Binary Image Classifier

A Python-based binary image classifier using PyTorch, designed to detect wildfires in images.

Dataset

This project uses the Wildfire Detection Image Data from Kaggle.

Before running the project:

1. Download the dataset from Kaggle.
2. Unzip the contents.
3. Place the forest_fire directory in the root of this repository.

Prerequisites

• uv – Python version and dependency manager

Installation

Install dependencies with:

uv sync

Usage

Run the classifier with:

uv run main.py

On the first run, the model is trained, and inference is performed.
On subsequent runs, the weights are loaded and only inference is performed.