privcdd/lorawan

gr3ntwi Wan 2.2 Character LoRA

Prompt

-

Model description

Character LoRA for gr3ntwi trained on Wan 2.2 I2V 14B (dual transformer) using ostris ai-toolkit.

Trained for 3000 steps at learning rate 0.0001 with sigmoid timestep, EMA enabled, rank 32. Designed for use with WanAnimate workflows.

Includes both high_noise and low_noise variants. For best results, load both in your WanAnimate workflow — low_noise at strength 0.85 (identity), high_noise at strength 0.5 (motion).

Trigger words

You should use gr3ntwi to trigger the image generation.

Download model

Download them in the Files & versions tab.