HNHacker News
TopNewBestAskShowJobs

MrUssek

208 karma · joined April 1, 2019

submissionscomments

China has trained a 10 trillion parameter language model

twitter.com·4 pts·MrUssek·
0

What is your backup if the tech industry crashes?

4 pts·MrUssek·
10

The Future of Deep Learning Is Photonic

spectrum.ieee.org·1 pts·MrUssek·
0

Separating MNIST digits using Optimal Transport

mrussek.com·1 pts·MrUssek·
0

Enigma: GPT-2 trained on 10K Nature Papers: Can you spot the difference?

stefanzukin.com·183 pts·MrUssek·
105

GShard: Scaling giant models with conditional computation and automatic sharding

arxiv.org·112 pts·MrUssek·
35