TVM性能评估分析（四）

TVM性能评估分析（四）

Figure 1.  Efficient Privacy-Preserving ML Using TVM

Figure 2.  Motivation: Privacy-Preserving ML

Figure 3.  Backend

Figure 4. Differential privacy (DP) provides a formal guarantee that models trained on similar datasets are indistinguishable Informally

Figure 5. System Overview

Figure 6. Comparison of Traditional Auto-tuning and AutoTVM

Figure 7.  ARM CPU

Figure 8.  Mali GPU

Figure 9. NVIDIA GPU

Figure 10. AMD GPU

Figure 11.  Automatic Kernel Optimization for Deep Learning on All Hardware Platforms