Computer-Vision-2020-1

2020-1st Computer Vision Project

2020학년도 1학기 컴퓨터비전 기말프로젝트로 Weapon Classifier를 만들기를 위한 레포지토리입니다.

1. Contributor

• 양재원, 유재원, 부준영

2. Version

• python3.6
• torch1.5 and torchvision 0.6
• Cuda >=10.2
• Ubuntu 18.04
• Multiple GPU: Nvidia Titan X

3. Train

• Must to do: You need to unzip the dataset.zip into the project root dir. The dataset folder should be in same dir as train.py
• command: python train.py --datapath (Root Dir of this Project) --epoch (epochs to train) --pretrained (set True for Pretranied ResNet Model)

4. Test

• Command: python test.py --datapath (Root Dir of this Project)

5. Classify your Weapon(knife,axe,pistol,rifle)

• Commnd: python demo.py
• Put your weapon data in the my_dataset dir and put it in each subfolder. It will classifiy the data with the trained model "cvd.pth" inside the "trained_model" dir.