Cover image suffices what this LORA does
There are versions which performed better in output testing than others, so here are a few recommendations:
-.-.-.-(V1.5)-.-.-.-
Should generate a more diversified output, considering the dataset has more variation.
-.-.-.-(V1.6.1 & 1.6.2)-.-.-.-
For anyone trying to recreate the cover. Add "squatting, from side" for better results.
Enjoy
