Implement L2 regularization (weight decay) penalty.
The L2 regularization term is:
$$\text{L2} = \frac{\lambda}{2} \sum_{i} w_i^2$$
where $\lambda$ is the regularization strength and $w$ are the model weights.
Input:
weights: 1D tensor of model weights lambda_: regularization strength (float) Output: A scalar representing the L2 penalty