vrr/prima.cpp
2
0
Fork 0
mirror of https://github.com/Lizonghang/prima.cpp.git synced 2025-09-09 20:24:34 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

add arg --cuda-mem

Browse source
This commit is contained in:
Zonghang Li 2025-01-16 09:15:34 +04:00
parent dab6b2e1c2
commit 46e99218b4
5 changed files with 19 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

7
common/arg.cpp
View file

@ -730,6 +730,13 @@ gpt_params_context gpt_params_parser_init(gpt_params & params, llama_example ex,
params.unload = true;
}
).set_env("LLAMA_ARG_UNLOAD"));
add_opt(llama_arg(
{"-cm", "--cuda-mem"}, "N",
format("maximum cuda memory to use (default: %d)", params.cuda_mem),
[](gpt_params & params, int value) {
params.cuda_mem = value; // in GiB
}
).set_env("LLAMA_ARG_CUDA_MEM"));
add_opt(llama_arg(
{"-n", "--predict", "--n-predict"}, "N",
format("number of tokens to predict (default: %d, -1 = infinity, -2 = until context filled)", params.n_predict),