File size: 2,485 Bytes
0f07ba7 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 |
#include <iostream>
#include <tuple>
#include "bark.h"
#include "gobark.h"
#include "common.h"
#include "ggml.h"
struct bark_context *c;
void bark_print_progress_callback(struct bark_context *bctx, enum bark_encoding_step step, int progress, void *user_data) {
if (step == bark_encoding_step::SEMANTIC) {
printf("\rGenerating semantic tokens... %d%%", progress);
} else if (step == bark_encoding_step::COARSE) {
printf("\rGenerating coarse tokens... %d%%", progress);
} else if (step == bark_encoding_step::FINE) {
printf("\rGenerating fine tokens... %d%%", progress);
}
fflush(stdout);
}
int load_model(char *model) {
// initialize bark context
struct bark_context_params ctx_params = bark_context_default_params();
bark_params params;
params.model_path = model;
// ctx_params.verbosity = verbosity;
ctx_params.progress_callback = bark_print_progress_callback;
ctx_params.progress_callback_user_data = nullptr;
struct bark_context *bctx = bark_load_model(params.model_path.c_str(), ctx_params, params.seed);
if (!bctx) {
fprintf(stderr, "%s: Could not load model\n", __func__);
return 1;
}
c = bctx;
return 0;
}
int tts(char *text,int threads, char *dst ) {
ggml_time_init();
const int64_t t_main_start_us = ggml_time_us();
// generate audio
if (!bark_generate_audio(c, text, threads)) {
fprintf(stderr, "%s: An error occurred. If the problem persists, feel free to open an issue to report it.\n", __func__);
return 1;
}
const float *audio_data = bark_get_audio_data(c);
if (audio_data == NULL) {
fprintf(stderr, "%s: Could not get audio data\n", __func__);
return 1;
}
const int audio_arr_size = bark_get_audio_data_size(c);
std::vector<float> audio_arr(audio_data, audio_data + audio_arr_size);
write_wav_on_disk(audio_arr, dst);
// report timing
{
const int64_t t_main_end_us = ggml_time_us();
const int64_t t_load_us = bark_get_load_time(c);
const int64_t t_eval_us = bark_get_eval_time(c);
printf("\n\n");
printf("%s: load time = %8.2f ms\n", __func__, t_load_us / 1000.0f);
printf("%s: eval time = %8.2f ms\n", __func__, t_eval_us / 1000.0f);
printf("%s: total time = %8.2f ms\n", __func__, (t_main_end_us - t_main_start_us) / 1000.0f);
}
return 0;
}
int unload() {
bark_free(c);
}
|