MGRQ: Post-Training Quantization For Vision Transformer With Mixed Granularity Reconstruction.
Lianwei YangZhikai LiJunrui XiaoHaisong GongQingyi GuPublished in: CoRR (2024)
Keyphrases
- training process
- image reconstruction
- fuzzy logic
- computer vision
- vision system
- training set
- three dimensional
- high voltage
- high resolution
- compressed sensing
- online learning
- reconstruction method
- computational complexity
- image processing
- reconstructed image
- real time
- power system
- quantization noise
- training phase
- training samples
- decision trees
- image classification
- image sequences