This is a collection containing multiple variants for a Character LoRA. See Description for more information. This version was previously called SangoV2.5.

If you like my work and would like to leave a review, I mean.. feel free to. I do this for purely for fun - nothing else.

Trained on The Deep Dark v6.0

Dataset is from Inuyasha: The Final Act. Tagged backgrounds. Dataset consists of 188 images. Trained on site, used 25 epochs, 7 repeats, 4 batch size or 8225 steps. It is overtrained. As a result of overtraining, bath scenes are dark.

Weight/Setting Suggestions (Estimate)

• Lora Weight: 0.5 ~ 0.8

• CFG Scale: 5 ~ 10

• Sampler: Euler A, DPM++ series (unsure)

• Steps: 20 ~ 40

• Denoise: 0.5 ~ 0.75

• Resolution: 512x512 (use upscaler/hires)

See the gallery for prompt guidance. I have provided some x/y/z plots in the gallery, should give you an idea.

Prompt Usage

The trigger word is Sango. I didn't use unique tags for clothing. Here are two prompt examples. Hair color, breast size, eye color, hair style can be modified. Hiraikotsu wasn't tagged.

Kimono Outfit

Sango, 1girl, solo, brown eyes, brown hair, black hair, long hair, low-tied long hair, makeup, bangs, kimono, skirt, green skirt, anime coloring, retro artstyle

Armor Outfit

Sango, 1girl, solo, brown eyes, brown hair, black hair, ponytail, makeup, bangs, armor, shoulder pads, anime coloring, retro artstyle

Be sure to make use of quality tags, body tags, expressions. Have some decent negatives. You can reference my gallery for examples.

Anything else I didn't mention is probably not worth mentioning.