Is Integer Arithmetic Enough for Deep Learning Training?
Alireza GhaffariMarzieh S. TahaeiMohammadreza TayaranianMasoud AsgharianVahid Partovi NiaPublished in: NeurIPS (2022)
Keyphrases
- deep learning
- integer arithmetic
- deep architectures
- restricted boltzmann machine
- unsupervised learning
- deep belief networks
- unsupervised feature learning
- machine learning
- mental models
- training set
- supervised learning
- training samples
- gray scale
- training examples
- euclidean distance transform
- model selection
- principal component analysis