[mlpack] GSOC'20 mlpack on constrained devices - weekly updates

Germán Lancioni gmansoft at hotmail.com
Wed May 27 14:07:15 EDT 2020


Sorry, I meant Omar, great work!
________________________________
From: Germán Lancioni <gmansoft at hotmail.com>
Sent: Wednesday, May 27, 2020 11:06 AM
To: Ryan Birmingham <rainventions at gmail.com>; Omar Shrit <omar at shrit.me>
Cc: mlpack at lists.mlpack.org <mlpack at lists.mlpack.org>
Subject: Re: [mlpack] GSOC'20 mlpack on constrained devices - weekly updates

Hey Ryan,

Great progress! This is one of the projects that I believe is quite useful for IoT and cloud latency mitigation in RT apps, so exited to see development on this front.

A couple of notes that may help:

1) If you haven't tried yet, you can play a bit with the different dependencies to get an even smaller footprint. I.e. explore other BLAS/LAPACK implementations.
2) Disk footprint is composed by library size (~4MB in your case) + model size. I've recently discovered that some models (when saved) are 10x bigger than let's say sklearn models. Maybe it's a stretch goal, but you could probably explore how to downsize serialized models as well. After all, a real world application would need both the library and the model to fit into i.e. 10MB.

Please keep us updated, amazing work!

Regards,
German

________________________________
From: mlpack <mlpack-bounces at lists.mlpack.org> on behalf of Ryan Birmingham <rainventions at gmail.com>
Sent: Wednesday, May 27, 2020 10:19 AM
To: Omar Shrit <omar at shrit.me>
Cc: mlpack at lists.mlpack.org <mlpack at lists.mlpack.org>
Subject: Re: [mlpack] GSOC'20 mlpack on constrained devices - weekly updates

Cool project and update! Thanks for sharing!
-Ryan Birmingham


On Wed, May 27, 2020 at 12:38 PM Omar Shrit <omar at shrit.me<mailto:omar at shrit.me>> wrote:

Hello everyone,

You can find here my weekly updates on my GSOC project for the last week.

https://shrit.me/blog/

Best regards,

Omar
_______________________________________________
mlpack mailing list
mlpack at lists.mlpack.org<mailto:mlpack at lists.mlpack.org>
http://knife.lugatgt.org/cgi-bin/mailman/listinfo/mlpack
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://knife.lugatgt.org/pipermail/mlpack/attachments/20200527/69291773/attachment-0001.htm>


More information about the mlpack mailing list