News for hackers. Stuff that matters.

A comment on Optimizing ML training with metagradient descent. Parent comment

Re:Optimizing ML training with metagradient descent by lamename on Tuesday March 25, 2025 @10:14PM (#43478155)
You might find this helpful for prioritizing which knobs to turn first https://github.com/google-research/tuning_playbook
[ Parent | on HN ]
Re:Optimizing ML training with metagradient descent by OccamsMirror on Wednesday March 26, 2025 @01:04AM (#43479039)
Starting to get a bit out of date. Pity they stopped updating it.
[ Parent | on HN ]
Re:Optimizing ML training with metagradient descent by ayepif on Wednesday March 26, 2025 @06:49AM (#43480757)
Pity indeed! Do you have any suggested resources that are more up-to-date?
[ Parent | on HN ]
Re:Optimizing ML training with metagradient descent by lamename on Wednesday March 26, 2025 @08:42AM (#43481583)
Last commit 1 yr. They welcome your contributions in CONTRIBUTING.md
[ Parent | on HN ]