I'm getting weird results with this LoRA, way too many aerial views, even though there aren't many in the images, so, it will be a test version.
Forced to play with the negative prompts and sometimes AI doesn't even respect them.
Trigger words: dr3amc1ty, city, detailed, realistic screencap
Negative prompts: (high view), (aerial view), (bird view)
Use "fantasy town" instead of "city" and (street view), glass tower, glass building...
Strenght: 0.7 -> 1
