Fire
Dev Track
FR
Writing my own LLM: what I learned
"I don't understand what I can't build." So he wrote his own LLM. He shares what he learned from it.
Presented by
"What I cannot create, I do not understand," said Richard P. Feynman.
It's the same for me: if I want to really understand something, I have to rewrite it. That's what I did to try to understand LLMs.
In this session, we'll inspect the contents of a (small) LLM written in Python to really understand what it is, and what it isn't.
Dans le même parcours
