Trained from Kedama Milk's work, based on PonyXLv6.
v02 edit:
Concentrated dataset, more stable, coloring and shading is better as well.
v01:
A bit unstable on pony's "swinging style shift", may redo the dataset and re-train in the future.
Trained from Kedama Milk's work, based on PonyXLv6.
v02 edit:
Concentrated dataset, more stable, coloring and shading is better as well.
v01:
A bit unstable on pony's "swinging style shift", may redo the dataset and re-train in the future.