Cite
Notes
Only stored in your browser.
Attribution
Quartet II: Accurate LLM Pre-Training in NVFP4 by Improved Unbiased Gradient Estimation
arXiv 2026
DASH: Faster Shampoo via Batched Block Preconditioning and Efficient Inverse-Root Solvers
Hogwild! Inference: Parallel LLM Generation via Concurrent Attention
arXiv 2025
from 3 papers
Dan Alistarh
Alina Shutova
Andrei Panferov
Anton Sinitsin
Denis Kuznedelev
George Yakushev
Gleb Rodionov
Ionut-Vlad Modoranu
Mher Safaryan
Philip Zmushko