📌 Introduction

This project was built for learning and research purposes, focusing on the EfficientNet architecture in Computer Vision.
Specifically, the model is trained completely from scratch to solve the task of vehicle classification (bus, family sedan, fire engine, etc.).

---

📂 Dataset

• The project uses the Vehicle Dataset from Kaggle.
• The dataset contains various types of vehicles for image classification tasks.

---

🧠 Model Architecture

• Implements all EfficientNet versions (B0 → B7).
• The model is implemented entirely from scratch, without any pretrained weights.
• Focuses on correctly reproducing the EfficientNet design philosophy:
  • Compound scaling: scaling depth, width, and resolution simultaneously.
  • MBConv blocks: using Depthwise Separable Convolution combined with Squeeze-and-Excitation.

---

📊 Results

Model	Accuracy
B0	81.5

---

📎 References

• Original paper: EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
• Kaggle Dataset: Vehicle Dataset

---

⚙️ Installation

bash
conda env create -f environment.yml # Need to install anaconda/miniconda before
conda activate ai_env
git clone https://github.com/cor1211/efficentNet_from_scratch.git

---

🚀 Using

bash
python test.py --image_path "your_image_path"

✨ This repository is for learning & research purposes, designed for anyone who wants to understand how to build EfficientNet from scratch.