carpentries-incubator / lesson-gpu-programming

GPU Programming with Python and CUDA.
https://carpentries-incubator.github.io/lesson-gpu-programming/
Other
20 stars 12 forks source link

Explain convolution #29

Closed isazi closed 2 years ago

isazi commented 3 years ago

In episode 2 (CuPy) we use a convolution operation on an image to explain CuPy. We need a sentence or two to explain what the convolution operation is.

Codemonk-adi commented 3 years ago

Would like to work on this issue, have experience in deep learning models including Convolution models

isazi commented 3 years ago

Sure, let me know if you write something.

Codemonk-adi commented 3 years ago

The Convolution Operation takes an input image and a feature detector(i.e, filter). The feature detector is placed on the input image and dot product of the feature detector and the overlapping image is calculated.

The filter is then slid across the image mapping it into a resultant output image.i.e, the output of the convolution Operation

Codemonk-adi commented 3 years ago

let me if know if this is okay and may I be assigned for the other issues as well?

isazi commented 3 years ago

Send me a pull request for this change and I'll review and approve it.

MaciejWas commented 3 years ago

Hi, I think that for convolutions it would be best to provide some sort of visualization. There are tons of those such as this one. Why not include it (of course with crediting the author) ?

isazi commented 3 years ago

This is also a good idea @MaciejWas

isazi commented 2 years ago

After all the contributions, I think we can close this issue.