Non-uniform DNN structured subnets sampling for dynamic inference

Keyphrases

• Adaptive Structures 33%
• Batch Size 33%
• Channel Selection 33%
• CIFAR-10 Dataset 33%
• Computing Accuracy 33%
• Computing Load 33%
• Deep Neural Network 100%
• Deep Neural Network Structure 33%
• Dynamic Deep Neural Networks 33%
• Dynamic Description 33%
• Dynamic Inference 100%
• Dynamic Requirements 33%
• Dynamic Workload 33%
• GPU 33%
• Hardware Accelerator 33%
• Hardware Resources 33%
• ImageNet 33%
• ImageNet Dataset 33%
• Latency 33%
• Low-rank Approximation 33%
• Model Compression 33%
• Network Sampling 33%
• Non-uniform Channel 33%
• Power Budget 33%
• Power-limited System 33%
• ResNet 33%
• ResNet18 33%
• Resource Allocation 33%
• Resource-constrained Systems 33%
• Sampling Methods 33%
• Time Dynamics 33%
• Time-adaptive 33%
• Via Model 33%

Computer Science

• Compression Technique 20%
• Deep Neural Network 100%
• Dynamic Workload 20%
• Graphics Processing Unit 20%
• Hardware Accelerator 20%
• Hardware Resource 20%
• Model Compression 20%
• Neural Network Architecture 20%
• Rank Approximation 20%
• Resource Allocation 20%
• Subnet 100%
• Subnetwork 20%