Trying sagemaker
Do the ./upload_training_container.sh to build an image and put it into ECR
Do the ./upload_inference_container.sh to build an image for inference and put it to ECR
Do test locally run upload_inference_container.sh and then do
docker run -p 8080:8080 -v "$(pwd)"/model:/opt/ml/model:ro test-sagemaker-deploy-container:latest serve
where model is a folder with model.joblib file
Or run a tests like that:
docker build -f test.Dockerfile . -t test:latest
docker run -v "$(pwd)"/model/:/opt/ml/model/:ro test:latest
Enjoy coverage reporting!