TIKAPP is becoming a collection of tools for simulation of neural networks. The first available part is an ANSI-C++ library with support for backpropation networks.
Release Date:
Topics:
License:
Ratings and Reviews
Be the first to post a text review of TIKAPP - Neural network library and to. Rate and review a project by clicking thumbs up or thumbs down in the right column.
Project Feed
-
Tracker artifact added
Anonymous created the hacMNQSt artifact -
libtikapp 0.4.0 file released: tikapp_0.4.0.tar.gz
libtikapp 0.4.0 - Switched from qmake to automake - Debianized the source package - Transformed those ugly static ...::load() functions into regular constructors - Introduced a parent-child relationship between TNetObject objects. Any object can now be delete with a simple 'delete ptr'. It deregisters itself from its parent while preserving data interity. - Removed size attribute of TLayers - Every module now has an empty input layer be default - Every TMLPModule now has an empty output layer by default - The last layer (meaning last in the list of layer pointers) is now automatically considered to be the output layer of an TMLPModule - Child storage is now implemented in a generic way for every network object using the template class TChildPtrVector - A TNeuron does not store any information about outgoing connections anymore. This will save a lot of memory. This information is now only stored in TFlexNeuron which need this for error calculation using backpropagation. - All constructor of network object now take a TPosition instead of three integer values - Bugfix: Corrected missing vector clearance in TNetwork::close()
-
libtikapp 0.4.0 file released: tikapp_0.4.0.zip
libtikapp 0.4.0 - Switched from qmake to automake - Debianized the source package - Transformed those ugly static ...::load() functions into regular constructors - Introduced a parent-child relationship between TNetObject objects. Any object can now be delete with a simple 'delete ptr'. It deregisters itself from its parent while preserving data interity. - Removed size attribute of TLayers - Every module now has an empty input layer be default - Every TMLPModule now has an empty output layer by default - The last layer (meaning last in the list of layer pointers) is now automatically considered to be the output layer of an TMLPModule - Child storage is now implemented in a generic way for every network object using the template class TChildPtrVector - A TNeuron does not store any information about outgoing connections anymore. This will save a lot of memory. This information is now only stored in TFlexNeuron which need this for error calculation using backpropagation. - All constructor of network object now take a TPosition instead of three integer values - Bugfix: Corrected missing vector clearance in TNetwork::close()
-
libtikapp 0.4.0 file released: tikapp_0.4.0_i386.deb
libtikapp 0.4.0 - Switched from qmake to automake - Debianized the source package - Transformed those ugly static ...::load() functions into regular constructors - Introduced a parent-child relationship between TNetObject objects. Any object can now be delete with a simple 'delete ptr'. It deregisters itself from its parent while preserving data interity. - Removed size attribute of TLayers - Every module now has an empty input layer be default - Every TMLPModule now has an empty output layer by default - The last layer (meaning last in the list of layer pointers) is now automatically considered to be the output layer of an TMLPModule - Child storage is now implemented in a generic way for every network object using the template class TChildPtrVector - A TNeuron does not store any information about outgoing connections anymore. This will save a lot of memory. This information is now only stored in TFlexNeuron which need this for error calculation using backpropagation. - All constructor of network object now take a TPosition instead of three integer values - Bugfix: Corrected missing vector clearance in TNetwork::close()
-
libtikapp 0.4.0 file released: libtikapp0-dev_0.4.0_i386.deb
libtikapp 0.4.0 - Switched from qmake to automake - Debianized the source package - Transformed those ugly static ...::load() functions into regular constructors - Introduced a parent-child relationship between TNetObject objects. Any object can now be delete with a simple 'delete ptr'. It deregisters itself from its parent while preserving data interity. - Removed size attribute of TLayers - Every module now has an empty input layer be default - Every TMLPModule now has an empty output layer by default - The last layer (meaning last in the list of layer pointers) is now automatically considered to be the output layer of an TMLPModule - Child storage is now implemented in a generic way for every network object using the template class TChildPtrVector - A TNeuron does not store any information about outgoing connections anymore. This will save a lot of memory. This information is now only stored in TFlexNeuron which need this for error calculation using backpropagation. - All constructor of network object now take a TPosition instead of three integer values - Bugfix: Corrected missing vector clearance in TNetwork::close()
-
libtikapp 0.4.0 file released: libtikapp0_0.4.0_i386.deb
libtikapp 0.4.0 - Switched from qmake to automake - Debianized the source package - Transformed those ugly static ...::load() functions into regular constructors - Introduced a parent-child relationship between TNetObject objects. Any object can now be delete with a simple 'delete ptr'. It deregisters itself from its parent while preserving data interity. - Removed size attribute of TLayers - Every module now has an empty input layer be default - Every TMLPModule now has an empty output layer by default - The last layer (meaning last in the list of layer pointers) is now automatically considered to be the output layer of an TMLPModule - Child storage is now implemented in a generic way for every network object using the template class TChildPtrVector - A TNeuron does not store any information about outgoing connections anymore. This will save a lot of memory. This information is now only stored in TFlexNeuron which need this for error calculation using backpropagation. - All constructor of network object now take a TPosition instead of three integer values - Bugfix: Corrected missing vector clearance in TNetwork::close()
-
File released: /libtikapp/0.4.0/tikapp_0.4.0.zip
-
File released: /libtikapp/0.4.0/tikapp_0.4.0.tar.gz
-
File released: /libtikapp/0.4.0/tikapp_0.4.0_i386.deb
-
File released: /libtikapp/0.4.0/libtikapp0_0.4.0_i386.deb
Rate and Review
Be the first person to add a text review.
Related Projects
-
niftilib
niftilib is a collection of i/o routines for the nifti1 neuroimage data format. C (nifticlib), Java (niftijlib), Matlab (niftimatlib), and Python (pynifti) code is available. Information about the nifti format is available at http://nifti.nimh.nih.gov/df
-
rt0d4
Random Thoughts On Distributed Algorithms: a collection of tools and ideas for the development of distributed applications. The first released part is an RxRPC security model based on OpenSSL and elliptic curve criptography. Please check docs for info...
-
TIKAPP - Neural network library and to Actions
About SourceForge
Develop Software
Copyright © 2009 Geeknet, Inc. All rights reserved. Terms of Use
