vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-10 09:04:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fixes for vulkan multigpu

Browse source
This commit is contained in:
Concedo 2024-02-09 14:42:27 +08:00
parent fe424a5466
commit 992eea71d7
3 changed files with 34 additions and 18 deletions
Download patch file Download diff file Expand all files Collapse all files

5
gpttype_adapter.cpp
View file

@ -961,7 +961,12 @@ ModelLoadResult gpttype_load_model(const load_model_inputs inputs, FileFormat in
}
#endif
model_params.main_gpu = cu_parseinfo_maindevice;
#if defined(GGML_USE_CUBLAS)
model_params.split_mode = llama_split_mode::LLAMA_SPLIT_ROW;
#else
model_params.split_mode = llama_split_mode::LLAMA_SPLIT_LAYER;
#endif
llama_ctx_params.n_batch = kcpp_params->n_batch;
llama_ctx_params.n_threads = kcpp_params->n_threads;