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?
10 / 31
BackNext

Cleaning the data

The problem is that most of the internet is garbage. Spam. Broken pages. The same article copy-pasted ten thousand times across different websites. Comment sections that would make your grandmother faint. written by bots, for bots. You do not want a model that has absorbed all of that uncritically.

So before any training begins, someone has to clean it. Harmful content gets removed. Duplicates get stripped out. Low-quality text gets discarded. Personal information gets filtered. It is slow, unglamorous work, and it matters more than almost anything else in the process.

What the model learns depends entirely on what it trains on. Garbage in, garbage out.

Once the data is clean, there is still one more problem. The network can't actually read any of it.

Citations(1)↓
  1. 1. huggingface.co
Citations(1)↓
  1. 1. huggingface.co