LoRa is not very stable in motion yet for generating videos of dressed and undressed characters Because the training was done on 30 images, LoRa likes to become static. Perhaps I'll try to retrain LoRa with videos of different lengths. Prompt template: Two adjacent identical videos located side by side about [Here we need a description of the girl's appearance and background.]. On the left she is [Here you need to describe the clothes of the dressed girl]. On the right she is completely naked.

Files with SHA256: 560427c74a28e7156eaad4bbc99a6a1496fce98e5304d6606195731fe6d060a6