Wan-dancer

Instructions

Upload a portrait image (9:16 vertical, full-body/half-body facing front) and a music file (wav/mp3, 10s-30s).
Select a dance genre and customize generation parameters if desired, then click Generate.
ZeroGPU Optimization: To stay within Hugging Face ZeroGPU time limits (120s max), keep Inference Steps around 12-15 and Max Audio Duration around 10-15s.

Portrait Image (will be auto-cropped to 9:16 vertical)

Music (wav or mp3)

Max Audio Duration (seconds)

Video generation is computed sequentially in 5-second segments. Standard length is 10s.

10 30

Audio Start Time (seconds)

Choose where in the song the dance should start (e.g. when the beat drops).

0 120

Dance Genre

Generated Video

Status / Log

Examples

Portrait Image (will be auto-cropped to 9:16 vertical)	Music (wav or mp3)	Dance Genre