DNNZip: Selective Layers Compression Technique in Deep Neural Network Accelerators.
Habiba LahdhiriMaurizio PalesiSalvatore MonteleoneDavide PattiGiuseppe AsciaJordane LorandelEmmanuelle BourdelVincenzo CataniaPublished in: DSD (2020)
Keyphrases
- neural network
- single layer
- multiple layers
- multi layer
- data compression
- artificial neural networks
- back propagation
- neural network is trained
- image compression
- compression ratio
- compression scheme
- feed forward neural networks
- fuzzy artmap
- bp neural network
- single chip
- pattern recognition
- deep learning
- fuzzy neural network
- computing systems
- multiresolution
- neural network model
- auto associative
- genetic algorithm
- deep belief networks
- real time
- computing platform
- field programmable gate array
- hidden layer
- network model
- neural nets
- general purpose
- compression algorithm
- feed forward
- fault diagnosis