This LoRA allows you to make images of drivers with their driver airbag inflated.
If you're wondering what "Bing Style" in the title means, essentially Bing's AI Image Creator handled the concept of airbags way better than Stable Diffusion does. However, it started to get overly sensitive with flagging content. So I trained this model using a collection of the best images I gathered from Bing's AI Image Creator.
While it's hit or miss, general phrases it's trained on and understands are "hugging the airbag", "face pressing against the airbag", "smoke coming out from the dashboard", and "pushing his/her hand against the airbag".
It generates good results between CFG: 3-5, 30 steps seems adequate enough, and all ckpts and renderers seem fine, but the best result (imo) has been using either Juggernaut XL or EpicrealismXLv7, karras and dpmpp_3m