https://github.com/exitare/mlflow-wrapper

MLFlow wrapper for handling experiments and runs with ease
https://github.com/exitare/mlflow-wrapper

helper mlflow python

Last synced: 26 days ago
JSON representation

MLFlow wrapper for handling experiments and runs with ease

• Host: GitHub
• URL: https://github.com/exitare/mlflow-wrapper
• Owner: Exitare
• License: mit
• Created: 2022-08-04T03:50:47.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2022-08-14T03:35:30.000Z (over 2 years ago)
• Last Synced: 2024-12-12T04:34:29.013Z (about 1 month ago)
• Topics: helper, mlflow, python
• Language: Python
• Homepage:
• Size: 20.5 KB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# MLFlow Wrapper

Mlflow Wrapper is a python library intended to abstract some functionality away from the developer

when interacting with the mlflow library.

The library supports/ improves the handling of experiments and offers helper functions which are not available in mlflow

by default.

Furthermore, the library offers functions to interact with Runs itself, and uploading/downloading files.

# Installation

```pip install mlflow-wrapper```

# Quick start

Some quick start examples. Please refer to the [wiki](https://github.com/Exitare/mlflow-wrapper/wiki)

for a complete overview of all functions.

## Experiment Handler

```

from mlflow_wrapper.experiment_handler import ExperimentHandler

# Connect to a local mlflow server

exp_handler:ExperimentHandler = ExperimentHandler()

# Get the id of a new experiment which does no exist so far

exp_id = exp_handler.get_experiment_by_name(experiment_name="New Experiment")

```

## Run Handler

```

from mlflow_wrapper.run_handler import RunHandler

run_handler: RunHandler = RunHandler()

# Delete a parent run and all associated children run. 

# Does only delete the first occurence of the given run name. If multiple runs do have the same name,

# this command needs to be executed multiple times

run_handler.delete_run(experiment_id=exp_id, run_name="My Run")

```

# Bugs & Issues

Please use the GitHub issue tracker for issues. I will try to get to them asap.

# Feedback

Feedback is most welcomed and I will respond asap.