Trigger Word = DCBDLNDS.
. This lora was trained on my old bing images a dataset of 217 images @10,000 Steps Rank 32, If you cant get what you want try dropping the weight of lora to .65, Captioned with gemini3Flash
. More CGI style than realism (I need to sort out how to get this better for v2 as sample pictures are not what i expected"rushed lora"), excels at minotaurs and other animals, Warriors etc. Warriors can sometimes have painted lines over there body without prompting for it so i wll try weed out in v2 of the model when i get time, Dragons are average!
. Here's 2 old video's i made which will give you an idea of what this lora is about:
https://www.reddit.com/r/aivideo/s/YOeqqBbfu5 and https://www.reddit.com/r/aivideo/s/hwF8rrW2eM which were made over 2 years ago.