vrr/prima.cpp
2
0
Fork 0
mirror of https://github.com/Lizonghang/prima.cpp.git synced 2025-09-10 17:44:33 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

keep the output layer weights in shared memory by default

Browse source
This commit is contained in:
Lizonghang 2025-01-25 23:31:43 +04:00
parent f3dd5776eb
commit 1ca9a43bd1
4 changed files with 16 additions and 16 deletions
Download patch file Download diff file Expand all files Collapse all files

8
common/common.cpp
View file

@ -1691,11 +1691,11 @@ static ggml_type kv_cache_type_from_str(const std::string & s) {
struct llama_context_params llama_context_params_from_gpt_params(const gpt_params & params) {
auto cparams = llama_context_default_params();
cparams.n_world = params.n_world;
cparams.rank = params.rank;
cparams.unload = params.unload;
cparams.n_world = params.n_world;
cparams.rank = params.rank;
cparams.unload = params.unload;
cparams.keep_out_in_metal = params.keep_out_in_metal;
cparams.n_gpu_layers = params.n_gpu_layers;
cparams.n_gpu_layers = params.n_gpu_layers;
std::copy(std::begin(params.n_layer_window), std::end(params.n_layer_window), cparams.n_layer_window);
if (cparams.master_ip != nullptr) {