MEGA: A Memory-Efficient GNN Accelerator Exploiting Degree-Aware Mixed-Precision Quantization.
Zeyu ZhuFanrong LiGang LiZejian LiuZitao MoQinghao HuXiaoyao LiangJian ChengPublished in: CoRR (2023)
Keyphrases
- memory efficient
- external memory
- iterative deepening
- parallel implementation
- multiple sequence alignment
- high precision
- neural network
- search algorithm
- multi dimensional
- precision and recall
- high recall
- pattern growth
- data sets
- integral image
- quantization error
- field programmable gate array
- inter frame
- detection method