A Python-based binary image classifier using PyTorch, designed to detect wildfires in images.
This project uses the Wildfire Detection Image Data from Kaggle.
Before running the project:
- Download the dataset from Kaggle.
- Unzip the contents.
- Place the
forest_firedirectory in the root of this repository.
- uv – Python version and dependency manager
Install dependencies with:
uv syncRun the classifier with:
uv run main.pyOn the first run, the model is trained, and inference is performed.
On subsequent runs, the weights are loaded and only inference is performed.