
Prompts
-
hiodoshi_ao, blue_hair, medium_hair, blue_eyes, eyeliner, single_side_up
-
otonose_kanade, blonde_hair, long_hair, red_eyes, side_ponytail,
-
ichijou_ririka, red_hair, long_hair, red_eyes, two_side_up, eyeliner
-
todoroki_hajime, blonde_hair, medium_hair, purple_eyes, beret
-
juufuutei_raden, black_hair (or brown), long_hair, blue_eyes, twintails, multicolored_hair
Weight: 0.5 - 0.8
Old stuff:
(applies only to the initial versions) I wanted to see how see how low dataset amount will effect the quality.
Below are the amount of images for each lora:
-
hiodoshi_ao: 23
-
otonose_kanade: 10
-
ichijou_ririka: 13
-
todoroki_hajime: 10
-
juufuutei_raden: 10
Each lora are subjected to 10 image repeats and 8 epoch training. Although, I will upload the epoch 7 for most (only juufuutei_raden will be of epoch 8).
Note: I have a history of uploading the wrong lora file, so.. do tell if I repeat the same mistake ^^".
Note (8/7): yea... updating later on lmao. Will wait for new outfit images to appear.
Note (9/12): updated, I'm quite happy with it. If this one is still lacking for you, I suggest to wait a little longer. I'm sure better versions will pop up with more images becoming available.
描述:
训练词语: otonose_kanade
名称: otonose_kanade-07.safetensors
大小 (KB): 18551
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success