At the moment I cannot upload any larger workflows to OpenArt. You can find information about the current status here:

https://youtu.be/dDIKGomah3Q

Please note that this workflow is deprecated and will not receive any further updates for compatibility with newer versions of ComfyUI. We recommend switching to an alternative workflow to ensure that you can take advantage of the latest features and improvements offered by ComfyUI.

Try out my new workflow; it's incredibly user-friendly, well-organized, faster, more efficient, and modular. Not everything I have planned has been implemented yet, but you should definitely check it out.

Automatic_comfyui_sdxl_modul_img2img_v21

or

https://openart.ai/workflows/murphylanga/automatic_comfyui_sdxl_modul-incl-faceswap/2Bsh5QwsrQkKFx6CLCz7

This workflows are fully automated and come with helpful training videos

Thank you for your understanding, and we hope you find the suggested workflows useful for your projects.

_________________________________________________________________________________________________________________

The workflow is now integrated with the latest custom nodes in:

https://openart.ai/workflows/murphylanga/murphylanga_sdxl_new_ultimate-v105---merged-collection/LVygANrUJuJbifw8JTcb

-------

(This template is used for Workflow Contest)

What this workflow does

👉Workflow with integrated automation and translation, expanded to include various face optimizations

Many built-in automations, various options for face optimization and face swapping

Version 3:0

Adaptation to new custom nodes December 15, 2023

Translated nodes summarized

The Reactor accounts have been adapted to version Beta 4.01

Additionally in version 2.0 - Pipe installed without and with controlnet

Added nodes to put in Information when saving images and can display the information (model, steps, sampler, Clip L Prompt, Clip G Prompt, Refiner...

The saving of intermediate steps can now be switched on and off together. The paths can be preset if the steps are to be saved

How to use this workflow

👉 Make the settings in the sampler group.

The application is like my basic workflow

Find a detailed video description in the link below

https://openart.ai/workflows/tamarin_extra-large_98/basic-workflow-with-built-in-automation-and-translation-to-expand/5udJAmfyerDJerfAekVq

Select the Base Model, VAE, and Refiner checkpoints.

Checkpointloader Base Preview allows you to display preview images for the models after saving them to the location of the model of the same name.

Choose the SDXL ratio.

All important settings can be selected in this group. For testing purposes, you can simply leave the default settings unchanged.

Select refiner on or off.

With Global Seed, seeds can be planted anywhere at the same time.

In the Styler Select group, stylers can be selected and easily replaced. There are various “parked” there.

The language can be selected in the Prompt + Styler group, English is always possible using the switch.

You can also specify how, which and where the styles should work.

The prompts used will appear there for review.

An automatic calculation for the ClipEncoder is preset there.

The Lora group can be switched on and off via bypass.

(Right click on the group)

Also individual Loras in the group.

Preview images are possible for the Loras as described at the checkpoint.

The Controlnet group can also be switched on and off via bypass.

Selection and strength can be controlled via CR Multicontrolnet Stack. Other controlnets can also be installed and replaced in this group.

SDXL Sampler Base and Sampler Refiner should always remain switched on.

Optimizations can be made via the “highresfix”, “ultimate upscale image1” and “2” groups. These groups are normally switched off via bypass and can easily be switched on again after the basic image has been generated. Step by step or even partially.

New images are only calculated if the global seed is not set or if something is changed in the command prompt or basic settings.

Otherwise, the optimizations can be made step by step.

By changing the noodles, the different knots can be combined in different ways.

Bypass all groups that you don't need.

At the end, additional information about the image is saved.

The storage name and location can be set via the imageOutput nodes. The PlaySounds Node gives an acoustic signal with each “finished” generation.

Some statements for further development are already attached.

Tips about this workflow

👉 Tips for good results with the following noodle order:

whithout optimize to highresfix

from highresfix to upscaler step1

from upscaler step1 to detailer 1

from detailer 1 to detailer 2 use image sharpen

from detailer 2 to upscale step 2 use image sharpen

from upscale step 2 to exit

or

whithout optimize to highresfix

from highresfix to to detailer 2

from detailer 2 to upscale step 1 use image sharpen

from upscaler step1 to face crop

from face crop to upscale step 2 use image sharpen

from upscale step 2 to exit

or

whithout optimize to highresfix

from highresfix to upscale step 1 use image sharpen

from upscaler step1 to face crop

from face crop to face swap

from face swap to exit use image sharpen

or, or ,or.....

🎥 Video demo link (optional)

👉  https://youtu.be/tqSJnHbUCSA