Verified: 2 years ago
SafeTensor
Type | |
Stats | 749 0 |
Reviews | (153) |
Published | Apr 8, 2023 |
Base Model | |
Training | Steps: 1,800 Epochs: 1 |
Trigger Words | 6o9 |
Hash | AutoV2 0C67C4304F |
This LORA is based on Bocchi's classic 6o9 face.
Trained on anythingV5, but other anime-style models should also worked.
Best result by weighting around 0.8-1.1, lower will not generate 6o9 face and higher the image will be distorted.
Known issues: the hair will automatically become pink if weighting more than 1, no matter what prompts is used. I will fix it in later version.
It's a LOCON model, so make sure you already have LOCON installed.