This LoRA was trained on SDXL 1.0 and uses the same training images as my first Jonvelle LoRA. The training images in the first LoRA were rather small, this LoRA uses larger training images and should give better results.Thank you to @countlippe for pointing out the need for better training images.

Files with SHA256: 42af81513573a5b2d6e5a64f31e2427b0fba1836248cf07295315f175b20d6d4