This is my first published LoRA, so please don't be too harsh! Naturally, some people will like it while others might not appreciate it — and that's completely okay!

For image generation, I used a standard basic workflow without any additional custom nodes, except for the SeedVR2 upscaler (though you can use whichever upscaler you prefer)!

I also tested various sampler combinations: res_multistep/simple, dpm++sde/beta, and res2s/beta. In my opinion, dpm++sde/beta delivers the best results!

The model strength works from 0.60 and above. I personally used 0.80!

This model also works well in combination with other LoRAs and can achieve truly excellent results when stacked. However, for the examples shown, I used only this model to avoid any confusion about image quality!

The trigger word is, of course... vortex