The dataset consists of 5 classes:
- Bus
- Truck
- Car
- Traffic signal
- Truck
- Cloning the YoloV5 file from official repository.
- Changing the directory of yolov5
- Installing the dependencies
- Download all versions pre-trained weights.
- Go to yolov5/data/.
- Open data.yaml
- Edit the following inside it:
- Training and Validation file path
- Number of classes and Class names.
- Set images size 640 with batch of 8.
- Total 633 images for training and 165 images for validation present in 5 classes.
- Train model around 600 epochs .Stopping training early as no improvement was observed in last 100 epochs. Best results observed at epoch 201, best model saved as best.pt.
- Visualise the training metrics with the help of tensorboard.