This is the first LoRa model trained by me. I wanted to find out if it was possible to train a good LoRa model without captions. I noticed that with captions the model tends to reproduce the subjects in the dataset too closely. I think I did a decent job that obviously can be improved. The base model is stable diffusion 1.5, and I have tried using LoRa on various models with satisfactory results (generally better than the base model). I would be interested if you would share the images produced and perhaps try to give me some tips to improve the results. I hope you enjoy it, please let me know.