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