ERNIE ImageI used the same dataset as the Flux lora. I trained this locally on my GPU using Ostris AI toolkit.FLUXLightning lora trained on Flux.1 DSimilar dataset as SD1.5. I did not resize images to 512x512. They are their original sizes from web searches. Updated captioning using Florence. Reviewed and manually edited the captions. Lora was trained on <a target="_blank" rel="ugc" href="http://fal.ai">fal.ai</a> 1000 steps at 4e-4 training rate. This was mostly an experiment to see how training a flux lora would work.Use lora strength of 0.5 to 0.7. But feel free to experiment.There is no specific trigger word. But it's helpful to include "Lightning from Final Fantasy" in the prompt.SD 1.5Use weight of around 0.67 for best results.LoRA trained on SD 1.5.Images trained on game official art and ingame screenshots. Note the "cat ears" in the negative prompt. I only had one training image where Claire had cat ears, yet it spontaneously shows up more often than not in the generated images. Might be a positive for some people. It is necessary to add "pink hair" to the prompt to get the Lightning look.Images uploaded were generated using GalaxyTimeMachine Photo model. However, also works well with DreamShaper and chilloutmix.

Files with SHA256: b3db33eade21b3a1837d9e513b8729e3a5065451b20e32b9de689953682ea13f