From 821b28a4fa3ce17dd5bb5e2c72392e5b89bd4e14 Mon Sep 17 00:00:00 2001 From: Aaron Miller Date: Mon, 8 May 2023 14:52:17 -0700 Subject: [PATCH] mpt: allow q4_2 quantized models to load --- llmodel/mpt.cpp | 1 + 1 file changed, 1 insertion(+) diff --git a/llmodel/mpt.cpp b/llmodel/mpt.cpp index c8f3c230..6e4c5761 100644 --- a/llmodel/mpt.cpp +++ b/llmodel/mpt.cpp @@ -236,6 +236,7 @@ bool mpt_model_load(const std::string &fname, std::istream &fin, mpt_model & mod case 1: wtype = GGML_TYPE_F16; break; case 2: wtype = GGML_TYPE_Q4_0; break; case 3: wtype = GGML_TYPE_Q4_1; break; + case 5: wtype = GGML_TYPE_Q4_2; break; default: { fprintf(stderr, "%s: invalid model file '%s' (bad f16 value %d)\n",