<ul><li>V1: Trained with Adamw8bit optimizer </li><li>V2: Same dataset but using Prodigy optimizer , giving increased flexibilty and more cohesive style transfer</li></ul>All V2 example images also use the <a target="_blank" rel="ugc" href="https://civarchive.com/models/705653/flux-detail-and-style-aiai-glimra-lora">aiai-Glimra</a> stylizer LoRAA Flux LoRA based on Disco Diffusion images, mainly spaces such as scenery and buildings.Toss any sort of prompt at it and see what sticks. Should work with very short as well as elaborate prompts.Trained in AI Toolkit on Flux Dev with ~20 LLM captioned images.Trigger is aiai-DiscoSpaces, but works without it.If you like what I do, feel free to tag me on social media; I'd love to see your creations.

Files with SHA256: 3d09acfdb88497986fe79bf02e91c7555b60d093ba4c13d17e4ac319864460f9