Priya Kasimbeg (@kasimbegpriya) 's Twitter Profile
Priya Kasimbeg

@kasimbegpriya

ID: 1609300739720941571

calendar_today31-12-2022 21:30:10

9 Tweet

37 Followers

47 Following

Zachary Nado (@zacharynado) 's Twitter Profile Photo

tl;dr submit a training algorithm* that is faster** than Adam*** and win $10,000 πŸ’ΈπŸš€ *a set of hparams, self-tuning algorithm, and/or update rule **see rules for how we measure speed ***beat all submissions, currently the best is NAdamW in wallclock and DistShampoo in steps

MLCommons (@mlcommons) 's Twitter Profile Photo

MLCommons #AlgoPerf results are in! 🏁 $50K prize competition yielded 28% faster neural net training with non-diagonal preconditioning beating Nesterov Adam. New SOTA for hyperparameter-free algorithms too! Full details in our blog. mlcommons.org/2024/08/mlc-al… #AIOptimization #AI

Google AI (@googleai) 's Twitter Profile Photo

Congratulations to everyone who submitted to theΒ MLCommons AlgoPerf training algorithms competition! We were delighted to provide compute resources for evaluating so many exciting submissions.

AlgoPerf (@algoperf) 's Twitter Profile Photo

Hi there! This account will post about the AlgoPerf benchmark and leaderboard updates for faster neural network training via better training algorithms. But let's start with what AlgoPerf is, what we have done so far, and how you can train neural nets ~30% faster.

Damek (@damekdavis) 's Twitter Profile Photo

Lecture 11: benchmarking optimizers 1. the problem: comparing optimizers (sgd, adam, etc.) in deep learning is tricky. 2. challenge 1: defining "speed". curves cross, so use time-to-result. 3. challenge 2: hyperparameter tuning trap. protocol matters more than algo? (choi et

Lecture 11: benchmarking optimizers
1.  the problem: comparing optimizers (sgd, adam, etc.) in deep learning is tricky.
2.  challenge 1: defining "speed". curves cross, so use time-to-result.
3.  challenge 2: hyperparameter tuning trap. protocol matters more than algo? (choi et