Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
SDXS: Real-Time One-Step Latent Diffusion Models with Image Conditions
| Date | Stars |
|---|---|
| 2026-07-31 | 664 |
| 2026-08-06 | 664 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
<div align="center"> ## SDXS: Real-Time One-Step Latent Diffusion Models with Image Conditions [](https://idkiro.github.io/sdxs) [](https://arxiv.org/abs/2403.16627) [](https://huggingface.co/IDKiro/sdxs-512-0.9) [](https://huggingface.co/IDKiro/sdxs-512-dreamshaper) [](https://huggingface.co/IDKiro/sdxs-512-dreamshaper-anime) [](https://huggingface.co/IDKiro/sdxs-512-dreamshaper-sketch) [](https://huggingface.co/spaces/IDKiro/SDXS-512-DreamShaper) [](https://huggingface.co/spaces/IDKiro/SDXS-512-DreamShaper-Anime) [](https://huggingface.co/spaces/IDKiro/SDXS-512-DreamShaper-Sketch) *Yuda Song, Zehao Sun, Xuanwu Yin* </div> We present two models, SDXS-512 and SDXS-1024, achieving inference speeds of approximately <b>100 FPS</b> (30x faster than SD v1.5) and <b>30 FPS</b> (60x faster than SDXL) on a single GPU. Assuming the image generation time is limited to <b>1 second</b>, then SDXL can only use 16 NFEs to produce a slightly blurry image, while SDXS-1024 can generate 30 clear images.  Moreover, our proposed method can also train ControlNet, offering promising applications in image-conditioned control and facilitating efficient image-to-image translation. <p align="left" > <img src="images\sketch.gif" width="800" /> </p> ## 🔥News - **April 11, 2024:** [SDXS-512-DreamShaper-Anime](https://huggingface.co/IDKiro/sdxs-512-dreamshaper-anime) is released. We also create some Gradio demo on Hugging Face. - **April 10, 2024:** [SDXS-512-DreamShaper](https://huggingface.co/IDKiro/sdxs-512-dreamshaper) and [SDXS-512-DreamShaper-Sketch](https://huggingface.co/IDKiro/sdxs-512-dreamshaper-sketch) are released. We also upload our demo code. - **March 25, 2024:** [SDXS-512-0.9](https://huggingface.co/IDKiro/sdxs-512-0.9) is released, it is an old version of SDXS-512. ## ⚡️Demo Create a new environment: ```sh conda create -n sdxs ``` Activate the new environment: ```sh conda activate sdxs ``` Install requirements: ```sh conda install python=3.10 pytorch=2.2.1 torchvision torchaudio pytorch-cuda=11.8 xformers=0.0.25 -c pytorch -c nvidia -c xformers pip install -r requirements.txt ``` Run text-to-image demo: ```sh python demo.py ``` Run anime-style text-to-image (LoRA) demo: ```sh python demo_anime.py ``` Run sketch-to-image (ControlNet) demo: ```sh python demo_sketch.py ``` ## 💡Train I found that [DMD2](https://github.com/tianweiy/DMD2) release the training code, and its training scheme is identical to the new version of SDXS, so you can refer to it. Unfortunately, the SDXS training code is not allowed to be open-sourced and will most likely not be updated again. ## ✒️Method ### Model Acceleration We train an extremely light-weight image decoder to mimic the original VAE decoder’s output through a combination of output distillation loss and GAN loss. We also leverage the block removal distillation strategy to efficiently transfer the knowledge from the original U-Net to a more compact version.  SDXS demonstrates efficiency far surpassing that of the base models, even achieving image generation at 100 FPS for 512x512 images and 30 FPS for 1024x1024 images on the GPU.  ### Text-to-Image To reduce
Excerpt of 5,514 characters
Read on GitHub
32
Ikko Eltociear Ashimine · Japan
1
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:e62a6ffed2ef24d3, desc:latent diffusion