2

I installed mlflow via pip and opened it in the browser through the terminal. The tab Experiments displays normally; however, when I switch to Models tab, the app crashes due to the following error:

INVALID_PARAMETER_VALUE: Model registry functionality is unavailable; 
got unsupported URI './mlruns' for model registry data storage. 
Supported URI schemes are: ['postgresql', 'mysql', 'sqlite', 'mssql']. 
See https://www.mlflow.org/docs/latest/tracking.html#storage for how to run 
an MLflow server against one of the supported backend storage locations.

I would like to log model data locally without connecting to external servers or databases. I thank you in advance for any help!

Matin Zivdar
  • 593
  • 5
  • 20
Sanjin Juric Fot
  • 121
  • 1
  • 6

1 Answers1

0

According to MLflow documentations

If running your own MLflow server, you must use a database-backed backend store in order to access the model registry via the UI or API.

So you should use --backend-store-uri to configure the database-backed store. MLflow supports the database dialects MySQL, MS SQL, SQLite, and PostgreSQL. You can read more here.

For example, --backend-store-uri sqlite:///mlflow.db would use a local SQLite database.

mlflow ui --backend-store-uri sqlite:///mlflow.db
Matin Zivdar
  • 593
  • 5
  • 20