
Software Translations:
Artist Rockin' Jelly Bean's lora. Seriously, I don't know this artist at all.
The reason why I thought of making this artist's lora is because someone in the comments on my other lora's page seemed to really want this artist's lora. Well, I searched for it.
I find I have little interest in his art style, and with the diversity of his art style, I don't know how to make lora.
Secondly, the quality of the images of his work that I could find on the web was unprecedentedly bad - the average pixel count was almost 600 X 600. And the compression is not good enough. Some images do not even reach 400 pixels in length... I gave up.
But today, because of something, as a change of mood, I made this lora. Firstly, I collect the images and did a basic repair of the image. Zooming in on small images and looking to improve the detail and quality a little. Then, I tagged his more realistic-looking works with "realistic", so if you want to create a more realistic-looking image, try writing "realistic" too. What are the rest of the styles? ....... I don't know. Maybe try "cartoon", or "source_cartoon", "flat art" anything you can think of. Even removing "score_9, score_8_up, score_7_up," and so on.
Then When I run the program, and the program gives me terrible feedback values, too. I thought the lora might fail, but luckily the file was tested and it worked. Realistic style images are generated more finely, as this is the type of image where I find some large images of passable quality. But the rest of the images are very, very, very bad, so don't even think about having high detail images.
Well, that's it.
描述:
训练词语:
名称: Rockin'_Jelly_Bean_Style_PonyXL.safetensors
大小 (KB): 166554
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success