ARLP: Automatic multi-agent transformer reinforcement learning pruner for one-shot neural network pruning

Bowen Guo, Xiang Chang, Fei Chao, Xiawu Zheng, Chih-Min Lin, Yanjie Chen, Changjing Shang, Qiang Shen

Abstract: Highlights•ARLP uses reinforcement learning to autonomously optimize neural network pruning.•ARLP uses zero-cost pruning proxies to streamline pruning and save resources.•ARLP uses Transformers to enhance pruning adaptability and efficiency.