Beyond Homogeneous Attention: Memory-Efficient LLMs via Fourier-Approximated KV Cache

Add code
Jun 13, 2025

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: