OnlyTerp/turboquant

First open-source implementation of Google TurboQuant (ICLR 2026) -- near-optimal KV cache compression for LLM inference. 5x compression with near-zero quality loss.

⭐ 78 ⑂ 11 Python MIT · 2026-05-25推送

attention compression deep-learning llm llm-inference machine-learning pytorch quantization transformer vllm kv-cache kv-cache-compression iclr google-research memory-optimization vector-quantization

78

Watchers

0

贡献者

0

Commits

0

Releases

1

Open Issues

2026-05-25

最近推送

原文中文

暂无 README