diff --git a/doc/en/Kimi-K2.md b/doc/en/Kimi-K2.md index b5f546a..1839ede 100644 --- a/doc/en/Kimi-K2.md +++ b/doc/en/Kimi-K2.md @@ -5,6 +5,9 @@ ### Overview We are very pleased to announce that Ktransformers now supports Kimi-K2. +On a single-socket CPU with one consumer-grade GPU, running the Q4_K_M model yields roughly 10 TPS and requires about 600 GB of VRAM. +With a dual-socket CPU and sufficient system memory, enabling NUMA optimizations increases performance to about 14 TPS. + ### Model & Resource Links - Official Kimi-K2 Release: