What’s the meaning of “T” in the screenshot?

Hi @Jinyan_Liu.

T here represents the transpose operation.

`Transposing a matrix involves flipping its rows and columns, effectively turning rows into columns and columns into rows.`

In the exercise,

X.shape = (4778, 10)

W.shape = (443, 10)

So it matches: without T added there.

Isn’t x(0).shape = (1, 10)? So x(0)T.shape should be (10,1)? Then X becomes (10, 4778)?

Even later in the exercise, only W is transposed. I don’t understand what is transposed here in the picture.

The figure comes from an older Machine Learning lecture, where the ‘x’ features and ‘w’ weights are column vectors.

It doesn’t apply if the features and weights are stored as row vectors.

You will get used to applying transpositions as necessary, as there isn’t a universal agreement or consistency about the orientation of X or W.

Thanks! Now I understand!