Model to fake images of Phillip The 2 artist
Trained on Pony6, but applicable on other models. Trained on 2 batch size, then increased to 8 to improve speed as an experiment (seems to work fine)
Model doesn't need any trigger word, but it was tagged: phillipthe2
Important: not every image generated would be fine, many deformed and weird may come along, so to get 1 good you may (or may not, it depends) need to generate 2..3..5..7 on Pony. More styled models work better, but lora style itself slightly changes if that's the case. Reason for the above is very limited data set (only 48 images including crops).
Description
Model to fake images of Phillip The 2 artist
Trained on Pony6, but applicable on other models. Trained on 2 batch size, then increased to 8 to improve speed as an experiment (seems to work fine)
Model doesn't need any trigger word, but it was tagged: phillipthe2
Important: not every image generated would be fine, many deformed and weird may come along, so to get 1 good you may (or may not, it depends) need to generate 2..3..5..7 on Pony. More styled models work better, but lora style itself slightly changes if that's the case. Reason for the above is very limited data set (only 48 images including crops).