Re: Bug#217648: ITP: Torch3 -- libtorch3: machine learning library version 3
On Sun, 2003-10-26 at 09:15, Cosimo Alfarano wrote:
> * Package name : libtorch3c102
Correct me if I'm wrong, but I don't think the suffix c102 is required
for packages which didn't have a non-c102 version in Debian.
> Version : 0.0
> Upstream Author : Ronan Collobert, Samy Bengio and Johnny Marithoz
> * URL : http://www.torch.ch/
> * License : BSD
> Description : Torch3: machine learning library version 3
"machine learning library" should be sufficient for the short
description. The version is already encoded in the package name (SONAME)
and the Version field.
> This package is the Torch runtime library.
> Torch is a machine-learning library, written in C++. Its aim is to
> provide the state-of-the-art of the best algorithms.
Best algorithms for what? How are they the best?
> It is, and it
> will be, in development forever.
I'm torn here as to whether this should be in the description. Does this
mean that the interface is in flux? Is the library itself unstable? If
so, that definitely belongs in the description.
If it just means that new versions are constantly coming out, most
likely it can be dropped. That's true of most software.
> Main features:
> * Many gradient-based methods, including multi-layered perceptrons,
> radial basis functions, and mixtures of experts. Many small "modules"
> (Linear module, Tanh module, SoftMax module, ...) can be plugged
> * Support Vector Machine, for classification and regression.
> * Distribution package, includes Kmeans, Gaussian Mixture Models,
> Hidden Markov Models, and Bayes Classifier, and classes for speech
> recognition with embedded training.
> * Ensemble models such as Bagging and Adaboost.
> * Non-parametric models such as K-nearest-neighbors, Parzen Regression
> and Parzen Density Estimator.
> Torch is an open library whose authors encourage everybody to develop
> new packages to be included in future versions on the official website.
The bullets here aren't necessary, as each of the items is already
either a sentence or a clause in a list. I'd roll all of them into two
or three paragraphs.
Joe Drew <firstname.lastname@example.org> <email@example.com>
My weblog doesn't detail my personal life: http://me.woot.net