This is a LoRA trained on about 80 images of Negev from Girls Frontline, the results are quite mixed and likely I didn't use the best training parameters but it's at least better than base model :)