A BETTER VERSION IS INCOMMING I am removing text from the dataset
as you read
Vtuber portrait Chibi
Vtuber portrait Chibi (vtubeportchibi.safetensors)
Trained with 35 threshold on characters
10th epoch
10 repeats
locon
32 / 16 dim/alpha
Best around 0.3-75 0.85~ (going higher is still fine)
00050-based64_v3_1860615929.png
MY oc
00084-HoloKukiv2.2-fp16_470034664.png
00080-HoloKukiv2.2-fp16_3991821192.png
vs
Vtuber Portrait Chibi Char (vtubeportchibichar.safetensoros)
trained with
75 trheshold on characters
20th epoch
7 repeats
locon
32 / 16 dim/alpha
Best around 0.3-75 0.85~ (going higher is still fine)
