Convenient
Containerization with
MLFLow Project
Presented By:
Sudeep James TIrkey
Software Consultant
MachineX by Knoldus
Our Agenda
01 Problems with Packaging
02 What is MLFlow
03 Benefits of MLFlow
04 MLFlow Projects
05 Demo
Some major practical challenges in Machine Learning models packaging and
deployment that can be handled through docker are:
● Non-uniform environments across models
● Non-uniform library requirements across models
● Non-uniform resource requirements across models
● Scaling at model level
Problems with Packaging
● MLflow is a platform to streamline machine learning development, including
tracking experiments, packaging code into reproducible runs, and sharing
and deploying models.
● It’s built around REST APIs and simple data formats that can be used from a
variety of tools, instead of only providing a small set of built-in functionality.
● MLflow is designed to work with any ML library, algorithm, deployment tool or
language.
● MLflow is an open source project that users and library developers can
extend.
● MLflow’s open format makes it very easy to share workflow steps and models
across organizations if you wish to open source your code.
What is MLFlow
MLflow currently offers four components:
● MLflow Tracking
● MLflow Projects
● MLflow Models
● MLflow Model Registry
MLFlow Components
● Integrates with existing Code; requires little change.
● Platform Independent.
● Can be used by 1 to 1000 people.
● Scales to big data with Apache Spark.
● Throws metrics in Prometheus format.
Benefits of MLFlow
● Just a convention for organizing and describing your code.
● Each project is simply a directory of files, or a Git repository, containing your
code.
● Includes an API and command-line tools for running projects, making it
possible to chain together projects into workflows.
MLFlow Projects
Demo
● https://mlflow.org/docs/latest/index.html
● https://github.com/mlflow/mlflow
● https://github.com/mlflow/mlflow-example
References
Thank You !

Convenient Containerization with MLFLow Project

  • 1.
    Convenient Containerization with MLFLow Project PresentedBy: Sudeep James TIrkey Software Consultant MachineX by Knoldus
  • 2.
    Our Agenda 01 Problemswith Packaging 02 What is MLFlow 03 Benefits of MLFlow 04 MLFlow Projects 05 Demo
  • 3.
    Some major practicalchallenges in Machine Learning models packaging and deployment that can be handled through docker are: ● Non-uniform environments across models ● Non-uniform library requirements across models ● Non-uniform resource requirements across models ● Scaling at model level Problems with Packaging
  • 4.
    ● MLflow isa platform to streamline machine learning development, including tracking experiments, packaging code into reproducible runs, and sharing and deploying models. ● It’s built around REST APIs and simple data formats that can be used from a variety of tools, instead of only providing a small set of built-in functionality. ● MLflow is designed to work with any ML library, algorithm, deployment tool or language. ● MLflow is an open source project that users and library developers can extend. ● MLflow’s open format makes it very easy to share workflow steps and models across organizations if you wish to open source your code. What is MLFlow
  • 5.
    MLflow currently offersfour components: ● MLflow Tracking ● MLflow Projects ● MLflow Models ● MLflow Model Registry MLFlow Components
  • 6.
    ● Integrates withexisting Code; requires little change. ● Platform Independent. ● Can be used by 1 to 1000 people. ● Scales to big data with Apache Spark. ● Throws metrics in Prometheus format. Benefits of MLFlow
  • 7.
    ● Just aconvention for organizing and describing your code. ● Each project is simply a directory of files, or a Git repository, containing your code. ● Includes an API and command-line tools for running projects, making it possible to chain together projects into workflows. MLFlow Projects
  • 8.
  • 9.
  • 10.