@article{swiftkvfastprefilloptimizedinferencewith, title = {SwiftKV: Fast Prefill-Optimized Inference with Knowledge-Preserving Model Transformation}, author = {Aurick Qiao and Zhewei Yao and Samyam Rajbhandari and Yuxiong He}, year = {2024}, eprint = {2410.03960}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2410.03960v2}, }