EngineX-Mthreads/enginex-mthreads-llama.cpp
4
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

sync from b7516

Browse Source
This commit is contained in:
xiezhongtao
2026-01-16 11:16:14 +08:00
parent f4ae4cc7da
commit 6ee41dd9e3
380 changed files with 18435 additions and 38806 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
tools/fit-params/fit-params.cpp
View File

@@ -26,16 +26,16 @@ int main(int argc, char ** argv) {
llama_numa_init(params.numa);
auto mparams = common_model_params_to_llama(params);
auto cparams = common_context_params_to_llama(params);
const llama_params_fit_status status = llama_params_fit(params.model.path.c_str(), &mparams, &cparams,
params.tensor_split, params.tensor_buft_overrides.data(), params.fit_params_target.data(), params.fit_params_min_ctx,
const bool success = llama_params_fit(params.model.path.c_str(), &mparams, &cparams,
params.tensor_split, params.tensor_buft_overrides.data(), params.fit_params_target, params.fit_params_min_ctx,
params.verbosity >= 4 ? GGML_LOG_LEVEL_DEBUG : GGML_LOG_LEVEL_ERROR);
if (status != LLAMA_PARAMS_FIT_STATUS_SUCCESS) {
if (!success) {
LOG_ERR("%s: failed to fit CLI arguments to free memory, exiting...\n", __func__);
exit(1);
}
LOG_INF("%s: printing fitted CLI arguments to stdout...\n", __func__);
common_log_flush(common_log_main());
std::this_thread::sleep_for(10ms); // to avoid a race between stderr and stdout
printf("-c %" PRIu32 " -ngl %" PRIu32, cparams.n_ctx, mparams.n_gpu_layers);
size_t nd = llama_max_devices();