Dnotitia's STAR-KV cuts KV cache by up to 20x, earns ICML 2026 Spotlight selection
July 02, 2026 | 18:16
Seoul-based Dnotitia released STAR-KV, a low-rank KV cache compression method achieving up to 20x reduction, with the paper selected as a Spotlight at ICML 2026, one of the top venues in machine learning research.
Latest News
Mobile Version