This model generates candid images of people in a Target store. The model is trained on images of people's backs facing the camera, but can produce people facing in any direction. It was trained mostly on the backsides of women in Target stores. Asses from small to large sizes were in this dataset, so you can use a range between any. Additionally, shorts and leggings were included, so either can be generated. For shorts, butt cheeks can be hanging out of the shorts if written into the prompt.

This is my first LORA, so I apologize for the jankiness. Please see below for my findings for best results:

Model: This LORA was trained using BigLove as a point of reference. For best results, I would suggest using that model. If using with other checkpoints, I would suggest trying at a lower weight like 0.5.

Strength: 0.6-0.8. From my experience, the LORA is pretty strong and high weights can distort the results. I would try at 0.8 and lower it.

Prompt examples:

Target store, candid, indoors, from behind, a woman in a short pink shorts is standing in a store, tight shorts, huge round butt, sneakers, socks, low angle

Target store, candid, indoors, from behind, a woman in tight blue shorts is standing in a store, small round ass, small body, petite, close-up

Prompt suggestions:

You can use the suggestions above or remix the prompts in my posted examples. Trying out different angles, such as "from above", "from below", "from side", as well as close-ups can produce different results. To generate full body, I suggest including something about feet, ankles, shoes, or socks into the prompt.

I would suggest resolutions 1024x1024 and higher. Portrait, landscape, and square can all produce different results.