About

Images generated using AOM3 and Cetus-mix models.

• Trained on AOM3 model using fast kohya trainer colab

• Very small dataset for training

• Recommended weight: 0.6-0.8

• Highly recommend using high res fix

Author Notes

• This LoRA was made with the intentions of trying out pruning tags for training and whether or not a very small dataset can still get relatively good results.

• I would be more than thrilled to have another person make a better LoRA in the future.

描述:

训练词语: tlok_asami_sato,main_outfit,green eyes,black hair