Skip to content

Latest commit

 

History

History

resnet_v1

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
Model Source Size, MB Quality: Accuracy, % Latency, ms
CoreML Source iPhone 11 Pro iPhone 7
Top-1 Top-5 Top-1 Top-5 CPU GPU ANE CPU GPU
ResNet18_V1 PyTorch 45 69.76 89.08 69.76 89.08 20 35 8 78 49
ResNet34_V1 PyTorch 83 73.32 91.42 73.30 91.42 35 47 9 150 50
ResNet50_V1 PyTorch 98 76.13 92.86 76.15 92.87 70 50 10 186 70
ResNet50_V1 Keras 98 74.73 91.95 74.93 92.06 66 45 9 170 62
ResNet101_V1 PyTorch 170 77.37 93.55 77.37 93.56 165 58 11 327 119
ResNet101_V1 Keras 165 76.18 92.58 76.42 92.79 151 55 9 313 135
ResNet152_V1 PyTorch 230 78.31 94.05 78.31 94.06 216 75 14 490 204
ResNet152_V1 Keras 231 76.40 92.92 76.60 93.12 214 73 12 441 165

Reference: https://arxiv.org/abs/1512.03385

Inputs:

  • 224x224
  • 3-channel RGB
  • [0-255] pixel range

Dataset:

Versions:

  • iOS 13.5.1
  • XCode 11.5