Hybrid Precision Floating-Point (HPFP) Selection to Optimize Hardware-Constrained Accelerator for CNN Training.
Muhammad JunaidHayotjon AlievSangbo ParkHyungWon KimHoyoung YooSanghoon SimPublished in: Sensors (2024)
Keyphrases
- floating point
- floating point arithmetic
- instruction set
- fixed point
- floating point unit
- square root
- hardware and software
- field programmable gate array
- sparse matrices
- low cost
- computing systems
- cellular neural networks
- interval arithmetic
- image processing
- graphics processing units
- neural network
- training set
- computer systems
- scheduling problem