Doubled the training data size and refined the prompts. It feels like it gives more details while the prompt gets similar results to v1.0










A Genshin Impact style 'Portrait' creator that creates portrait images in the style of Genshin characters.
Has the trigger word genshinportrait, but not sure it's needed. Write in descriptive english form such as
genshinportrait, A male with short tussled black hair which gradients into dark blue tips and a pair of square spectacles. He has a deep maroon outfit withblack and grey detailing and a small, blueish-black crow perched on his shoulder. On the opposite shoulder is a black feathered armor plate. He has deep red eyes. <lora:flux_genshinportraits_000002400:1>
There's a weird issue in Flux.D that means it sometimes produces blurry images on white/color backgrounds. You can usually fix this with a new seed, or by adding something like 'plain bg' or 'plain white bg' to your prompt.
See the example images for more prompt ideas on descriptions.
Basically makes everything much more consistent and keeps to the front-on, flat style:
Doubled the training data size and refined the prompts. It feels like it gives more details while the prompt gets similar results to v1.0
1. 재게시된 모델의 권리는 원 제작자에게 있습니다.
2. 모델 원작자가 모델을 인증받으려면 공식 채널을 통해 SeaArt.AI 직원에게 문의하세요. 저희는 모든 창작자의 권리를 보호하기 위해 노력합니다. 인증하러 이동
