deepanshu-yadav / spacesense_mlops_challenge Goto Github PK

This repository is the solution to the hiring challenge for mlops intern at Spacesense.

1. It expected the image segmentation code should be served through a FastAPI server.
2. Dockerize the whole FastAPI server.

The solution can executed by two ways

1. Locally
2. Docker

For running it locally create an anaconda environment

1. conda create -n spacesense python=3.9

2. conda activate spacesense

3. pip install -r requirements.txt

4. uvicorn server:app --host 0.0.0.0 --port 8000

5. In another terminal run

curl -L -F "file=@resources/dog.jpg" http://127.0.0.1:8000/segmentation -o "test.png"

You will get the ouput image as test.png.

For Docker

1. Install docker first in your machine.

2. Go to docker directory using cd docker

3. Now build the docker image using docker build -t imageseg .

4. Run the container docker container run -d -p 8000:8000 imageseg

5. Test the server by cd .. and then curl -L -F "file=@resources/dog.jpg" http://127.0.0.1:8000/segmentation -o "test.png"

6. You can even directly run the already built image docker container run -d -p 8000:8000 pronoob007/spacesense-image-seg:latest Obviously it should be tested the same way you built the container.