Function main

subprojects/llama.cpp/tests/test-tokenizer-1-spm.cpp:17–125

#include <atomic>

int main(int argc, char ** argv) {
    if (argc < 2) {
        fprintf(stderr, "Usage: %s <vocab-file>\n", argv[0]);
        return 1;
    }

    const std::string fname = argv[1];

    fprintf(stderr, "%s : reading vocab from: '%s'\n", __func__, fname.c_str());

    llama_model * model;
    llama_context * ctx;

    llama_backend_init();

    // load the vocab
    {
        auto mparams = llama_model_default_params();

        mparams.vocab_only = true;

        model = llama_model_load_from_file(fname.c_str(), mparams);

        if (model == NULL) {
            fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
            return 1;
        }

        auto cparams = llama_context_default_params();

        ctx = llama_init_from_model(model, cparams);

        if (ctx == NULL) {
            fprintf(stderr, "%s: error: failed to load vocab '%s'\n", __func__, fname.c_str());
            llama_model_free(model);
            return 1;
        }
    }

    const llama_vocab * vocab = llama_model_get_vocab(model);

    //GGML_ASSERT(llama_vocab_type(model) == LLAMA_VOCAB_TYPE_SPM);
    if (llama_vocab_type(vocab) != LLAMA_VOCAB_TYPE_SPM) {
        return 99;
    }

#ifdef _WIN32
    // We need this for unicode console support
    console::init(false, false);
    atexit([]() { console::cleanup(); });
#endif

    const int n_vocab = llama_vocab_n_tokens(vocab);

    for (int i = 0; i < n_vocab; ++i) {
        std::string str = common_detokenize(ctx, std::vector<int>(1, i), true);
        std::vector<llama_token> tokens = common_tokenize(ctx, str, false, true);
        std::string check = common_detokenize(ctx, tokens);