Weight Adjustment Challenge

Tune weights and biases, inspect neuron activations, lower model loss, and solve progressively harder prediction challenges with immediate visual feedback, clear progress, and rewards.

Challenge Command Centre

Configure a challenge, start the round, then improve the network using manual or guided adjustments.


Score0
Moves0 / 40
Time02:00
Stars☆☆☆
Loss
Accuracy
Best loss
Hints used0
Challenge progress0%

Interactive Network Canvas

Connection colour shows sign. Thickness shows magnitude. Neuron fill shows activation.

Sample 1
Positive Negative Locked

Prediction Inspector

Compare the active sample, weighted sums, outputs, and target.

Coach and Diagnostics

Request direction, inspect gradients, or apply an optimiser step.

Start a challenge to receive coaching.

Weight and Bias Laboratory

Move sliders, enter exact values, lock parameters, or adjust whole layers.

Dataset and Sample Evaluation

Inspect inputs, targets, predictions, and sample-level loss.

Plotly Performance Dashboard

Track loss, accuracy, parameter movement, predictions, activations, and decision boundaries.

Forward Propagation Calculations

Follow the current sample through weighted sums, biases, and activations.

Start a challenge to see calculations.

Challenge Log

Review moves, improvements, warnings, achievements, and optimiser actions.

Custom Challenge Builder

Achievements and Results

Unlock rewards for accuracy, efficiency, manual solving, and stable performance.

Save, Export, and Accessibility

Keep configurations, export evidence, and adapt the interface.


H Hint A Auto step Space Pause Ctrl Z Undo

Related Calculators

Build a Neural NetworkNeuron Activation GameActivation Function MatchBackpropagation PuzzleDense Layer Output GameVanishing Gradient EscapeNeural Network Architecture BuilderDropout DefenderLoss Function Challenge

Important Note: All the Calculators listed in this site are for educational purpose only and we do not guarentee the accuracy of results. Please do consult with other sources as well.