MiCRO: Near-Zero Cost Gradient Sparsification for Scaling and Accelerating Distributed DNN Training.
Daegun YoonSangyoon OhPublished in: CoRR (2023)
Keyphrases
- training process
- communication cost
- distributed systems
- cooperative
- lightweight
- distributed environment
- training phase
- distributed data
- online learning
- image segmentation
- least squares
- training algorithm
- peer to peer
- test set
- image processing
- cost sensitive
- data sets
- high cost
- mobile agents
- supervised learning
- multi class
- training set
- neural network