Optimizing Multi-GPU Parallelization Strategies for Deep Learning Training.
Saptadeep PalEiman EbrahimiArslan ZulfiqarYaosheng FuVictor ZhangSzymon MigaczDavid W. NellansPuneet GuptaPublished in: IEEE Micro (2019)
Keyphrases
- deep learning
- deep architectures
- restricted boltzmann machine
- parallel processing
- unsupervised feature learning
- machine learning
- unsupervised learning
- deep belief networks
- training examples
- training samples
- supervised learning
- mental models
- weakly supervised
- general purpose
- shared memory
- active learning
- training set
- data sets