Castling-ViT: Compressing Self-Attention via Switching Towards Linear-Angular Attention at Vision Transformer Inference

Documenti analoghi

• INGeo: Accelerating Instant Neural Scene Reconstruction with Noisy Geometry Priors
  di: Li, Chaojian, et al.
  Pubblicazione: (2022)
• ShiftAddViT: Mixture of Multiplication Primitives Towards Efficient Vision Transformer
  di: You, Haoran, et al.
  Pubblicazione: (2023)
• FBNetV5: Neural Architecture Search for Multiple Tasks in One Run
  di: Wu, Bichen, et al.
  Pubblicazione: (2021)
• ViT-AdaLA: Adapting Vision Transformers with Linear Attention
  di: Li, Yifan, et al.
  Pubblicazione: (2026)
• MixRT: Mixed Neural Representations For Real-Time NeRF Rendering
  di: Li, Chaojian, et al.
  Pubblicazione: (2023)