Managing Machine Learning Experiments

Experiment Management 101

Get started now and watch the video part 1.

Experiment Management 101

2021 Comet, Inc. All Rights Reserved

Comet powers data scientists at these great companies

Part 1: Defining Scope & Creating Baselines

• Key principles of machine learning experiment management
• How to define and document the scope and success criteria for your project
• Approaches to setting up dumb baselines that take production and real-world variables into account
  

hello@comet.ml

Managing your machine learning experiments can be a painful and tedious process. But the good news is that it doesn’t have to be. Managing experiments can be straightforward and simple. 

This introductory course will help you understand the key concepts of experiment management and how to implement it at your organization.

Part 3: Tuning & Serving Your Model

• Process for iterating across different types of models
• Approaches to fine tuning hyperparameters 
• Key considerations for tracking and preparing your model as it goes into production
  

Start Standardizing Your ML Experiments 

Understanding Experiment Management is just the beginning. This ebook provides:

• Overview of how standardizing your approach to ML experiments can provide improved reproducibility
• Step-by-Step guide for standardizing your approach to ML experiments
• Questions you should ask at every step of the process to ensure you're making the appropriate considerations
  

Part 2: Data Validation & Preprocessing

• Tips for understanding and validating your data as you get started
• Steps you should take to ensure data is prepared for your experiment
• Considerations for splitting your data for training, validation, test sets, and more