#include "api.h"
#include "chat_gguf.h"
#include "draft_table.h"
#include "gguf_parser.h"
#include "model_macros.h"
#include <algorithm>
#include <chrono>
#include <cmath>
#include <cstdint>
#include <filesystem>
#include <fstream>
#include <functional>
#include <iostream>
#include <map>
#include <memory>
#include <mutex>
#include <nlohmann/json.hpp>
#include "graph_ir.h"
#include <numeric>
#include <random>
#include <type_traits>
#include <variant>
#include <stdexcept>
#include <string>
#include <vector>
#include <iomanip>
#include <sstream>
#include "logger.h"
#include "model.h"
#include "safetensors_loader.h"
#include "tokenizer.h"
#include "utf8_text.h"
Go to the source code of this file.
|
| static void | tinyllama::log_vector_summary_detailed (const std::string &name, const std::vector< float > &v, int current_pos, int current_layer, int N) |
| |
| static std::string | tinyllama::read_file_api (const std::string &path) |
| |
| static int | tinyllama::argmax (const std::vector< float > &v) |
| |
| static std::vector< int > | tinyllama::prompt_lookup_draft (const std::vector< int > &h, int k) |
| |
| static const ember_draft::DraftTable * | tinyllama::corpus_draft_table (int vocab_size) |
| |
| static void | tinyllama::apply_repetition_penalty (std::vector< float > &logits, const std::vector< int > &generated, float penalty, size_t window=64) |
| |
| template<class IsStop > |
| static void | tinyllama::apply_contract_penalty (std::vector< float > &lg, const std::vector< int > &a, const std::vector< int > &b, float p, int last_n, IsStop is_stop, std::vector< char > &seen) |
| |
| static void | tinyllama::debug_dump_topk (const std::vector< float > &logits) |
| |
| static bool | tinyllama::logits_debug_on () |
| |
| static int | tinyllama::sample_top_k_top_p_temperature (const std::vector< float > &logits, float temperature, int top_k, float top_p, std::mt19937 &rng) |
| |
◆ SAFE_SQRT