Training Visualizer
Learn how ML models train through hands-on experiments
Your First Dataset - Iris Flowers
What is a dataset? How do we represent data for ML?
The Problem
Can we teach a computer to identify flower species from measurements? Each flower is measured by petal length and width, and we want to predict whether it's a Setosa or Versicolor iris.
What You'll Learn
- •How data is represented for machine learning (features and labels)
- •How a neural network draws a "decision boundary" to separate classes
- •The concept of training a model to fit data
Data source: UCI ML Repository
Loading model...
Network Architecture
Blue = positive weights, Red = negative. Thickness shows magnitude.
Each output node shows the probability that the input belongs to that class. The model predicts the class with the highest probability.
| Petal Length (cm) | Petal Width (cm) | Species |
|---|---|---|
| 1.4 | 0.2 | Setosa |
| 1.3 | 0.4 | Setosa |
| 1.5 | 0.2 | Setosa |
| 1.4 | 0.3 | Setosa |
| 1.7 | 0.4 | Setosa |
| 1.6 | 0.2 | Setosa |
| 1.4 | 0.1 | Setosa |
| 1.5 | 0.4 | Setosa |
| 4.0 | 1.3 | Versicolor |
| 4.5 | 1.5 | Versicolor |
| 3.9 | 1.4 | Versicolor |
| 4.2 | 1.3 | Versicolor |
| 4.7 | 1.4 | Versicolor |
| 4.3 | 1.6 | Versicolor |
| 3.8 | 1.1 | Versicolor |
| 4.4 | 1.5 | Versicolor |
What to Notice
The model learned to separate the two species based on their measurements. Each point is a real flower, and the colored regions show where the model thinks each species belongs.