This repository has been archived on 2025-08-26. You can view files and clone it, but cannot push or open issues or pull requests.
Files
enginex_bi_series-sherpa-onnx/c-api-examples/streaming-zipformer-c-api.c
2024-10-23 16:40:12 +08:00

146 lines
4.6 KiB
C

// c-api-examples/streaming-zipformer-c-api.c
//
// Copyright (c) 2024 Xiaomi Corporation
//
// This file demonstrates how to use streaming Zipformer with sherpa-onnx's C
// API.
// clang-format off
//
// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2
// tar xvf sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2
// rm sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2
//
// clang-format on
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "sherpa-onnx/c-api/c-api.h"
int32_t main() {
const char *wav_filename =
"sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/test_wavs/0.wav";
const char *encoder_filename =
"sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/"
"encoder-epoch-99-avg-1.onnx";
const char *decoder_filename =
"sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/"
"decoder-epoch-99-avg-1.onnx";
const char *joiner_filename =
"sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/"
"joiner-epoch-99-avg-1.onnx";
const char *tokens_filename =
"sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/tokens.txt";
const char *provider = "cpu";
const SherpaOnnxWave *wave = SherpaOnnxReadWave(wav_filename);
if (wave == NULL) {
fprintf(stderr, "Failed to read %s\n", wav_filename);
return -1;
}
// Zipformer config
SherpaOnnxOnlineTransducerModelConfig zipformer_config;
memset(&zipformer_config, 0, sizeof(zipformer_config));
zipformer_config.encoder = encoder_filename;
zipformer_config.decoder = decoder_filename;
zipformer_config.joiner = joiner_filename;
// Online model config
SherpaOnnxOnlineModelConfig online_model_config;
memset(&online_model_config, 0, sizeof(online_model_config));
online_model_config.debug = 1;
online_model_config.num_threads = 1;
online_model_config.provider = provider;
online_model_config.tokens = tokens_filename;
online_model_config.transducer = zipformer_config;
// Recognizer config
SherpaOnnxOnlineRecognizerConfig recognizer_config;
memset(&recognizer_config, 0, sizeof(recognizer_config));
recognizer_config.decoding_method = "greedy_search";
recognizer_config.model_config = online_model_config;
const SherpaOnnxOnlineRecognizer *recognizer =
SherpaOnnxCreateOnlineRecognizer(&recognizer_config);
if (recognizer == NULL) {
fprintf(stderr, "Please check your config!\n");
SherpaOnnxFreeWave(wave);
return -1;
}
const SherpaOnnxOnlineStream *stream =
SherpaOnnxCreateOnlineStream(recognizer);
const SherpaOnnxDisplay *display = SherpaOnnxCreateDisplay(50);
int32_t segment_id = 0;
// simulate streaming. You can choose an arbitrary N
#define N 3200
fprintf(stderr, "sample rate: %d, num samples: %d, duration: %.2f s\n",
wave->sample_rate, wave->num_samples,
(float)wave->num_samples / wave->sample_rate);
int32_t k = 0;
while (k < wave->num_samples) {
int32_t start = k;
int32_t end =
(start + N > wave->num_samples) ? wave->num_samples : (start + N);
k += N;
SherpaOnnxOnlineStreamAcceptWaveform(stream, wave->sample_rate,
wave->samples + start, end - start);
while (SherpaOnnxIsOnlineStreamReady(recognizer, stream)) {
SherpaOnnxDecodeOnlineStream(recognizer, stream);
}
const SherpaOnnxOnlineRecognizerResult *r =
SherpaOnnxGetOnlineStreamResult(recognizer, stream);
if (strlen(r->text)) {
SherpaOnnxPrint(display, segment_id, r->text);
}
if (SherpaOnnxOnlineStreamIsEndpoint(recognizer, stream)) {
if (strlen(r->text)) {
++segment_id;
}
SherpaOnnxOnlineStreamReset(recognizer, stream);
}
SherpaOnnxDestroyOnlineRecognizerResult(r);
}
// add some tail padding
float tail_paddings[4800] = {0}; // 0.3 seconds at 16 kHz sample rate
SherpaOnnxOnlineStreamAcceptWaveform(stream, wave->sample_rate, tail_paddings,
4800);
SherpaOnnxFreeWave(wave);
SherpaOnnxOnlineStreamInputFinished(stream);
while (SherpaOnnxIsOnlineStreamReady(recognizer, stream)) {
SherpaOnnxDecodeOnlineStream(recognizer, stream);
}
const SherpaOnnxOnlineRecognizerResult *r =
SherpaOnnxGetOnlineStreamResult(recognizer, stream);
if (strlen(r->text)) {
SherpaOnnxPrint(display, segment_id, r->text);
}
SherpaOnnxDestroyOnlineRecognizerResult(r);
SherpaOnnxDestroyDisplay(display);
SherpaOnnxDestroyOnlineStream(stream);
SherpaOnnxDestroyOnlineRecognizer(recognizer);
fprintf(stderr, "\n");
return 0;
}