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

ChatON:WIP:Add c api wrapper for chat_template_apply

Browse source 
Initial skeletons

Update existing logics to help with same. Also the inbetween helper
was having a bad signature wrt returning status and data, thats also
fixed.
This commit is contained in:
HanishKVC 2024-04-26 07:58:40 +05:30
parent e62699f923
commit 308d3bf3ff
1 changed files with 29 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

34
common/chaton.hpp
View file

@ -306,6 +306,9 @@ inline bool chaton_tmpl_apply_ex(
std::vector<int> &lens, std::vector<int> &lens,
bool alertAssistantAtEnd bool alertAssistantAtEnd
) { ) {
if (!chaton_tmpl_exists(tmpl)) {
return false;
}
ChatParts cp = {}; ChatParts cp = {};
std::stringstream ss; std::stringstream ss;
std::string globalBegin = chaton_tmpl_role_kv(tmpl, K_GLOBAL, {K_BEGIN}); std::string globalBegin = chaton_tmpl_role_kv(tmpl, K_GLOBAL, {K_BEGIN});
@ -390,16 +393,37 @@ inline bool chaton_tmpl_apply_ex(
// global-begin + [[role-begin] + [role-prefix] + msg + role-suffix] + global-end // global-begin + [[role-begin] + [role-prefix] + msg + role-suffix] + global-end
// if there is a combination of system-user messages, // if there is a combination of system-user messages,
// then 1st user message will have user-prefix only if systemuser-1st-user-has-prefix is true // then 1st user message will have user-prefix only if systemuser-1st-user-has-prefix is true
inline std::string chaton_tmpl_apply( inline int32_t chaton_tmpl_apply(
const std::string &tmpl, const std::string &tmpl,
const std::vector<llama_chat_message> &msgs, const std::vector<llama_chat_message> &msgs,
bool alertAssistantAtEnd bool alertAssistantAtEnd,
std::string &tagged
) { ) {
std::string tagged;
std::string types; std::string types;
std::vector<int> lens; std::vector<int> lens;
chaton_tmpl_apply_ex(tmpl, msgs, tagged, types, lens, alertAssistantAtEnd); if (!chaton_tmpl_apply_ex(tmpl, msgs, tagged, types, lens, alertAssistantAtEnd)) {
return tagged; return -1;
}
return tagged.size();
}
inline int32_t chaton_tmpl_apply_capi(
const char *tmpl,
const struct llama_chat_message *msgs,
const size_t numMsgs,
bool alertAssistantAtEnd,
char *dest,
int32_t destLength
) {
if ((tmpl == nullptr) || (dest == nullptr)) {
return -1;
}
std::vector<const llama_chat_message *> vMsgs;
for(size_t i=0; i<numMsgs; i++) {
vMsgs.push_back(vMsgs[i]);
}
std::string taggedMsgs;
int32_t taggedLength = chaton_tmpl_apply(tmpl, vMsgs, alertAssistantAtEnd, taggedMsgs);
} }