MyTorch is a comprehensive deep learning library that offers all essential functionalities required for building and training neural networks. It is built with NumPy and offers a wide range of features including simple MLP activations, loss functions, and batch normalization. Additionally, it provides advanced capabilities such as Convolutional Neural Networks libraries, Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) in a similar style to PyTorch. With Mytorch, users can efficiently develop and train deep learning models using a familiar API and take advantage of the powerful numerical computation capabilities of NumPy.
This library was developed as a part of the 11785 Introduction to Deep Learning course at CMU