Initial-Key Cache: An Efficient KV Cache Strategy Focusing on Initial and Key Tokens for LLMs.
Zhongyi Tang, Zejiang He, Junzhong Shen, Yiyue Hu, Luchen Zhou, Yongzhang Nie, Yongwen Wang
Browse the full IJCNN paper archive.
Zhongyi Tang, Zejiang He, Junzhong Shen, Yiyue Hu, Luchen Zhou, Yongzhang Nie, Yongwen Wang
Browse the full IJCNN paper archive.