test-codes for azure nnunet pipeline inference

1. Build dockerfile

   docker build -t nnunet_local_inference_env .
   

2. Run dockerfile

   docker run -v ${pwd}:/app -p 5001:5001 nnunet_local_inference_env
   

Note : If you are running from docker desktop, don't forget to mount project root directory to /app directory

Quick note : Workflow tested with python 3.10.6

1. Clone this repository

   git clone https://github.com/tekmen0/azure-inference-test.git
   

2. Move to repository folder

   cd azure-inference-test
   

3. (Optional) Activate python virtual environment.

4. Install the inference test server

   python -m pip install azureml-inference-server-http
   

5. Install requirements.

   pip3 install requirements.txt 
   

6. Move to server directory

   cd server
   

7. Start inference server

   azmlinfsrv --entry_script score.py
   

Now your scoring is deployed locally, running at 127.0.0.1:5001

You can send request to endpoint 127.0.0.1:5001/score for executing the 'run' function in score.py

Example request using python can be found in request.py, you may also want to use curl or postman