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

args: main & server now call gpt_params_handle_model_default

Browse source
This commit is contained in:
Olivier Chafik 2024-04-26 00:50:29 +01:00
parent 40a961db60
commit 9c0db4dd9d
3 changed files with 28 additions and 20 deletions
Download patch file Download diff file Expand all files Collapse all files

44
common/common.cpp
View file

@ -1310,6 +1310,29 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa
return false; return false;
} }
void gpt_params_handle_model_default(gpt_params & params) {
if (!params.hf_repo.empty()) {
// short-hand to avoid specifying --hf-file -> default it to --model
if (params.hf_file.empty()) {
if (params.model.empty()) {
throw std::invalid_argument("error: --hf-repo requires either --hf-file or --model\n");
}
params.hf_file = params.model;
} else if (params.model.empty()) {
params.model = "models/" + string_split(params.hf_file, '/').back();
}
} else if (!params.model_url.empty()) {
if (params.model.empty()) {
auto f = string_split(params.model_url, '#').front();
f = string_split(f, '?').front();
f = string_split(f, '/').back();
params.model = "models/" + f;
}
} else if (params.model.empty()) {
params.model = "models/7B/ggml-model-f16.gguf";
}
}
bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params) { bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params) {
bool invalid_param = false; bool invalid_param = false;
std::string arg; std::string arg;
@ -1338,26 +1361,7 @@ bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params) {
throw std::invalid_argument("error: --prompt-cache-all not supported in interactive mode yet\n"); throw std::invalid_argument("error: --prompt-cache-all not supported in interactive mode yet\n");
} }
if (!params.hf_repo.empty()) { gpt_params_handle_model_default(params);
// short-hand to avoid specifying --hf-file -> default it to --model
if (params.hf_file.empty()) {
if (params.model.empty()) {
throw std::invalid_argument("error: --hf-repo requires either --hf-file or --model\n");
}
params.hf_file = params.model;
} else if (params.model.empty()) {
params.model = "models/" + string_split(params.hf_file, '/').back();
}
} else if (!params.model_url.empty()) {
if (params.model.empty()) {
auto f = string_split(params.model_url, '#').front();
f = string_split(f, '?').front();
f = string_split(f, '/').back();
params.model = "models/" + f;
}
} else if (params.model.empty()) {
params.model = "models/7B/ggml-model-f16.gguf";
}
if (params.escape) { if (params.escape) {
process_escapes(params.prompt); process_escapes(params.prompt);

2
common/common.h
View file

@ -170,6 +170,8 @@ struct gpt_params {
std::string image = ""; // path to an image file std::string image = ""; // path to an image file
}; };
void gpt_params_handle_model_default(gpt_params & params);
bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params); bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params);
bool gpt_params_parse(int argc, char ** argv, gpt_params & params); bool gpt_params_parse(int argc, char ** argv, gpt_params & params);

2
examples/server/server.cpp
View file

@ -2846,6 +2846,8 @@ static void server_params_parse(int argc, char ** argv, server_params & sparams,
} }
} }
gpt_params_handle_model_default(params);
if (!params.kv_overrides.empty()) { if (!params.kv_overrides.empty()) {
params.kv_overrides.emplace_back(); params.kv_overrides.emplace_back();
params.kv_overrides.back().key[0] = 0; params.kv_overrides.back().key[0] = 0;