VRDL Lab2: Object Detection

This is homework 2 in NYCU Selected Topics in Visual Recognition Deep Learning class.

NOTIFICATION: Because the main code are too big, follow the link below to download it. https://drive.google.com/file/d/120wryn8IRAiVxciyYO6KGjOj7Es5HzOa/view?usp=sharing

model weights link: https://drive.google.com/file/d/19BDS4q4flsyEAiyRenIpNlEqeUWAVTgz/view?usp=sharing

Hardware

The following specs were used to create the original solution.

• Google Colab
• PC
  • Intel(R) Core(TM) i7-7700 CPU @ 3.60GHz
  • NVIDIA RTX 1080 Ti

Environment

cd yolov5
pip install mmcv
pip install -U pycocotools
pip install -r requirements.txt

Dataset Preparation

Download this folder including dataset and put it in the same directory with yolov5 folder. https://drive.google.com/file/d/1NEXWWJoLRW9R4RpK9EYYvPfvFFDPRTEE/view?usp=sharing

Requirement

• Download and unzip 'datasets_final.zip'.
• There will include './my_datasets/images' and './my_datasets/labels'
• Download and unzip the main code: https://drive.google.com/file/d/120wryn8IRAiVxciyYO6KGjOj7Es5HzOa/view?usp=sharing

Train and Valid

You can customize your own batch size and epochs.

python train.py --img 640 --batch {} --epochs {} --data yrdl.yaml --weights yolov5l.pt

Inference

The code not only trains, but also valid and test the model. You can run the model by following:

python detect.py --weights {} --conf 0.1 --source {}

Replacing the weights for './runs/train/exp/weights/best.pt' and source for your test data directory.

References

YOLOv5: https://github.com/ultralytics/yolov5