Watch Loss decrease and Accuracy increase as the neural network learns epoch by epoch
Loss & Accuracy Curves
Loss (lower is better)
Accuracy (higher is better)
Training Controls
Total Epochs20
Learning Rate0.010
Noise Level0.020
Playback Speed1.00
Live Metrics
Current Epoch
0
Current Loss
0.680
Accuracy
50.0%
LR Schedule
Warmup
Ready: Click "Start Training" to watch the loss curve and accuracy evolve dynamically.
Reading Guide
Slow convergence with stalled accuracy: Learning rate is likely too small. Wildly oscillating curves: Learning rate is too large or noise level is too high. Architecture: Input(400) - Hidden1(25) - Hidden2(15) - Output(1) with 10,431 total parameters.