[mlpack] GSoC 17: better benchmarking

Ryan Curtin ryan at ratml.org
Fri Mar 3 10:28:41 EST 2017


On Fri, Mar 03, 2017 at 03:51:02PM +0800, Thyrix wrote:
> Hi,
> 
> My name is Jiaqi Yang(Thyrix on irc). I'm a third year electric engineering
> undergraduate at Wuhan University, China.
> 
> I'm very interested in machine learning and I have finished a pattern
> recognition course in my university. I also took the open course by Andrew
> Ng on machine learning, and searched a lot in this area.
> 
> I have read through the idea list, and I think I can work on the "better
> benchmarking" task.
> 
> I have been participating in competitive programming contest(won ACM\ICPC
> China regional gold medal)in the last two years. I'm proficient in c++, and
> also have some python experience.
> 
> I have seen the issue about model zoo. Can this be a idea of GSoC? I'd like
> to work on this idea too.
> 
> Please give me some advice.

Hi Jiaqi,

Thanks for getting in touch.  If you are interested in the benchmarking
project, I'd suggest getting familiar with the code:

https://github.com/zoq/benchmarks

There is also an issue open now that could provide as a good starting
point:

https://github.com/mlpack/mlpack/issues/886

The model zoo could also be made into a GSoC project (or part of a GSoC
project); if you'd like to propose something in this direction, be sure
that whatever you propose has a clear plan associated with it.  I think
that the idea of the model zoo is still not fully formalized, so there
would still be some design work to do.  Marcus might already have
further ideas about how that could work, so I'll defer to his
perspective on how the whole thing should be put together.

Let me know if I can clarify anything I've written here.

Thanks,

Ryan

-- 
Ryan Curtin    | "I am the luckiest man alive!"
ryan at ratml.org |   - General Borzov


More information about the mlpack mailing list