lucataco / flux-rplctcpl

Flux LoRA Training Experiment - Training two people in one LoRA with two images

  • Public
  • 127 runs
  • Fine-tune

About

This is a proof of concept that trains two models in a single LoRA. Featuring Replicate’s AI characters 0_1 and Violeta, the trigger word is ReplicateCouple (rplctcpl)

Training

Only two photoshopped images were used for this training, at 1200 steps:

img1

img1

Next steps

Next steps would be to generate a dozen or so images to use as a synthetic dataset to train a better LoRA