
Azure Innovations with Mark Russinovich
RunAs Radio
00:00
The Differences Between Language Models and Video Models
There are some differences depending on the type of model you're going to train because large language models, you give it a huge amount of text and it processes the text in tiny chunks. When you talk about video models, yeah, like for economist vehicle driving, those are megabytes, tens of megabytes in size of chunks. 30, 60 frames a second multiple cameras, every frame matters. And so in fact, you might need a petabyte of data to train a model. So it's the gears not specific to the workload per se. They there's a shape to this workload that it's good at running out.
Transcript
Play full episode