AI Research 技术调研

标签: multi-view-transformer

此标签下有4条笔记。

  • 2026年7月16日

    RVT: Robotic View Transformer for 3D Object Manipulation

    • manipulation
    • multi-view-transformer
    • orthographic-rendering
    • point-cloud
    • virtual-views
    • keypose-prediction
    • rlbench
    • behavior-cloning
    • language-conditioned
    • view-based-policy
  • 2026年7月16日

    3D-MVP: 3D Multiview Pretraining for Robotic Manipulation

    • 3d-pretraining
    • masked-autoencoder
    • multi-view-transformer
    • rvt
    • objaverse
    • representation-learning
    • rlbench
    • colosseum
    • sample-efficiency
  • 2026年7月16日

    RVT-2: Learning Precise Manipulation from Few Demonstrations

    • manipulation
    • multi-view-transformer
    • coarse-to-fine
    • point-renderer
    • keypose-prediction
    • rlbench
    • few-shot
    • precision-manipulation
    • behavior-cloning
  • 2026年7月16日

    SAM-E: Leveraging Visual Foundation Model with Sequence Imitation for Embodied Manipulation

    • manipulation
    • segment-anything
    • visual-foundation-model
    • lora
    • multi-view-transformer
    • action-chunking
    • rlbench
    • few-shot-adaptation
    • behavior-cloning
    • keypose-prediction

  • GitHub