Python Docs
R Docs
Admin Docs
Python Docs
Welcome to Metaflow for Python
Introduction
Why Metaflow
What is Metaflow
Release Notes
Roadmap
Contributing to Metaflow
Get in Touch
Getting Started
Installing Metaflow
Tutorials
Metaflow on AWS
Metaflow on AWS
Metaflow Sandbox
Deploying to AWS
Developing with Metaflow
Basics of Metaflow
Inspecting Flows and Results
Debugging with Metaflow
Scaling Out and Up
Loading and Storing Data
Managing External Libraries
Dealing with Failures
Organizing Results
Going to Production with Metaflow
Scheduling Metaflow Flows
Internals of Metaflow
Technical Overview
Testing Philosophy
Why Metaflow
​
Here is a data scientist.
Data scientists solve important business problems.
A good way to start solving a problem is to prototype a solution in a notebook.
There are great libraries available for machine learning that make prototyping fun.
After the first experiments, it is a good idea to start keeping track of models and data.
It can take a lot of work to get the latest data and keep the models up to date reliably.
Production workflows should run on servers, not on a laptop.
These days, servers execute containers. Workflows need to be packaged for execution.
Where should the results go? Sometimes they are deployed as containers too.
Containerized models can be consumed by business applications.
Stakeholders evaluate the results. They want more models and better models!
Taking care of all this can be hard. Metaflow can help.
​
Previous
Welcome to Metaflow for Python
Next - Introduction
What is Metaflow
Last updated
1 year ago
Export as PDF