Galore pytorch
1.0.0
Inoffizielle WIP-WIP -Implementierung von Galore: Gedächtniseffizientes LLM-Training durch Gradienten-niedrig-Rang-Projektion
@article{zhao2024galore,
title = {GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection},
author = {Jiawei Zhao and Zhenyu Zhang and Beidi Chen and Zhangyang Wang and Anima Anandkumar and Yuandong Tian},
year = {2024},
journal = {arXiv preprint arXiv: 2403.03507}
}