Verified: a year ago
SafeTensor
Type | |
Stats | 3,284 8,791 |
Reviews | (472) |
Published | Jul 31, 2023 |
Base Model | |
Training | Steps: 150 Epochs: 2 |
Usage Tips | Clip Skip: 2 |
Hash | AutoV2 CB4A25F2C0 |
V1
Stylisation LoRA trained on Caravaggio's oil paintings.
face generation results are generally good under 1024,1024 resolution (use HiRes fix)
Hands are often twisted/blurry; eye generation occasionally yields crossed eyes/unmatching eye size or pupil colours. Use TI/Adetailer or inpainting to fix these issues.
recommended weight for this LoRA: 0.4-0.7