[mlpack-git] [mlpack/mlpack] Added the discretize function. Design changes as discussed. (#703)

[Marcus Edel]

> +    size_t y = loc(i, 1);
> +
> +    /*(x,y) in channels, is ((x+p), (y+p)) in incCh*/
> +    CubeType patch = incCh.tube((x + p) - p, (y + p) - p,
> +                          (x + p) + p - 1, (y + p) + p - 1);
> +    // since each patch has 13 channel we have to increase the index by 13    
> +    features.slices(channel, channel + 12) = patch;
> +    channel += 13;
> +  }
> +}
> +
> +template<typename MatType, typename CubeType>
> +void StructuredForests<MatType, CubeType>::
> +Rearrange(const CubeType& channels, CubeType& ch)
> +{
> +  //we do (16,16,13*1000) to 256, 1000, 13, in vectorized code.

Can you explain why we use 256, I think we could also use a value > 256.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/pull/703/files/72eb1ef22c7db1ea33af3de1cd043cdb277ec562#r74704389
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160814/b32c7a0a/attachment.html>