Focusing on interaction between giant and the objects in city, the giant size is expected to be relatable with surrounding building. So you should not use this LoRa if you want to generate very large scale giant.

Trained with around 120 images, mainly consist of giant male (human & furry), with a part of image being giantess / giantess futa.

Prompt:

giant male, 1giant, building size

I labelled the images that has 2 giants as "2giants", but they are much harder to generate.

You can add prompts to describe the position of giant in the city. For example:

sitting on crosswalk, sitting on building, leaning on tower

Also blended concept of building ??? into the Lora, refer to this Pony Lora for sample.

Prompt:

male ????????????, holding building, ??????? building, building penetrated

For ???????????? using vehicles, use the following prompt:

male ????????????, holding vehicle, ??????? vehicle, vehicle penetrated

and add the object:

holding train, ??????? train, train penetrated

This blended concept is not stable and its performance varying on base model, for example, I could not generate it on coco-illustrious at all.

Trained with Anyillustrious https://civitai.com/models/886917?modelVersionId=992802