下载后放进对应的 ComfyUI 目录。models/ 指 ComfyUI 的模型根目录。

DC Cyberpunk Style 1

作者：dkpc69

v1 MossBase Trained @7000 Steps, Rank 32, Captioned with Gpt-6.1SolMed, Dataset of 206 Images using my midjourney. Trained with Ai-Toolkit, Works with MossSFT, With Moss you may have to generate 4 images to get a really good one.

v1 krea2raw Trained @10,094 Steps, Rank 32, Captioned with Gpt-6.1SolMed, Dataset of 206 Images using my midjourney. Trained with Fizgig lora trainer, Crank lora up to 1.2 strength to get a stronger aesthetic.

DC Cyberpunk Style 1 Helps to create cinematic cyberpunk images featuring heavily modified characters, armored streetwear, elaborate vehicles, megacities, and atmospheric industrial scenes, with realistic and hyper-realistic plastic skin lol, worn materials, and a retro and modern science-fiction film aesthetic.

v1.0 [Krea2Raw] Negatives: yes this lora was a bit of a fail it will generate basic looking normal humans and normal cars and items if prompted in very basic terms, i must have left a few normal looking cars and normal people in the dataset, BUT it does create some amazing cyberpunk images if you follow my prompts, I have also added a custom instruction txt file you can give to Chatgpt (only one i have tested it on) to create you prompts which are featured in alot of the samples here. When you add any of these images to comfy it will say v2 at the end that's because this was my second fail of trying this lora lol i have renamed to v1 when uploading here.

Check the bottom of this lora page for more examples of this lora or check my profile gallery to see more examples and loras.

This lora needs weird prompts but if you are having troubles, Prompt the way i have prompted and expand from there and you shouldn't go wrong.

IM POSTING IMAGES 12 A DAY FOR THE NEXT WEEK FOR MORE REFERENCES SORRY IN ADVANCE FOR THE SPAM OF CYBERPUNK.

If you use this lora in merges or upload to other generation sites please mention my civitai page on your post, Thankyou.

触发词