FARNN: FPGA-GPU Hybrid Acceleration Platform for Recurrent Neural Networks.
Hyungmin ChoJeesoo LeeJaejin LeePublished in: IEEE Trans. Parallel Distributed Syst. (2022)
Keyphrases
- recurrent neural networks
- real time
- neural network
- recurrent networks
- echo state networks
- feed forward
- parallel architecture
- complex valued
- long short term memory
- reservoir computing
- low cost
- cascade correlation
- artificial neural networks
- feedforward neural networks
- reconfigurable hardware
- gpu implementation
- field programmable gate array
- neural model
- high speed
- parallel processing
- hardware implementation
- graphics hardware
- genetic algorithm
- nonlinear dynamic systems
- graphics processing units
- parallel computing
- parallel implementation
- communication protocol
- parallel computation
- hardware architecture
- control system
- hybrid learning
- chaotic time series
- signal processing
- learning algorithm
- pose estimation