Replace Q3_K_S, Q3_K_M quants with new IQ3_S, IQ3_M versions

See https://github.com/ggerganov/llama.cpp/commit/4c4cb30736582cacb1a164a9d4bc8e17b1014be7

mistral-7b-instruct-v0.2.Q3_K_M.gguf → mistral-7b-instruct-v0.2.IQ3_M.gguf

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:37d6224de3a16f6d939242a003144ca5029599327bfff62b5909ba809beae1e3
-size 3518986432
+oid sha256:853da015d147c00a95b657ab1cbbcdbf5f300e77392bc6a7f038336b770066b1
+size 3284891840

mistral-7b-instruct-v0.2.Q3_K_S.gguf → mistral-7b-instruct-v0.2.IQ3_S.gguf

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d6ef365507adb83b71e440c6c9d850d826dbfb34db15db4549e57cb23fd89a1a
-size 3164567744
+oid sha256:ccefafe3bd387fac4a393c24b4e5974071d67cb43df4d7d166bc81cdf5eb711f
+size 3182393536