YouTube Tools
Support

Stanford CS25: Transformers United: study plan and how long it takes

The Stanford Online playlist is 50 videos and 55.6 hours long. Here's how long it takes at different paces, a week-by-week schedule built from the real videos, and a planner that turns it into your own daily plan.

50 videos · 55.6 hours · Stanford Online

Stanford CS25 is a seminar series rather than a lecture course: each session is a talk on transformers from researchers at leading labs, from DeepMind's Perceiver to LLM reasoning, video generation and AGI. The playlist collects several years of the seminar (V1 to V6).

Talks are independent, so you can watch them in any order.

Who it's for: ML engineers and researchers who already understand transformers and want to keep up with frontier research.

How long does it take?

The full playlist is 55.6 hours across 50 videos. Here's roughly how many days that is at different daily watch times and speeds, counting video time only:

Days to finish the playlist by daily watch time and playback speed
Daily timeAt 1xAt 1.5xAt 2x
30 min a day112 days75 days56 days
1 h a day56 days38 days28 days
2 h a day28 days19 days14 days

Make your own schedule

The playlist link is already filled in. Press Plan, set how much time you have each day and your speed, and you'll get a day-by-day schedule with a finish date and a calendar file. It reads the live playlist, so new videos are included.

Open on YouTube Track watched videos

Week-by-week schedule

A sample plan at 2 hours a day and normal speed, using whole videos so none is split across days. That's 7 weeks in total. For a different pace, start date or speed, use the planner above.

Week 1Videos 1–9 · 8h 39m

  1. Day 1 · 1h 11m

    • 1Stanford CS25: V1 I Transformers United: DL Models that have revolutionized NLP, CV, RL23m
    • 2Stanford CS25: V1 I Transformers in Language: The development of GPT Models, GPT349m

  2. Day 2 · 1h 9m

    • 3Stanford CS25: V1 I Transformers in Vision: Tackling problems in Computer Vision1h 9m

  3. Day 3 · 1h 21m

    • 4Stanford CS25: V1 I Decision Transformer: Reinforcement Learning via Sequence Modeling1h 21m

  4. Day 4 · 1h 6m

    • 5Stanford CS25: V1 I Mixture of Experts (MoE) paradigm and the Switch Transformer1h 6m

  5. Day 5 · 59m

    • 6Stanford CS25: V1 I DeepMind's Perceiver and Perceiver IO: new data family architecture59m

  6. Day 6 · 1h 6m

    • 7Stanford CS25: V1 I Self Attention and Non-parametric transformers (NPTs)1h 6m

  7. Day 7 · 1h 48m

    • 8Stanford CS25: V1 I Transformer Circuits, Induction Heads, In-Context Learning1h
    • 9Stanford CS25: V1 I Audio Research: Transformers for Applications in Audio, Speech, Music48m
Week 2Videos 10–16 · 7h 43m

  1. Day 8 · 53m

    • 10Stanford CS25: V2 I Represent part-whole hierarchies in a neural network, Geoff Hinton53m

  2. Day 9 · 1h 12m

    • 11Stanford CS25: V2 I Introduction to Transformers w/ Andrej Karpathy1h 12m

  3. Day 10 · 1h 6m

    • 12Stanford CS25: V2 I Language and Human Alignment1h 6m

  4. Day 11 · 1h 8m

    • 13Stanford CS25: V2 I Emergent Abilities and Scaling in LLMs1h 8m

  5. Day 12 · 53m

    • 14Stanford CS25: V2 I Strategic Games53m

  6. Day 13 · 1h 16m

    • 15Stanford CS25: V2 I Robotics and Imitation Learning1h 16m

  7. Day 14 · 1h 15m

    • 16Stanford CS25: V2 I Common Sense Reasoning1h 15m
Week 3Videos 17–24 · 8h 59m

  1. Day 15 · 1h 8m

    • 17Stanford CS25: V2 I Biomedical Transformers1h 8m

  2. Day 16 · 1h 22m

    • 18Stanford CS25: V2 I Neuroscience-Inspired Artificial Intelligence1h 22m

  3. Day 17 · 1h 18m

    • 19Stanford CS25: V3 I Low-level Embodied Intelligence w/ Foundation Models1h 18m

  4. Day 18 · 48m

    • 20Stanford CS25: V3 I Generalist Agents in Open-Ended Worlds48m

  5. Day 19 · 1h 21m

    • 21Stanford CS25: V3 I How I Learned to Stop Worrying and Love the Transformer1h 21m

  6. Day 20 · 1h 9m

    • 22Stanford CS25: V3 I Recipe for Training Helpful Chatbots1h 9m

  7. Day 21 · 1h 53m

    • 23Stanford CS25: V3 I No Language Left Behind: Scaling Human-Centered Machine Translation52m
    • 24Stanford CS25: V3 I Beyond LLMs: Agents, Emergent Abilities, Intermediate-Guided Reasoning, BabyLM1h
Week 4Videos 25–31 · 8h 55m

  1. Day 22 · 1h 19m

    • 25Stanford CS25: V3 I Retrieval Augmented Language Models1h 19m

  2. Day 23 · 1h 17m

    • 26Stanford CS25: V4 I Overview of Transformers1h 17m

  3. Day 24 · 1h 17m

    • 27Stanford CS25: V4 I Jason Wei & Hyung Won Chung of OpenAI1h 17m

  4. Day 25 · 1h 16m

    • 28Stanford CS25: V4 I Aligning Open Language Models1h 16m

  5. Day 26 · 1h 5m

    • 29Stanford CS25: V4 I Demystifying Mixtral of Experts1h 5m

  6. Day 27 · 1h 20m

    • 30Stanford CS25: V4 I Transformers that Transform Well Enough to Support Near-Shallow Architectures1h 20m

  7. Day 28 · 1h 20m

    • 31Stanford CS25: V4 I From Large Language Models to Large Multimodal Models1h 20m
Week 5Videos 32–39 · 8h 22m

  1. Day 29 · 1h 38m

    • 32Stanford CS25: V4 I Behind the Scenes of LLM Pre-training: StarCoder Use Case1h 2m
    • 33Stanford CS25: V4 I Hyung Won Chung of OpenAI37m

  2. Day 30 · 1h 1m

    • 34Stanford CS25: V5 I Overview of Transformers1h 1m

  3. Day 31 · 1h 12m

    • 35Stanford CS25: V5 I RL as a Co-Design of Product and Research, Karina Nguyen1h 12m

  4. Day 32 · 1h 1m

    • 36Stanford CS25: V5 I The Advent of AGI, Div Garg1h 1m

  5. Day 33 · 1h 6m

    • 37Stanford CS25: V5 I Large Language Model Reasoning, Denny Zhou of Google Deepmind1h 6m

  6. Day 34 · 1h 13m

    • 38Stanford CS25: V5 I On the Biology of a Large Language Model, Josh Batson of Anthropic1h 13m

  7. Day 35 · 1h 11m

    • 39Stanford CS25: V5 I Multimodal World Models for Drug Discovery, Eshed Margalit of Noetik.ai1h 11m
Week 6Videos 40–47 · 9h 19m

  1. Day 36 · 1h 15m

    • 40Stanford CS25: V5 I Transformers in Diffusion Models for Image Generation and Beyond1h 15m

  2. Day 37 · 1h 14m

    • 41Stanford CS25: V5 I Transformers for Video Generation, Andrew Brown of Meta1h 14m

  3. Day 38 · 1h 17m

    • 42Stanford CS25: Transformers United V6 I Overview of Transformers1h 17m

  4. Day 39 · 1h 11m

    • 43Stanford CS25: Transformers United V6 I From Representation Learning to World Modeling1h 11m

  5. Day 40 · 1h 17m

    • 44Stanford CS25: Transformers United V6 I On the Tradeoffs of State Space Models and Transformers1h 17m

  6. Day 41 · 2h

    • 45Stanford CS25: Transformers United V6 I The Ultra-Scale Talk: Scaling Training to Thousands of GPUs1h 2m
    • 46Stanford CS25: Transformers United V6 I From Next-Token Prediction to Next-Generation Intelligence58m

  7. Day 42 · 1h 7m

    • 47Stanford CS25: Transformers United V6 I Advancing Science and Medicine with Collaborative AI Agents1h 7m
Week 7Videos 48–50 · 3h 40m

  1. Day 43 · 1h 13m

    • 48Stanford CS25: Transformers United V6 I Distinct Modes of Generalization from Parameters and Context1h 13m

  2. Day 44 · 1h 23m

    • 49Stanford CS25: Transformers United V6 I Serving Transformers: Lessons from the Trenches1h 23m

  3. Day 45 · 1h 5m

    • 50Stanford CS25: Transformers United V6 I From Language Models to Native Multimodal Intelligence1h 5m

Tips for finishing

  • Watch the overview talk at the start of each version first (V2, V4), then pick talks on topics you care about.
  • Tick off talks you're not interested in using the playlist length calculator; the planner will schedule only the rest.

Frequently asked questions

How long does it take to finish Stanford CS25: Transformers United?

The playlist has 50 videos and 55.6 hours of content. At 1 hour a day that's about 56 days at normal speed, or 38 days at 1.5x. At 2 hours a day it's about 28 days.

More study plans

See all 166 study plans

Tools for any playlist