vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Remove mtest (#3177)

Browse source
This commit is contained in:
Branden Butler 2023-09-25 10:05:42 -05:00
parent 1e78fa4f91
commit 78112ab5c2
1 changed files with 0 additions and 17 deletions
Download patch file Download diff file Expand all files Collapse all files

17
examples/mpi/mpi.cpp
View file

@ -233,23 +233,6 @@ int main(int argc, char ** argv) {
params.n_threads, std::thread::hardware_concurrency(), llama_print_system_info());
}
// determine the maximum memory usage needed to do inference for the given n_batch and n_ctx parameters
// uncomment the "used_mem" line in llama.cpp to see the results
if (params.mem_test) {
{
fprintf(stderr, "%s: testing memory usage for n_batch = %d, n_ctx = %d\n", __func__, params.n_batch, params.n_ctx);
const std::vector<llama_token> tmp(params.n_batch, llama_token_bos(ctx));
llama_eval(ctx, tmp.data(), tmp.size(), params.n_ctx, params.n_threads);
}
llama_print_timings(ctx);
llama_free(ctx);
llama_free_model(model);
return 0;
}
// export the cgraph and exit
if (params.export_cgraph) {
llama_eval_export(ctx, "llama.ggml");