From a5aa793ed4cb2a3231738f74eb1305000cfeb17f Mon Sep 17 00:00:00 2001 From: ymcui <16095339+ymcui@users.noreply.github.com> Date: Mon, 5 Feb 2024 15:14:40 +0800 Subject: [PATCH] quantize: add iq3_xxs warning iq3_xxs should not be used without imatrix. --- examples/quantize/quantize.cpp | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/examples/quantize/quantize.cpp b/examples/quantize/quantize.cpp index 85f403ffc..3960e42b8 100644 --- a/examples/quantize/quantize.cpp +++ b/examples/quantize/quantize.cpp @@ -287,9 +287,9 @@ int main(int argc, char ** argv) { } } - if ((params.ftype == LLAMA_FTYPE_MOSTLY_IQ2_XS || params.ftype == LLAMA_FTYPE_MOSTLY_IQ2_XXS || params.ftype == LLAMA_FTYPE_MOSTLY_Q2_K_S) && imatrix_data.empty()) { + if ((params.ftype == LLAMA_FTYPE_MOSTLY_IQ2_XS || params.ftype == LLAMA_FTYPE_MOSTLY_IQ2_XXS || params.ftype == LLAMA_FTYPE_MOSTLY_Q2_K_S || params.ftype == LLAMA_FTYPE_MOSTLY_IQ3_XXS) && imatrix_data.empty()) { fprintf(stderr, "\n===============================================================================================\n"); - fprintf(stderr, "Please do not use IQ2_XXS, IQ2_XS or Q2_K_S quantization without an importance matrix\n"); + fprintf(stderr, "Please do not use IQ2_XXS, IQ2_XS, Q2_K_S or IQ3_XXS quantization without an importance matrix\n"); fprintf(stderr, "===============================================================================================\n\n\n"); return 1; }