PackQViT: Faster Sub-8-bit Vision Transformers via Full and Packed Quantization on the Mobile.
Peiyan DongLei LuChao WuCheng LyuGeng YuanHao TangYanzhi WangPublished in: NeurIPS (2023)
Keyphrases
- uniform quantization
- mobile devices
- computer vision
- successive approximation
- vision system
- adaptive quantization
- real time
- mobile networks
- autonomous mobile
- image processing
- mobile phone
- mobile learning
- quantization scheme
- mobile communication
- quantization error
- memory efficient
- location aware
- mobile environments
- context aware
- subband
- visual perception
- mobile computing
- color quantization
- highly efficient
- gray code
- mobile applications
- multiresolution
- computational complexity
- partial discharge