MLP(3, [4, 4, 1]) Visualizer

Built by Lijian Liu
to visualize Karpathy's Micrograd
Network Architecture
Input
Layer 0
Layer 1
Layer 2
Loss
Bias
Select Input Samples (click to toggle)
Predictions vs Targets
Sample 0: 0.0000 → 1.0
PARAMETER COUNT
Layer W shape b shape Calculation Params
Layer 0 (4, 3) (4,) 4×3 + 4 16
Layer 1 (4, 4) (4,) 4×4 + 4 20
Layer 2 (1, 4) (1,) 1×4 + 1 5
Total 16+20+5 41
Parameters (hover to highlight, click to lock)
Computation Graph (DAG)
Built by Lijian Liu to visualize Karpathy's Micrograd
Loss —
Iter 0
Train:
← Scroll horizontally to see full graph →
■ Selected → Inputs → Outputs (arrows show data flow)