Image Classification on CIFAR-10 (test) (Latency and Memory Usage)
8.86Amortized Latency (s)Optimized Pipeline Architecture
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Optimized Pipeline ArchitectureModel=ResNet-20, Batch Size=512, Processor=Intel(R) Xeon(R) Gold 5418Y CPU, 32 cores, 256 GB of RAM2026.04 | 8.86 | 98.96 | |
| Cheon et al.Model=ResNet-20, Batch Size=512, Processor=2× Intel Xeon Gold 6248 CPUs, 40 cores, and 1 TB of RAM2026.04 | 15.76 | 370.28 | |
| AutoFHEModel=ResNet-20, Batch Size=512, Processor=AWS Instances r5.24xlarge, 96 CPU cores, 768 GB RAM2026.04 | 75 | — |