LoRA ผู้ใช้
cloudlora
Trained on 25 images of figures in clouds generated by dalle3. Works great between 0,5 and 1