130: Bumblebee Buzz with José Valim
Thinking Elixir Podcast
00:00
Axon Serving Is a Critical Piece in This Workflow Right?
The serving is the idea of encapsulating. So if you think about like a neural network task, there are things that you want to do as part of the neural network and like you want toDo that in the GPU. Right, but there areThings that you cannot do in theGPU. There are things like for example, like I was saying like if you want todo like sentiment analysis, you have like to tokenize the text and the GPUs can't do that right now. We want to figure out like a model that the things that run in the neural network, we are we can batch it because the thing about CPU is that it's really parallizable so what you do
Transcript
Play full episode
