First release










Once again not entirely happy with how it turned out - probably because I neede/wanted this one to be perfect. It does a good job still and I feel like it might bring something new to the table compared to LoRA already out there, so here it is.
Refined/pruned the dataset several times and did multiple runs with varying training parameters. Due to the nature of the training images it has a moderate tendency to go - so use appropriate negative prompts to stay clear.
Recommended weight from my tests: between 0.85 and all the way up to 1, depending on the checkpoint with a CFG scale of between 5 and 7.
Most sample images were done using various checkpoints with DPM++ 3M SDE Karras (and other DPM variants as well as Euler Max) at a CFG scale around 5 at a base resolution of 512x768 with a HighRes-Fix of 2 with a denoising strength between 0.2 and 0.6.
Other settings may work better or worse - feel free to share yours!
First release
1. The rights to reposted models belong to original creators.
2. Original creators should contact SeaArt.AI staff through official channels to claim their models. We are committed to protecting every creator's rights. Click to Claim
