https://github.com/erikbavenstrand/neural-network-implementation
An Artificial Neural Network Implementation in Numpy to work with MNIST dataset.
https://github.com/erikbavenstrand/neural-network-implementation
ai feedforward-neural-network machine-learning mnist neural-network numpy python
Last synced: about 2 months ago
JSON representation
An Artificial Neural Network Implementation in Numpy to work with MNIST dataset.
- Host: GitHub
- URL: https://github.com/erikbavenstrand/neural-network-implementation
- Owner: ErikBavenstrand
- License: mit
- Created: 2019-11-20T13:10:41.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2024-01-27T15:45:00.000Z (over 2 years ago)
- Last Synced: 2025-03-27T16:41:11.555Z (about 1 year ago)
- Topics: ai, feedforward-neural-network, machine-learning, mnist, neural-network, numpy, python
- Language: Python
- Homepage: https://bavenstrand.se
- Size: 11.2 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Neural Network Implementation with Numpy
> The project was done for an introductory course in artificial intelligence. The work was done in groups of two.
The project aimed to implement a simple artificial neural network in Python using Numpy. To then evaluate the implementation, the famous MNIST dataset was used where we achieved a **96%** accuracy. To further investigate the topic, we created our own very small dataset using paint and with that, we achieved a **70%** accuracy. More information about how the neural network was implemented can be found in the project [report](Neural-Network_Report.pdf).
## Usage example
First, you need to install the requirements found in the requirements.txt file. This is preferably done inside a virtual environment.
```sh
pip install -r requirements.txt
```
After that, the following command can be run to construct and train a network.
```sh
python main.py 784 100 10 NN.bin
```
## Development setup
Python 3 is required.
## Meta
Erik Båvenstrand – [Portfolio](https://bavenstrand.se) – erik@bavenstrand.se
Distributed under the MIT license. See ``LICENSE`` for more information.
[github.com/ErikBavenstrand](https://github.com/ErikBavenstrand)