any divergence (even if the benchmark is better) from full precision is error | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

electroglyph 29 days ago | parent | context | favorite | on: KVarN: Native vLLM backend for KV-cache quantizati...

any divergence (even if the benchmark is better) from full precision is error

7e 29 days ago [–]

Just pretend that it is the next step update when training. You didn’t train your model to step=inf, I hope?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact