Hi Everyone, this is my first post or model upload, I'm doing this as a hobby and I'm not an expert by any means.
There are two versions of this LoKr:
V1 : Was made with a dataset of about 90% in-game screenshots from the game's model viewer.
So, it's visibly a bit darker and raw on textures.
V2 : Was made with a dataset of 30% in-game screenshots from the game's model viewer and 70% of cinematics of the game.
Trigger words :
V1 : CL41R3R3DF13LD
V2 : CL41R3R3D_3.0
All the images showcased were made with a strength of 0.90.
All images include at some point of the prompt :
"In-game screenshot, RE Engine render, 3D digital art, Subsurface scattering skin, ambient occlusion, Substance Painter textures, ray-traced cinematic lighting, video game character model, high-fidelity CGI asset, real-time viewport render, Capcom character aesthetic."
That's what I use to get CGI like renders, but this LoKrs do very well on Photorealistic images as well.
All the images were generated using the Moody Krea Mix of catlover1937.
7.0 Version.
And the videos with LTX 10Eros from tenstrip.
https://civarchive.com/models/2447875/ltx-10eros
(I apologize if there is another way to tag authors or their models)
I'm planning to upload several LoKr if this one does well.
Basically ALL Resident Evil Characters, Final Fantasy, Lies Of P, Mortal Kombat and Street Fighter among other franchises.
Some Examples : Ada, Alcina, Alyssa Ashcroft, Annette Birkin, Ashley, Bela Dimitrescu, Cassandra Dimitrescu, Claire CODE VERONICA, Daniela Dimitrescu, Donna Beneviento, Elena (Village), Excella, Grace Ashcroft, Ingrid Hunnigan, Jill, Katherine Warren, Leon, Mother Miranda, Rebecca, Sheeva Alomar and Sherry (Adult from Requiem of course).
The majority of Final Fantasy characters and Lies Of P as well.
If I'm not wrong it's possible to drag&drop the downloaded images to your Comfyui if you want to, (Except the ones with split screen since those do not have the metadata). But my workflow might be a mess lol, and I generate images of plus 12.000x7600 more or less, plus my prompting is not great. If for some reason the metadata is not on the pics let me know.
I hope you all like this LoKr and feel free to ask anything!
Description
V1 : Was made with a dataset of about 90% in-game screenshots from the game's model viewer.
So, it's visibly a bit darker and raw on textures.
V1 : CL41R3R3DF13LD
FAQ
Comments (8)
Holy shit she looks like an actual princess with pigtails, anyway thanks and i'm already making a anim with her :)
Glad you liked it!
You can clearly distinguish between versions by adding V2 using the "Add Version" button located to the left of the blue version label at the top of the page.
I don't see them as "V2 is better that V1" , they're just different because of the dataset used, maybe some ppl prefer the V1 because is more Raw and blender like and V2 is a mix. Thanks for the comment!
@Li3sOfPaul I am not saying that "V2 is better than V1." The way you are currently offering your LoRA forces users to identify the version by looking at small filenames before downloading, which results in very poor usability. If you use the "Add Version" feature to separate V1 and V2, distinct download buttons will appear for each, eliminating the need for users to read small filenames—a method used by almost all creators on Civitai.
@fantaseed Oh I get what you are saying now, that way I coud even put the images under the exact version instead of having them all mixed up on the showcase. I'll edit the post, Thanks!.
You can also use Version A, Version B etc. If you want to differentiate them.
WHAAAAAAAAAAAAAA 😍😍😍