Type | |
Stats | 307 |
Reviews | (47) |
Published | Jul 19, 2023 |
Base Model | |
Usage Tips | Clip Skip: 2 |
Hash | AutoV2 8E533D06A8 |
这个模型主要是想针对naifu进行改进,风格偏向干净线条清晰,使用了部分mj图片和一些自合的lora
This model is mainly intended to improve on naifu, with a clean style and clear lines, using some mj pictures and some self-synthesizing lora
目前解决的问题:
Problems currently solved:
1.画扫把,目前可以画不从中间断开的扫把(类似的长杆物品都可以试试)
1. Draw a broom, currently you can draw a broom without a break in the middle (similar long rod items can try)
2.效果更好一些的倒影
2. A better reflection
3.可以刻画男性角色
3. You can portray male characters
4.部分画风优化,可以进行中国水墨画、水彩、素描等画风tag
4. Part of the painting style optimization, can carry out Chinese ink painting, watercolor, sketch and other painting style tag
目前没有解决的问题:
Currently unresolved issues:
1.原生的吃面条的构图和大部分使用筷子的进食方式(太难了T。T)
1. The original composition of eating noodles and most of the way of eating using chopsticks (too difficult to T. T)
2.英文text构建会有错别字
2. English text construction will have errors
3.无法写中文
3. Unable to write Chinese
4.横向长图会有部分肢体崩坏(需要roll图)
4. Lateral long image will have some limb collapse (require roll image)