A Hardware-Centric Approach to Increase and Prune Regular Activation Sparsity in CNNs.
Tim HotfilterJulian HöferFabian KreßFabian KempfLeonhard KraftTanja HarbaumJürgen BeckerPublished in: AICAS (2023)
Keyphrases
- low cost
- hardware and software
- real time
- cellular neural networks
- search space
- embedded systems
- reduce the energy consumption
- computing systems
- hardware implementation
- image classification
- image processing
- sparse representation
- computer systems
- data processing
- massively parallel
- case study
- circuit design
- parallel hardware