#634 · Primary category： Education & Research
train-deepseek-r1
Building DeepSeek R1 from Scratch
Project last updated：03/21/25
GitHub Stars
793
Forks
135
Contributors
1
License
MIT
Why we included this project
If you learn best by building, this notebook is a great way to understand DeepSeek-R1. It walks through the training pipeline from the technical report, covering GRPO reinforcement learning, reward functions, cold-start SFT, rejection sampling, and distillation, all on a tiny base model that runs locally. Hand-drawn flowcharts and plain-language explanations sit alongside the code, so even newcomers to reinforcement learning can follow. For anyone thinking about replicating this training recipe on their own data, it's a concrete starting point.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
prompts.chat
f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.
JavaGuide
Java Interview & Backend General Interview Guide, covering computer fundamentals, databases, distributed systems, high concurrency, system design, and AI application development.
system-prompts-and-models-of-ai-tools
A curated collection of system prompts, internal tools, and AI models from popular AI assistants and coding agents.
30-seconds-of-code
Coding articles to level up your development skills
generative-ai-for-beginners
21 Lessons, Get Started Building with Generative AI