Trained on 34 images.
The eyes are wrong this needs more training
I agree in the example images, but may be worth trying in other models and/or playing with the weight.
ch4rl13hunn4m.safetensors
