llama-cpp-capacitor 0.0.6 → 0.0.7

package/cpp/llama-vocab.h

@@ -0,0 +1,176 @@
+#pragma once
+
+#include "llama.h"
+
+#include <string>
+#include <vector>
+#include <memory>
+
+// pre-tokenization types
+enum llama_vocab_pre_type {
+    LLAMA_VOCAB_PRE_TYPE_DEFAULT        = 0,
+    LLAMA_VOCAB_PRE_TYPE_LLAMA3         = 1,
+    LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_LLM   = 2,
+    LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_CODER = 3,
+    LLAMA_VOCAB_PRE_TYPE_FALCON         = 4,
+    LLAMA_VOCAB_PRE_TYPE_MPT            = 5,
+    LLAMA_VOCAB_PRE_TYPE_STARCODER      = 6,
+    LLAMA_VOCAB_PRE_TYPE_GPT2           = 7,
+    LLAMA_VOCAB_PRE_TYPE_REFACT         = 8,
+    LLAMA_VOCAB_PRE_TYPE_COMMAND_R      = 9,
+    LLAMA_VOCAB_PRE_TYPE_STABLELM2      = 10,
+    LLAMA_VOCAB_PRE_TYPE_QWEN2          = 11,
+    LLAMA_VOCAB_PRE_TYPE_OLMO           = 12,
+    LLAMA_VOCAB_PRE_TYPE_DBRX           = 13,
+    LLAMA_VOCAB_PRE_TYPE_SMAUG          = 14,
+    LLAMA_VOCAB_PRE_TYPE_PORO           = 15,
+    LLAMA_VOCAB_PRE_TYPE_CHATGLM3       = 16,
+    LLAMA_VOCAB_PRE_TYPE_CHATGLM4       = 17,
+    LLAMA_VOCAB_PRE_TYPE_VIKING         = 18,
+    LLAMA_VOCAB_PRE_TYPE_JAIS           = 19,
+    LLAMA_VOCAB_PRE_TYPE_TEKKEN         = 20,
+    LLAMA_VOCAB_PRE_TYPE_SMOLLM         = 21,
+    LLAMA_VOCAB_PRE_TYPE_CODESHELL      = 22,
+    LLAMA_VOCAB_PRE_TYPE_BLOOM          = 23,
+    LLAMA_VOCAB_PRE_TYPE_GPT3_FINNISH   = 24,
+    LLAMA_VOCAB_PRE_TYPE_EXAONE         = 25,
+    LLAMA_VOCAB_PRE_TYPE_CHAMELEON      = 26,
+    LLAMA_VOCAB_PRE_TYPE_MINERVA        = 27,
+    LLAMA_VOCAB_PRE_TYPE_DEEPSEEK3_LLM  = 28,
+    LLAMA_VOCAB_PRE_TYPE_GPT4O          = 29,
+    LLAMA_VOCAB_PRE_TYPE_SUPERBPE       = 30,
+    LLAMA_VOCAB_PRE_TYPE_TRILLION       = 31,
+    LLAMA_VOCAB_PRE_TYPE_BAILINGMOE     = 32,
+    LLAMA_VOCAB_PRE_TYPE_LLAMA4         = 33,
+    LLAMA_VOCAB_PRE_TYPE_PIXTRAL        = 34,
+    LLAMA_VOCAB_PRE_TYPE_SEED_CODER     = 35,
+    LLAMA_VOCAB_PRE_TYPE_HUNYUAN        = 36,
+    LLAMA_VOCAB_PRE_TYPE_KIMI_K2        = 37,
+    LLAMA_VOCAB_PRE_TYPE_HUNYUAN_DENSE  = 38,
+};
+
+struct LLM_KV;
+struct llama_model_loader;
+
+struct llama_vocab {
+    struct token_data {
+        std::string      text;
+        float            score;
+        llama_token_attr attr;
+    };
+
+    llama_vocab();
+    ~llama_vocab();
+
+    void load(llama_model_loader & ml, const LLM_KV & kv);
+
+    std::string get_tokenizer_model() const;
+    std::string get_tokenizer_pre() const;
+
+    enum llama_vocab_type     get_type()     const;
+    enum llama_vocab_pre_type get_pre_type() const;
+
+    uint32_t n_tokens() const;
+    uint32_t n_token_types() const;
+
+    std::string type_name() const;
+
+    bool is_normal      (llama_token id) const;
+    bool is_unknown     (llama_token id) const;
+    bool is_control     (llama_token id) const;
+    bool is_byte        (llama_token id) const;
+    bool is_user_defined(llama_token id) const;
+    bool is_unused      (llama_token id) const;
+    bool is_eog         (llama_token id) const;
+
+    uint8_t     token_to_byte(llama_token id) const;
+    llama_token byte_to_token(uint8_t ch)     const;
+
+    llama_token text_to_token(const std::string & text) const;
+
+    const token_data & get_token_data(llama_token id) const;
+
+    const char *     token_get_text (llama_token id) const;
+    float            token_get_score(llama_token id) const;
+    llama_token_attr token_get_attr (llama_token id) const;
+
+    llama_token token_bos() const;
+    llama_token token_eos() const;
+    llama_token token_eot() const;
+    llama_token token_eom() const;
+    llama_token token_unk() const;
+    llama_token token_sep() const;
+    llama_token token_nl () const;
+    llama_token token_pad() const;
+    llama_token token_mask() const;
+
+    llama_token token_prefix() const;
+    llama_token token_middle() const;
+    llama_token token_suffix() const;
+
+    llama_token token_fim_pre() const;
+    llama_token token_fim_suf() const;
+    llama_token token_fim_mid() const;
+    llama_token token_fim_pad() const;
+    llama_token token_fim_rep() const;
+    llama_token token_fim_sep() const;
+
+    bool get_add_space_prefix          () const;
+    bool get_add_bos                   () const;
+    bool get_add_eos                   () const;
+    bool get_add_sep                   () const;
+    bool get_ignore_merges             () const;
+    bool get_clean_spaces              () const;
+    bool get_remove_extra_whitespaces  () const;
+    bool get_escape_whitespaces        () const;
+    bool get_treat_whitespace_as_suffix() const;
+
+    int max_token_len() const;
+
+    int find_bpe_rank(const std::string & token_left, const std::string & token_right) const;
+    std::vector<std::string> get_bpe_merges() const;
+
+    std::vector<char> get_precompiled_charsmap() const;
+
+    int32_t tokenize(
+                   const char * text,
+                      int32_t   text_len,
+                  llama_token * tokens,
+                      int32_t   n_tokens_max,
+                         bool   add_special,
+                         bool   parse_special) const;
+
+    std::vector<llama_token> tokenize(
+            const std::string & raw_text,
+                         bool   add_special,
+                         bool   parse_special = false) const;
+
+    // does not write null-terminator to buf
+    int32_t token_to_piece(
+                  llama_token   token,
+                         char * buf,
+                      int32_t   length,
+                      int32_t   lstrip,
+                         bool   special) const;
+
+    // use cached data
+    const std::string & token_to_piece(llama_token token) const;
+
+    int32_t detokenize(
+            const llama_token * tokens,
+                      int32_t   n_tokens,
+                         char * text,
+                      int32_t   text_len_max,
+                         bool   remove_special,
+                         bool   unparse_special) const;
+
+    std::string detokenize(
+            const std::vector<llama_token> & tokens,
+                                      bool   special) const;
+
+    void print_info() const;
+
+private:
+    struct impl;
+    std::unique_ptr<impl> pimpl;
+};

package/cpp/llama.cpp

@@ -0,0 +1,358 @@
+#include "llama-impl.h"
+
+#include "llama-chat.h"
+#include "llama-mmap.h"
+#include "llama-vocab.h"
+#include "llama-model-loader.h"
+#include "llama-model-saver.h"
+#include "llama-model.h"
+
+#include "ggml.h"
+#include "ggml-backend.h"
+
+#include <algorithm>
+#include <cstddef>
+#include <cstdint>
+#include <cstdio>
+#include <cstring>
+#include <ctime>
+
+#if defined(_MSC_VER)
+#pragma warning(disable: 4244 4267) // possible loss of data
+#endif
+
+//
+// interface implementation
+//
+
+struct llama_sampler_chain_params llama_sampler_chain_default_params() {
+    struct llama_sampler_chain_params result = {
+        /*.no_perf                     =*/ true,
+    };
+
+    return result;
+}
+
+size_t llama_max_devices(void) {
+    return 16;
+}
+
+bool llama_supports_mmap(void) {
+    return llama_mmap::SUPPORTED;
+}
+
+bool llama_supports_mlock(void) {
+    return llama_mlock::SUPPORTED;
+}
+
+bool llama_supports_gpu_offload(void) {
+    return lm_ggml_backend_dev_by_type(LM_GGML_BACKEND_DEVICE_TYPE_GPU) != nullptr ||
+           llama_supports_rpc();
+}
+
+bool llama_supports_rpc(void) {
+    return lm_ggml_backend_reg_by_name("RPC") != nullptr;
+}
+
+void llama_backend_init(void) {
+    lm_ggml_time_init();
+
+    // needed to initialize f16 tables
+    {
+        struct lm_ggml_init_params params = { 0, NULL, false };
+        struct lm_ggml_context * ctx = lm_ggml_init(params);
+        lm_ggml_free(ctx);
+    }
+}
+
+void llama_numa_init(enum lm_ggml_numa_strategy numa) {
+    if (numa != LM_GGML_NUMA_STRATEGY_DISABLED) {
+        auto * dev = lm_ggml_backend_dev_by_type(LM_GGML_BACKEND_DEVICE_TYPE_CPU);
+        LM_GGML_ASSERT(dev && "CPU backend is not loaded");
+        auto * reg = lm_ggml_backend_dev_backend_reg(dev);
+        auto * numa_init_fn = (decltype(lm_ggml_numa_init) *) lm_ggml_backend_reg_get_proc_address(reg, "lm_ggml_backend_cpu_numa_init");
+        numa_init_fn(numa);
+    }
+}
+
+void llama_backend_free(void) {
+    lm_ggml_quantize_free();
+}
+
+int64_t llama_time_us(void) {
+    return lm_ggml_time_us();
+}
+
+// Returns 0 on success, -1 on error, and -2 on cancellation via llama_progress_callback
+static int llama_model_load(const std::string & fname, std::vector<std::string> & splits, llama_model & model, llama_model_params & params) {
+    // loading time will be recalculated after the first eval, so
+    // we take page faults deferred by mmap() into consideration
+    model.t_load_us = 0;
+    time_meas tm(model.t_load_us);
+
+    model.t_start_us = tm.t_start_us;
+
+    try {
+        llama_model_loader ml(fname, splits, params.use_mmap, params.check_tensors, params.kv_overrides, params.tensor_buft_overrides);
+
+        ml.print_info();
+
+        model.hparams.vocab_only = params.vocab_only;
+
+        try {
+            model.load_arch(ml);
+        } catch(const std::exception & e) {
+            throw std::runtime_error("error loading model architecture: " + std::string(e.what()));
+        }
+        try {
+            model.load_hparams(ml);
+        } catch(const std::exception & e) {
+            throw std::runtime_error("error loading model hyperparameters: " + std::string(e.what()));
+        }
+        try {
+            model.load_vocab(ml);
+        } catch(const std::exception & e) {
+            throw std::runtime_error("error loading model vocabulary: " + std::string(e.what()));
+        }
+
+        model.load_stats(ml);
+        model.print_info();
+
+        if (params.vocab_only) {
+            LLAMA_LOG_INFO("%s: vocab only - skipping tensors\n", __func__);
+            return 0;
+        }
+
+        if (!model.load_tensors(ml)) {
+            return -2;
+        }
+    } catch (const std::exception & err) {
+        LLAMA_LOG_ERROR("%s: error loading model: %s\n", __func__, err.what());
+        return -1;
+    }
+
+    return 0;
+}
+
+static struct llama_model * llama_model_load_from_file_impl(
+        const std::string & path_model,
+        std::vector<std::string> & splits,
+        struct llama_model_params params) {
+    lm_ggml_time_init();
+
+    if (!params.vocab_only && lm_ggml_backend_reg_count() == 0) {
+        LLAMA_LOG_ERROR("%s: no backends are loaded. hint: use lm_ggml_backend_load() or lm_ggml_backend_load_all() to load a backend before calling this function\n", __func__);
+        return nullptr;
+    }
+
+    unsigned cur_percentage = 0;
+    if (params.progress_callback == NULL) {
+        params.progress_callback_user_data = &cur_percentage;
+        params.progress_callback = [](float progress, void * ctx) {
+            unsigned * cur_percentage_p = (unsigned *) ctx;
+            unsigned percentage = (unsigned) (100 * progress);
+            while (percentage > *cur_percentage_p) {
+                *cur_percentage_p = percentage;
+                LLAMA_LOG_CONT(".");
+                if (percentage >= 100) {
+                    LLAMA_LOG_CONT("\n");
+                }
+            }
+            return true;
+        };
+    }
+
+    llama_model * model = new llama_model(params);
+
+    // create list of devices to use with this model
+    if (params.devices) {
+        for (lm_ggml_backend_dev_t * dev = params.devices; *dev; ++dev) {
+            model->devices.push_back(*dev);
+        }
+    } else {
+        std::vector<lm_ggml_backend_dev_t> rpc_servers;
+        // use all available devices
+        for (size_t i = 0; i < lm_ggml_backend_dev_count(); ++i) {
+            lm_ggml_backend_dev_t dev = lm_ggml_backend_dev_get(i);
+            switch (lm_ggml_backend_dev_type(dev)) {
+                case LM_GGML_BACKEND_DEVICE_TYPE_CPU:
+                case LM_GGML_BACKEND_DEVICE_TYPE_ACCEL:
+                    // skip CPU backends since they are handled separately
+                    break;
+
+                case LM_GGML_BACKEND_DEVICE_TYPE_GPU:
+                    lm_ggml_backend_reg_t reg = lm_ggml_backend_dev_backend_reg(dev);
+                    if (lm_ggml_backend_reg_name(reg) == std::string("RPC")) {
+                        rpc_servers.push_back(dev);
+                    } else {
+                        model->devices.push_back(dev);
+                    }
+                    break;
+            }
+        }
+        // add RPC servers at the front of the list
+        if (!rpc_servers.empty()) {
+            model->devices.insert(model->devices.begin(), rpc_servers.begin(), rpc_servers.end());
+        }
+    }
+
+    // if using single GPU mode, remove all except the main GPU
+    if (params.split_mode == LLAMA_SPLIT_MODE_NONE) {
+        if (params.main_gpu < 0) {
+            model->devices.clear();
+        } else {
+            if (params.main_gpu >= (int)model->devices.size()) {
+                LLAMA_LOG_ERROR("%s: invalid value for main_gpu: %d (available devices: %zu)\n", __func__, params.main_gpu, model->devices.size());
+                llama_model_free(model);
+                return nullptr;
+            }
+            lm_ggml_backend_dev_t main_gpu = model->devices[params.main_gpu];
+            model->devices.clear();
+            model->devices.push_back(main_gpu);
+        }
+    }
+
+    for (auto * dev : model->devices) {
+        size_t free, total; // NOLINT
+        lm_ggml_backend_dev_memory(dev, &free, &total);
+        LLAMA_LOG_INFO("%s: using device %s (%s) - %zu MiB free\n", __func__, lm_ggml_backend_dev_name(dev), lm_ggml_backend_dev_description(dev), free/1024/1024);
+    }
+
+    const int status = llama_model_load(path_model, splits, *model, params);
+    LM_GGML_ASSERT(status <= 0);
+    if (status < 0) {
+        if (status == -1) {
+            LLAMA_LOG_ERROR("%s: failed to load model\n", __func__);
+        } else if (status == -2) {
+            LLAMA_LOG_INFO("%s: cancelled model load\n", __func__);
+        }
+
+        llama_model_free(model);
+        return nullptr;
+    }
+
+    return model;
+}
+
+// deprecated
+struct llama_model * llama_load_model_from_file(
+        const char * path_model,
+        struct llama_model_params params) {
+    return llama_model_load_from_file(path_model, params);
+}
+
+struct llama_model * llama_model_load_from_file(
+        const char * path_model,
+        struct llama_model_params params) {
+    std::vector<std::string> splits = {};
+    return llama_model_load_from_file_impl(path_model, splits, params);
+}
+
+struct llama_model * llama_model_load_from_splits(
+        const char ** paths,
+        size_t n_paths,
+        struct llama_model_params params) {
+    std::vector<std::string> splits;
+    if (n_paths == 0) {
+        LLAMA_LOG_ERROR("%s: list of splits is empty\n", __func__);
+        return nullptr;
+    }
+    for (size_t i = 0; i < n_paths; ++i) {
+        splits.push_back(paths[i]);
+    }
+    return llama_model_load_from_file_impl(splits.front(), splits, params);
+}
+
+void llama_model_save_to_file(const struct llama_model * model, const char * path_model) {
+    llama_model_saver ms(*model);
+    ms.add_kv_from_model();
+    ms.add_tensors_from_model();
+    ms.save(path_model);
+}
+
+//
+// chat templates
+//
+
+int32_t llama_chat_apply_template(
+                              const char * tmpl,
+         const struct llama_chat_message * chat,
+                                  size_t   n_msg,
+                                    bool   add_ass,
+                                    char * buf,
+                                 int32_t   length) {
+    const std::string curr_tmpl(tmpl == nullptr ? "chatml" : tmpl);
+
+    // format the chat to string
+    std::vector<const llama_chat_message *> chat_vec;
+    chat_vec.resize(n_msg);
+    for (size_t i = 0; i < n_msg; i++) {
+        chat_vec[i] = &chat[i];
+    }
+
+    std::string formatted_chat;
+    llm_chat_template detected_tmpl = llm_chat_detect_template(curr_tmpl);
+    if (detected_tmpl == LLM_CHAT_TEMPLATE_UNKNOWN) {
+        return -1;
+    }
+    int32_t res = llm_chat_apply_template(detected_tmpl, chat_vec, formatted_chat, add_ass);
+    if (res < 0) {
+        return res;
+    }
+    if (buf && length > 0) {
+        strncpy(buf, formatted_chat.c_str(), length);
+    }
+    return res;
+}
+
+//
+// model split
+//
+
+int llama_split_path(char * split_path, size_t maxlen, const char * path_prefix, int split_no, int split_count) {
+    static const char * const SPLIT_PATH_FORMAT = "%s-%05d-of-%05d.gguf";
+    if (snprintf(split_path, maxlen, SPLIT_PATH_FORMAT, path_prefix, split_no + 1, split_count)) {
+        return strlen(split_path);
+    }
+    return 0;
+}
+
+int llama_split_prefix(char * split_prefix, size_t maxlen, const char * split_path, int split_no, int split_count) {
+    std::string str_split_path(split_path);
+    char postfix[32];
+    snprintf(postfix, 32, "-%05d-of-%05d.gguf", split_no + 1, split_count);
+    std::string str_postfix(postfix);
+
+    // check if split_prefix ends with postfix
+    int size_prefix = str_split_path.size() - str_postfix.size();
+    if (size_prefix > 0 && str_split_path.find(str_postfix, size_prefix) != std::string::npos) {
+        snprintf(split_prefix, std::min((size_t) size_prefix + 1, maxlen), "%s", split_path);
+        return size_prefix;
+    }
+
+    return 0;
+}
+
+const char * llama_print_system_info(void) {
+    static std::string s;
+    s.clear(); // Clear the string, since it's static, otherwise it will accumulate data from previous calls.
+
+    for (size_t i = 0; i < lm_ggml_backend_reg_count(); i++) {
+        auto * reg = lm_ggml_backend_reg_get(i);
+        auto * get_features_fn = (lm_ggml_backend_get_features_t) lm_ggml_backend_reg_get_proc_address(reg, "lm_ggml_backend_get_features");
+        if (get_features_fn) {
+            lm_ggml_backend_feature * features = get_features_fn(reg);
+            s += lm_ggml_backend_reg_name(reg);
+            s += " : ";
+            for (; features->name; features++) {
+                s += features->name;
+                s += " = ";
+                s += features->value;
+                s += " | ";
+            }
+        }
+    }
+
+    return s.c_str();
+}
+