CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-31 08:51:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'master' into xsn/private_batch_api

Browse Source
This commit is contained in:
Xuan Son Nguyen
2025-03-13 15:55:18 +01:00
parent 46596caf6d be7c303410
commit 17f954c8e2
173 changed files with 26425 additions and 16117 deletions
Download Patch File Download Diff File Expand all files Collapse all files

34
common/common.cpp
View File

@@ -10,7 +10,6 @@
// Change JSON_ASSERT from assert() to GGML_ASSERT:
#define JSON_ASSERT GGML_ASSERT
#include "json.hpp"
#include "json-schema-to-grammar.h"
#include "llama.h"
#include <algorithm>
@@ -483,6 +482,11 @@ void string_replace_all(std::string & s, const std::string & search, const std::
s = std::move(builder);
}
std::string regex_escape(const std::string & s) {
static const std::regex special_chars("[.^$|()*+?\\[\\]{}\\\\]");
return std::regex_replace(s, special_chars, "\\$0");
}
std::string string_join(const std::vector<std::string> & values, const std::string & separator) {
std::ostringstream result;
for (size_t i = 0; i < values.size(); ++i) {
@@ -953,8 +957,8 @@ struct common_init_result common_init_from_params(common_params & params) {
return iparams;
}
if (params.ctx_shift && !llama_kv_cache_can_shift(lctx)) {
LOG_WRN("%s: KV cache shifting is not supported for this model, disabling KV cache shifting\n", __func__);
if (params.ctx_shift && !llama_kv_self_can_shift(lctx)) {
LOG_WRN("%s: KV cache shifting is not supported for this context, disabling KV cache shifting\n", __func__);
params.ctx_shift = false;
}
@@ -1060,7 +1064,7 @@ struct common_init_result common_init_from_params(common_params & params) {
llama_batch_ext_ptr batch(llama_batch_ext_init_from_text(tmp.data(), std::min(tmp.size(), (size_t) params.n_batch), 0, 0));
llama_decode_ext(lctx, batch.get());
}
llama_kv_cache_clear(lctx);
llama_kv_self_clear(lctx);
llama_synchronize(lctx);
llama_perf_context_reset(lctx);
}
@@ -2032,3 +2036,25 @@ common_control_vector_data common_control_vector_load(const std::vector<common_c
return result;
}
template <>
json common_grammar_trigger::to_json() const {
json out {
{"type", (int) type},
{"value", value},
};
if (type == COMMON_GRAMMAR_TRIGGER_TYPE_TOKEN) {
out["token"] = (int) token;
}
return out;
}
template <>
common_grammar_trigger common_grammar_trigger::from_json(const json & in) {
common_grammar_trigger out;
out.type = (common_grammar_trigger_type) in.at("type").get<int>();
out.value = in.at("value").get<std::string>();
if (out.type == COMMON_GRAMMAR_TRIGGER_TYPE_TOKEN) {
out.token = (llama_token) in.at("token").get<int>();
}
return out;
}