GLP4NN: A Convergence-invariant and Network-agnostic Light-weight Parallelization Framework for Deep Neural Networks on Modern GPUs.
Hao FuShanjiang TangBingsheng HeCe YuJizhou SunPublished in: ICPP (2018)
Keyphrases
- lightweight
- neural network
- wireless sensor networks
- artificial neural networks
- multi layer
- neural network model
- weight update
- back propagation
- communication infrastructure
- fuzzy logic
- general purpose
- feed forward
- network architecture
- feed forward neural networks
- single layer
- network structure
- computer networks
- parallel processing
- communication networks
- embedded systems
- nearest neighbor
- pattern recognition