. Krea2 Trigger Word = dcd@rkf@ntsy. Use 4 Step generations for intended output and 8 steps for more HD looking images, Use the embedded workflows which provide amazing results when grain is added. Trained @ 9000 Steps with a 379 Image Dataset, with ostris Ai-Toolkit, Captioned with Gpt 4.1. Check out my profile gallery for more examples of what this lora can do.
. Flux,ZImage Turbo and base V1 Trigger Word = dcdrkfntsy.
. This Dark Fantasy Lora was to the older dark fantasy surreal style of @shuj1nk0 on tiktok he makes amazing dark fantasy, surreal and medieval images with midjourney, this lora can get close to his style of images with the correct prompting but its hard to get the midjourney creativity with this turbo model, also adding a good upscale to your workflows will make the images soo much better than the previews.
. Trained with AI-Toolkit for 10,000 steps Rank 32 Tagged with detailed captions consisting of 60-100 words with JoyCaption Beta One (377 Images Total)
. All images have generation details/ Workflow for ComfyUi just click the image you want, then save, then drop into ComfyUI or Open the image with notepad on pc and you can search all the metadata there.
https://civarchive.com/user/dkpc69
Description
Original Filename = DC Dark Fantasy Style 2 [Flux.2-KleinBase9B][JoyCaptionBetaOne][Trigger=dcdrkfntsy.] v1
Trained for 10,000 Steps
Had to quickly generate these preview images all one shot first pick you can get much better by running prompts again, Have fun
FAQ
Comments (2)
SO COOL!
You did such a fantastic job on this, I know the Krea 2 version is going to get even more use than the ZiT.