cenconq25/delta-compress-llm

Proof of concept: Exploiting temporal coherence in LLM inference-- delta encoding for KV cache compression and weight-skip prediction. Achieves F16-quality KV cache at Q4_0 compression ratios with zero perplexity loss on llama.cpp.

41 best points · 7 comments · 1 Hacker News post · 51 GitHub stars · View on GitHub

Hacker News discussions

• Apply video compression on KV cache to 10,000x less error at Q4 quant (41 points, 7 comments, by polymorph1sm, 2026-03-22)