This LoRA is designed to generate a viewpoint in which the character's penis, torso, and face fit within the frame, making it appear as though the viewer is looking up while performing fellatio on that character.

The dataset consists of 108 images generated using ControlNet to control the generation process. Each image was generated using line art that I manually composited in Photoshop. The goal is to achieve consistent angle and composition across a various female/futanari body types.

While it was entirely trained on futanari-focus images, it could likely be used to generate male-focus images as well. Although, I'm not sure the composition would be as consistent (need further testing).

Trigger Prompts:

fellatio taker pov, standing, 

fellatio taker pov, sitting, 

Customization tags:

arms behind back, outstretched arm, outstretched arms, saliva on penis, ejaculation, cum on penis, veiny penis

Tips & Known Quirks:

• In the prompt, you can omit the following tags: taker pov, implied fellatio, straight-on, from below, close-up, foreshortening. As these tags were used to generate each image in the dataset, but omitted during the training process. So in theory, the model should essentially recognize the trigger words: fellatio taker pov as these tags.

• incoming after more test...