Loading slide

What a Trained Model Actually Is

  1. 01What a Trained Model Actually Is
  2. 02Memory card
  3. 03A model is a file
  4. 04Two files
  5. 05How the two fit together
  6. 06A Piano
  7. 07The valuable file
  8. 08You connect to it
  9. 09Where does training data come from?
  10. 10Cleaning the data
  11. 11The network can't read
  12. 12Guess the next word
  13. 13Right or wrong
  14. 14One nudge, a billion times
  15. 15Take a breath
  16. 16Where your memories live
  17. 17A base model is not an assistant
  18. 18Why it's called pre-training
  19. 19The weights are the model
  20. 20The scale of it
  21. 21Why so many?
  22. 22What a weight actually is
  23. 23Everywhere and nowhere
  24. 24Using the model: inference
  25. 25The whole arrangement
  26. 26Reinforce your understanding
  27. 27Question: What is a model, really?
  28. 28Question: Learning from raw text
  29. 29Question: The model doesn't know what it doesn't know
  30. 30Quiz: answer
  31. 31The machine learned something real.
  32. 32Want to go deeper?
8 / 31
BackNext

You connect to it

In theory, you could download both files and run a model on your own computer. But the weights file alone is enormous. A typical model is tens or hundreds of gigabytes, and running one requires a huge amount of memory and processing power, far more than most personal computers have.

So instead, we connect to it.

The file lives on servers (powerful computers somewhere else) owned by the company that built the model. When you type a message to an AI assistant, you are not running anything on your own machine. Your message travels over the internet to their computers. The model processes it there. The answer travels back to you.

The model never leaves their computer. You just get the response.

But that raises an obvious question. Those weights, hundreds of gigabytes of numbers sitting in a file, how did they end up knowing anything at all?

Citations
Citations