[mlpack-git] [mlpack] add train test split (#523)

[Marcus Edel]

I guess a simple example would make things more clear. Imagine you have a dataset with a bunch of colour images (e.g. CIFAR-10), so more than one channel per image or dataset, something like that:

<p align="center">
<img src="https://urgs.org/marcus/media/we_need_to_go_deeper.png" alt="we need to go deeper color">
</p>

So, I guess I would store that image in a cube with 3 slices one slice per channel. But now I would like to store 10 images in the same data structure that would result in a cube with 10 * 3 slices.

As far as I can see is that your code works with data sets that contains one channel per image, something like that would be totally fine:

<p align="center">
<img src="https://urgs.org/marcus/media/we_need_to_go_deeper_gray.png" alt="we need to go deeper gray">
</p>

As I said I'm not sure we should handle such datasets with the code. I guess, we have to specify the number of channels/slices per dataset somehow to make it work.

I hope this makes some sense.

---
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/pull/523#issuecomment-188945938
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160225/a941208a/attachment.html>