Full-duplex speech models post-trained with reinforcement learning for improved conversational interactivity.
-
Multi-Faceted Interactivity Alignment in Full-Duplex Speech Models
Paper • 2606.11167 • Published • 2 -
kyutai/moshika-rl-seamless
Audio-to-Audio • 8B • Updated • 20 • 14 -
kyutai/personaplex-rl-seamless
Audio-to-Audio • 8B • Updated • 18 • 9 -
kyutai/interactivity-alignment-samples
Viewer • Updated • 5.56k • 1.63k • 6