Sign In
0bba20aa-e949-45a0-9b44-9ddc675581ff.jpeg
c75b9994-0759-41b7-aa95-220b9e345736.jpeg
91ed8b14-de1a-4830-a100-48b96507a830.jpeg
bc0e3bb3-8fcc-4578-b07c-9180c4fa356e.jpeg
22f54194-55a5-4eab-a1c8-69cecaa4b815.jpeg
5806897a-04b8-4a46-9f0e-3d79a51b8876.jpeg
58a179f6-56c3-489b-8499-824de0421d27.jpeg
ec617287-0795-4d32-8b3e-25581c5597f4.jpeg
6acea413-61f8-41c3-84d2-63576e21ce1f.jpeg
d7e160ae-51e4-4a8c-87a3-65e7206a9390.jpeg
97acfffa-c7d5-4428-b5c4-547c99cbf1a0.jpeg
ae0f04ac-9c22-4cf6-a8c3-7376a2916388.jpeg
30e71ce2-cd51-46d0-b333-d7451bb7afc6.jpeg
01634996-1d08-43c5-95ab-c99dafe7e9e3.jpeg
8964c16a-9d72-4dab-8b16-c03deeb43385.jpeg

Comments

PERSONAfied Semi-Real(LJA) - v1.0 Showcase

I put a good amount of effort into showing the capabilities and limitations of this model. It is strongest at doing what is designed to do, take an anime Lora and get a 2.8 semi realistic portrait with good detailed backgrounds and aesthetics with minimal additional lora. It follows direction decently for lighting, camera, fantastical elements such as warrior armor. My full prompts are available and I put a variety of use including ones with no Loras. It can come out a bit too HDR for my tastes when you pump up the Sampling steps. You'll find very detailed results and a good 2.5-2.8D look with steps from 30-59. 70 is way overkill but you can see the results here. I use ADetailer sometimes and use other checkpoints for it and it looks ok. I've provided examples of both. Typically frames the subject center, and cowboy shot style without prompt. I hope you enjoy the results you get and I would love to see your creations.