LORA trained on 70 publicly available images of Dafne Keen. Trained on base SD 1.5, should work with most realistic models. Use ADetailer if details are lost and strength of 1 should work on most prompts. V2 LoRA trained on 38 publicly available images of Dafne Keen. Trained on base SD 1.5, should work with most models.Use ADetailer if details are lost and strength of 0.74 should work on most prompts. Use lower strength for other models. (23, 15 images) - (24, 16 repeats). Picked the 7th epoch. Consider leaving a review if you like the model. All feedback is appreciated. Note: For the images without prompts in the details, I am using ELLA for better consistency/details when generating in ComfyUI. You can find the workflow itself attached to each image - the trigger word was wrong in some but the lora worked regardless.

Files with SHA256: 6629982a83f6693331e055e6740b2826e82b85961896e573d5234936f5704315