Completed
25:30 DeepSeek’s Aha Moment
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Understanding DeepSeek R1 and GRPO - A Technical Deep Dive
Automatically move to the next video in the Classroom when playback concludes
- 1 0:00 Preview
- 2 0:31 Intro to Arxiv Dives
- 3 3:42 Why is R1 Important?
- 4 6:38 What is a Reasoning Model?
- 5 8:55 What are DeepSeek R1’s Contributions?
- 6 12:27 How DeepSeek-v3 Works
- 7 16:01 What Hardware do You Need?
- 8 16:50 How DeepSeek-R1-Zero Works
- 9 17:23 How GRPO works
- 10 25:30 DeepSeek’s Aha Moment
- 11 29:06 R1 on ARC-AGI Benchmark
- 12 30:20 Self-Hosting DeepSeek
- 13 31:38 How DeepSeek-R1 Works
- 14 34:05 What was the Cold Start Data
- 15 36:58 Rejection Sampling and Supervised Fine Tuning
- 16 38:30 Helpfulness and Harmlessness Reinforcement Learning
- 17 39:45 Distilling Smaller Models
- 18 41:25 Distillation vs. Reinforcement Learning
