Fixed typo in dimensions of bias term

[deepam20050]

Description of changes:

By submitting this pull request, I confirm that you can use, modify,
copy, and redistribute this contribution, under the terms of your
choice.

In the Vectorization section,

We have
$\mathbf{X} \in \mathbb{R}^{n \times d}$, $\mathbf{W} \in \mathbb{R}^{d \times q}$.

Now, since:
$$\mathbf{O} = \mathbf{X} \mathbf{W} + \mathbf{b}$$
and
$$\hat{\mathbf{Y}} = \mathrm{softmax}(\mathbf{O}),$$

we require $\mathbf{b} \in \mathbb{R}^{n \times q}$. This ensures that the softmax operation can be computed row-wise on $\mathbf{O}$.