Modify the weights of the LoRA to achieve the desired result. The base model for SD 1.5 will look the best. However I had great results with Dreamlike Photoreal, and Realistic Vision as well, and various blends.
This was also trained with the expanded VAE located here. Please use this VAE when inferring with this LoRA.
Description
Similar to the Lauren Boebert LoRA, faces might not look great at a distance, so you'll have to inpaint them. Hands/arms are still weird. Proportions in general need some work.
Changelog:
Initial release
Training Data:
Used bmaltais's implementation of kohya_ss LoRA training scripts.
Total Images: 15 total high resolution images
Regularization Images: 300 total "sexy woman" regularization images from ProGamerGov's dataset at huggingface.
Trigger word: tomi lahren woman
Training Steps: ~2700
Unet Training Rate: 5.0e-06 (0.000005)
Text Encoder Training Rate: 5.0e-05 (0.00005)