Verified: 9 months ago
SafeTensor
Type | |
Stats | 45 0 |
Reviews | (1) |
Published | Jul 22, 2024 |
Base Model | |
Training | Steps: 448 Epochs: 12 |
Usage Tips | Clip Skip: 1 |
Hash | AutoV2 3C88EB9779 |
Created on Civitai
This is LoRA for Bollywood sweetheart Kiara Advani.
Based on the training and my testing, these are the settings I have found to have given me the best results:
Model: NightVisionXL
Sampler: DP ++2M Karras
Steps: 20-30
CFG Scale: 6-7
Also found the following negative prompts to be useful based on the issues with training data and model bias:
long face, triangle face shape
Obviously feel free to experiment with other photorealistic models and do share your work on this page so I can appreciate them too :)