Code for AI Ukraine 2018 workshop "Object detector creation from scratch".
It is a basic object detector tutorial to understanding single-stage object detector workflow.
To prepare for the workshop, please:
- Clone code from this repository.
- Copy dataset from https://drive.google.com/open?id=1XUsg8sdsscWoeR0Ea9Ns0y4PKPVyxT8W
- Unzip dataset to the folder one level upper than the repository
- Copy model from https://drive.google.com/open?id=1G8DxBNwbKBtdD7UE07dwpSynaomDCa-h to the repository root
- Install Anaconda for Python 3 from https://www.anaconda.com/download/
- Open Open Anaconda Prompt and type 'pip install tensorflow'
- Type "jupyter notebook" in Anaconda Prompt to run notebooks to train, run and evaluate the model
- Use "YOLO1_train.ipynb" to train the model
- Use "YOLO1_evaluation.ipynb" to run and evaluate the model