Update README.md

README.md

@@ -2,4 +2,4 @@ Custom GGUF quants of Hermes-3-Llama-3.1-8B, where the Output Tensors are quanti
 
 Update: This repo now contains OF32.EF32 GGUF IQuants for even more accuracy. Enjoy! 😋 
 
-UPDATE: This repo now contains updated O.E.IQuants, which were quantized using the llama.cpp version: 4067 (54ef9cfc). This particular version of llama.cpp made it so all K*Q mat_mul computations were done in F32 vs BF16. This change, plus the other very impactful prior change, which made all K*Q mat_muls be computed with F32 (float32) precision, have enhanced this O.E.IQuants and made it necessary for this update. Cheers! 
+UPDATE: This repo now contains updated O.E.IQuants, which were quantized, using a new F32-imatrix, using llama.cpp version: 4067 (54ef9cfc). This particular version of llama.cpp made it so all K*Q mat_mul computations were done in F32 vs BF16, when using FA (Flash Attention). This change, plus the other very impactful prior change, which made all K*Q mat_muls be computed with F32 (float32) precision for CUDA-Enabled devices, has compoundedly enhanced the O.E.IQuants and has made it excitingly necessary for this update to be pushed. Cheers!