DARKSIDE: A Heterogeneous RISC-V Compute Cluster for Extreme-Edge On-Chip DNN Inference and Training.
Angelo GarofaloYvan TortorellaMatteo PerottiLuca ValenteAlessandro NadaliniLuca BeniniDavide RossiFrancesco ContiPublished in: CoRR (2023)
Keyphrases
- training process
- low cost
- edge detection
- structured prediction
- training set
- training phase
- probabilistic inference
- clustering algorithm
- low power consumption
- high speed
- training samples
- edge information
- weighted graph
- data clustering
- analog vlsi
- heterogeneous networks
- application specific
- edge detector
- real time
- cluster analysis
- clustering method
- data points
- image processing