https://github.com/ncqxm/cifar10-cnn

Image classification on CIFAR-10 using a custom CNN built with TensorFlow/Keras. Visualized performance with confusion matrix and training curves.
https://github.com/ncqxm/cifar10-cnn

cifar10 cnn deep-learning keras machine-learning tensorflow

Last synced: 3 months ago
JSON representation

Image classification on CIFAR-10 using a custom CNN built with TensorFlow/Keras. Visualized performance with confusion matrix and training curves.

• Host: GitHub
• URL: https://github.com/ncqxm/cifar10-cnn
• Owner: ncqxm
• Created: 2025-03-22T13:53:51.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2025-03-22T14:26:55.000Z (over 1 year ago)
• Last Synced: 2025-04-01T02:45:58.712Z (over 1 year ago)
• Topics: cifar10, cnn, deep-learning, keras, machine-learning, tensorflow
• Language: Jupyter Notebook
• Homepage: https://colab.research.google.com/github/ncqxm/cifar10-cnn/blob/main/CIFAR_10_cnn.ipynb
• Size: 33.9 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# 🧠 CIFAR-10 Image Classification using CNN

[![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/ncqxm/cifar10-cnn/blob/main/CIFAR_10_cnn.ipynb)

This project implements a custom Convolutional Neural Network (CNN) using TensorFlow/Keras to classify images in the CIFAR-10 dataset.

---

## Features

- Custom CNN with BatchNormalization and Dropout

- Trained on 10 classes of CIFAR-10 dataset

- Achieved up to 87% test accuracy

- Visualizations:

  - Training Accuracy & Loss

  - Confusion Matrix

  - Classification Report

---

## **Dataset**

- CIFAR-10 (60,000 images, 10 classes)

- Available via: `tensorflow.keras.datasets.cifar10`

---

## Tech Stack

![Python](https://img.shields.io/badge/Python-3.x-blue)

![TensorFlow](https://img.shields.io/badge/TensorFlow-2.x-orange)

![Jupyter](https://img.shields.io/badge/Jupyter-Notebook-informational)

---

## Results

### Confusion Matrix  

![Confusion Matrix](images/confusion_matrix.png)

### Training Curves  

![Training Metrics](images/training_metrics.png)

---

## How to Run

1. Clone this repo:

   ```bash

   git clone https://github.com/yourusername/cifar10-cnn.git

   cd cifar10-cnn

   

2. Install dependencies:

   ```bash

   pip install -r requirements.txt

3. Run the notebook:

   Open cifar10_cnn.ipynb using Jupyter or Colab

---

## Trained Model

The trained model is saved as `models/cifar10_cnn_v2.h5`.

```python

from tensorflow.keras.models import load_model

model = load_model("models/cifar10_cnn_v2.h5")