Speech Gesture Generation on TED (test)
22.083FGDJoint embedding model
Evaluation Results
| Method | Links | |
|---|---|---|
| Joint embedding modelModality=Text, Output Poses=30, Seed Poses=42020.09 | 22.083 | |
| Speech2GestureModality=Audio, Output Poses=30, Seed Poses=42020.09 | 19.254 | |
| Attentional Seq2SeqModality=Text, Output Poses=30, Seed Poses=42020.09 | 18.154 | |
| Trimodal Gesture Generation ModelModality=Trimodal (Text, Audio, Speaker Identity), Output Poses=30, Seed Poses=42020.09 | 3.729 |