#pragma once

#include "ggml.h" // for ggml_log_level

#include <string>

#include <type_traits>

#include <vector>

#ifdef __GNUC__

#    if defined(__MINGW32__) && !defined(__clang__)

#        define LLAMA_ATTRIBUTE_FORMAT(...) __attribute__((format(gnu_printf, __VA_ARGS__)))

#    else

#        define LLAMA_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))

#    endif

#else

#    define LLAMA_ATTRIBUTE_FORMAT(...)

#endif

//

// logging

//

LLAMA_ATTRIBUTE_FORMAT(2, 3)

void llama_log_internal        (ggml_log_level level, const char * format, ...);

void llama_log_callback_default(ggml_log_level level, const char * text, void * user_data);

#define LLAMA_LOG(...)       llama_log_internal(GGML_LOG_LEVEL_NONE , __VA_ARGS__)

#define LLAMA_LOG_INFO(...)  llama_log_internal(GGML_LOG_LEVEL_INFO , __VA_ARGS__)

Unexecuted instantiation: llama-model.cpp:llama_model::print_info() const::$_0::operator()(unsigned int) const

Unexecuted instantiation: llama-model.cpp:llama_model::print_info() const::$_1::operator()(unsigned int) const

Unexecuted instantiation: llama-model.cpp:llama_model::print_info() const::$_2::operator()(unsigned int) const

Unexecuted instantiation: llama-model.cpp:llama_model::print_info() const::$_3::operator()(unsigned int) const

Unexecuted instantiation: llama-model.cpp:llama_model::print_info() const::$_4::operator()(unsigned int) const

Unexecuted instantiation: llama-model.cpp:llama_model::print_info() const::$_5::operator()(unsigned int) const

Unexecuted instantiation: llama-model.cpp:llama_model::print_info() const::$_6::operator()(unsigned int) const

#define LLAMA_LOG_WARN(...)  llama_log_internal(GGML_LOG_LEVEL_WARN , __VA_ARGS__)

#define LLAMA_LOG_ERROR(...) llama_log_internal(GGML_LOG_LEVEL_ERROR, __VA_ARGS__)

#define LLAMA_LOG_DEBUG(...) llama_log_internal(GGML_LOG_LEVEL_DEBUG, __VA_ARGS__)

#define LLAMA_LOG_CONT(...)  llama_log_internal(GGML_LOG_LEVEL_CONT , __VA_ARGS__)

//

// helpers

//

template <typename T>

struct no_init {

    T value;

    no_init() = default;

};

template <typename dst_t, typename src_t>

static inline dst_t llama_cast(src_t v) {

    if constexpr (std::is_same_v<src_t, dst_t>) {

        return v;

    } else if constexpr (std::is_same_v<src_t, ggml_fp16_t> && std::is_same_v<dst_t, float>) {

        return ggml_fp16_to_fp32(v);

    } else if constexpr (std::is_same_v<src_t, float> && std::is_same_v<dst_t, ggml_fp16_t>) {

        return ggml_fp32_to_fp16(v);

    } else {

        static_assert(std::is_same_v<dst_t, void>, "unsupported type combination");

    }

}

Unexecuted instantiation: llama-graph.cpp:unsigned short llama_cast<unsigned short, float>(float)

Unexecuted instantiation: llama-graph.cpp:float llama_cast<float, unsigned short>(unsigned short)

Unexecuted instantiation: llama-graph.cpp:float llama_cast<float, float>(float)

Unexecuted instantiation: llama-kv-cache.cpp:unsigned short llama_cast<unsigned short, float>(float)

Unexecuted instantiation: llama-kv-cache.cpp:float llama_cast<float, float>(float)

struct time_meas {

    time_meas(int64_t & t_acc, bool disable = false);

    ~time_meas();

    const int64_t t_start_us;

    int64_t & t_acc;

};

template <typename T>

struct buffer_view {

    T * data;

    size_t size = 0;

    bool has_data() const {

        return data && size > 0;

    }

Unexecuted instantiation: buffer_view<float>::has_data() const

Unexecuted instantiation: buffer_view<int>::has_data() const

};

void replace_all(std::string & s, const std::string & search, const std::string & replace);

// TODO: rename to llama_format ?

LLAMA_ATTRIBUTE_FORMAT(1, 2)

std::string format(const char * fmt, ...);

std::string llama_format_tensor_shape(const std::vector<int64_t> & ne);

std::string llama_format_tensor_shape(const struct ggml_tensor * t);

std::string gguf_kv_to_str(const struct gguf_context * ctx_gguf, int i);

#define LLAMA_TENSOR_NAME_FATTN   "__fattn__"

#define LLAMA_TENSOR_NAME_FGDN_AR "__fgdn_ar__"

#define LLAMA_TENSOR_NAME_FGDN_CH "__fgdn_ch__"