2. Start and Check a RunΒΆ
This section serves as explanation on how to start a run locally. The run will be tracked in your mantik account.
First install the mantik package. You will also need mlflow for this.
pip install mantik
pip install mlflow
Setup the necessary environment variables
export MANTIK_USERNAME=<MANTIK USERNAME>
export MANTIK_PASSWORD=<MANTIK PASSWORD>
eval $(mantik init)
2.1. Setup the Wine Quality Estimator projectΒΆ
Clone the tutorial and build the relevant docker image
git clone https://gitlab.com/mantik-ai/tutorials.git
docker build -t wine-quality-executor tutorials/wine-quality-estimator/mlproject
2.2. Start a RunΒΆ
How to submit a run:
mantik runs local "wine-quality-estimator/mlproject/MLProject" --name "Local: wine quality estimator" --project-id <PROJECT ID> --code-repository-id <CODE REPOSITORY ID> --branch main --experiment-repository-id <EXPERIMENT REPOSITORY ID> --parameter alpha=0.1
The different ids can be found using the Details
action button of the entities in the Mantik UI.
This will execute the project locally. The run is book-kept in the platform, so when navigating in the UI to
Projects -> <PROJECT NAME> -> Runs -> Submissions
you will see the run.
Your system info such as
GPU info
CPU model
Platform
CPU core count
Total Memory
Will automatically be recorded as part of the backend config. It can be disabled by setting
the flag --backend-config-system-info=False
when executing the run.