✦
Summary
Viggle-Animate replaces characters in a video by propagating a change made to one of the video frames. It takes a driving clip and a repainted reference frame, aiming to preserve the original movement, camera, and timing while transferring the appearance of the new character throughout the shot.
Key Features
The model takes a driving video and one repainted frame from that video.
The video stage uses fixed conditioning rather than a user-supplied text prompt.
Motion, timing, and camera behavior are inherited from the driving clip.
No separate pose estimator is needed in the video-generation stage.
Character replacement does not require a separate segmentation-mask channel.
The distilled sampler uses three forward passes in the upstream setup.
It fine-tunes the 33.1B MiniMax-H3 reference-to-video Transformer.
The model card links community ComfyUI nodes for the workflow.
About Viggle-Animate
The model is a 33.1B full fine-tune of MiniMax-H3 reference-to-video architecture, distilled to three forward passes. Because the edited reference already matches the source pose and framing, the video stage does not need a separate pose estimator, segmentation model, face tracker, or text encoder.
Viggle-Animate suits animation and character-replacement workflows where a prepared reference image is available. The release includes weights, examples, and links to community ComfyUI integration. Reported speed measurements use a B200 GPU; the MiniMax-H3 community license governs usage and should be reviewed for deployment.