Optimizer: Prodigy










V1: Trained with Adamw8bit optimizer
V2: Same dataset but using Prodigy optimizer , giving increased flexibilty and more cohesive style transfer
All V2 example images also use the aiai-Glimra stylizer LoRA
A 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.
Optimizer: Prodigy
1.転載モデルは学習・共有目的のみで使用し、著作権は原作者に帰属します
2.モデルの認証は公式チャンネルでご連絡ください。クリエイターの権利保護に努めています クリックして認証
