5. MSE Loss

Easy

Implement Mean Squared Error loss.

Signature: def mse_loss(y_pred: np.ndarray, y_true: np.ndarray) -> float

Math

L_{MSE} = \frac{1}{n} i = 1 \sum n (\overset{y}{^}_{i} - y_{i})^{2}

Related problems

Asked at

Python 30/10 runs today

Output

Anything you print() in your code will show up here after you click Run.

Test Results

○perfect

○typical

○large error🔒 Premium

○loss is non-negative

○loss is non-negative (negative residuals)

○symmetric: mse(a,b) == mse(b,a)

○symmetric (swapped): mse(b,a) == mse(a,b)

○2D batch (B=2, D=3) — must normalize over all elements (mean), not sum/B🔒 Premium