How does "tiling" in [6] work?

In [6]: Tile/copy our data to increase the training set size and reduce the number of training epochs.

How does it work? Does it just copy (row-wise) each batch of the training data (X of shape (200, 2), and Y of shape (200, 1)) 1,000 times, i.e. adding vertically in the matrix 999 more batches of the same data (stretching it by extra 199,800 rows)?

1 Like

Hello @Robert_Ascan

I suggest you to find it out yourself. You may start with a toy dataset with just 2 rows and 3 columns, use tile the same way the lab does, but with less repeating times (change from 1000 to maybe just 3), then print the tile result.

You may also want to read the docs page for tile for what else it can do.