Pytorch implementation of Siamese-Neural-Network. Finds similarity in faces using Siamese-Neural-Network. Siamese-Neural-Network is used for one-shot learning. It means you don't need a huge dataset for it's training, one example in each class is enough for it's working, though in this implementation more thatn one example has been used but it is still very less compared to what the state of art algorithm requires.
Git clone the repository and run the ipynb file in Google Colab or Jupyter Notebook(some importing changes will be required).
Google Colab: No installations required
Jupyter Notebook: Install Pytorch and torchvision using pip
pip install pytorch
pip install torchvision
AT & T. You can get the dataset by cloning this repository
You can run the code to train the model and save it for later use. Just uncomment the loading model part to use the pre-trained model. Or you can download the pretrained model from Here though keep in mind that it is a GPU model so it won't work on a CPU. If you're using this pretrained model no need to run the training part of the code. You can directly load the model and test the code.
- Pytorch - The Deep Learning library used
See also the list of contributors who participated in this project.