vid-TLDR: Training Free Token merging for Light-weight Video Transformer.
Joonmyung ChoiSanghyeok LeeJaewon ChuMinhyuk ChoiHyunwoo J. KimPublished in: CoRR (2024)
Keyphrases
- lightweight
- multimedia
- video data
- fuzzy logic
- video sequences
- video content
- wireless sensor networks
- training process
- key frames
- video streams
- fault diagnosis
- space time
- video processing
- training set
- real time
- video database
- visual data
- communication infrastructure
- video analysis
- video segmentation
- power transformers
- learning algorithm
- video clips
- video retrieval
- low cost
- distribution network