Back to Mmdetection

CSPNeXt ImageNet Pre-training

configs/rtmdet/classification/README.md

3.3.02.5 KB
Original Source

CSPNeXt ImageNet Pre-training

In this folder, we provide the imagenet pre-training config of RTMDet's backbone CSPNeXt.

Requirements

To train with these configs, please install MMPreTrain first.

Install by MIM:

shell
mim install mmpretrain

or install by pip:

shell
pip install mmpretrain

Prepare Dataset

To pre-train on ImageNet, you need to prepare the dataset first. Please refer to the guide.

How to Train

You can use the classification config in the same way as the detection config.

For single-GPU training, run:

shell
python tools/train.py \
    ${CONFIG_FILE} \
    [optional arguments]

For multi-GPU training, run:

shell
bash ./tools/dist_train.sh \
    ${CONFIG_FILE} \
    ${GPU_NUM} \
    [optional arguments]

More details can be found in user guides.

Results and Models

ModelresolutionParams(M)Flops(G)Top-1 (%)Top-5 (%)Download
CSPNeXt-tiny224x2242.730.3469.4489.45model
CSPNeXt-s224x2244.890.6674.4192.23model
CSPNeXt-m224x22413.051.9379.2794.79model
CSPNeXt-l224x22427.164.1981.3095.62model
CSPNeXt-x224x22448.857.7682.1095.69model