joapolarbear / dl_notes

1 stars 1 forks source link

HAQ: Hardware-Aware Automated Quantization with Mixed Precision #16

Open joapolarbear opened 4 years ago

joapolarbear commented 4 years ago

CVPR 2019

Hardware-Aware Automated Quantization (HAQ) framework for quantization strategy search, PDF

Contribution

  1. leverages the reinforcement learning to automatically determine the quantization policy

and model size, we employ a hardware simulator to generate direct feedback signals (latency and energy) to the RL agent.

  1. fully automated and can specialize the quantization policy for different neural network architectures and hardware architectures.

Large search space

with M different neural network models, each with N layers, on H different hardware platforms, there are in total O(H × M × 8^{2N} ) possible solutions (Assuming the bitwidth is 1 to 8 for both weights and activations)

Limitation