Scalable high-performance architecture for convolutional ternary neural networks on FPGA.
Adrien Prost-BoucleAlban BourgeFrédéric PétrotHande AlemdarNicholas CaldwellVincent LeroyPublished in: FPL (2017)
Keyphrases
- neural network
- network architecture
- hardware design
- hardware implementation
- hardware architecture
- software implementation
- multi layer
- real time
- fpga implementation
- artificial neural networks
- hardware architectures
- pattern recognition
- pipelined architecture
- learning capabilities
- dedicated hardware
- associative memory
- scalable distributed
- back propagation
- management system
- real time image processing
- fpga technology
- multiprocessor architecture
- field programmable gate array
- data intensive
- data flow
- high speed
- fuzzy logic
- low power consumption
- training process
- recurrent neural networks
- reconfigurable hardware
- xilinx virtex
- power consumption
- systolic array
- genetic algorithm