Neural Network Playground

Rust+WebAssembly

01 Data Generation

Pattern Type

Sample Size

Noise Level

0 points generated

Max 5 hidden layers • 12 neurons per layer

Input

Hidden 1

Hidden 2

Output

Activation Function

Algorithm

Learning Rate

Type

2 → 8 → 8 → 1 neurons • RELU

Neurons

Positive weights

Negative weights

Epoch

Loss

0.0000

Accuracy

0.00%

Validation Loss

—

Epochs

Batch Size

Validation Split (20%)

Early Stopping

Train the model to visualize loss curves

Generate a dataset and ensure the WASM module has loaded to enable training.

> Neural network initialized

> Ready to train