today
- +1
- Created
- 8d
Ranking data as of October 12, 2026 (UTC).
Star History
Today, hour by hour
01
Overview
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. It ranks #2989 on GitTiger, gaining +1 star on October 12, 2026 (UTC).
The project is written in Python and has 2 forks. It was created 8d ago.
Installation
git clone https://github.com/speridlabs/iris-3b.git
cd iris-3b
# see README for setup