Type | |
Stats | 383 2,683 |
Reviews | (79) |
Published | Dec 3, 2023 |
Base Model | |
Trigger Words | qtcg licg gfcg xscg |
Hash | AutoV2 12ECFA9E97 |
I found a new AI tool Shakker, a best image to image tool. You can try it via https://www.shakker.ai
This model is still suitable for purposes such as tweets for ancient-style cultivation novels, character designs for Orange Light games, and illustrations for ancient-style anime.
Everyone should be aware that this model has trigger words,
Four trigger words represent four different stylistic directions.
In other words, a large model can generate images in four different and distinct styles,
Do not understand this as a strong generalization of the model.
The corresponding LORA has also been trained, but its image generation results are not ideal, so it will not be uploaded for release.
I hope that each uploaded model has its own strengths, and there is no need to add a negative review for models that are just passable.
If you download it locally, use EULER A for steps 27-29, or RESTART sampler for 36 steps +
For resolution, it is recommended to switch between 960X536 aspect ratio and 960X768.
Be sure to enable ADetailer!!!