The Gothic style LoRa was trained on the SDXL Illustrious model. In the learning process, a dataset of 42 images generated by me was used. The training took 6300stages (15 repetitions for 10 epochs), loss: 0.144, Dim-64, Alpha-32, Sampling Method-DPM++ 2M Karras.

Training prompt: "(Anime style illustration: 1.3), young woman (ethnicity: 1.2), (age: 1.1), with black hair, red eyes and pale skin (facial features: 1.3), wearing a black blazer, white ruffled shirt, black trousers and black shoes (detailed description of clothing:1.2), (accessories: red beret, sword, skeleton wings: 1.1), holds a burning sword, stands over a pile of skulls, (ferocious facial expression: 1.2), (dynamic pose: 1.2), taken from a low angle, looks down, central frame, red background, fire elements, spectacular lighting, digital graphics, clear focus, bright color palette, 8k resolution, artstation quality, (detailed skulls: 1.1), (blazing fire: 1.3), (detailed wings: 1.3)

Thus, the presented LoRa demonstrates a very specific Gothic punk style. As the influence increases, the characteristic detail increases, and the figure of the main character becomes more stylish and leggy, and the poses become more interesting. Try my new LoRa. I wish you a pleasant creativity.