Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-19 01:05:09 +02:00
Code Issues Packages Projects Releases Wiki Activity

prevent gpu offload if kquant is selected with clblast for now

Browse Source
This commit is contained in:
Concedo
2023-06-12 21:57:31 +08:00
parent 215edf420b
commit 120851df53
1 changed files with 5 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
llama.cpp
+5 -1
View File
@@ -1077,7 +1077,11 @@ static void llama_model_load_internal(
#if defined(GGML_USE_CLBLAST)
if (file_version == LLAMA_FILE_VERSION_GGJT_V3) {
if (hparams.ftype >= LLAMA_FTYPE_MOSTLY_Q2_K && hparams.ftype <= LLAMA_FTYPE_MOSTLY_Q6_K) {
printf("\n===\nK-Quants are currently not supported with CLBlast!!!\nPlease select a q4_0, q4_0, q5_0 or q5_1 format instead!\n=====\n");
if(n_gpu_layers>0)
{
n_gpu_layers = 0;
printf("\n===\nCLBlast cannot offload layers for K-Quants!\nPlease select a q4_0, q4_0, q5_0 or q5_1 format instead!\n=====\n");
}
}
}
#endif