This repository uses docker to build MLFlow Tracking Server in a very simplified manner. For this purpose, one need to have Docker Engine installed.
Prerequisites: Docker Engine.
First, check the docker-compose.yaml to find the env variables that are needed to be set up. You'd need to create .env file and add all the following environment variables with appropriate values.
POSTGRES_PASSWORD=POSTGRES_PASSWORD
POSTGRES_USER=POSTGRES_USER
POSTGRES_DB=POSTGRES_DB
ARTIFACT_ROOT=ARTIFACT_ROOT
MLFLOW_TRACKING_USERNAME=MLFLOW_TRACKING_USERNAME
MLFLOW_TRACKING_PASSWORD=MLFLOW_TRACKING_PASSWORD
MLFLOW_UPDATED_USERNAME=MLFLOW_UPDATED_USERNAME
MLFLOW_UPDATED_PASSWORD=MLFLOW_UPDATED_PASSWORD
MLFLOW_TRACKING_URI=http://0.0.0.0:5000Once these variables are set, follow these commands to run the tracking server:
docker-compose up --build
Afterward, run the authentication script in the docker container to change the default credentials.
<https://github.com/MahdyShirdel/postgres/>