IKL 20250303 - LCPP b3637 - pre GGML file version V2

Author: Nexesenex

README.md

@@ -1,5 +1,23 @@
 # ik_llama.cpp: llama.cpp fork with better CPU performance
 
+IK_Llama.cpp.nxs is my fork of Ikawrakow's clone of Llama_cpp.
+
+The objective of this fork is to adapt it to my quantization needs mainly, because Ikawrakow's quants are SOTA.
+Then, to be able to use them on Croco.cpp, my fork of Kobold.cpp.
+And also to amuse myself learning more about merging relatively complex stuff.
+
+Current IK_Llama commit : 
+
+Current Llama.cpp mainline base : b3570 -> b3637-3 (pre GGML sync "File version V2")
+
+
+
+
+----------
+
+
+# llama.cpp clone with better CPU performance
+
 [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](https://opensource.org/licenses/MIT)
 
 ## TL;DR