arXiv2
Search
Dark
/ Date
/ Name
Aa
W
/ Date
/ Name
"au:"Maximilian Golub"" — arXiv2 Search
Showing 1–8 of 8 results
/ Date
/ Name
Jun 11, 2018
Full deep neural network training on a pruned weight budget
Jul 7, 2025
Helix Parallelism: Rethinking Sharding Strategies for Interactive Multi-Million-Token LLM Decoding
Jan 26, 2026
LatentMoE: Toward Optimal Accuracy per FLOP and Parameter in Mixture of Experts
Oct 16, 2023
Microscaling Data Formats for Deep Learning
Feb 16, 2023
With Shared Microexponents, A Little Shifting Goes a Long Way
Jun 5, 2025
Beyond the Buzz: A Pragmatic Take on Inference Disaggregation
Feb 18, 2024
Shaping Human-AI Collaboration: Varied Scaffolding Levels in Co-writing with Language Models
Sep 23, 2020
Procrustes: a Dataflow and Accelerator for Sparse Deep Neural Network Training