moDNN: Memory optimal DNN training on GPUs.
Xiaoming ChenDanny Z. ChenXiaobo Sharon HuPublished in: DATE (2018)
Keyphrases
- training process
- general purpose
- optimal solution
- computational power
- dynamic programming
- training samples
- early stopping
- memory usage
- training phase
- main memory
- worst case
- training set
- neural network
- training examples
- closed form
- active learning
- memory requirements
- data structure
- machine learning
- memory size
- real time