helpers.hpp File Reference

Helper Utilities. More...

#include "common.hpp"
#include "minja/chat-template.hpp"
#include "minja/minja.hpp"
#include <cassert>
#include <chrono>
#include <llama/ggml.h>
#include <llama/llama.h>
#include <lloyal/nlohmann/json.hpp>
#include <memory>
#include <sstream>
#include <string>
#include <vector>

Go to the source code of this file.

Classes
struct	lloyal::ChatTemplateResult
	Result from complete chat template processing. More...

Namespaces
namespace	lloyal
	JSON Schema to Grammar Converter (Header-Only)
namespace	lloyal::detail

Typedefs
using	lloyal::json = nlohmann::ordered_json

Functions
std::string	lloyal::detail::common_token_to_piece (const struct llama_vocab *vocab, llama_token token, bool special)
std::string	lloyal::detail::get_token_safe (const llama_model *model, llama_token token)
const char *	lloyal::detail::get_chatml_template ()
std::string	lloyal::detail::apply_chat_template_helper (const std::string &template_str, const nlohmann::ordered_json &messages, const std::string &bos_token, const std::string &eos_token, bool add_generation_prompt, bool add_bos, bool add_eos)
void	lloyal::batch_clear (llama_batch &batch)
	Clear batch to empty state.
void	lloyal::batch_add (llama_batch &batch, llama_token id, int32_t pos, const std::vector< llama_seq_id > &seq_ids, bool logits, int32_t capacity=-1)
	Add single token to batch with position and sequence info.
std::string	lloyal::format_chat_template_from_model (const llama_model *model, const std::string &messages_json, const std::string &template_override="")
	Format chat messages using model's built-in template.
std::vector< std::string >	lloyal::extract_template_stop_tokens (const llama_model *model, const std::string &template_str)
	Dynamically detect stop tokens from chat template.
ChatTemplateResult	lloyal::format_chat_template_complete (const llama_model *model, const std::string &messages_json, const std::string &template_override="")
	Complete chat template processing with stop token detection.
bool	lloyal::validate_chat_template_helper (const std::string &template_str)
	Validate chat template syntax.
const std::vector< ggml_type > &	lloyal::get_kv_cache_types ()
	Get list of supported KV cache types.
ggml_type	lloyal::kv_cache_type_from_str (const std::string &s)
	Convert cache type string to ggml_type enum.
bool	lloyal::is_truthy (const std::string &value)
	Check if string represents a truthy value.
bool	lloyal::is_falsey (const std::string &value)
	Check if string represents a falsey value.
bool	lloyal::is_autoy (const std::string &value)
	Check if string represents an auto value.
std::string	lloyal::string_repeat (const std::string &str, size_t n)
std::string	lloyal::string_join (const std::vector< std::string > &values, const std::string &separator)
std::vector< std::string >	lloyal::string_split (const std::string &str, const std::string &delimiter)

Detailed Description

Helper Utilities.

Collection of utility functions for common llama.cpp operations:

• Batch operations: Build and manage token batches for decoding
• Chat template processing: Format messages, extract stop tokens, validate templates
• Parameter conversion: KV cache type mapping, string validation helpers
• String utilities: Repeat, join, split operations

Source: Vendored from llama.cpp/common/ License: MIT License - Copyright (c) 2023-2024 The ggml.ai team

Definition in file helpers.hpp.