ResNet-18 is a relatively compact deep neural network.The ResNet-18 architecture consists of 18 layers, including convolutional, pooling, and fully connected layers. It incorporates residual blocks, a key innovation that utilizes shortcut connections to facilitate the flow of information through the network.
| GPU | IXUCA SDK | Release | Branch |
|---|---|---|---|
| MR-V100 | 4.4.0 | 26.03 | release/26.03 |
| MR-V100 | 4.3.0 | 25.12 | release/25.12 |
Note: 请切换到与您的 SDK 版本对应的 Release 分支进行测试。请勿直接在 master 分支上运行测试，因为 master 分支可能包含与您的本地 SDK 版本不兼容的最新更改。
切换分支命令示例：
git checkout release/26.03
Pretrained model: https://download.pytorch.org/models/resnet18-f37072fd.pth
Dataset: https://www.image-net.org/download.php to download the validation dataset.
pip3 install -r ../../igie_common/requirements.txtpython3 ../../igie_common/export.py --model-name resnet18 --weight resnet18-f37072fd.pth --output resnet18.onnxexport DATASETS_DIR=/Path/to/imagenet_val/
export RUN_DIR=../../igie_common/# Accuracy
bash scripts/infer_resnet18_fp16_accuracy.sh
# Performance
bash scripts/infer_resnet18_fp16_performance.sh# Accuracy
bash scripts/infer_resnet18_int8_accuracy.sh
# Performance
bash scripts/infer_resnet18_int8_performance.sh| Model | BatchSize | Precision | FPS | Top-1(%) | Top-5(%) |
|---|---|---|---|---|---|
| ResNet18 | 32 | FP16 | 9592.98 | 69.77 | 89.09 |
| ResNet18 | 32 | INT8 | 21314.55 | 69.53 | 88.97 |