Counter Propagation based on Prototorch.
Project description
About
This is the implementation of Counterpropagation Models based on Prototorch by SICIM.
The implemented model is mainly based on the report by Villmann, Schubert and Kaden (2021) and the paper Kaden et. al. (2021). However, the original Counterpropagation model can be found in Hecht-Nielsen (1987).
Note, that the original Model is not implemented here (i.e. non-differentiable SOM initialization + perceptron-layer), since this implementation is focused on influencing the response by the subsequent supervised layer.
However, a simple approach to implement the Hecht-Nielsen model is to initialize a SOM on the data and use the assignments to create the subsequent perceptron layer.
Contact
bugs, fixes, ideas etc. can be send to trebuhcsynnor@gmail.com
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for counterprop_prototorch-1.0.1.tar.gz
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | 04825cd114ac901d4d55f4a81df748e07c0e5b920f4e12d69ec401e249f92fae |
|
| MD5 | 16753421b8ecc16518b441643587b728 |
|
| BLAKE2b-256 | 8e8048e86d807022d5ad228658f3a0f5093a9363a32852f1c7a57c2a3e5028a2 |
Hashes for counterprop_prototorch-1.0.1-py3-none-any.whl
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | 51bf81c54a76236c93e4544d145513521f1dc228bc92f5010d4d22acd018b6d6 |
|
| MD5 | 40d2207e804fec2035d61e7fabe81c18 |
|
| BLAKE2b-256 | b34a347dfc68ef0378d9894b8cb9b20841317776f630161ba675e350cfc78890 |
