Spiking Neural Network Image Classification with SNN-Torch and Gradio This project demonstrates the use of Spiking Neural Networks (SNNs) for image classification, by utilizing the SNN-Torch library. SNNs are biologically inspired models that emulate the spiking behavior of neurons, providing energy-efficient computation. The repository includes a pre-trained SNN model and an interactive demo powered by Gradio.
The model was trained on the CIFAR 10 dataset, as such to test the model's abilities, the photos used for testing should belong to one of the
classes the model can recognize. The available classes are:
["airplane", "automobile", "bird", "cat", "deer", "dog", "frog", "horse", "ship", "truck"]
The demo can be found in this Demo Google Colab Notebook
The Training process can be found in this Training Google Colab Notebook
- Model_App.ipynb: This notebook contains the code to deploy the SNN model using Gradio.
- README.md: The readme file you are currently reading, it provides an overview of the project.
- SNN_Image_Classifier.ipynb: This notebook contains the code for creating and training an SNN model for image classification.
- SNN_Model.pth: The pre-trained SNN model saved in PyTorch's .pth format. This model is used in the Gradio app for image classification.
- requirements.txt: A list of required Python packages and libraries needed to run the notebooks locally. Install these dependencies using pip install -r requirements.txt.
SNN-Torch Documentation: Learn more about SNNs and how to use SNN-Torch in the official documentation.
SNN-Torch GitHub Repository: Explore the SNN-Torch library on GitHub.