Giuseppe Cuccu

icon indicating a website link exascale.info/giuse icon indicating an email [email protected]

icon company Exascale Infolab, University of Fribourg icon location Switzerland icon location Bringing state-of-the-art machine learning into real world solutions

Results 21 comments of Giuseppe Cuccu

Welcome Note

Sounds awesome! Count me in! On Sun, Apr 23, 2017 at 8:23 AM, Sameer Deshmukh wrote: > Thanks for this initiative! I'll have a look and any libraries that I...

Feature/docker support

Awesome!

Bug in `Numo::DFloat#sort`

The latest published gem version is 0.9.1.2, which does not include the commit e3d14227e9574ccf8c8415751c8a6a97c6da3dd0 which fixes this bug. Could you please release an hotfix? I have my reasons for not...

API and method return types unification by making Numo a tensor library [ex: Arithmetic overflow]

I would not suggest to add overflow checks. I agree on the purpose of speed. But if `#sum` returns `Integer`, I do not expect overflow. If `#sum` returns `Numo::UInt8` (or...

API and method return types unification by making Numo a tensor library [ex: Arithmetic overflow]

I would not check for overflow. I feel like when the return type is `Integer` there should not be any of the problems associated with the original type. What I...

API and method return types unification by making Numo a tensor library [ex: Arithmetic overflow]

If it is "typed" it should return a `NArray`. Why does it return an `Integer`? This way I would know there is overflow, and I can do the cast to...

API and method return types unification by making Numo a tensor library [ex: Arithmetic overflow]

Thank you for taking this issue under consideration. I sincerely care for Numo, I hope my enthusiasm correctly transpires as my fullest and most serious support. I think using zero-dimensional...

[Propose] Return 0-dimensional array instead of Ruby numeric object

I second this proposition. Had a long discussion on the mathematical reasons behind it in issue #91 (which incidentally would be solved by this).

Numo::NArray#var

In my opinion, I prefer Numo's choice of default. The standard use of Numo is for data samples, for which the correct choice of variance is the sample variance (denominator...

Optimize marshaling in parallel computation by allowing constructing Numo arrays from pre-allocated data pointers, such as coming from views.

You can see it is not as easy as defining the returning data structure as a closure and send it to all children (you will need to `$ gem install...