
Episode 13: Jonathan Frankle, MIT, on the lottery ticket hypothesis and the science of deep learning
Generally Intelligent
How to Train 700 Networks on Image Net?
Tuctrt s i mi o go to tech out your open sourcean tpeado eatsin school. When you want to train 700 networks on image net could you actually pull that off if i gave you the resources to do it? It's besus. Hilarius josh said exactly the same thing the other day. He was like, our internal test should be given a cluster of hundreds of gpus. We keep all those machines busy. And an te all the were lik machine. Wellyil mose unders, that's what my research is. I keep the t p s busy. My whole research group now uses the darned spreadto al it the darned
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.