This version was trained for 16 epochs. There is a good chance that it is over-fit, but that might be what you want with this LoRA. I plan to upload some of the earlier checkpoints as well, since they also produced some good looking images.










You can use this to generate flag designs for all your piracy needs.
This LoRA has been trained on renderings of Jolly Roger flag designs.
I used the tag "jolly roger" on all training images, use that as the main trigger word. However, you can improve the results you get by using other more specific terms that were common elements in the training set, these include:
"<color> field background" (e.g. "black field background")
"death's head" or "skull"
"crossed bones" or "crossbones"
"cutlass" (although most depictions of pirate flags, including the ones in the training set, show scimitars rather than cutlasses)
This version was trained for 16 epochs. There is a good chance that it is over-fit, but that might be what you want with this LoRA. I plan to upload some of the earlier checkpoints as well, since they also produced some good looking images.
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
