TopGolf Girl (WAN 2.2 T2V) is a motion-tuned LoRA trained on 700+ short clips capturing the natural movement, form, and expression of a blonde woman playing golf across various lighting conditions and outfits.
The model enhances realism, body control, and camera alignment within the WAN 2.2 T2V pipeline — producing stable, lifelike swing sequences with minimal jitter or motion collapse.
⚡ Support Future Versions (20–25 Epoch Target)
I’m continuing to train additional versions of this model for improved fidelity and motion consistency. Run or fine-tune it yourself on RunPod and help fund future releases:
https://runpod.io?ref=c3tt3sd8
💼 Want to Collaborate?
I’m available for AI model development, training pipeline builds, and creative LoRA collaborations. Message me here: linkedin.com/in/christabb
Description
Training extended to 16 epochs using WAN 2.2 A14B Dual-Phase LoRA config
Improved arm articulation, stretch motions, and upper-body continuity
Enhanced motion stability in side-view sequences and golf-swing transitions
Added subtle breathing and pre-swing posture nuance for more lifelike pacing
Maintains dual variant balance:
🟣 HighNoise — dynamic realism, increased motion detail
🩵 LowNoise — smoother portrait tone and controlled body motion
Recommended LoRA strength: 0.65–0.8, Clip Skip 2
⚡ Next up: continuing refinement toward Epoch 20–25 for complete swing-cycle mastery.
Run your own version or support future training via RunPod → https://runpod.io?ref=c3tt3sd8
FAQ
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.