This LoRA is trained on a black MMA/BJJ grappling dummy (zkxdummy) with multiple controlled body positions.
The model focuses on maintaining a stable object identity across different poses rather than generating a character.
It supports multiple body configurations:
- Standing fighting stance
- Sitting position (neutral posture)
- Lying supine (on back)
- Lying prone (face down)
Trigger word: zkxdummy
Recommended usage:
Use the trigger word together with pose descriptions such as:
"standing fighting stance", "sitting position", "lying prone", "lying supine"
This LoRA is best used for:
- MMA / BJJ visualization
- martial arts concept art
- grappling position studies
- sports illustration workflows
- pose reference generation
It is not a character LoRA, but an object-based multi-pose training model designed for consistent structural control.
Side note : its my first LoRa and i'm proud of myself lol. enjoy !
Description
First release of a black MMA/BJJ grappling dummy LoRA trained on a consistent reference background and multiple controlled poses.
The model focuses on maintaining a stable object identity across different body positions rather than generating a character.
Supported poses:
- Standing fighting stance
- Sitting position
- Lying prone (face down)
- Lying supine (on back)
Designed for pose control, martial arts visualization, and sports concept generation.
Trigger word: zkxdummy
Comments (1)
Nice, keep the good work.