Adaptive Blockwise 8-bit Floating Point Quantization of Parameters for Communication-Efficient Large Language Model Training
Procedia Computer Science · 2026 · 292 · 3218-3227
DOI: 10.1016/j.procs.2026.09.398 ↗Kurz gefasst
Fachartikel in der Zeitschrift Procedia Computer Science, 2026. Verfasst von Igor Polyakov und 1 weiteren Autor:innen. Die Arbeit nennt 28 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: ITMO University.
Abstract (Originalsprache)
Kein Abstract verfügbar.
Förderung
- ITMO University
Zitieren
APA
Polyakov, I., & Dukhanov, A. (2026). Adaptive Blockwise 8-bit Floating Point Quantization of Parameters for Communication-Efficient Large Language Model Training. Procedia Computer Science, 292, 3218-3227. https://doi.org/10.1016/j.procs.2026.09.398
Vancouver
Polyakov I, Dukhanov A. Adaptive Blockwise 8-bit Floating Point Quantization of Parameters for Communication-Efficient Large Language Model Training. Procedia Computer Science. 2026;292:3218-3227. doi:10.1016/j.procs.2026.09.398
BibTeX
@article{Polyakov2026,
author = {Polyakov, Igor and Dukhanov, Alexey},
title = {Adaptive Blockwise 8-bit Floating Point Quantization of Parameters for Communication-Efficient Large Language Model Training},
journal = {Procedia Computer Science},
year = {2026},
volume = {292},
pages = {3218--3227},
doi = {10.1016/j.procs.2026.09.398}
}
EUVIMED