Convolutional Networks

CNN Explorer

Follow a small image through convolution, ReLU, pooling and scores.

y[r,c] = Σ input[r×stride+i−pad,c×stride+j−pad] × kernel[i,j]

0×1 + 0×0 + 1×-1 + 0×1 + 1×0 + 1×-1 + 1×1 + 1×0 + 1×-1

Output = -2

Output size = floor((input + 2×padding − kernel) / stride) + 1floor((5 + 2×0 − 3) / 1) + 1 = 3

Speed ×Step 0 / 8

End-to-end toy CNN

Fixed educational filters and dense weights, not a pretrained recognizer. Its three scores demonstrate the calculation only.

Flattened features · 1 × 4

Dense logits · 1 × 3

Softmax outputs · 1 × 3

Opening this lab…

No account required. Local images and model files stay in your browser. Educational models demonstrate mechanisms; they are not production predictors.

CNN Explorer

End-to-end toy CNN

Continue exploring

End-to-end toy CNN