Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-20 01:31:42 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'upstream' into concedo_experimental

Browse Source 
# Conflicts:
#	.github/workflows/build.yml
#	ci/run.sh
#	docs/function-calling.md
#	examples/gritlm/gritlm.cpp
#	ggml/CMakeLists.txt
#	ggml/cmake/common.cmake
#	ggml/src/CMakeLists.txt
#	ggml/src/ggml-cpu/CMakeLists.txt
#	ggml/src/ggml-cpu/ggml-cpu.c
#	ggml/src/ggml-hip/CMakeLists.txt
#	ggml/src/ggml-vulkan/CMakeLists.txt
#	ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt
#	requirements/requirements-compare-llama-bench.txt
#	scripts/compare-llama-bench.py
#	tests/CMakeLists.txt
This commit is contained in:
Concedo
2025-06-18 00:16:54 +08:00
parent f07434f4c1 860a9e4eef
commit 4356a00f4a
47 changed files with 1526 additions and 301 deletions
Download Patch File Download Diff File Expand all files Collapse all files
common/arg.cpp
+3 -6
View File
@@ -990,10 +990,6 @@ static bool common_params_parse_ex(int argc, char ** argv, common_params_context
params.tensor_buft_overrides.push_back({nullptr, nullptr});
}
if (params.reranking && params.embedding) {
throw std::invalid_argument("error: either --embedding or --reranking can be specified, but not both");
}
if (!params.chat_template.empty() && !common_chat_verify_template(params.chat_template, params.use_jinja)) {
throw std::runtime_error(string_format(
"error: the supplied chat template is not supported: %s%s\n",
@@ -2749,9 +2745,10 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
).set_examples({LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_EMBEDDINGS"));
add_opt(common_arg(
{"--reranking", "--rerank"},
string_format("enable reranking endpoint on server (default: %s)", params.reranking ? "enabled" : "disabled"),
string_format("enable reranking endpoint on server (default: %s)", "disabled"),
[](common_params & params) {
params.reranking = true;
params.embedding = true;
params.pooling_type = LLAMA_POOLING_TYPE_RANK;
}
).set_examples({LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_RERANKING"));
add_opt(common_arg(
common/chat-parser.cpp
+5
View File
@@ -49,6 +49,7 @@ bool common_chat_msg_parser::add_tool_call(const std::string & name, const std::
// LOG_DBG("Tool call arguments:\n\traw: %s\n\tresult: %s\n", arguments.c_str(), tool_call.arguments.c_str());
result_.tool_calls.emplace_back(tool_call);
return true;
}
bool common_chat_msg_parser::add_tool_call(const json & tool_call) {
@@ -378,3 +379,7 @@ std::optional<common_chat_msg_parser::consume_json_result> common_chat_msg_parse
/* .is_partial = */ found_healing_marker,
};
}
void common_chat_msg_parser::clear_tools() {
result_.tool_calls.clear();
}
common/chat-parser.h
+2
View File
@@ -115,4 +115,6 @@ class common_chat_msg_parser {
const std::vector<std::vector<std::string>> & args_paths = {},
const std::vector<std::vector<std::string>> & content_paths = {}
);
void clear_tools();
};
common/chat.cpp
+4 -2
View File
@@ -1838,7 +1838,7 @@ static common_chat_params common_chat_templates_apply_legacy(
if (res < 0) {
// if the custom "tmpl" is not supported, we throw an error
// this is a bit redundant (for good), since we're not sure if user validated the custom template with llama_chat_verify_template()
throw std::runtime_error("this custom template is not supported");
throw std::runtime_error("this custom template is not supported, try using --jinja");
}
// if it turns out that our buffer is too small, we resize it
@@ -1921,7 +1921,9 @@ common_chat_msg common_chat_parse(const std::string & input, bool is_partial, co
} catch (const common_chat_msg_partial_exception & ex) {
LOG_DBG("Partial parse: %s\n", ex.what());
if (!is_partial) {
throw std::runtime_error(ex.what());
builder.clear_tools();
builder.move_to(0);
common_chat_parse_content_only(builder);
}
}
auto msg = builder.result();
common/common.cpp
+41 -37
View File
@@ -775,6 +775,9 @@ bool fs_validate_filename(const std::string & filename) {
return true;
}
#include <iostream>
// returns true if successful, false otherwise
bool fs_create_directory_with_parents(const std::string & path) {
#ifdef _WIN32
@@ -792,9 +795,16 @@ bool fs_create_directory_with_parents(const std::string & path) {
// process path from front to back, procedurally creating directories
while ((pos_slash = path.find('\\', pos_slash)) != std::string::npos) {
const std::wstring subpath = wpath.substr(0, pos_slash);
const wchar_t * test = subpath.c_str();
const bool success = CreateDirectoryW(test, NULL);
pos_slash += 1;
// skip the drive letter, in some systems it can return an access denied error
if (subpath.length() == 2 && subpath[1] == ':') {
continue;
}
const bool success = CreateDirectoryW(subpath.c_str(), NULL);
if (!success) {
const DWORD error = GetLastError();
@@ -808,8 +818,6 @@ bool fs_create_directory_with_parents(const std::string & path) {
return false;
}
}
pos_slash += 1;
}
return true;
@@ -905,34 +913,6 @@ struct common_init_result common_init_from_params(common_params & params) {
const llama_vocab * vocab = llama_model_get_vocab(model);
if (params.reranking) {
bool ok = true;
if (llama_vocab_bos(vocab) == LLAMA_TOKEN_NULL) {
LOG_WRN("%s: warning: vocab does not have a BOS token, reranking will not work\n", __func__);
ok = false;
}
bool has_eos = llama_vocab_eos(vocab) != LLAMA_TOKEN_NULL;
bool has_sep = llama_vocab_sep(vocab) != LLAMA_TOKEN_NULL;
if (!has_eos && !has_sep) {
LOG_WRN("%s: warning: vocab does not have an EOS token or SEP token, reranking will not work\n", __func__);
ok = false;
} else if (!has_eos) {
LOG_WRN("%s: warning: vocab does not have an EOS token, using SEP token as fallback\n", __func__);
} else if (!has_sep) {
LOG_WRN("%s: warning: vocab does not have a SEP token, reranking will not work\n", __func__);
ok = false;
}
if (!ok) {
llama_model_free(model);
return iparams;
}
}
auto cparams = common_context_params_to_llama(params);
llama_context * lctx = llama_init_from_model(model, cparams);
@@ -974,6 +954,35 @@ struct common_init_result common_init_from_params(common_params & params) {
}
}
if (llama_pooling_type(lctx) == LLAMA_POOLING_TYPE_RANK) {
bool ok = true;
if (llama_vocab_bos(vocab) == LLAMA_TOKEN_NULL) {
LOG_WRN("%s: warning: vocab does not have a BOS token, reranking will not work\n", __func__);
ok = false;
}
bool has_eos = llama_vocab_eos(vocab) != LLAMA_TOKEN_NULL;
bool has_sep = llama_vocab_sep(vocab) != LLAMA_TOKEN_NULL;
if (!has_eos && !has_sep) {
LOG_WRN("%s: warning: vocab does not have an EOS token or SEP token, reranking will not work\n", __func__);
ok = false;
} else if (!has_eos) {
LOG_WRN("%s: warning: vocab does not have an EOS token, using SEP token as fallback\n", __func__);
} else if (!has_sep) {
LOG_WRN("%s: warning: vocab does not have a SEP token, reranking will not work\n", __func__);
ok = false;
}
if (!ok) {
llama_free(lctx);
llama_model_free(model);
return iparams;
}
}
// load and optionally apply lora adapters
for (auto & la : params.lora_adapters) {
llama_adapter_lora_ptr lora;
@@ -1151,11 +1160,6 @@ struct llama_context_params common_context_params_to_llama(const common_params &
cparams.op_offload = !params.no_op_offload;
cparams.swa_full = params.swa_full;
if (params.reranking) {
cparams.embeddings = true;
cparams.pooling_type = LLAMA_POOLING_TYPE_RANK;
}
cparams.type_k = params.cache_type_k;
cparams.type_v = params.cache_type_v;
common/common.h
-1
View File
@@ -351,7 +351,6 @@ struct common_params {
int32_t embd_normalize = 2; // normalisation for embeddings (-1=none, 0=max absolute int16, 1=taxicab, 2=euclidean, >2=p-norm)
std::string embd_out = ""; // empty = default, "array" = [[],[]...], "json" = openai style, "json+" = same "json" + cosine similarity matrix
std::string embd_sep = "\n"; // separator of embeddings
bool reranking = false; // enable reranking support on server
// server params
int32_t port = 8080; // server listens on this network port