AIDAdiffusion

Concept Art/Prototyping faster with ourbunka internal tool AIDA (AIDAdiffusion), "All-In-one" app for running Stable Diffusion on windows PC locally, with easy to use web-based UI 🎉 !

Powered by Stable Diffusion V1-4.

V0.1.3 New Features

-new Advance mode UI

-more tweakable parameters such as custom seed, scheduler/sampler, guidance_scale etc.

-new supported schedulers : LMS, DDIM,LMSD,PNDM

-all dependencies included (If you download V0.1.3-Prepack)

Requirements

-modern gpu with DirectML support (Tested on AMD RX5700, newer nvidia & amd gpu should be supported too)

-8GB of VRAM or more

-32GB of RAM or more

-windows 10

-Python 3.10.7

Preview

AIDA GUI github pages preview

Watch vdeo Demo here.

*AIDA GUI required "AIDAdiffusion.exe" running in your local pc or cloud server to function properly

*currently github pages version seem to have issues connecting with locally hosted AIDAdiffusion.exe(github pages seem to automatically convert http to https?), we recommend using locally hosted AIDA GUI http://localhost:9998/files/build.

ToS

BY USING AIDA(AIDADIFFUSION) APP, YOU ARE AGREEING WITH OURBUNKA MOTION AIDA(AIDADIFFUSION) TERMS OF SERVICES, LICENSE AND STABLE DIFFUSION LICENSE.

Download

Download AIDA(AIDAdiffusion) Latest release here. Also available on Itch.io.

V0.1.3 Dependency installation

If you download Latest V0.1.3-PrePack releases, ignore this. All python, stablediffusion & onnx dependencies in included in the release.

If you gitclone this repo, you need to copy these python libraries into aidadiffusion/diffusers/examples/inference/vendor directory -diffusers -ftfy -numpy -onnxruntime -scipy -torch -tqdm -transformers

you also need to accept stablediffusion license on huggingface, download stable-diffusion-v-1-4 repo. Then, copy to aidadiffusion/diffusers/examples/inference/ . After that, go to aidadiffusion/diffusers/examples/inference/ in terminal, run "python save_onnx.py" script to convert the pretrained model to onnx runtime.

no huggingface token auth required after these step completed.

Credits

We want to thanks authors, researchers and engineers of following repos for their research and implementations for making this app possible.

-@compvis, Robin Rombach, Andreas Blattmann, Domninik Lorenz, Patrick Esser, Björn Ommer, Stablility AI and Runway works on stable-diffusion https://github.com/CompVis/stable-diffusion

-@harishanand95's diffusers https://github.com/harishanand95/diffusers

-@huggingface's diffusers https://github.com/huggingface/diffusers

-@chakra-ui's chakra-ui https://github.com/chakra-ui/chakra-ui