Step 4: Dataset Shape

1,797 samples x 64 features

1 ExplorePlay below
2 ReadUnderstand
3 BuildHands-on lab
4 CompareSolution
💡 ReflectThink deeper

Dataset dimensions

Before training any model, you must know the shape of your data.

0
Samples
0
Features (pixels)
0
Classes (digits 0-9)
8 x 8
Image size

Explore any sample

Enter an index (0-1796) or use arrow keys to browse.

Sample 0
Loading...
Loading...
Loading...

Think Deeper

In the Explore tab, try entering index 1797. What happens? Why?

Index out of range — there are only 1,797 samples (indices 0–1796). In production, your model only knows what it was trained on. Data outside that range is out of distribution.
Cybersecurity tie-in: The shape of your data tells you everything about the problem. 1,797 samples x 64 features. In a network dataset it might be 500,000 connections x 12 features. Same structure. Same tools.

Loading...