Knowledge Graph — Coursera Notes › Academic disciplines › Computer Science / Information Technology › Artificial Intelligence › Deep Learning

PyTorch

service · part of Deep Learning

A deep learning framework chosen for dynamic graphs and Pythonic integration, used by Walmart for recommendation systems. It offers more control and flexibility by requiring explicit definition of layers, forward pass, optimizer, loss function, and manual training loops. PyTorch is popular in research and has better integration with HuggingFace's transformer library.

Inside PyTorch (6)

• DataLoader — A PyTorch utility that handles batching and shuffling of data during training and testing.
• nn.Module — Base class in PyTorch for defining neural network architectures by subclassing.
• state_dict — A PyTorch object containing model weights, saved separately from the model architecture.
• TorchScript — A PyTorch feature to serialize and optimize models for deployment without Python.
• TorchServe — PyTorch's serving system for deploying models in production.
• torchvision — A PyTorch library for loading and preprocessing image datasets like CIFAR-10.

Connections

• Alternative to TensorFlow
• Alternative to TensorFlow
• Alternative to Scikit-learn
• Related to TensorFlow
• Used for Deep Learning
• Uses SGD
• Uses Deep Learning
• Uses TorchScript
• Related to TorchServe
• Related to torchvision
• Related to DataLoader
• Related to nn.Module
• Related to state_dict
• Related to TorchScript
• Related to Walmart
• Related to Deep Learning
• Used for Model versioning
• Uses Python
• Alternative to HuggingFace

This is the text view of an interactive 3D knowledge graph — open this page with JavaScript enabled to explore it visually.