═══════════✧═══════════
✦ L I N E A V E R I S ✦
═══════════✧═══════════
V4 I fixed the previous upload...
Use normal CFG ranges test at will
If you want an anime blend use the casual illustrious quality tags "Masterpiece, awa" etc.
If you want a more western style, don't use quality tags and just use
Western comics (style)
COMIC STYLE "Western comics (style)"
ANIME STYLE : TAGS "Masterpiece, awa"
Masterpiece, awa
The last images that have a highly textured and detailed style are of the version under the files listed as FP16 here is the recommendation for that version ↓
V4 "FP16"
Use low cfg 3.0 - 5 otherwise image will be overbaked and saturated
It's experimental but unstable
Steps 20 - 45
When using the recommended tags on the info or add a ":0.4 to 0.6"
weight to the tag "Sharp Tapered Lineart:0.4" for example
V3
I have achieved a proper LEAN comic + anime blend
Mixing various models to make a base Bacanora Comics (Might upload in a few days) with
Indexed model (I just love the prompt adherence that model has)
Jorge Jimenez style + Clusterfuck + Marker & Ink Lycoris (It's supposed to be for Pony but I somehow made it work, just needed high weights) + AyakaStar_ST and various detail enhances to make the linework stand out.
The tags western comics (style), hatching \(texture\), marker (texture)
Are optional.
An image with tags and without the tags right after (same seed and settings) Upscale no facedetailer:
Veritas in Lineis — Truth within the lines
A fusion of Soft_Sketch Illustrious-XL @Rin_x0
and COMMIX @phinjo
woven together into a vision of
clean outlines, and painterly finesse.
───────────✧───────────
⚜ Features ⚜
・ Finely detailed semi-realistic rendering
・ Crisp, almost engraved outlines
───────────✧───────────
I am still learning, while enjoying the road. Hope you like it :D
═══════════✧═══════════
Description
Used Klatah to improve previous model
FAQ
Comments (5)
@solss_ Thx for the tip♥
According to the example images, it seems this is trained to be specifically a face close-up model?
No I just uploaded similar images. I use them for drawing, I had them pre-generated when I was uploading the model, so I used those.
Some context.
I've been tweaking my workflow so I won't have to have to write a lot, when I was working with this model, I had set up a "scene composer" node, but for some reason the randomize option was not working. I set a huge batch of various characters running while I worked and they all ended up under the same parameters and since I liked them, I uploaded those.
I've been using multiple view generations on my recent uploads, I'll get back to this one and add more variation of images.