common : refactor cli arg parsing (#7675)

* common : gpt_params_parse do not print usage

* common : rework usage print (wip)

* common : valign

* common : rework print_usage

* infill : remove cfg support

* common : reorder args

* server : deduplicate parameters

ggml-ci

* common : add missing header

ggml-ci

* common : remote --random-prompt usages

ggml-ci

* examples : migrate to gpt_params

ggml-ci

* batched-bench : migrate to gpt_params

* retrieval : migrate to gpt_params

* common : change defaults for escape and n_ctx

* common : remove chatml and instruct params

ggml-ci

* common : passkey use gpt_params

1 files changed, 17 insertions, 31 deletions

diff --git a/examples/llama-bench/llama-bench.cpp b/examples/llama-bench/llama-bench.cpp
index fa7ad1bd..5c31548a 100644
--- a/examples/llama-bench/llama-bench.cpp
+++ b/examples/llama-bench/llama-bench.cpp
@@ -41,20 +41,6 @@ static std::string join(const std::vector<T> & values, const std::string & delim
     return str.str();
 }
 
-template<class T>
-static std::vector<T> split(const std::string & str, char delim) {
-    std::vector<T> values;
-    std::istringstream str_stream(str);
-    std::string token;
-    while (std::getline(str_stream, token, delim)) {
-        T value;
-        std::istringstream token_stream(token);
-        token_stream >> value;
-        values.push_back(value);
-    }
-    return values;
-}
-
 template<typename T, typename F>
 static std::vector<std::string> transform_to_str(const std::vector<T> & values, F f) {
     std::vector<std::string> str_values;
@@ -322,28 +308,28 @@ static cmd_params parse_cmd_params(int argc, char ** argv) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<std::string>(argv[i], split_delim);
+            auto p = string_split<std::string>(argv[i], split_delim);
             params.model.insert(params.model.end(), p.begin(), p.end());
         } else if (arg == "-p" || arg == "--n-prompt") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<int>(argv[i], split_delim);
+            auto p = string_split<int>(argv[i], split_delim);
             params.n_prompt.insert(params.n_prompt.end(), p.begin(), p.end());
         } else if (arg == "-n" || arg == "--n-gen") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<int>(argv[i], split_delim);
+            auto p = string_split<int>(argv[i], split_delim);
             params.n_gen.insert(params.n_gen.end(), p.begin(), p.end());
         } else if (arg == "-pg") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<std::string>(argv[i], ',');
+            auto p = string_split<std::string>(argv[i], ',');
             if (p.size() != 2) {
                 invalid_param = true;
                 break;
@@ -354,21 +340,21 @@ static cmd_params parse_cmd_params(int argc, char ** argv) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<int>(argv[i], split_delim);
+            auto p = string_split<int>(argv[i], split_delim);
             params.n_batch.insert(params.n_batch.end(), p.begin(), p.end());
         } else if (arg == "-ub" || arg == "--ubatch-size") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<int>(argv[i], split_delim);
+            auto p = string_split<int>(argv[i], split_delim);
             params.n_ubatch.insert(params.n_ubatch.end(), p.begin(), p.end());
         } else if (arg == "-ctk" || arg == "--cache-type-k") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<std::string>(argv[i], split_delim);
+            auto p = string_split<std::string>(argv[i], split_delim);
             std::vector<ggml_type> types;
             for (const auto & t : p) {
                 ggml_type gt = ggml_type_from_name(t);
@@ -384,7 +370,7 @@ static cmd_params parse_cmd_params(int argc, char ** argv) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<std::string>(argv[i], split_delim);
+            auto p = string_split<std::string>(argv[i], split_delim);
             std::vector<ggml_type> types;
             for (const auto & t : p) {
                 ggml_type gt = ggml_type_from_name(t);
@@ -400,14 +386,14 @@ static cmd_params parse_cmd_params(int argc, char ** argv) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<int>(argv[i], split_delim);
+            auto p = string_split<int>(argv[i], split_delim);
             params.n_threads.insert(params.n_threads.end(), p.begin(), p.end());
         } else if (arg == "-ngl" || arg == "--n-gpu-layers") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<int>(argv[i], split_delim);
+            auto p = string_split<int>(argv[i], split_delim);
             params.n_gpu_layers.insert(params.n_gpu_layers.end(), p.begin(), p.end());
         } else if (arg == "-rpc" || arg == "--rpc") {
             if (++i >= argc) {
@@ -420,7 +406,7 @@ static cmd_params parse_cmd_params(int argc, char ** argv) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<std::string>(argv[i], split_delim);
+            auto p = string_split<std::string>(argv[i], split_delim);
             std::vector<llama_split_mode> modes;
             for (const auto & m : p) {
                 llama_split_mode mode;
@@ -442,13 +428,13 @@ static cmd_params parse_cmd_params(int argc, char ** argv) {
                 invalid_param = true;
                 break;
             }
-            params.main_gpu = split<int>(argv[i], split_delim);
+            params.main_gpu = string_split<int>(argv[i], split_delim);
         } else if (arg == "-nkvo" || arg == "--no-kv-offload") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<bool>(argv[i], split_delim);
+            auto p = string_split<bool>(argv[i], split_delim);
             params.no_kv_offload.insert(params.no_kv_offload.end(), p.begin(), p.end());
         } else if (arg == "--numa") {
             if (++i >= argc) {
@@ -466,28 +452,28 @@ static cmd_params parse_cmd_params(int argc, char ** argv) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<bool>(argv[i], split_delim);
+            auto p = string_split<bool>(argv[i], split_delim);
             params.flash_attn.insert(params.flash_attn.end(), p.begin(), p.end());
         } else if (arg == "-mmp" || arg == "--mmap") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<bool>(argv[i], split_delim);
+            auto p = string_split<bool>(argv[i], split_delim);
             params.use_mmap.insert(params.use_mmap.end(), p.begin(), p.end());
         } else if (arg == "-embd" || arg == "--embeddings") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            auto p = split<bool>(argv[i], split_delim);
+            auto p = string_split<bool>(argv[i], split_delim);
             params.embeddings.insert(params.embeddings.end(), p.begin(), p.end());
         } else if (arg == "-ts" || arg == "--tensor-split") {
             if (++i >= argc) {
                 invalid_param = true;
                 break;
             }
-            for (auto ts : split<std::string>(argv[i], split_delim)) {
+            for (auto ts : string_split<std::string>(argv[i], split_delim)) {
                 // split string by ; and /
                 const std::regex regex{R"([;/]+)"};
                 std::sregex_token_iterator it{ts.begin(), ts.end(), regex, -1};