ShaohonChen/cifar10_with_resnet50

resnet50 classification using mmengine

train resnet50 using mmengine

descript

this project is a implement for cifar10 classification mission. Use resnet50 as our classifier and use mmengine with our training farmwork.

We use SwanLab as our training tracker, hope you like it. ; )

install

Refer to mmengine official document.

Install the environment with following command.

# with cuda12.1 or you can find torch version you want at pytorch.org
pip install torch torchvision --index-url https://download.pytorch.org/whl/cu121

pip install -U openmim
mim install mmengine
pip install swanlab

train

Just use following command run and it will auto download datasets.

python train.py

result

As a 5 epoch train (using about 5min in simple RTX3090), we got 49.15% acc. Not bad.

You can find my train log in SwanLab Project