README.md

Docker

Build Image

Navigate to /docker and run - this will infer python and mlflow versions from using poetry.lock and pyproject.toml. Requires poetry.

./build_image.sh

Optional run help

./build_image.sh --help

Custom registry and image name (tag is inferred).

./build_image.sh \
--repository docker.io/clemens33/mlflow \
--push

Building image for mlflow deployments server

./build_image.sh \
--repository docker.io/clemens33/mlflow \
--genai \
--push

Build and push to repository

./build_image.sh \
--repository docker.io/clemens33/mlflow \
--tag latest \
--push

./build_image.sh \
--repository docker.io/clemens33/mlflow \
--genai \
--tag latest-deployments-server \
--push

Run Container

docker run -d --name mlflow -p 5000:5000 localhost/mlflow:2.13.0-py3.11

docker run -d --name mlflow -p 5000:5000 docker.io/clemens33/mlflow:2.13.0-py3.11

Persist mlruns and mlartifacts on host machine (not using any backend)

if [ ! -d ./data ]; then mkdir ./data; fi && \
docker run -d --name mlflow -p 5000:5000 \
-v $(pwd)/data:/home/mlflow \
localhost/mlflow:2.13.0-py3.11

Run container for mlflow deployments server. First create a .env-deployments-server.sh file with your API keys which are set as environment variables within the container.

echo '
# openai
OPENAI_API_KEY=xxx
OPENAI_API_KEY2=xxx

# anthropic
ANTHROPIC_API_KEY=xxx' > .env-deployments-server.sh

docker run --name mlflow-deployments-server \
-p 5000:5000 \
-v "$(pwd)/../samples/config.yaml:/home/mlflow/config.yaml" \
--env-file "$(pwd)/.env-deployments-server.sh" \
docker.io/clemens33/mlflow:2.13.0-deployments-server-py3.11 deployments-server