ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-02-19 04:40:09 +00:00
Code Issues Packages Projects Releases Wiki Activity

A few server commits from mainline. (#872)

Browse Source 
server : handle models with missing EOS token (#8997)

server : fix segfault on long system prompt (#8987)
* server : fix segfault on long system prompt
* server : fix parallel generation with very small batch sizes
* server : fix typo in comment

server : init stop and error fields of the result struct (#9026)

server : fix duplicated n_predict key in the generation_settings (#8994)

server : support reading arguments from environment variables (#9105)
* server : support reading arguments from environment variables
* add -fa and -dt
* readme : specify non-arg env var

server : add some missing env variables (#9116)
* server : add some missing env variables
* add LLAMA_ARG_HOST to server dockerfile
* also add LLAMA_ARG_CONT_BATCHING

Credits are to the respective authors.
Not a single merge conflict occurred.
Compiled, then tested without bug.
This commit is contained in:
Nexes the Elder
2025-10-28 08:58:31 +01:00
committed by GitHub
parent 0a80135392
commit e68dabc242
9 changed files with 142 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files

71
common/common.cpp
View File

@@ -87,6 +87,41 @@
#endif
using json = nlohmann::ordered_json;
//
// Environment variable utils
//
template<typename T>
static typename std::enable_if<std::is_same<T, std::string>::value, void>::type
get_env(std::string name, T & target) {
char * value = std::getenv(name.c_str());
target = value ? std::string(value) : target;
}
template<typename T>
static typename std::enable_if<!std::is_same<T, bool>::value && std::is_integral<T>::value, void>::type
get_env(std::string name, T & target) {
char * value = std::getenv(name.c_str());
target = value ? std::stoi(value) : target;
}
template<typename T>
static typename std::enable_if<std::is_floating_point<T>::value, void>::type
get_env(std::string name, T & target) {
char * value = std::getenv(name.c_str());
target = value ? std::stof(value) : target;
}
template<typename T>
static typename std::enable_if<std::is_same<T, bool>::value, void>::type
get_env(std::string name, T & target) {
char * value = std::getenv(name.c_str());
if (value) {
std::string val(value);
target = val == "1" || val == "true";
}
}
//
// CPU utils
//
@@ -239,12 +274,6 @@ static std::string parse_device_list(const std::string& value) {
// CLI argument parsing
//
void gpt_params_handle_hf_token(gpt_params & params) {
if (params.hf_token.empty() && std::getenv("HF_TOKEN")) {
params.hf_token = std::getenv("HF_TOKEN");
}
}
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
@@ -292,7 +321,9 @@ bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params) {
gpt_params_handle_model_default(params);
gpt_params_handle_hf_token(params);
if (params.hf_token.empty()) {
get_env("HF_TOKEN", params.hf_token);
}
if (params.escape) {
if (!params.prompt_is_binary) {
@@ -329,6 +360,32 @@ bool gpt_params_parse_ex(int argc, char ** argv, gpt_params & params) {
return true;
}
void gpt_params_parse_from_env(gpt_params & params) {
// we only care about server-related params for now
get_env("LLAMA_ARG_MODEL", params.model);
get_env("LLAMA_ARG_MODEL_URL", params.model_url);
get_env("LLAMA_ARG_MODEL_ALIAS", params.model_alias);
get_env("LLAMA_ARG_HF_REPO", params.hf_repo);
get_env("LLAMA_ARG_HF_FILE", params.hf_file);
get_env("LLAMA_ARG_THREADS", params.n_threads);
get_env("LLAMA_ARG_CTX_SIZE", params.n_ctx);
get_env("LLAMA_ARG_N_PARALLEL", params.n_parallel);
get_env("LLAMA_ARG_BATCH", params.n_batch);
get_env("LLAMA_ARG_UBATCH", params.n_ubatch);
get_env("LLAMA_ARG_N_GPU_LAYERS", params.n_gpu_layers);
get_env("LLAMA_ARG_THREADS_HTTP", params.n_threads_http);
get_env("LLAMA_ARG_CHAT_TEMPLATE", params.chat_template);
get_env("LLAMA_ARG_N_PREDICT", params.n_predict);
get_env("LLAMA_ARG_ENDPOINT_METRICS", params.endpoint_metrics);
get_env("LLAMA_ARG_ENDPOINT_SLOTS", params.endpoint_slots);
get_env("LLAMA_ARG_EMBEDDINGS", params.embedding);
get_env("LLAMA_ARG_FLASH_ATTN", params.flash_attn);
get_env("LLAMA_ARG_DEFRAG_THOLD", params.defrag_thold);
get_env("LLAMA_ARG_CONT_BATCHING", params.cont_batching);
get_env("LLAMA_ARG_HOST", params.hostname);
get_env("LLAMA_ARG_PORT", params.port);
}
bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
const auto params_org = params; // the example can modify the default params