I see no benefit of this over most standard Illustrious/XL models, like NovaFurry or NovaAnime, which output giantess/macro in about the same quality. Your Lora still shows the typical issues with cars, street markings, and mixed up building sizes. Maybe it still can be improved, trained with a larger variety of images (GT/GTS/Furry), to make the results more robust. If I had better hardware, I'd do it myself.