Change n to m for Y in Enc-Dec section

[adenhaus]

In the Encoder-Decoder section of this post, the output sequence $Y$ is introduced as being of length $m$, i.e. \mathbf{Y}_{1:m} but then inside the following equations, it is denoted as \mathbf{Y}_{1:n}, using the input sequence $X$'s length $n$ instead. But $Y$ does not need to be the same length as $X$. Later in the article, in the decoder section, the formulas do use $m$ instead of $n$ as the length for $Y$.

So I believe this is just a mistake and have proposed a fix. Please do let me know if I am mistaken.

Post author: @patrickvonplaten