Good ideas and conversation. No ads, no tracking. Login or Take a Tour!
alpha0 · 645 days ago · link · · parent · post: Stephen Wolfram: What Is ChatGPT Doing … and Why Does It Work?
Andreij Karpathy has a nice zero to hero lecture series that you can follow along and it will end with building your own simple GPT. First lecture is building your own MLP (multi-layer perceptron). At the end of that you have down backtracing and will finally understand what it means to train a model. https://karpathy.ai/zero-to-hero.html And LangChain is your friend if you want to use GPT as a component in a processing pipeline (as in integrating with Wolfram's alpha, etc.)