The Transformer Architecture: A New Architecture for Deep Learning

There was a school of thought that says that the large language models are kind of topping out and we're reaching the limits of that approach. And we need new like fundamentally new things. So, there is a way there's a framework for looking at this, which is that what we the things that we do with the transformer architecture. It basically tells the neural network for every word in us for every for every token in a sequence or suit.

Transcript

Play full episode

Transcript

Episode notes

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!

Get the app