Practical AI: Machine Learning, Data Science, LLM

Evaluating models without test data

Sep 20, 2022

44:55

Creator website

Highlights

AI Chapters

Episode notes

How to Evaluate Model Convergence Without Human Judgment

01:40

How Do I Train a Model?

03:24

Deep Learning - Is There a Problem With Scaling Properties?

03:15

Introduction

2min

The Weight Watcher Tool, a Data Science Consultant With Calculation Consulting

4min

How to Evaluate Model Convergence Without Human Judgment

2min

Deep Learning - Is There a Problem With Scaling Properties?

3min

How Do I Train a Model?

3min

Can You Give Us an Overview of the Weight Watcher Tool?

4min

The Engineering Tool

3min

How Deep Learning Can Be Used in the Sotware?

4min

Is It Necessary to Run a Weight Watcher?

2min

10.

How Easy Is It to Get It Running on a Model?

4min

11.

Do You Have a Workflow?

3min

12.

The Complex Optimization Problem

3min

13.

Deep Learning

5min

14.

Is There a Connection Between Theory and Engineering?

4min

WeightWatcher, created by Charles Martin, is an open source diagnostic tool for analyzing Neural Networks without training or even test data! Charles joins us in this episode to discuss the tool and how it fills certain gaps in current model evaluation workflows. Along the way, we discuss statistical methods from physics and a variety of practical ways to modify your training runs.

Join the discussion

Changelog++ members support our work, get closer to the metal, and make the ads disappear. Join today!

Featuring: