🧠 Image Classifier Challenge

Build a deep neural network and train it to classify images!

🏗️ Build Your Network

Click to add layers:

Dense (128)
Fully connected layer
➕
Dense (64)
Smaller dense layer
➕
Dropout (30%)
Prevents overfitting
➕
Layers:2
Parameters:0
Training:Not Started
Accuracy:0%

💡 Tips:

  • More layers = deeper learning
  • Dropout helps prevent overfitting
  • Balance depth vs training time
  • Aim for 90%+ accuracy!

Your Score

0
Beginner

Build a network with 90%+ accuracy to level up!

Current Network Architecture

Input Layer (28×28 = 784 neurons)
Output Layer (3 classes: Circle, Square, Triangle)

Training Progress

0%

Test Results

Train your model first, then test it!