https://github.com/balos1/pyann
ANN in Python for understanding how ANNs work.
https://github.com/balos1/pyann
ann artificial-neural-networks perceptron-learning-algorithm perceptron-learning-rule
Last synced: 3 months ago
JSON representation
ANN in Python for understanding how ANNs work.
- Host: GitHub
- URL: https://github.com/balos1/pyann
- Owner: balos1
- Created: 2017-12-09T08:15:16.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2017-12-09T08:15:43.000Z (over 7 years ago)
- Last Synced: 2025-02-09T09:43:14.740Z (4 months ago)
- Topics: ann, artificial-neural-networks, perceptron-learning-algorithm, perceptron-learning-rule
- Language: Python
- Homepage:
- Size: 16.4 MB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
Artificial Neural Networks in Python
===
This was written as ANN learning experience project, and is by no means a super efficient ANN module (yet).
It currently can implement a perceptron model or a ANN with a single hidden layer.
# For Best Results
Use the weights file `weights_perceptron.npy` if using the perceptron network.
Use the weights file `weights_multilayer.npy` if using the multilayer network.
# Dependencies
Requires the numpy module and python 3.6 or newer. This can be installed with pip:
```
> pip install numpy
```
If you don't have pip, here are details on installing numpy: https://scipy.org/install.html
# How to Use
```
usage: pyann.py [-h] [--weights WEIGHTS]
{train,test} {perceptron,multi-layer}
positional arguments:
{train,test} command to run
{perceptron,multi-layer}
what type of ANN to run
optional arguments:
-h, --help show this help message and exit
--weights WEIGHTS load a weights file instead of rng weights
```
note: Some settings such as learning rate, train/test file names are located at the top of project2.py as constants.
### Data File Format
Data files must be tab delimited and in the format ` `