Image size up, dataset and training improvements
Use to 1.0 ~ 0.8
The learning model is animefull-final-pruned.
It works better if you type some prompts.
If you want a black and white picture.
Promt : monochrome, grayscale,
Negative : color,
