OmniHuMo

AnyMo framework generates human motion from diverse inputs

Researchers have introduced AnyMo, a novel framework for generating human motion conditioned on various modalities like text, speech, and music. This approach utilizes a masked modeling transformer and a motion tokenize…