GhostNetV2: Enhance Cheap Operation with Long-Range Attention

Installation

conda create -n PyTorch python=3.10.10
conda activate PyTorch
conda install python=3.10.10 pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia
pip install opencv-python
pip install pyyaml
pip install timm
pip install tqdm

Note

• The test results including accuracy, params and FLOP are obtained by using fused model

Train

• Configure your IMAGENET dataset path in main.py for training
• Run bash main.sh $ --train for training, $ is number of GPUs

Test

• Configure your IMAGENET path in main.py for testing
• Run python main.py --test for testing

Results

Version	Epochs	Top-1 Acc	Top-5 Acc	Params (M)	FLOP (M)	Download
GhostNetV2-1.0	450	-	-	6.126	167.689	-
GhostNetV2-1.0*	450	75.15	92.25	6.126	167.689	model
GhostNetV2-1.3*	450	76.67	93.32	8.920	270.156	model
GhostNetV2-1.6*	450	77.76	93.97	12.343	399.636	model

• * means that weights are ported from original repo, see reference

Reference

• https://github.com/huawei-noah/Efficient-AI-Backbones