Precision-aware Latency and Energy Balancing on Multi-Accelerator Platforms for DNN Inference.
Matteo RissoAlessio BurrelloGiuseppe Maria SardaLuca BeniniEnrico MaciiMassimo PoncinoMarian VerhelstDaniele Jahier PagliariPublished in: CoRR (2023)
Keyphrases
- total energy
- energy minimization
- energy consumption
- inference process
- bayesian networks
- average precision
- response time
- precision and recall
- real time
- probabilistic inference
- parallel implementation
- neural network
- failure modes
- field programmable gate array
- bayesian model
- energy efficient
- bayesian inference
- high throughput
- high precision
- energy function
- probabilistic model