Training Deep Neural Networks in 8-bit Fixed Point with Dynamic Shared Exponent Management.
Hisakatsu YamaguchiMakiko ItoKatsuhiro YodaAtsushi IkePublished in: DATE (2021)
Keyphrases
- fixed point
- neural network
- cellular neural networks
- training process
- lookup table
- fixed point theorem
- floating point
- sufficient conditions
- dynamical systems
- belief propagation
- variational inequalities
- image processing
- multi layer perceptron
- feed forward
- post processing
- constraint databases
- least squares
- artificial neural networks