This LoRA aims to recreate the ambiance of 1980s from Japan. It has been trained on Modelscope using a curated set of 100 publicly available images depicting Japan's fashion trends, architecture, technology and urban life.

The images used in the training dataset, have a vintage, grainy, flat look to them; furthermore, most of them are captured by either reporters or hobbyist travellers (barring the fashion portfolio pics, and there too, those are mass media pics, not your Vogue types). The LoRA has been specifically trained to internalize this look as a feature. So, let's not expect extremely crispy digital images.

The LoRA aims to be faithful to the 1980s look, in its entirety.

Feel free to give it a try.