Verified: a year ago
SafeTensor
Type | |
Stats | 316 966 |
Reviews | (67) |
Published | Mar 13, 2024 |
Base Model | |
Training | Steps: 2,000 |
Trigger Words | ankha |
Hash | AutoV2 DA0F56BC6C |
Ankha LoRA trained on NovelAI with Kohya training scripts using the DAdaptAdam optimizer at 512x512 resolution. Works great, main issues are her necklace appearing randomly and tail placement and count. Can sometimes add bracers and bracelets without prompting, add (bracelet, armband) to your negative if you run into that issue. For her default outfit, add (white tunic, necklace) or a similar variant to your prompt. Should have wide compatibility with model mixes based on NovelAI, which is most of them.