AND THE BIT GOES DOWN: REVISITING THE QUANTIZATION OF NEURAL NETWORKS

AND THE BIT GOES DOWN: REVISITING THE QUANTIZATION OF NEURAL NETWORKS

• ABSTRACT
• • 什么是码本？
• 1 INTRODUCTION
• 2 RELATED WORK
• • Low-precision training.
  • Quantization.
  • Pruning.
  • Dedicated architectures.(专用架构)
• 3 OUR APPROACH
• • 3.1 QUANTIZATION OF A FULLY-CONNECTED LAYER(量化全连接层)
  • • Product Quantization (PQ).
    • Our algorithm.