This Hunyuan Video LoRA model is portraying a first-person boxing point of view, providing an experience where the viewer sees the action through the eyes of the boxer perspective.

The model was trained using the method I outlined in this article.

This LoRA is also compatible with character LoRAs, enabling integration with custom character LoRAs seamlessly into the first-person boxing scenario (Hopefully \m/), try it yourself.

Suggestion:

LoRA Strength: 1

Sampler/Scheduler: DPM++2M/Beta.

Guidance: 7

Model Sampling: 7

No Trigger Words.

Base Prompt:

First-person POV of a [man/woman] boxing directly toward the camera, as if sparring with the viewer. [Her facial expression is determined and focused]. [She wears a peach-colored sports bra, dark navy high-waisted leggings, and gray boxing gloves]. Her movements are energetic. With each punch exchanged, the camera shakes slightly in sync, adding a cinematic realism to the motion. The handheld camera follows with smooth, natural movement, capturing her dynamic footwork and fluid body motion. [Her physique is athletic, with medium-sized breasts]. [The background is a bright, well-lit outdoor paved area with soft greenery in the distance]. The mood is energetic and immersive, as if the viewer is inside a live-action scene.

You can modify the prompts.