Batch Size = 4

TensorFlow AI_Matrix_ResNet50 from AI-Matrix Batch Size = 1

• Layer Information
• Layer Latency
• Layer Allocated Memory
• Layer Aggregated Information
• Layer Aggregated Latency Percentage
• Layer Aggregated Allocated Memory Percentage
• Layer Occurrence Percentage
• GPU Kernel Information
• GPU Kernel Information Aggregated by Layer
• Layer GPU vs CPU Latency
• GPU Kernel flops per Layer
• GPU Kernel DRAM Reads per Layer
• GPU Kernel DRAM Writes per Layer
• GPU Kernel Achieved Occupancy per Layer
• GPU Kernel Information Aggregated by Model
• GPU Kernel Information Aggregated by Name