Home/Stanford Online/CS231n: Deep Learning for Computer Vision
Stanford University · on Stanford Online

CS231n: Deep Learning for Computer Vision

4.9(9,000)·1.5M enrolled
Advanced 50 hours English None CertificateFREE
Editor's Pick
The course that trained a generation of computer vision engineers — still the best academic deep dive into CNNs and vision available.

About this course

CS231n is one of the most influential academic courses ever taught in machine learning — the course where convolutional neural networks for image recognition became mainstream, first delivered by Fei-Fei Li and Andrej Karpathy at Stanford. It covers convolutional network architecture in depth, backpropagation, training techniques (batch normalization, dropout, learning rate scheduling), object detection frameworks (YOLO, Faster R-CNN), image segmentation, recurrent networks for sequence tasks, and generative models (GANs, VAEs).

The course materials — lecture videos, slides, and assignments — have been publicly available for years and have trained more practicing ML engineers than possibly any other single resource outside of a formal degree. Karpathy's lecture style and the depth of the assignments set a benchmark that few online courses match.

What you'll learn

Implement convolutional neural networks from scratch and understand their mechanics
Apply training techniques: batch normalization, dropout, and learning rate scheduling
Build and use object detection frameworks including YOLO and Faster R-CNN
Implement image segmentation models
Understand and implement generative models (GANs and VAEs)

This course includes

50h
On-demand video
Yes
Mobile access
English
Language
Comparison · LBS

Compare alternatives for CS231n: Deep Learning for Computer Vision

Same topic, different options. We surface the trade-offs others hide so you can pick the course that actually fits your time, budget, and goals.
Stanford Online4.9(9,000)
CS231n: Deep Learning for Computer Vision
Price
Free
Free lecture materials; some versions paid
Duration
50 hrs
Level
Advanced
Certificate
MIT OpenCourseWare4.9(15,000)
Linear Algebra (18.06)
Price
Free
Completely free, openly licensed — no certificate
Duration
34 hrs
Level
Intermediate
Certificate
Stanford Online4.9(7,000)
CS224n: Natural Language Processing with Deep Learning
Price
Free
Free lecture materials; some versions paid
Duration
50 hrs
Level
Advanced
Certificate
Stanford Online4.9(11,000)
CS229: Machine Learning
Price
Free
Free lecture materials available online
Duration
55 hrs
Level
Advanced
Certificate
Prices & availability can change — confirm on the provider's site. We're not affiliated with any single provider.

Instructor

FL
Fei-Fei Li / Andrej Karpathy
Stanford Online instructor
1.5M+ learners5 courses4.9 instructor rating

Originally taught by Fei-Fei Li (Stanford CS Professor, former Google Cloud AI Chief Scientist) and Andrej Karpathy (former Tesla AI Director, OpenAI founding member, now independent).

Requirements

  • Solid Python and NumPy skills
  • Basic understanding of neural networks and backpropagation
  • Linear algebra and calculus comfort

Who this course is for

  • ML engineers and researchers who want depth in computer vision specifically
  • Data scientists moving from tabular data into image and vision applications
  • Anyone who has taken Andrew Ng's ML courses and wants the next level of depth

About this provider

SO
Stanford Online
Stanford University's online learning platform offering free and paid courses from Stanford faculty across AI, ML, medicine, and computer science.
Visit Stanford Online

Frequently asked questions

The CNN fundamentals, backpropagation mechanics, and training techniques are foundational and don't change. Specific architectures evolve rapidly — the course provides the foundation to understand new architectures as they emerge.
The lecture videos and slides are freely available on Stanford Online and YouTube; assignments have been publicly shared by previous students. Some versions include paid certificate options.
The assignments are computationally intensive — Google Colab or similar GPU-enabled environments are typically used for free access to sufficient compute.
Free
to audit
Enroll now