Step-by-Step Guide Series:
SDXL - TXT to IMG Workflow
This article accompanies this workflow: link
This guide is intended to be as simple as possible, and certain terms will be simplified.
Workflow description :
This workflow allows you to generate images from text. It is optimized to integrate all metadata for simplified online publishing.
It integrates an upscaler/sampler which, although it takes longer than a classic upscaler, allows for much better quality.
Prerequisites :
📂Files :
An SDXL model : like this one
in ComfyUI/models/checkpoints/An upscale model (optional) : like this one
in ComfyUI/models/upscale_models/As many LoRAs as you like (optional) : like this one
in ComfyUI/models/loras/For add detail fonction (download automatically by my toolkit) :
A BBOX detector model : like this one
in ComfyUI/models/ultralytics/bbox/A SAM model : like this one
in ComfyUI/models/sams/A SEGM detector model : like this one
in ComfyUI/models/ultralytics/segm/
📦Only one custom node to install:
Don't forget to close the workflow and open it again once the nodes have been installed.
Usage :
In this new version of the workflow everything is organized by color:
Green is what you want to create, also called prompt,
Red is what you dont want,
Yellow is all the parameters to adjust the image,
Blue are the model files used by the workflow,
Light blue are post-processing nodes for upscale or add detail,
Purple is for LoRA.
Nodes list:
Positive prompt input :
Write what you want, in the form of a tag, not a complete sentence.
Negative prompt input:
Write what you dont want, always in the form of a tag.
Generation settings :
This node lets you configure the settings for your generation. It depends largely on the model you choose and what you want to create.
For SDXL it is recommended :
Start at 1024x1024. Dimensions too far apart can cause problems,
Steps between 30 and 50
The higher the number, the better the quality, but the longer it takes to get an image.cfg arround 5
the lower the number, the freer you leave the model. The higher the number, the more the image will resemble what you “strictly” asked for.),Sampler normal, karras or dpmpp,
Scheduler simple.
The seed is the unique identification number of the random starting point of an image. With exactly the same settings (prompt, step, etc.) changing the seed allows you to obtain a new image and not exactly the same one.
LoRA :
LoRAs are sort of "mini models" that are added to a base model. They can be used to create certain characters or styles. If you dont know what is LoRA just dont active any.
If you need more LoRA nodes, simply click on the “loras” connector on the left and add a new node—such as “LoRA 3x.” All LoRA nodes in the chain will be loaded.
Checkpoint loader :
This is the model file that will be used to generate your image.
There are many models available on Civitai. Choose the one that best suits what you want to generate.
UltimateSD Upscale :
This node is a tiled upscaler. To produce a better image, it generates your image at a higher resolution, but piece by piece. While this takes longer, it results in much higher quality.
In pass-through mode, the node will do nothing.
In addition to your own upscaling models, this node offers some that will be automatically downloaded.
The slider lets you choose the upscaling ratio; a value of 2 will produce an image that is twice as large.
Subject detailler :
In pass-through mode, the node will do nothing.
Here, you can choose the face, the hands, or both. This node will detect the relevant areas in your image and generate a new sample to achieve better results.
The slider defines the strength of this new generation over the first.
Now you're ready to create your image.
Just click on the “Queue” button to start:
While the image is being generated, you'll see a real-time preview in the sampler:
Once rendering is complete, the image appears in the “image saver” node.
All generated files are stored in comfyui/output/SDXL/YYYY-MM-DD/YYYY-MM-DD-HH-MM_SEED.