Hassaku aims to be a anime model with a bright and distinct anime style.
My Discord for everything related to anime models and art.
If you'd like to support my work, you can do so at: SubscribeStar, every bit helps and is truly appreciated!
____________________________________________________________
Supporters:
Thanks to my supporters tamashiicolle, pttcot, SETI and Kodokuna
____________________________________________________________
Using the model:
The model was trained primarily on images with minimal disruptive elements, such as floating text, logos, speech bubbles, or signatures. If any of these elements appear in an image, please include "signature" in your negative prompt.
Danbooru Metadata and franchise tags are not used in training, so please avoid using them in prompts. Metadata such as "highres" and franchise tags like "re:zero kara hajimeru isekai seikatsu" are not used in training.
A CFG scale between 3 and 7 is recommended. For best results, use CFG 6 with the Euler a sampler.
Refer to the example images, to understand, how the model should be used.
The model is designed to be simple and straightforward to prompt. Use the following tag order:
Number of characters
Character Tags
All remaining tags
Example: 1girl, rem \(re:zero\), standing, masterpiece, upper body
Here are some resolution options for SDXL:
1536 x 640
1344 x 768
1216 x 832
1152 x 896
1024 x 1024
896 x 1152
832 x 1216 (most recommended)
768 x 1344
640 x 1536
Lora use:
NoobAIs Loras working for the most part better in compare to Illustrious loras
Versions 2 and 3 is less compatible to most loras trained on Illustrious or Noob, because it got further trained away of both. Style A is still the most compatible for it.
______________________________________________________
Version and License info:
Below V1 merges ANIMAGINE XL 3.0
Version V1 use Illustrious-XL & WAI-NSFW-illustrious-SDXL with additional training
V2 is trained on its own and don't include any extra merge. Base was V1, it is trained to include newer or missing characters. It was also used for further training tests
V3 is a merge of Illustrious-XL and V2.2, to fix some issues and was also trained to include newer or missing characters
All Models using the Fair AI Public License 1.0-SD license.
Description
V9: Images have a slightly blurry effect and some concept have strong tendencies to create really basic 2d images now. It is sometimes a bit of a gamble between basic 2d images and more detailed ones.
Blurry because i trained some days on the edge of possibility, to make it more anime now. More basic images are in a learning transition, those images simply doesn't looks great, because of it.
FAQ
Comments (15)
V9: Images have a slightly blurry effect and some concept have strong tendencies to create really basic 2d images now. It is sometimes a bit of a gamble between basic 2d images and more detailed ones.
Blurry because i trained some days on the edge of possibility, to make it more anime now. More basic images are in a learning transition, those images simply doesn't looks great, because of it.
is there anything special to concider for use of this model? I get stuff thats not even anime and really have to over-engineer prompts to get remotely what I want - still with body deformations and weird artstyle but everyone else seems to have such a good time so am I doing something wrong? Do i have to use some special SDXL stuff in Automatic1111 like refiner or sth?
do you work with danbooru tags and the right dimentions? there isn't anything big to consider. SDXL is stubborn on training, don't want to let go stuff like to be realistic in some cases. That is a good thing to don't let go stuff, but it makes training new concepts harder.
I pruned artist tags until now, seeing what naiv3 can do and also things like cross-sections without mixing it up like 1.5 models do. I implement them now into my training.
Do you have enough vram? sdxl is much bigger then 1.5 models
@Ikena 32 Gb DDR4 3600 Mhz
Next question: Did it download incorrectly? I've had it where even in huggingface batchlinks will miss like 2kb off a model and just yeet it.
@ApexThunder Your gpu ram, not your pc ram. But in consideration of your highend ram, you propably have a good gpu. Maybe load it again, like duskfallcrew suggest
@Ikena My GPU: RTX 4070 OC Gigabyte 12 GB Aero
@ApexThunder Yep more then enough, is it only my sdxl model? is your gui of choice uptodate?
@Ikena the strange thing is that when I go to: stable diffusion checkpoint to load the checkpoint, it doesn't load it, and I don't understand why
@Ikena Anyway,my GPU is up to date
Does your dataset contain any 1.5 images? If so, I want to say that it results in a lesser model.
I disagree, my dataset contains some 1.5 images, who are more drawn looking and feature good backgrounds. To fight a bit the big majority of simple or blurry backgrounds of most images.