Before doing anything, you must clone this repos and install requirements
$ git clone https://github.com/Surayuth/csgo_yolov5
$ cd csgo_yolov5
$ pip install -r requirements.txt
1. Create dataset for training, validation and testing by
$ python3 create_dataset.py 0.2 0.1 --resize 512 512 --download -1
The command above creates a directory, csgo_dataset, having the right structure for being trained by yolov5 model. The val_ratio and test_ratio are set to 0.2 and 0.1, respectively. You can also resize all the images to specific size by using --resize w h. The optional argument --download -1 will download all datasets
available to this repos right now. You can change -1 to the numbers of datasets you want to download, i.e., --download 1 2.
2. Install yolov5 (inside csgo_yolov5)
$ git clone https://github.com/ultralytics/yolov5
$ cd yolov5
$ pip install -r requirements.txt
3. Copy YAML file of csgo_dataset to yolov5 directory
$ cp ../csgo.yaml ./
4. Train model
$ python3 train.py --img 512 512 --batch 4 --epochs 5 --data csgo.yaml --weights yolov5s.pt
The trained model will be in yolov5/runs/train/exp/weights/. The lastest model may be in different exp.
5. export to onnx
$ python3 models/export.py --weights runs/train/exp/weights/best.pt --img 512 512 --batch 1