MFlux.ai

Easy to get started

MFlux.ai makes it easy to get a head start for beginners and experts. It is easily integrated with existing ML projects. You are in control. Tutorials help beginners hit the ground running.

Get better results faster

Doing many experiments is a key to success in machine learning projects. MFlux.ai can track your experiments, help you reproduce them and pick the best models.

Built around open source components

MFlux.ai is a powerful combination of the best open source machine learning tools available. This means that the components have been battle-tested by many people, and that bugs are quickly found and fixed.

Ideal for small teams

Our mission is to enable small teams at any experience level to succeed in building robust, powerful, scalable and maintainable machine learning solutions

MLflow

Spin up your own private MLflow server at the click of a button

Experiment tracking

Track your experiment runs (parameters, descriptions, metrics, models, dataset versions)

Collaboration

Create a team and add collaborators to give them access

Object storage for AI

Store multiple versions of datasets and machine learning models. 10 GB of secure object storage is included, and you can expand it up to 10 TB

Example usage

$ pip install mflux-ai mlflow[extras]

import mlflow.sklearn
import mflux_ai

mflux_ai.init("your_project_key_goes_here")

mflux_ai.put_dataset(dataset, dataset_filename)
dataset = mflux_ai.get_dataset(dataset_filename)

mlflow.log_metric("validation_accuracy", validation_accuracy)
mlflow.log_param("model_type", "DecisionTreeClassifier")
mlflow.sklearn.log_model(model, "model")