代码拉取完成，页面将自动刷新
Residual Networks, or ResNets, learn residual functions with reference to the layer inputs, instead of learning unreferenced functions. Instead of hoping each few stacked layers directly fit a desired underlying mapping, residual nets let these layers fit a residual mapping.
# Install libGL
## CentOS
yum install -y mesa-libGL
## Ubuntu
apt install -y libgl1-mesa-dev
pip3 install tqdm
pip3 install onnx
pip3 install onnxsim
pip3 install tabulate
pip3 install ppq
pip3 install mmpretrain
pip3 install mmcv-lite
Dataset: https://www.image-net.org/download.php to download the validation dataset.
mkdir checkpoints
python3 export_onnx.py --output_model checkpoints/resnet_v1_d50.onnx
export DATASETS_DIR=/path/to/imagenet_val/
export CHECKPOINTS_DIR=./checkpoints
export RUN_DIR=./
export CONFIG_DIR=config/RESNET_V1_D50_CONFIG
# Accuracy
bash scripts/infer_resnet_v1_d50_fp16_accuracy.sh
# Performance
bash scripts/infer_resnet_v1_d50_fp16_performance.sh
# Accuracy
bash scripts/infer_resnet_v1_d50_int8_accuracy.sh
# Performance
bash scripts/infer_resnet_v1_d50_int8_performance.sh
| Model | BatchSize | Precision | FPS | Top-1(%) | Top-5(%) |
|---|---|---|---|---|---|
| ResNet_V1_D50 | 32 | FP16 | 3887.55 | 0.77544 | 0.93568 |
| ResNet_V1_D50 | 32 | INT8 | 7148.58 | 0.7711 | 0.93514 |
此处可能存在不合适展示的内容，页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容，可点击提交进行申诉，我们将尽快为您处理。
马建仓 AI 助手
