CIFAR-10 Image Classification (CNN)

This project focuses on building and training a Convolutional Neural Network (CNN) from scratch using PyTorch to classify 32x32 color images into 10 distinct categories.

---

🧠 Model Architecture (MyNet)

The network is built by subclassing nn.Module and features a robust custom structure:

• Convolutional Layers: 2 layers with a progressive filter design (128 and 512 channels) and a 5x5 kernel [📌, 📌].
• Pooling: nn.MaxPool2d layers for effective spatial downsampling [📌].
• Fully Connected Layers: A sequence of 4 dense layers (18432 -> 128 -> 128 -> 64 -> 10) for final class score calculation [📌].

---

📊 Results & Performance

• Average Test Accuracy: 73.74% [📌]

Class-by-Class Accuracy Breakdown:

• 🚗 Automobile: 87%
• 🐎 Horse: 83%
• 🚢 Ship: 83%
• 🚚 Truck: 81%
• ✈️ Airplane: 80%
• 🐸 Frog: 80%
• 🦌 Deer: 70%
• 🐶 Dog: 68%
• 🦅 Bird: 53%
• 🐱 Cat: 48% (Main area identified for future architecture tuning) [📌]

---

🛠️ Installation & Setup

1. Clone this repository:

   git clone https://github.com
   cd your-repo-name

2. Install the required dependencies:

   pip install -r requirements.txt