Resnet18

This is an exported version of Resnet18 from Aidge.

The original version trained on Imagenet and the finetuned version on CIFAR100.

Aidge support

Note: We tested this network for the following features. If you encounter any error please open an issue. Features not tested in CI may not be functional.

Feature	Tested in CI
ONNX import	✔
Backend CPU	✔
Export CPP	❌

Model

• Operators: 171 (11 types)
  • Add: 8
  • BatchNorm2D: 20
  • Conv2D: 3
  • FC: 1
  • Flatten: 1
  • GlobalAveragePooling: 1
  • Identity: 3
  • PaddedConv2D: 17
  • PaddedMaxPooling2D: 1
  • Producer: 99
  • ReLU: 17

CIFAR100

• Opset: 18
• Source: PyTorch
• Input
  • size: [N, 3, 224, 224]
  • format: [N, C, H, W]
  • preprocessing:
    • ?
• Output
  • size: [N, 100]

ImageNet1k

• Opset: 8
• Source: ?
• Input
  • size: [N, 3, 224, 224]
  • format: [N, C, H, W]
  • preprocessing:
    • ?
• Output
  • size: [N, 1000]