This is the basic build of FLUX
Missing nodes can be downloaded through the manager in the missing nodes list.
You can download the basic models yourself, depending on your system.
Or use the ready-made ComfyUI assembly from ELPA
Add loras to the CR LoRA Stack nodes in the LOAD LoRA Models group
Links to LoRAs
-Midjourney V6.1 meets FLUX
https://civarchive.com/models/646411/midjourney-v61-meets-flux-illustrious-sdxl
- FLUX [pro] 1.1 Style Lora
https://civarchive.com/models/832683/flux-pro-11-style-lora-extreme-detailer-for-flux1-dev
- FLUX Image Upgrader
-Flux [pro] ULTRA Style Lora
https://civarchive.com/models/995060/flux-pro-ultra-style-lora-extreme-detailer-for-flux1-dev
The Sampling group performs generation based on the input data
This group needs a blank canvas, a hint, and a model.
A blank canvas, or rather, empty noise, can be configured in the Flux Resolution Calc node.
The megapixel widget is an incoming upscale multiplier, we can factor in the base resolution by two and a half times.
aspect_ratio select the sides of the frame, vertical and horizontal orientations
Empty Latent Image-batch_size - the number of generated images for the entire pass
We can enter our own hint or generate a hint based on the incoming image.
Add the image to the Load Image node in the Input Image group
Select the prompt detection mode in the Switch (Any) node in the Prompt or Caption group
input1 - manual prompt
input2 - prompt from image (Caption)
It is set to Caption by default.
If all the models are loaded correctly, you're done.
Blue Queue button - start generation