Animal Detection Model

This repository contains code for a Python-based animal detection model using a pre-trained video classifier. The model can detect various animals in a given video.

Getting Started

To use this code, follow these instructions:

1. Clone this repository to your local machine or download the code files.
2. Ensure you have Python installed on your system.
3. Install the required dependencies using pip:

  pip install pytorch
  pip install pytorchvideo
  pip install gradio

4. Download the pre-trained model weights and class names file by running wget https://dl.fbaipublicfiles.com/pyslowfast/dataset/class_names/kinetics_classnames.json.
5. Open the provided Jupyter notebook (animal_model_with_interface.ipynb) in your preferred environment.

Usage

To use the model:

1. Instantiate the VideoClassifier class provided in the notebook.
2. Call the classify_video() method, passing the path to the video file you want to analyze.
3. The method returns a dictionary indicating whether any animals were found in the video.

Example usage:

# Instantiate your VideoClassifier
classifier = VideoClassifier()

# Call the classify_video method with the path to your video
result = classifier.classify_video("path/to/your/video.mp4")

# Print the result
print(result)

Gradio Interface

Alternatively, you can use the Gradio interface for easy interaction with the model. The notebook provides instructions and code snippets for setting up and launching the interface.

License

This code is provided under the MIT License. Feel free to modify and distribute it as needed.