OmniGen - A Diffusion Model You Might Not Need Photoshop Anymore
Added 2024-11-03 13:30:01 +0000 UTC
Video : https://youtu.be/DuY_wlU8FgM
Today, I'm thrilled to bring you an exclusive look at OmniGen, the revolutionary new diffusion model that's changing the game in AI image generation! We're diving deep into this groundbreaking technology that combines the power of Microsoft's Phi 3 large language models with SDXL VAE to create something truly extraordinary. Unlike traditional diffusion models that rely on multiple components like ControlNet or IPAdapter, OmniGen offers an all-in-one solution that's blowing minds in the AI community!
OmniGen
https://github.com/VectorSpaceLab/OmniGen
https://huggingface.co/spaces/Shitao/OmniGen
https://huggingface.co/Shitao/OmniGen-v1/tree/main
OmniGen-ComfyUI - TextPrompt Fix version , Forked by me
https://github.com/benjiyaya/OmniGen-ComfyUI/
Original GitHub project : https://github.com/AIFSH/OmniGen-ComfyUI
Model from Hugging Face save in 
Models/AIFSH/Shitao/OmniGen-v1
Comments
Never see this error before. i just install how I recorded in the video and run the workflow. Phi3Transformer, looks like you have something difference in your system.
Benjamin Law
2024-11-09 09:41:40 +0000 UTCRun the workflow ang get an Error message: Phi3Transformer does not support an attention implementation through torch.nn.functional.scaled_dot_product_attention yet
guni
2024-11-09 04:52:49 +0000 UTC
