UNO ComfyUI plugin synchronized open source
https://github.com/HM RunningHub/ComfyUI_RH_UNO
Features
Supports flux dev fp8 and flux schnell fp8
Supports running bf16 flux dev and flux schnell on 24g GPU

Progress bar
Real-time display of denoising progress
Local model loading. Does not force downloading models from Huggingface, making it more friendly to CN environments

Default usage of schnell model, 4-step inference, can generate relatively quickly
If switched to dev model, requires 25-step inference

https://github.com/bytedance/UNO/tree/main

In this research, we propose a highly consistent data synthesis pipeline to address this challenge. This pipeline leverages the intrinsic generative capabilities of diffusion transformers and produces high-harmonic multi-subject paired data. Additionally, we introduce UNO, which includes progressive cross-modal alignment and universal rotary position embedding. It is a multi-image object model iteratively trained from a text-to-image model. Extensive experiments show that our approach can achieve high consistency while ensuring controllability in both single-subject and multi-subject driven generation.

In this study, we propose a highly consistent data synthesis pipeline to tackle this challenge. This pipeline harnesses the intrinsic in context generation capabilities of diffusion transformers and generates high consistency multi subject paired data. Additionally, we introduce UNO, which consists of progressive cross modal alignment and universal rotary position embedding. It is a multi image conditioned subject to image model iteratively trained from a text to image model. Extensive experiments show that our method can achieve high consistency while ensuring controllability in both single subject and multi subject driven generation.