common.cpp

#include "common.h"
#include "llama.h"

#include <algorithm>
#include <cassert>
#include <cmath>
#include <cstring>
#include <ctime>
#include <fstream>
#include <iterator>
#include <iostream>
#include <regex>
#include <sstream>
#include <string>
#include <unordered_map>
#include <unordered_set>
#include <vector>
#include <cinttypes>

#if defined(__APPLE__) && defined(__MACH__)
#include <sys/types.h>
#include <sys/sysctl.h>
#endif

#if defined(_WIN32)
#define WIN32_LEAN_AND_MEAN
#ifndef NOMINMAX
#   define NOMINMAX
#endif
#include <codecvt>
#include <locale>
#include <windows.h>
#include <fcntl.h>
#include <io.h>
#else
#include <sys/ioctl.h>
#include <sys/stat.h>
#include <unistd.h>
#endif

#if defined(_MSC_VER)
#pragma warning(disable: 4244 4267) // possible loss of data
#endif

#if (defined(GGML_USE_CUBLAS) || defined(GGML_USE_SYCL))
#define GGML_USE_CUBLAS_SYCL
#endif

int32_t get_num_physical_cores() {
    unsigned int n_threads = std::thread::hardware_concurrency();
    return n_threads > 0 ? (n_threads <= 4 ? n_threads : n_threads / 2) : 4;
}

void process_escapes(std::string& input) {
    std::size_t input_len = input.length();
    std::size_t output_idx = 0;

    for (std::size_t input_idx = 0; input_idx < input_len; ++input_idx) {
        if (input[input_idx] == '\\' && input_idx + 1 < input_len) {
            switch (input[++input_idx]) {
                case 'n':  input[output_idx++] = '\n'; break;
                case 'r':  input[output_idx++] = '\r'; break;
                case 't':  input[output_idx++] = '\t'; break;
                case '\'': input[output_idx++] = '\''; break;
                case '\"': input[output_idx++] = '\"'; break;
                case '\\': input[output_idx++] = '\\'; break;
                case 'x':
                    // Handle \x12, etc
                    if (input_idx + 2 < input_len) {
                        const char x[3] = { input[input_idx + 1], input[input_idx + 2], 0 };
                        char *err_p = nullptr;
                        const long val = std::strtol(x, &err_p, 16);
                        if (err_p == x + 2) {
                            input_idx += 2;
                            input[output_idx++] = char(val);
                            break;
                        }
                    }
                    // fall through
                default:   input[output_idx++] = '\\';
                           input[output_idx++] = input[input_idx]; break;
            }
        } else {
            input[output_idx++] = input[input_idx];
        }
    }

    input.resize(output_idx);
}

std::string get_system_info(const gpt_params & params) {
    std::ostringstream os;

    os << "system_info: n_threads = " << params.n_threads;
    if (params.n_threads_batch != -1) {
        os << " (n_threads_batch = " << params.n_threads_batch << ")";
    }
    os << " / " << std::thread::hardware_concurrency() << " | " << llama_print_system_info();

    return os.str();
}

//
// Model utils
//

struct llama_model_params llama_model_params_from_gpt_params(const gpt_params & params) {
    auto mparams = llama_model_default_params();

    if (params.n_gpu_layers != -1) {
        mparams.n_gpu_layers = params.n_gpu_layers;
    }
    mparams.main_gpu        = params.main_gpu;
    mparams.split_mode      = params.split_mode;
    mparams.tensor_split    = params.tensor_split;
    mparams.use_mmap        = params.use_mmap;
    mparams.use_mlock       = params.use_mlock;
    //if (params.kv_overrides.empty()) {
        mparams.kv_overrides = NULL;
    /*} else {
        GGML_ASSERT(params.kv_overrides.back().key[0] == 0 && "KV overrides not terminated with empty key");
        mparams.kv_overrides = params.kv_overrides.data();
    }*/

    return mparams;
}

struct llama_context_params llama_context_params_from_gpt_params(const gpt_params & params) {
    auto cparams = llama_context_default_params();

    cparams.n_ctx             = params.n_ctx;
    cparams.n_batch           = params.n_batch;
    cparams.n_threads         = params.n_threads;
    cparams.n_threads_batch   = params.n_threads_batch == -1 ? params.n_threads : params.n_threads_batch;
    cparams.mul_mat_q         = params.mul_mat_q;
    cparams.seed              = params.seed;
    cparams.logits_all        = params.logits_all;
    cparams.embedding         = params.embedding;
    cparams.rope_scaling_type = params.rope_scaling_type;
    cparams.rope_freq_base    = params.rope_freq_base;
    cparams.rope_freq_scale   = params.rope_freq_scale;
    cparams.yarn_ext_factor   = params.yarn_ext_factor;
    cparams.yarn_attn_factor  = params.yarn_attn_factor;
    cparams.yarn_beta_fast    = params.yarn_beta_fast;
    cparams.yarn_beta_slow    = params.yarn_beta_slow;
    cparams.yarn_orig_ctx     = params.yarn_orig_ctx;
    cparams.offload_kqv       = !params.no_kv_offload;

    cparams.type_k = params.cache_type_k;
    cparams.type_v = params.cache_type_v;

    return cparams;
}

void llama_batch_clear(struct llama_batch & batch) {
    batch.n_tokens = 0;
}

void llama_batch_add(
                 struct llama_batch & batch,
                        llama_token   id,
                          llama_pos   pos,
    const std::vector<llama_seq_id> & seq_ids,
                               bool   logits) {
    batch.token   [batch.n_tokens] = id;
    batch.pos     [batch.n_tokens] = pos;
    batch.n_seq_id[batch.n_tokens] = seq_ids.size();
    for (size_t i = 0; i < seq_ids.size(); ++i) {
        batch.seq_id[batch.n_tokens][i] = seq_ids[i];
    }
    batch.logits  [batch.n_tokens] = logits;

    batch.n_tokens++;
}

std::tuple<struct llama_model *, struct llama_context *> llama_init_from_gpt_params(gpt_params & params) {
    auto mparams = llama_model_params_from_gpt_params(params);

    llama_model * model  = llama_load_model_from_file(params.model, mparams);
    if (model == NULL) {
        fprintf(stderr, "%s: error: failed to load model '%s'\n", __func__, params.model);
        return std::make_tuple(nullptr, nullptr);
    }

    auto cparams = llama_context_params_from_gpt_params(params);

    llama_context * lctx = llama_new_context_with_model(model, cparams);
    if (lctx == NULL) {
        fprintf(stderr, "%s: error: failed to create context with model '%s'\n", __func__, params.model);
        llama_free_model(model);
        return std::make_tuple(nullptr, nullptr);
    }

    /*for (unsigned int i = 0; i < params.lora_adapter.size(); ++i) {
        const std::string& lora_adapter = std::get<0>(params.lora_adapter[i]);
        float lora_scale = std::get<1>(params.lora_adapter[i]);
        int err = llama_model_apply_lora_from_file(model,
                                             lora_adapter.c_str(),
                                             lora_scale,
                                             ((i > 0) || params.lora_base.empty())
                                                ? NULL
                                                : params.lora_base.c_str(),
                                             params.n_threads);
        if (err != 0) {
            fprintf(stderr, "%s: error: failed to apply lora adapter\n", __func__);
            llama_free(lctx);
            llama_free_model(model);
            return std::make_tuple(nullptr, nullptr);
        }
    }*/

    /*if (params.ignore_eos) {
        params.sparams.logit_bias[llama_token_eos(model)] = -INFINITY;
    }*/

    {
        LOG("warming up the model with an empty run\n");

        std::vector<llama_token> tmp = { llama_token_bos(model), llama_token_eos(model), };
        llama_decode(lctx, llama_batch_get_one(tmp.data(), std::min(tmp.size(), (size_t) params.n_batch), 0, 0));
        llama_kv_cache_clear(lctx);
        llama_reset_timings(lctx);
    }

    return std::make_tuple(model, lctx);
}

//
// Vocab utils
//

std::vector<llama_token> llama_tokenize(
  const struct llama_context * ctx,
           const std::string & text,
                        bool   add_bos,
                        bool   special,
                        bool   no_space) {
    return llama_tokenize(llama_get_model(ctx), text, add_bos, special, no_space);
}

std::vector<llama_token> llama_tokenize(
    const struct llama_model * model,
           const std::string & text,
                        bool   add_bos,
                        bool   special,
                        bool   no_space) {
    // upper limit for the number of tokens
    int n_tokens = text.length() + add_bos;
    std::vector<llama_token> result(n_tokens);
    n_tokens = llama_tokenize(model, text.data(), text.length(), result.data(), result.size(), add_bos, special, no_space);
    if (n_tokens < 0) {
        result.resize(-n_tokens);
        int check = llama_tokenize(model, text.data(), text.length(), result.data(), result.size(), add_bos, special, no_space);
        GGML_ASSERT(check == -n_tokens);
    } else {
        result.resize(n_tokens);
    }
    return result;
}

std::string llama_token_to_piece(const struct llama_context * ctx, llama_token token) {
    std::vector<char> result(8, 0);
    const int n_tokens = llama_token_to_piece(llama_get_model(ctx), token, result.data(), result.size());
    if (n_tokens < 0) {
        result.resize(-n_tokens);
        int check = llama_token_to_piece(llama_get_model(ctx), token, result.data(), result.size());
        GGML_ASSERT(check == -n_tokens);
    } else {
        result.resize(n_tokens);
    }

    return std::string(result.data(), result.size());
}

std::string llama_detokenize_spm(llama_context * ctx, const std::vector<llama_token> & tokens) {
    const llama_token bos_id = llama_token_bos(llama_get_model(ctx));

    std::string piece;
    std::string result;

    for (size_t i = 0; i < tokens.size(); ++i) {
        piece = llama_token_to_piece(ctx, tokens[i]);

        // remove the leading space of the first non-BOS token
        if (((tokens[0] == bos_id && i == 1) || (tokens[0] != bos_id && i == 0)) && piece[0] == ' ') {
            piece = piece.substr(1);
        }

        result += piece;
    }

    return result;
}

std::string llama_detokenize_bpe(llama_context * ctx, const std::vector<llama_token> & tokens) {
    std::string piece;
    std::string result;

    for (size_t i = 0; i < tokens.size(); ++i) {
        piece = llama_token_to_piece(ctx, tokens[i]);

        result += piece;
    }

    // NOTE: the original tokenizer decodes bytes after collecting the pieces.
    return result;
}

bool llama_should_add_bos_token(const llama_model * model) {
    const int add_bos = llama_add_bos_token(model);

    return add_bos != -1 ? bool(add_bos) : (llama_vocab_type(model) == LLAMA_VOCAB_TYPE_SPM);
}