Type | |
Stats | 458 |
Reviews | (61) |
Published | Aug 17, 2024 |
Base Model | |
Hash | AutoV2 A42DE5F16C |
Anime I Like The Most
PLEASE READ THE DESCRIPTION (Promise .. its kinda short )
I Do NOT recommend you to copy paste all of your quality prompt at once in the v0.45 bcz its not good for overall asthetics of images . but if you want then you can do it , its just my recommendation
My goal for making this was to make a anime model high quality as the original anime,
This model is good for both NSFW AND SFW images and quality prompt are needed like (high quality, masterpiece) etc,
VERSION ALPHA 0.9
The v0.9alpha make makes great images with some specific styles ,
Pros - it makes vibrant images without any external lora or embedding, the overall quality of images is good (kinda ok ) . good for testing if you want .
Cons - Its limited to making images good with specific styles , the concepts are limited . too much need of external loras is felt during use , the eyes and body proportion are also very limited and not great , not great at anatomy.
VERSION 0.45 BETA
This version is a bit more tolerant to different artstyles . For example you can see the lora which I trained using civitai trainer with this model as the base model , (i am not a lora training expert so don't take it as the best .)[https://civitai.com/models/663913/dinoartframe] .
I recommend using keyword : anime_screencap, if you like the studio like anime images .
Pros: better overall aesthetics of images , tolerant to different artstyles, and is a bit more of adaptive.