SGQuant: Squeezing the Last Bit on Graph Neural Networks with Specialized Quantization.
IEEE International Conference on Tools with Artificial Intelligence（2020）
关键词
SGQuant,Graph Neural Networks,specialized quantization,industry field,high memory footprints,node embedding features,high memory footprint hurdles,memory-constrained devices,IoT devices,specialized GNN quantization scheme,GNN memory consumption,GNN-tailored quantization algorithm design,GNN quantization fine-tuning scheme,multigranularity quantization strategy,graph topology,GNN computation,automatic bit-selecting,appropriate quantization bits,multigranularity quantizations,full-precision GNNs
AI 理解论文
溯源树
样例
生成溯源树，研究论文发展脉络
Chat Paper
正在生成论文摘要