MotionRL: Align Text-to-Motion Generation to Human Preferences with Multi-Reward Reinforcement Learning

Documenti analoghi

• Learning Generalizable Human Motion Generator with Reinforcement Learning
  di: Mao, Yunyao, et al.
  Pubblicazione: (2024)
• MASA: Motion-aware Masked Autoencoder with Semantic Alignment for Sign Language Recognition
  di: Zhao, Weichao, et al.
  Pubblicazione: (2024)
• Motion-aware 3D Gaussian Splatting for Efficient Dynamic Scene Reconstruction
  di: Guo, Zhiyang, et al.
  Pubblicazione: (2024)
• Enhancing the Outcome Reward-based RL Training of MLLMs with Self-Consistency Sampling
  di: Wang, Jiahao, et al.
  Pubblicazione: (2025)
• ReAlign: Text-to-Motion Generation via Step-Aware Reward-Guided Alignment
  di: Weng, Wanjiang, et al.
  Pubblicazione: (2025)