vrr/prima.cpp
2
0
Fork 0
mirror of https://github.com/Lizonghang/prima.cpp.git synced 2025-09-11 06:04:50 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

add metal mem limit

Browse source
This commit is contained in:
Lizonghang 2025-01-23 16:08:52 +04:00
parent 33429ec4e1
commit 78a544d716
5 changed files with 102 additions and 67 deletions
Download patch file Download diff file Expand all files Collapse all files

1
include/llama.h
View file

@ -323,6 +323,7 @@ extern "C" {
uint32_t n_layer_window[32];// number of layers to process in each compute
uint32_t n_gpu_layers; // number of layers to process on GPU
bool unload; // whether to unload layer weights after use
bool keep_inp_out_in_metal; // whether to keep input/output weight in metal
char * master_ip; // ip address of the master node
char * next_node_ip; // ip address of the next node
uint32_t n_ctx; // text context, 0 = from model