Hello there and thanks for checking out this workflow!
—Intro—
AuraFlow is a new and powerful flow-based text-to-image generation model currently being trained.
AuraFlow 0.2 was just released and the versions available right now are meant as an introduction to the model and to catch the community's eye. It is not a finished 1000-fold-fine-tuned project (yet), but rather a WIP release that just happens to perform greatly already. So keep in mind as you experiment that, this is not even its final form.
Being a true open source project, it is bound to become great with the community's involvement.
As of now the model is rather large and might be a little painful to run with less than 16gb of VRAM. However the devs mentioned in AuraFlow's introduction blog post that "a mini version of (the) model that is still as powerful yet much much faster to run" is in the making!
And according to the discussions on their HuggingFace page a method of quantization seems to be a viable option to downsize the model and speed up inference as well.
—Purpose—
This workflow was built to test the new AuraFlow model and its amazing prompt adherence in a showcase of multiple 2-Pass generation approaches alongside some initial insights I gathered to get you started.
Features :
Aura-SR upscale (super-resolution project by the same devs)
Tiled Diffusion
Ultimate SD Upscale
—Custom Nodes used—
All of which can be installed through the ComfyUI-Manager
If you encounter any nodes showing up red (failing to load), you can install the corresponding custom node packs through the 'Install Missing Custom Nodes' tab on the ComfyUI Manager as well.
I have also included detailed instructions and links to all required/recommended models along with the folders they need to be placed in.
—Thanks—
The workflow would not be possible as is without these custom node packs. If you want to show your appreciation to the node creators, drop them a ⭐ on their github repos! Thank you!
Feel free to ask any questions, share improvements or suggestions in the comment section!
Also let me know if you encounter any confusing points I can elaborate on and focus on improving for the next update!
Description
v1 — initial release