vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-11 01:24:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

pulled all Occam's fixes and the kquants are all working now

Browse source
This commit is contained in:
Concedo 2023-06-13 16:15:13 +08:00
parent 9b6c35b651
commit 9830871d0f
2 changed files with 19 additions and 37 deletions
Download patch file Download diff file Expand all files Collapse all files

12
llama.cpp
View file

@ -1074,18 +1074,6 @@ 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) {
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
if (vocab_only) {
return;
}