Workflow for XL/Illustrious/NoobAI Models
Any feedback would be appreciated.
See the Article for a Guide:
https://civarchive.com/articles/17339 (probably doesn't work rn)
https://civarchive.com/articles/17339
Key Features:
LoRA
TIPO Prompting
Detail Enhancers
ControlNet Upscale
IPAdapter
HiresFix
OpenPose
Wildcards
ControlNET
Color Match
FBCNN
Regional Prompting
Inpaint
Outpaint
Workflows included:
🟥Advanced: comprehensive text to image workflow
🟨Basic: more basic version of the Standard workflow
🟦Standard: more basic version of the Advanced workflow
🟩Detailer: detailing previously generated or existing images
Experiment and enjoy!
Description
Improvements and Fixes
Added Detailer only workflow
Changed Image Save nodes
FAQ
Comments (10)
For those who can't install TIPO here's a fix i found. go to ComfyUI_windows_portable\python_embeded folder and use cmd and then copy this in,
pip install llama-cpp-python==0.3.4 --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cu124/What version of ComfyUI are you using?
Please let me know which versions of CUDA and Python you're using.
@jbh0716526 Don't remember what i used bro that was 5 months ago. I don't use tipo anymore.
Thanks for this comprehensive but still easy-to-follow workflow! Just a recommendation: I think it would be great if there was an image filter node so that we have more control over which images we want to take to upscale.
This is freaking awesome. Is there a possibility of adding regional prompting instead of comfycouple? If not, do you do commission workflows? Please let me know
Yes, I plan to transition from ComfyCouple to regional prompting in the next version.
@Legendaer I LOVE hearing that! Will be looking forward to it!
Could you provide usage examples for Regional Prompting? When I test with red and green dual-color masks, they consistently blend together and merge into a single character.
@TTeslaThunder I have added an Attention Couple node to the new workflow version, which should prevent this kind of "blending". I also highly recommend pairing the Regional Prompting with a ControlNet, as this guides the model to create separate and correctly placed subjects from the start.
@Legendaer thank you!