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

correct buffer size

Browse source
This commit is contained in:
ngxson 2024-04-24 19:41:51 +02:00
parent 7f89803536
commit 476d319fde
3 changed files with 9 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

8
llama.cpp
View file

@ -17303,7 +17303,7 @@ LLAMA_API int32_t llama_chat_get_model_template(
return -1;
} else {
snprintf(buf, length, "%s", model_template.c_str());
return model_template.size();
return model_template.size() + 1;
}
}
@ -17469,7 +17469,7 @@ LLAMA_API int32_t llama_chat_get_prefix(
}
std::string output = ss.str();
snprintf(buf, length, "%s", output.c_str());
return output.size();
return output.size() + 1;
}
LLAMA_API int32_t llama_chat_get_postfix(
@ -17551,7 +17551,7 @@ LLAMA_API int32_t llama_chat_get_postfix(
}
std::string output = ss.str();
snprintf(buf, length, "%s", output.c_str());
return output.size();
return output.size() + 1;
}
LLAMA_API bool llama_chat_support_system_message(const enum llama_chat_template ttmpl) {
@ -17641,7 +17641,7 @@ LLAMA_API int32_t llama_chat_apply_template(
if (buf && length > 0) {
snprintf(buf, length, "%s", output.c_str());
}
return output.size();
return output.size() + 1;
}
LLAMA_API int llama_split_path(char * split_path, size_t maxlen, const char * path_prefix, int split_no, int split_count) {

8
llama.h
View file

@ -870,7 +870,7 @@ extern "C" {
/// @param add_ass Whether to end the prompt with the token(s) that indicate the start of an assistant message.
/// @param buf A buffer to hold the output formatted prompt. The recommended alloc size is 2 * (total number of characters of all messages)
/// @param length The size of the allocated buffer
/// @return The total number of bytes of the formatted prompt. If is it larger than the size of buffer, you may need to re-alloc it and then re-apply the template.
/// @return The total number of bytes of the formatted prompt (null terminator included). If is it larger than the size of buffer, you may need to re-alloc it and then re-apply the template.
LLAMA_API int32_t llama_chat_apply_template(
const struct llama_model * model,
const char * tmpl,
@ -885,7 +885,7 @@ extern "C" {
/// @param name Template name (can be a nullptr for default template). See: https://github.com/ggerganov/llama.cpp/pull/6588
/// @param buf The output buffer
/// @param length The size of the allocated buffer
/// @return The total number of bytes of the template. If a named template cannot be found, it will use default template. If no template can be found, it returns -1
/// @return The total number of bytes of the template (null terminator included). If a named template cannot be found, it will use default template. If no template can be found, it returns -1
LLAMA_API int32_t llama_chat_get_model_template(
const struct llama_model * model,
const char * name,
@ -903,7 +903,7 @@ extern "C" {
/// @param prev_role The role of the previous message, can be nullptr
/// @param buf The output buffer
/// @param length The size of the allocated buffer
/// @return The total number of bytes of the output string
/// @return The total number of bytes of the output string (null terminator included)
LLAMA_API int32_t llama_chat_get_prefix(
const enum llama_chat_template tmpl,
const char * role,
@ -917,7 +917,7 @@ extern "C" {
/// @param prev_role The role of the previous message, can be nullptr
/// @param buf The output buffer
/// @param length The size of the allocated buffer
/// @return The total number of bytes of the output string
/// @return The total number of bytes of the output string (null terminator included)
LLAMA_API int32_t llama_chat_get_postfix(
const enum llama_chat_template tmpl,
const char * role,

2
tests/test-chat-template.cpp
View file

@ -103,7 +103,7 @@ int main(void) {
formatted_chat.size()
);
formatted_chat.resize(res);
std::string output(formatted_chat.data(), formatted_chat.size());
std::string output(formatted_chat.data());
std::cout << output << "\n-------------------------\n";
assert(output == expected);
}