Reward Modulated Self-Organizing Recurrent Neural Networks
Project description
Reward Modulated Self-Organizing Recurrent Neural Networks
RMSORN is a subclass of neuro-inspired artificial network, Self Organizing Recurrent Neural Networks. With reward driven self-organization, this network achieves performance with networks trained with supervised learning algorithms.
To install the latest release:
pip install rmsorn
The library is still in alpha stage, so you may also want to install the latest version from the development branch:
pip install git+https://github.com/Saran-nns/rmsorn
Usage:
Update Network configurations
Navigate to home/conda/envs/ENVNAME/Lib/site-packages/rmsorn
or if you are unsure about the directory of rmsorn
Run
import rmsorn
rmsorn.__file__
to find the location of the rmsorn package
Then, update/edit the configuration.ini
from rmsorn.tasks import PatternRecognition
inputs, targets = PatternRecognitionTask.generate_sequence()
train_plast_inp_mat,X_all_inp,Y_all_inp,R_all, frac_pos_active_conn = SimulateRMSorn(phase = 'Plasticity',
matrices = None,
inputs = np.asarray(inputs),sequence_length = 4, targets = targets,
reward_window_sizes = [1,5,10,20],
epochs = 1).train_sorn()
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
rmsorn-0.0.6.tar.gz
(18.7 kB
view hashes)
Built Distribution
rmsorn-0.0.6-py3-none-any.whl
(19.8 kB
view hashes)