Deep learning courses and concepts
Deep learning is machine learning built on neural networks with many layers, in which the system works out for itself which features of the raw data matter instead of being told which to look at. This is why one family of methods now handles images, audio and text, problems that once needed separate hand-built approaches, and why its main costs are data and computing power. Courses use its technical terms from the start, because almost every idea has a name: a part of a network, or a step in training one.
Deep learning courses
Advanced: Generative AI for Developers
Google Skills
Building AI Agents with Multimodal Models
NVIDIA Deep Learning Institute
Community Computer Vision Course
Hugging Face
Computer Vision
Kaggle Learn
Computer Vision Fundamentals with Google Cloud
Google Skills
Deep Learning for Images with PyTorch
DataCamp
Elements of AI: Building AI
University of Helsinki & MinnaLearn
Google DeepMind: AI Research Foundations
Google Skills
Image Modeling with Keras
DataCamp
Introduction to Deep Learning with PyTorch
DataCamp
Rapid Application Development with Large Language Models (LLMs)
NVIDIA Deep Learning Institute
The Art of Compressing LLMs: Pruning, Distillation, and Quantization
NVIDIA Deep Learning Institute
Deep learning terms explained
Most of these terms fall into three groups: the building blocks and kinds of network (activation functions, attention, transformers, convolutional and recurrent networks), the mechanics of training one (loss function, learning rate, batch size, epoch), and ways of making a trained network smaller (pruning, knowledge distillation). Courses use most of them without explaining them, so reading a few first makes a course easier to follow.
Develop AI is where deep learning courses sit: building and training models, and the engineering that turns a trained network into something people can use. To understand what these systems do without building one, start with Learn AI.