Hydra: Deadline-Aware and Efficiency-Oriented Scheduling for Deep Learning Jobs on Heterogeneous GPUs.
Zichao YangHeng WuYuanjia XuYuewen WuHua ZhongWenbo ZhangPublished in: IEEE Trans. Computers (2023)
Keyphrases
- deep learning
- scheduling algorithm
- parallel machines
- unsupervised learning
- machine learning
- unsupervised feature learning
- resource constraints
- scheduling problem
- identical machines
- identical parallel machines
- mental models
- weakly supervised
- general purpose
- maximum likelihood
- single machine
- processing times
- grid environment
- response time
- face recognition