News for hackers. Stuff that matters.
Re:Neural Modules (NeMo): A toolkit for conversational AI by mijoharas on Monday September 16, 2019 @05:35AM (#20982391)
I was trying to understand the significance of this. There is a blog post here[0]. (I had a brief glance, please correct me if I'm wrong) It seems to be a framework allowing neural networks (and layers of neural networks) to be composed together more easily.

It's not completely clear to me what the advantages are without digging in, I assume to simplify/standardise the interface between the layers?

[0] https://devblogs.nvidia.com/neural-modules-for-speech-langua...

Re:Neural Modules (NeMo): A toolkit for conversational AI by option on Monday September 16, 2019 @10:28AM (#20984359)
it has “neural type” system which check what can be connected to what. Also, comes with several modules for speech recognition and NLP
Re:Neural Modules (NeMo): A toolkit for conversational AI by falsedan on Monday September 16, 2019 @08:55AM (#20983533)
> A “Neural Module” is a block of code that computes a set of outputs from a set of inputs.

I struggle to find any clarity in this description. You can omit most of it:

> [...] code [...] computes [...] output [...] from [...] input [...]

Which is true, and trivial. This important points for understanding (what kind of input? what kind of output? why sets? what's 'Neural' about functions?) is all missing.

Re:Neural Modules (NeMo): A toolkit for conversational AI by gawi on Monday September 16, 2019 @10:35AM (#20984436)
Further in the text:

> A Neural Module’s inputs/outputs have a Neural Type, that describes the semantics, the axis order, and the dimensions of the input/output tensor. This typing allows Neural Modules to be safely chained together to build applications, as in the ASR example below.

Re:Neural Modules (NeMo): A toolkit for conversational AI by nerdponx on Monday September 16, 2019 @12:43PM (#20985911)
Sounds like a typed Pytorch module then?
Re:Neural Modules (NeMo): A toolkit for conversational AI by falsedan on Tuesday September 17, 2019 @10:55AM (#20995887)
> inputs/outputs have a Neural Type, that describes the semantics, the axis order, and the dimensions of the input/output tensor

Does that really answer any of my questions? What's 'Neural' about the type?

Re:Neural Modules (NeMo): A toolkit for conversational AI by MrEldritch on Wednesday September 18, 2019 @10:38AM (#21005875)
Sounds kind of like what's described in Neural Networks, Types, and Functional Programming: http://colah.github.io/posts/2015-09-NN-Types-FP/
Re:Neural Modules (NeMo): A toolkit for conversational AI by falsedan on Thursday September 19, 2019 @09:47AM (#21016101)
Is that mentioned in the readme? Is this readme worth reading if you don't already know what the project is about?