FIXAR: A Fixed-Point Deep Reinforcement Learning Platform with Quantization-Aware Training and Adaptive Parallelism.

Je Yang, Seongmin Hong, Joo-Young Kim

Browse the full DAC paper archive.