MMExit: Enabling Fast and Efficient Multi-modal DNN Inference with Adaptive Network Exits.

Xiaofeng HouJiacheng LiuXuehan TangChao LiKwang-Ting ChengLi LiMinyi Guo
Published in: Euro-Par (2023)
Keyphrases
  • multi modal
  • audio visual
  • cross modal
  • multi modality
  • high dimensional
  • image annotation
  • fusing multiple
  • machine learning
  • image segmentation
  • data sources
  • training process
  • video search