WAN2.1 HuMo Изображение&Аудио в Видео -KJ

HuMo is an open‑source video generation project from teams at Tsinghua University and ByteDance, designed around human subjects.

• Highlights:
• Accepts multiple inputs: text, images, and audio
• Produces high‑fidelity, controllable human videos
• Strong adherence to textual instructions
• Maintains subject identity across frames; accurate audio‑driven motion
• Flexible pipelines: text+image, text+audio, or text+image+audio
• Exports up to 720p resolution
• Suited to content creation, virtual avatars, education, and artistic work