PyTorch Flashcards

An open-source deep learning framework developed by Facebook’s AI Research lab, widely used for research and production.

A tensor is a multi-dimensional array used as the fundamental building block in PyTorch computations.

Autograd is PyTorch’s automatic differentiation engine that powers neural network training.

A module in PyTorch that provides layers and loss functions to build deep learning models.

The step in neural networks where input data is passed through the layers to generate predictions.

The process of computing gradients of loss with respect to model weights to update them during training.

A PyTorch module containing optimization algorithms like SGD, Adam, and RMSprop.

A way to serialize PyTorch models so they can run outside Python environments, ideal for deployment.

A wrapper for PyTorch that provides structured training loops, making model training more readable and scalable.

A utility that provides batches of data efficiently for training and evaluation loops.