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

Revert "Support using mmap when applying LoRA (#2095)" (#2206)

Browse source 
Has perf regression when mlock is used.

This reverts commit 2347463201.
This commit is contained in:
Howard Su 2023-07-13 21:58:25 +08:00 • committed by GitHub
parent ff5d58faec
commit 32c5411631
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
5 changed files with 9 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

3
examples/common.cpp
View file

@ -285,6 +285,7 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
break;
}
params.lora_adapter = argv[i];
params.use_mmap = false;
} else if (arg == "--lora-base") {
if (++i >= argc) {
invalid_param = true;
@ -520,7 +521,7 @@ void gpt_print_usage(int /*argc*/, char ** argv, const gpt_params & params) {
fprintf(stderr, " --mtest compute maximum memory usage\n");
fprintf(stderr, " --export export the computation graph to 'llama.ggml'\n");
fprintf(stderr, " --verbose-prompt print prompt before generation\n");
fprintf(stderr, " --lora FNAME apply LoRA adapter\n");
fprintf(stderr, " --lora FNAME apply LoRA adapter (implies --no-mmap)\n");
fprintf(stderr, " --lora-base FNAME optional model to use as a base for the layers modified by the LoRA adapter\n");
fprintf(stderr, " -m FNAME, --model FNAME\n");
fprintf(stderr, " model path (default: %s)\n", params.model.c_str());