IMPORTANT: Please use the trigger word and check the sample images for prompts.
You can leave a comment/ post your images to the model page if you are facing issues and I will help in any way I can.
Do not leave bad reviews without reaching out first.

LORA trained on 52 publicly available images of Mia Khalifa.
Trained on base SD 1.5, should work with most models.

Use ADetailer if details are lost and strength of 0.7 should work on most prompts. Use lower strength for other models.

Will produce ?????? unless included in negative prompt.
(24, 20, 8 images) - (22, 18, 8 repeats) - the last one is 768x512 images. (needed landscape aspect ratio for some of the training images).
Picked the 7th epoch. The likeness is better in the later epochs but there are also signs of overfitting. Let me know if you guys want the other epochs, i can upload them as a different version.
Consider leaving a review if you like the model. All feedback is appreciated.