common : add -hfd option for the draft model (#11318)

* common : add -hfd option for the draft model

* cont : fix env var

* cont : more fixes
This commit is contained in:
Georgi Gerganov
2025-01-20 22:29:43 +02:00
committed by GitHub
parent aea8ddd516
commit 80d0d6b4b7
3 changed files with 24 additions and 6 deletions

View File

@ -133,7 +133,8 @@ static void common_params_handle_model_default(
const std::string & model_url,
std::string & hf_repo,
std::string & hf_file,
const std::string & hf_token) {
const std::string & hf_token,
const std::string & model_default) {
if (!hf_repo.empty()) {
// short-hand to avoid specifying --hf-file -> default it to --model
if (hf_file.empty()) {
@ -163,7 +164,7 @@ static void common_params_handle_model_default(
model = fs_get_cache_file(string_split<std::string>(f, '/').back());
}
} else if (model.empty()) {
model = DEFAULT_MODEL_PATH;
model = model_default;
}
}
@ -299,8 +300,9 @@ static bool common_params_parse_ex(int argc, char ** argv, common_params_context
}
// TODO: refactor model params in a common struct
common_params_handle_model_default(params.model, params.model_url, params.hf_repo, params.hf_file, params.hf_token);
common_params_handle_model_default(params.vocoder.model, params.vocoder.model_url, params.vocoder.hf_repo, params.vocoder.hf_file, params.hf_token);
common_params_handle_model_default(params.model, params.model_url, params.hf_repo, params.hf_file, params.hf_token, DEFAULT_MODEL_PATH);
common_params_handle_model_default(params.speculative.model, params.speculative.model_url, params.speculative.hf_repo, params.speculative.hf_file, params.hf_token, "");
common_params_handle_model_default(params.vocoder.model, params.vocoder.model_url, params.vocoder.hf_repo, params.vocoder.hf_file, params.hf_token, "");
if (params.escape) {
string_process_escapes(params.prompt);
@ -1629,6 +1631,13 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
params.hf_repo = value;
}
).set_env("LLAMA_ARG_HF_REPO"));
add_opt(common_arg(
{"-hfd", "-hfrd", "--hf-repo-draft"}, "<user>/<model>[:quant]",
"Same as --hf-repo, but for the draft model (default: unused)",
[](common_params & params, const std::string & value) {
params.speculative.hf_repo = value;
}
).set_env("LLAMA_ARG_HFD_REPO"));
add_opt(common_arg(
{"-hff", "--hf-file"}, "FILE",
"Hugging Face model file. If specified, it will override the quant in --hf-repo (default: unused)",