@article{tagcoptimizinggradientcommunicationin, title = {TAGC: Optimizing Gradient Communication in Distributed Transformer Training}, author = {Igor Polyakov and Alexey Dukhanov and Egor Spirin}, year = {2025}, eprint = {2504.05638}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2504.05638v1}, }