EDGE
Click to visit website
About
EDGE is a powerful tool developed for editable dance generation from music, enabling realistic and physically-plausible dance creation while being responsive to arbitrary music inputs. Powered by a transformer-based diffusion model combined with the Jukebox music feature extractor, EDGE allows significant editing capabilities including joint-wise conditioning, motion in-betweening, and dance continuation. It has been tested against recent methodologies and shows a strong preference from human raters for the dances it generates. EDGE encodes music into embeddings using the Jukebox model and can create multiple sequences of dance clips, ensuring temporal consistency for full video generation. It incorporates constraints for lower and upper body generation, emphasizes physical realism, avoids foot sliding, and applies a novel Contact Consistency Loss to enhance realism in its outputs. The design of the website draws inspiration from other exemplary designs.
Platform
Task
Features
• motion in-betweening
• dance continuation
• joint-wise conditioning
• transformer-based diffusion model
• physical realism with contact consistency loss
• generates arbitrary-length dances
• uses music embeddings from jukebox
• editable dance generation
Average Rating: 0.0
Average Rating: 0.0
5 Stars:
0 Ratings
4 Stars:
0 Ratings
3 Stars:
0 Ratings
2 Stars:
0 Ratings
1 Star:
0 Ratings
User Ratings
No ratings available.
Sign In to Rate this Tool
Alternatives
freebeat AI Workout App
AI-powered dancing app that generates personalized dance routines synced to your favorite music tracks.
View DetailsFeatured Tools
Scalytics Connect
A federated AI framework that integrates decentralized data sources for AI development.
View Details