A huge thank you to Datacrunch.io for generously providing their resources for training and support. I highly recommend them.
This is a new version LoRA for Flux_Dev. It was trained on a larger dataset compared to the LoRA for SDXL, with around 2,000 images focused on dark color tones and detailing. It now seems to recognize more shades than Flux_Dev. However, I would love to get your feedback. Please share in the comments how it responds to your prompts and examples of your work.
This LoRA will make you everything in noble dark colors. You can not use keywords, you can use them, the result will be different. But always interesting. Share your results, I'm curious to see what you will get
Description
FAQ
Comments (6)
Sample images look amazing, though I can't replicate them with exactly the same settings on ComfyUI. What did you use to generate these?
Samples are made through Automatic1111. Try it
@Sa_May hey - i'll give it a shot in A1111. Really digging the feel of the lora :)
@dingleberry_juice That's strange that it doesn't work in ComfyUI. I'll try to check it today.
Hey, I tried it again in ComfyUI but only with the trigger words "black and color" and it worked fine. I guess it didn't like the 'dark' part. :) I'll post some ComfyUI images later.
@dingleberry_juice This lora gives great results in Comfy UI. It's my favorite for SDXL.
Tested with juggernaut XL 5 and PhotoPedia XL 4.5.
You probably figured this out by now, but In Comfy UI, the loras are not in the prompt, they are loaded from a specific node.