DecentLaM: Decentralized Momentum SGD for Large-batch Deep Training.
Kun YuanYiming ChenXinmeng HuangYingya ZhangPan PanYinghui XuWotao YinPublished in: ICCV (2021)
Keyphrases
- stochastic gradient descent
- batch mode
- training set
- supervised learning
- cooperative
- data sets
- least squares
- training process
- online algorithms
- training phase
- training examples
- deep architectures
- batch processing
- deep learning
- training algorithm
- artificial neural networks
- learning algorithm
- genetic algorithm
- neural network