LogQuant: Log-Distributed 2-Bit Quantization of KV Cache with Superior Accuracy Preservation

Documenti analoghi

• QJL: 1-Bit Quantized JL Transform for KV Cache Quantization with Zero Overhead
  di: Zandieh, Amir, et al.
  Pubblicazione: (2024)
• Feather-SQL: A Lightweight NL2SQL Framework with Dual-Model Collaboration Paradigm for Small Language Models
  di: Pei, Wenqi, et al.
  Pubblicazione: (2025)
• RotateKV: Accurate and Robust 2-Bit KV Cache Quantization for LLMs via Outlier-Aware Adaptive Rotations
  di: Su, Zunhai, et al.
  Pubblicazione: (2025)
• PolarQuant: Quantizing KV Caches with Polar Transformation
  di: Han, Insu, et al.
  Pubblicazione: (2025)
• AhaKV: Adaptive Holistic Attention-Driven KV Cache Eviction for Efficient Inference of Large Language Models
  di: Gu, Yifeng, et al.
  Pubblicazione: (2025)