nikola/llama_cpp
1
0
Fork 0
Code Issues Pull Requests Actions 4 Packages Projects Releases Wiki Activity

app : add the llama download subcommand (#24982)

Browse Source 
* app : add the download command (with llama-download)

Signed-off-by: Adrien Gallouët <angt@huggingface.co>

* Remove llama-download tool for now

Signed-off-by: Adrien Gallouët <angt@huggingface.co>

---------

Signed-off-by: Adrien Gallouët <angt@huggingface.co>
This commit is contained in:
Adrien Gallouët
2026-06-25 13:36:36 +02:00
committed by GitHub
parent f728adab68
commit 683b04cc4a
5 changed files with 107 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
app/CMakeLists.txt
View File

@@ -1,6 +1,6 @@
set(TARGET llama-app)
add_executable(${TARGET} llama.cpp)
add_executable(${TARGET} llama.cpp download.cpp)
set_target_properties(${TARGET} PROPERTIES OUTPUT_NAME llama)
target_link_libraries(${TARGET} PRIVATE

70
app/download.cpp Normal file
View File

@@ -0,0 +1,70 @@
#include "arg.h"
#include "common.h"
#include "download.h"
#include "log.h"
#include <cstdio>
#include <filesystem>
static void print_usage(int /*argc*/, char ** argv) {
printf(
"\nexamples:\n"
" %s -hf ggml-org/gemma-3-4b-it-qat-GGUF\n"
" %s -hf ggml-org/gemma-3-4b-it-qat-GGUF:Q4_K_M\n"
" %s -hf ggml-org/models -hff model.gguf\n"
" %s -mu https://example.com/model.gguf -m model.gguf\n"
"\n",
argv[0], argv[0], argv[0], argv[0]
);
}
int llama_download(int argc, char ** argv);
int llama_download(int argc, char ** argv) {
common_init();
common_params params;
params.verbosity = LOG_LEVEL_ERROR;
if (!common_params_parse(argc, argv, params, LLAMA_EXAMPLE_DOWNLOAD, print_usage)) {
return 1;
}
const bool has_source = !params.model.hf_repo.empty() || !params.model.url.empty() ||
!params.model.path.empty() || !params.model.docker_repo.empty();
if (!has_source) {
fprintf(stderr, "error: no model source specified (use --hf-repo, --model-url, --model or --docker-repo)\n");
return 1;
}
try {
common_params_handle_models(params, LLAMA_EXAMPLE_DOWNLOAD, {});
} catch (const std::exception & e) {
fprintf(stderr, "error: %s\n", e.what());
return 1;
}
if (!params.models_preset.empty()) {
// -hf pointed at a preset repo: print the preset path and stop
printf("%s\n", params.models_preset.c_str());
return 0;
}
if (params.model.path.empty()) {
fprintf(stderr, "error: model download failed\n");
return 1;
}
if (!std::filesystem::exists(params.model.path)) {
fprintf(stderr, "error: model file does not exist: %s\n", params.model.path.c_str());
return 1;
}
printf("%s\n", params.model.path.c_str());
if (!params.mmproj.path.empty()) {
printf("%s\n", params.mmproj.path.c_str());
}
if (!params.speculative.draft.mparams.path.empty()) {
printf("%s\n", params.speculative.draft.mparams.path.c_str());
}
return 0;
}

2
app/llama.cpp
View File

@@ -19,6 +19,7 @@ int llama_batched_bench(int argc, char ** argv);
int llama_fit_params(int argc, char ** argv);
int llama_quantize(int argc, char ** argv);
int llama_perplexity(int argc, char ** argv);
int llama_download(int argc, char ** argv);
// Self-update is only supported for binaries built with llama-install.sh
static int llama_update(int argc, char ** argv) {
@@ -61,6 +62,7 @@ static const command cmds[] = {
{"serve", "HTTP API server", {"server"}, false, llama_server },
{"cli", "Command-line interactive interface", {"client"}, false, llama_cli },
{"update", "Update llama to the latest release", {}, UPDATE_HIDDEN, llama_update },
{"download", "Download a model", {"get"}, false, llama_download },
{"completion", "Text completion", {"complete"}, true, llama_completion },
{"bench", "Benchmark prompt processing and text generation", {}, true, llama_bench },
{"batched-bench", "Benchmark batched decoding performance", {}, true, llama_batched_bench},