Comment on Stable Diffusion 3 Medium Fine-tuning Tutorial — Stability AI

istanbullu@lemmy.ml ⁨1⁩ ⁨month⁩ ago

• Even_Adder@lemmy.dbzer0.com ⁨1⁩ ⁨month⁩ ago

  I don’t think so. They’re going to have to do a lot better than a tutorial to win people back.

  source

  • istanbullu@lemmy.ml ⁨1⁩ ⁨month⁩ ago

    kohya now supports flux fine tuning. I have seen nice examples in civitai.

    source

    • Even_Adder@lemmy.dbzer0.com ⁨1⁩ ⁨month⁩ ago

      Those might just be LoRA merged models, not full fine-tuning. From what I heard, fine-tuning doesn’t work because the models are distilled. You’d have to find a way to undistill them to train them.

      source
  • clb92@feddit.dk ⁨1⁩ ⁨month⁩ ago

    People have been training great Flux LoRAs for a while now, haven’t they?

    source

    • Even_Adder@lemmy.dbzer0.com ⁨1⁩ ⁨month⁩ ago

      Last I heard, LoRAs cause catastrophic forgetting in the model, and full fine-tuning doesn’t really work.

      source

      • clb92@feddit.dk ⁨1⁩ ⁨month⁩ ago

        Oh well, in practice I’ll just continue to enjoy this (possibly forgetful and not-fully-finetunable) model then, that still gives me amazing results 😊

        source
      • erenkoylu@lemmy.ml ⁨1⁩ ⁨month⁩ ago

        quite the opposite. Lora’s are very effective against catastrophic forgetting, and full finetuning is very dangerous.

        source