NVIDIA’s LoGRA Cuts RL Training Memory With Low-Rank Gradient Sketches
NVIDIA’s LoGRA paper shows low-rank gradient sketches cutting average RL training memory by up to 45.7% on tested Qwen reasoning setups—and making 27B RL runnable on one 8×H100 node where dense Adam OOMs. What the method is, what Table 1 shows, and what it does not prove.
Just in
Latest
No posts to display
Timeline
Every story in memory efficiency, grouped by month.
October 2026
- 7 Oct 2026: NVIDIA's LoGRA Cuts RL Training Memory With Low-Rank Gradient SketchesNVIDIA’s LoGRA paper shows low-rank gradient sketches cutting average RL training memory by up to 45.7% on tested Qwen reasoning setups—and making 27B RL…
Latest on TSN
Explore the TSN Atlas: Every topic we cover, mapped. Click a bubble to dive in.
