gneuralnetwork
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gneuralnetwork] Gneural Network 0.8.0 has been released!


From: Nan .
Subject: Re: [Gneuralnetwork] Gneural Network 0.8.0 has been released!
Date: Sat, 2 Apr 2016 12:25:07 -0500

yes, you are correct, my bad, sorry. :P

you use *wbest, which is double too.
usually I use double, because the memory is allocated for storing double numbers.


Date: Sat, 2 Apr 2016 18:40:40 +0200
Subject: Re: [Gneuralnetwork] Gneural Network 0.8.0 has been released!
From: address@hidden
To: address@hidden
CC: address@hidden

Thanks Nan!

As far as I can tell the two lines (I mean mine and the one you suggest) should be equivalent, in the sense that they allocate the same amount of memory.

JM


2016-04-02 16:51 GMT+02:00 Nan . <address@hidden>:
Congrats,

When I looked at the code, Monte Carlo seems have a small issue.

file gneural_network-0.8.0/src/msmco.c, line 38

should be:

wbest=malloc((NNUM*MAX_IN+1)*sizeof(double));


, if I understand the code right.

Thanks,

Nan.


Date: Sat, 2 Apr 2016 08:53:08 +0200
From: address@hidden
To: address@hidden
CC: address@hidden
Subject: [Gneuralnetwork] Gneural Network 0.8.0 has been released!


Hello Everyone,

I have the pleasure to let you know that a new version of Gneural Network has been released! You can find it here:


Gneural Network is the GNU package which implements a programmable neural network. In this version several important features/changes have been introduced. In particular:

- The previous genetic algorithm has been drastically improved.
- An advanced multi-scale Monte Carlo optimizer has been implemented to train the neural network.
- New scripts have been implemented and are provided in the folder "tests", one which fits a quandratic curve, another which trains a network to compute the modulus of a two-dimensional vector and, finally, one which trains a network to invert the order of numbers in a given couple (a,b).
- Several bugs have been fixed.

A big thanks goes to everyone who has contributed to this new version!

Happy Neural Hacking!

JM

--



--

reply via email to

[Prev in Thread] Current Thread [Next in Thread]