Sign In

六合一 SIX IN ONE BanG Dream! It's MyGO!!!!!

83
804
2.6k
13
Updated: Feb 2, 2024
character
Verified:
SafeTensor
Type
LoRA
Stats
536
2,564
Reviews
Published
Feb 2, 2024
Base Model
SDXL 1.0
Trigger Words
fanstyle
Hash
AutoV2
C9FE997FA3
Chenkin's Avatar
Chenkin

六位角色合一的lora,4dim训练

1.5版本更新

1.可以实现番剧画风和其他画风

触发词::fanstyle

2.更稳定的多人图

1.0版本更新:

触发词:

素世:Nagasaki Soyo,

猫猫:Kaname Rāna,heterochromia

灯:Takamatsu Tomori

祥子:Togawa Sakiko

立希:Shiina Taki

1.0正式版做出的变化:

1.训练素材

共2407张番剧截图,其中:

427 灯 466 爱音 315 猫猫 328 祥子 414 立希 457 素世

2.1将角色词进行前置,但将角色词放最前面

2.2 对角色特征进行删除,但保留了服装tag,便以换装

2.3 对词语重复的tag进行删除

3.多人

此版本相比上版本出双人图有比较大的进步,但仍存在混特征的情况,暂不能稳定出双人以上(比如三人)的图


-0.9版本因为是全标训练,需要手打角色特征,改进中

爱音:Anon Chihaya,pink hair, grey eyes,necktie,school uniform

素世:Nagasaki Soyo, blue eyes, brown hair, school uniform

猫猫:Kaname Rāna, grey eyes,yellow eyes,heterochromia,black shirt,off shoulder

灯:Takamatsu Tomori,brown hair, brown eyes,grey jacket,school uniform

祥子:Togawa Sakiko ,hair ribbon,long hair,yellow eyes,grey hair, yellow ribbon, long sleeves

立希:Shiina Taki , black hair, mole under eye, purple eyes, school uniform, jacket, purple necktie

训练素材支持:CyberHarem (DeepGHS CyberHarem) (huggingface.co)

Training model: Animagine XL V3,

Inference model: Animagine XL V3

Training resolution: 1024*1024

Recommended output resolution: 1024*1280

Recommended model weight: 0.8~1