https://github.com/profjuvii/neural-net-from-scratch
Repository for writing a Neural Network in the C Language
https://github.com/profjuvii/neural-net-from-scratch
c cpp library machine-learning mlp mnist neural-networks perceptron synapse
Last synced: 7 months ago
JSON representation
Repository for writing a Neural Network in the C Language
- Host: GitHub
- URL: https://github.com/profjuvii/neural-net-from-scratch
- Owner: profjuvii
- License: mit
- Created: 2024-07-19T10:56:14.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-01-17T03:08:37.000Z (9 months ago)
- Last Synced: 2025-01-17T04:17:47.981Z (9 months ago)
- Topics: c, cpp, library, machine-learning, mlp, mnist, neural-networks, perceptron, synapse
- Language: C
- Homepage:
- Size: 13.9 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Neural Network in the C Language
This repository was created to deepen my understanding of neural networks, enhance my programming skills, and gain experience in building libraries in C.
## Description
This project implements a library named **Synapse** for constructing various MLP (Multilayer Perceptron) architectures in C/C++. The library is optimized for fast training of large models and offers flexibility for easily adjusting hyperparameters. By adhering to encapsulation within the project’s modules, the library simplifies usage while maintaining full control over the training process.
## Installation
To clone this repository, execute the following bash command:
```bash
git clone https://github.com/profjuvii/neural-net-from-scratch.git
```
## Navigation
The repository is organized into several folders, each serving a specific purpose:
- **synapse**: Contains the core neural network library.
- **perceptron**: Provides an example of using the library to train a model for solving the logical AND problem.
- **mnist_preparation**: Includes a Python script for preparing handwritten digit data from the MNIST dataset in CSV format.
- **mnist_training**: Contains code for training a model on pre-processed MNIST data.
- **models**: Stores pre-trained models.
To train your own model, ensure that the necessary compiler extensions are included as follows:
```bash
-I../synapse/include -L../synapse/lib -lsynapse
```
> [!NOTE]
>
> This depends on the location of the project files in your environment.
## Requirements for Training the MNIST Model
To run the code for training a neural network on handwritten MNIST data, follow these steps:
1. **Prepare MNIST Data**:
- In the `mnist_preparation/` folder, you need to unpack the following two archives:
```bash
gunzip train-images-idx3-ubyte.gz
gunzip train-labels-idx1-ubyte.gz
```
- Next, set up a Python virtual environment:
```bash
python3 -m venv .venv
```
- Activate the virtual environment:
```bash
source .venv/bin/activate
```
- Install the necessary library:
```bash
pip install idx2numpy
```
- After this, you can run the MNIST data preparation script.
2. **Build and Run the Training Code**:
- In the `mnist_training/` folder, run the following command to compile the code:
```bash
make
```
- Once the build is complete, execute the program using:
```bash
./main
```
Once these steps are completed, the training of the neural network on the MNIST dataset will begin as intended.
Thanks to the hyperparameter settings defined in the code, I achieved an accuracy of 93.9%. The trained model has been saved in the `models/` folder.
> [!NOTE]
>
> You can also find all the library functions in the header files located in `synapse/include/`.
## Author
Created by [Denys Bondarchuk](https://github.com/profjuvii). Feel free to reach out or contribute to the project!