sampler.hpp File Reference

Token Sampling Operations. More...

#include "common.hpp"
#include "logits.hpp"
#include "tokenizer.hpp"
#include <cstdint>
#include <ctime>
#include <llama/llama.h>
#include <optional>
#include <stdexcept>
#include <type_traits>
#include <vector>

Go to the source code of this file.

Classes
struct	lloyal::detail::is_optional< T >
	Type trait to detect std::optional<T> More...
struct	lloyal::detail::is_optional< std::optional< T > >

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

Concepts
concept	lloyal::SamplingParamsLike
	C++20 concept: Any type with sampling parameter fields.

Functions
template<class X , class T >
constexpr T	lloyal::detail::as_value (const X &x, T def)
	Extract value from either T or std::optional<T> with fallback.
llama_token	lloyal::sampler::greedy (llama_context *ctx, const llama_vocab *vocab)
	Greedy sampling: Select token with highest probability.
template<SamplingParamsLike P>
llama_token	lloyal::sampler::sample_with_params (llama_context *ctx, const llama_vocab *vocab, const P &params, llama_sampler *grammarSampler=nullptr)
	Sample with configurable parameters (template accepts any SamplingParams type)
llama_token	lloyal::sampler::greedy (llama_context *ctx, const llama_model *model)
	Greedy sampling with automatic vocab extraction.
template<SamplingParamsLike P>
llama_token	lloyal::sampler::sample_with_params (llama_context *ctx, const llama_model *model, const P &params, llama_sampler *grammarSampler=nullptr)
	Parameterized sampling with automatic vocab extraction.
template<SamplingParamsLike P>
llama_sampler *	lloyal::sampler::create_chain (const P &params)
	Create a persistent sampler chain from parameters.
llama_sampler *	lloyal::sampler::clone_chain (llama_sampler *chain)
	Clone a sampler chain.
void	lloyal::sampler::reseed_chain (llama_sampler *chain, uint32_t new_seed)
	Reseed the dist sampler in a chain.
void	lloyal::sampler::free_chain (llama_sampler *chain)
	Free a sampler chain.
void	lloyal::sampler::apply (llama_sampler *chain, llama_token_data_array *cur_p)
	Apply a sampler chain to a candidate array.
void	lloyal::sampler::accept (llama_sampler *chain, llama_token token)
	Accept a token into the sampler chain.

Detailed Description

Token Sampling Operations.

Wraps llama.cpp sampling APIs with configurable sampling strategies. Uses C++20 concepts for generic parameter handling.

Architecture:

• Concept-constrained templates accept any Nitrogen-generated SamplingParams type
• No struct duplication or adapters required
• Supports greedy, temperature, top-k, top-p, min-p, grammar-constrained sampling

Definition in file sampler.hpp.