The dataset is made of anime screencaps.
I'm pretty happy with the outcome, being able to depict all the outfits i can remember. Her dresses are a little off, but close most enough of the time.
Using tags "anime screencap", or just "screencap" might help with matching the likeliness, but will impact quality to some extent.
I'm always open for feedback, so I can improve my work in the future.
An epoch with relatively good balance between likeliness to the original and overall quality, being able to depict the outfits.
Go ahead and upload yours!
Your query returned no results β please try removing some filters or trying a different term.