Article URL: https://www.nature.com/articles/d41586-025-00764-2
Comments URL: https://news.ycombinator.com/item?id=43426984
Points: 9
# Comments: 0
Created
2mo
|
Mar 21, 2025, 7:20:09 PM
Login to add comment
Other posts in this group

I discovered that in LLM inference, keys and values in the KV cache have very different quantization sensitivities. Keys need higher precision than values to maintain quality.
I patched llama.cp
Article URL: https://clojurescript.org/news/2025-05-16-release




Article URL: https://bobacollection.staxmuseum.org/
Comments URL: https://news.y