Amazon SageMaker - YOLOv4 End-to-End Example Notebook

Clone this repository

git clone https://github.com/kwwendt/sagemaker-yolov4-e2e-example.git

Training

For this example, you will need to supply your own dataset for training. There are free CV datasets available on Roboflow and other similar sites.

Building the custom training container

Make sure you have AWS CLI access to interact with Amazon Elastic Container Registry.

Execute the following commands:

cd training

chmod +x build_and_push_container.sh
./build_and_push_container.sh

This may take several minutes. Please wait until the container is fully pushed up to ECR to continue.

Execute the notebook

Execute the steps within the e2e_yolov4_training.ipynb notebook in Amazon SageMaker Studio.