@article{mixedsparsitytrainingachieving4times, title = {Mixed Sparsity Training: Achieving 4$\times$ FLOP Reduction for Transformer Pretraining}, author = {Pihe Hu and Shaolong Li and Longbo Huang}, year = {2024}, eprint = {2408.11746}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2408.11746v1}, }