🎨 Krea 2 LoRA Trainer

Trains on Krea 2 RAW, optionally validates on Turbo, runs on HF Jobs, pushed to your Hub. You only pay for the job's GPU minutes.

…or paste this to your coding agent:

The name of your LoRA

Has to be unique — used for your output model repo (you/<name>) and dataset repo.

What are you training?

Drives how images are captioned and what kind of trigger is suggested.

🎨 Style 🧸 Object or character ⚙️ Custom