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?
11 / 31
BackNext

The network can't read

Computers don't understand language. We covered this early on. They work with numbers, and nothing else. So before training can begin, all that text has to be converted into numbers.

Every sentence. Every word. Every punctuation mark. Each one gets assigned a numeric ID, and the whole training dataset becomes a very long sequence of those IDs. Billions of web pages, turned into a single river of numbers.

How exactly that conversion works, and what it means for the model, is a story worth telling in full. It comes later. For now, the important thing is just that it happens: text goes in, numbers come out, and the network trains on those numbers.

And the training task itself is simple.

Citations(2)↓
  1. 1. platform.openai.com
  2. 2. huggingface.co
Citations(2)↓
  1. 1. platform.openai.com
  2. 2. huggingface.co