Sign In

Miku [Twitch Streamer] (LoRA) - SD1.5

21
219
6
Verified:
SafeTensor
Type
LoRA
Stats
219
0
Reviews
Published
Feb 28, 2024
Base Model
SD 1.5
Usage Tips
Clip Skip: 1
Strength: 0.8
Trigger Words
dathi
Hash
AutoV2
960BA39DB0

Miku by π“π“žπ“‘π“žπ““π“¨

This LoRA was trained with 100 images of the Twitch Streamer Miku "AmandaThi", each sized 768x768.

The preview images were generated using the model epiCRealism - pure Evolution V5

Settings used in the preview images:

  • LoRA Weight: (Operates from 0.7 to 1.0) - Recommended: (0.8 to 1)

  • Default Resolution: 768x768

  • Recommended Resolution: Depending on what you want (NOTE BELOW)

  • Sampler: DPM++ 3M SDE Exponential

  • Steps: 20

  • Cfg Scale: 2 (Depending on your model)

  • Keyword: dathi

  • Other trigger: makeup, eyelashes, eyeliner,

Note:

This model works well at most resolutions, so you won't have many problems. However, in case you get distorted or warped faces, simply try in other resolution. In the preview images, you'll notice that lower resolutions are used for close-up faces, while higher resolutions are used for full-body shots. So, always try to play around the resolution and give a try to this LoRA.

HAVE FUN :)

Find my other LoRAs HERE! :)