Created with Google Colab and uses at around 40+ images of long hair model photos as datasets. Came with 2 different epoch levels, 10 Epochs and 20 Epochs. User may need to adjust the LORA strength below than value 1.

Works best when using Dreaming_In_Digital or DreamShaper model (see suggested models below) and also enabling these feature to get best image quality as possible:

1. Go to Settings> Optimizations and enable FP8 Weight (cuts down memory usage much further to 2.5-3GB VRAM, requires capable GPUs eg. Nvidia RTX series)

2. Also enable Cache FP16 Weight for LORA as well.