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

examples: cache hf model when --model not provided

Browse source
This commit is contained in:
Amir 2024-05-17 22:23:10 +00:00
parent 0fc1e820a9
commit 4679fde41b
2 changed files with 19 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

14
common/common.cpp
View file

@ -1354,7 +1354,8 @@ void gpt_params_handle_model_default(gpt_params & params) {
}
params.hf_file = params.model;
} else if (params.model.empty()) {
params.model = "models/" + string_split(params.hf_file, '/').back();
const char* cache_dir = getenv("LLAMA_CACHE") ? getenv("LLAMA_CACHE") : DEFAULT_LLAMA_CACHE;
params.model = cache_dir + string_split(params.hf_file, '/').back();
}
} else if (!params.model_url.empty()) {
if (params.model.empty()) {
@ -2076,6 +2077,17 @@ static bool llama_download_file(const std::string & url, const std::string & pat
}
}
// Create parent directories if not exist
const std::vector<std::string> path_parts = string_split(path_temporary, DIRECTORY_SEPARATOR);
std::string parent_dir = "";
struct stat st;
for (unsigned i = 0; i < path_parts.size() - 1; i++) {
parent_dir += path_parts[i] + DIRECTORY_SEPARATOR;
if (stat(parent_dir.c_str(), &st) != 0) {
mkdir(parent_dir.c_str(), S_IRWXU);
}
}
// Set the output file
std::unique_ptr<FILE, decltype(&fclose)> outfile(fopen(path_temporary.c_str(), "wb"), fclose);
if (!outfile) {

6
common/common.h
View file

@ -32,6 +32,12 @@
} while(0)
#define DEFAULT_MODEL_PATH "models/7B/ggml-model-f16.gguf"
#define DEFAULT_LLAMA_CACHE ".cache/"
#ifdef _WIN32
#include <direct.h>
#define mkdir(path, mode) _mkdir(path) // On Windows, _mkdir does not take mode
#endif
// build info
extern int LLAMA_BUILD_NUMBER;