EngineX-Iluvatar/enginex-mr_series-sherpa-onnx
Archived
16
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Add more C API examples (#1255)

Browse Source 
C API examples for zipformer, paraformer, and TeleSpeech-ASR CTC models.
This commit is contained in:
Han Zhu
2024-08-14 10:52:47 +08:00
committed by GitHub
parent 9f06b059d7
commit f300ec0f98
8 changed files with 689 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

160
.github/workflows/c-api.yaml vendored
View File

@@ -163,3 +163,163 @@ jobs:
./whisper-c-api
rm -rf sherpa-onnx-whisper-*
- name: Test non-streaming zipformer
shell: bash
run: |
gcc -o zipformer-c-api ./c-api-examples/zipformer-c-api.c \
-I ./build/install/include \
-L ./build/install/lib/ \
-l sherpa-onnx-c-api \
-l onnxruntime
ls -lh zipformer-c-api
if [[ ${{ matrix.os }} == ubuntu-latest ]]; then
ldd ./zipformer-c-api
echo "----"
readelf -d ./zipformer-c-api
fi
curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-zipformer-small-en-2023-06-26.tar.bz2
tar xvf sherpa-onnx-zipformer-small-en-2023-06-26.tar.bz2
rm sherpa-onnx-zipformer-small-en-2023-06-26.tar.bz2
ls -lh sherpa-onnx-zipformer-small-en-2023-06-26
echo "---"
ls -lh sherpa-onnx-zipformer-small-en-2023-06-26/test_wavs
export LD_LIBRARY_PATH=$PWD/build/install/lib:$LD_LIBRARY_PATH
export DYLD_LIBRARY_PATH=$PWD/build/install/lib:$DYLD_LIBRARY_PATH
./zipformer-c-api
rm -rf sherpa-onnx-zipformer-*
- name: Test streaming zipformer
shell: bash
run: |
gcc -o streaming-zipformer-c-api ./c-api-examples/streaming-zipformer-c-api.c \
-I ./build/install/include \
-L ./build/install/lib/ \
-l sherpa-onnx-c-api \
-l onnxruntime
ls -lh streaming-zipformer-c-api
if [[ ${{ matrix.os }} == ubuntu-latest ]]; then
ldd ./streaming-zipformer-c-api
echo "----"
readelf -d ./streaming-zipformer-c-api
fi
curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2
tar xvf sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2
rm sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2
ls -lh sherpa-onnx-streaming-zipformer-en-20M-2023-02-17
echo "---"
ls -lh sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/test_wavs
export LD_LIBRARY_PATH=$PWD/build/install/lib:$LD_LIBRARY_PATH
export DYLD_LIBRARY_PATH=$PWD/build/install/lib:$DYLD_LIBRARY_PATH
./streaming-zipformer-c-api
rm -rf sherpa-onnx-streaming-zipformer-*
- name: Test non-streaming paraformer
shell: bash
run: |
gcc -o paraformer-c-api ./c-api-examples/paraformer-c-api.c \
-I ./build/install/include \
-L ./build/install/lib/ \
-l sherpa-onnx-c-api \
-l onnxruntime
ls -lh paraformer-c-api
if [[ ${{ matrix.os }} == ubuntu-latest ]]; then
ldd ./paraformer-c-api
echo "----"
readelf -d ./paraformer-c-api
fi
curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-paraformer-zh-small-2024-03-09.tar.bz2
tar xvf sherpa-onnx-paraformer-zh-small-2024-03-09.tar.bz2
rm sherpa-onnx-paraformer-zh-small-2024-03-09.tar.bz2
ls -lh sherpa-onnx-paraformer-zh-small-2024-03-09
echo "---"
ls -lh sherpa-onnx-paraformer-zh-small-2024-03-09/test_wavs
export LD_LIBRARY_PATH=$PWD/build/install/lib:$LD_LIBRARY_PATH
export DYLD_LIBRARY_PATH=$PWD/build/install/lib:$DYLD_LIBRARY_PATH
./paraformer-c-api
rm -rf sherpa-onnx-paraformer-*
- name: Test streaming paraformer
shell: bash
run: |
gcc -o streaming-paraformer-c-api ./c-api-examples/streaming-paraformer-c-api.c \
-I ./build/install/include \
-L ./build/install/lib/ \
-l sherpa-onnx-c-api \
-l onnxruntime
ls -lh streaming-paraformer-c-api
if [[ ${{ matrix.os }} == ubuntu-latest ]]; then
ldd ./streaming-paraformer-c-api
echo "----"
readelf -d ./streaming-paraformer-c-api
fi
curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-paraformer-bilingual-zh-en.tar.bz2
tar xvf sherpa-onnx-streaming-paraformer-bilingual-zh-en.tar.bz2
rm sherpa-onnx-streaming-paraformer-bilingual-zh-en.tar.bz2
ls -lh sherpa-onnx-streaming-paraformer-bilingual-zh-en
echo "---"
ls -lh sherpa-onnx-streaming-paraformer-bilingual-zh-en/test_wavs
export LD_LIBRARY_PATH=$PWD/build/install/lib:$LD_LIBRARY_PATH
export DYLD_LIBRARY_PATH=$PWD/build/install/lib:$DYLD_LIBRARY_PATH
./streaming-paraformer-c-api
rm -rf sherpa-onnx-streaming-paraformer-*
- name: Test telespeech
shell: bash
run: |
gcc -o telespeech-c-api ./c-api-examples/telespeech-c-api.c \
-I ./build/install/include \
-L ./build/install/lib/ \
-l sherpa-onnx-c-api \
-l onnxruntime
ls -lh telespeech-c-api
if [[ ${{ matrix.os }} == ubuntu-latest ]]; then
ldd ./telespeech-c-api
echo "----"
readelf -d ./telespeech-c-api
fi
curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04.tar.bz2
tar xvf sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04.tar.bz2
rm sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04.tar.bz2
ls -lh sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04
echo "---"
ls -lh sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04/test_wavs
export LD_LIBRARY_PATH=$PWD/build/install/lib:$LD_LIBRARY_PATH
export DYLD_LIBRARY_PATH=$PWD/build/install/lib:$DYLD_LIBRARY_PATH
./telespeech-c-api
rm -rf sherpa-onnx-telespeech-*

15
c-api-examples/CMakeLists.txt
View File

@@ -30,6 +30,21 @@ target_link_libraries(whisper-c-api sherpa-onnx-c-api)
add_executable(sense-voice-c-api sense-voice-c-api.c)
target_link_libraries(sense-voice-c-api sherpa-onnx-c-api)
add_executable(zipformer-c-api zipformer-c-api.c)
target_link_libraries(zipformer-c-api sherpa-onnx-c-api)
add_executable(streaming-zipformer-c-api streaming-zipformer-c-api.c)
target_link_libraries(streaming-zipformer-c-api sherpa-onnx-c-api)
add_executable(paraformer-c-api paraformer-c-api.c)
target_link_libraries(paraformer-c-api sherpa-onnx-c-api)
add_executable(streaming-paraformer-c-api streaming-paraformer-c-api.c)
target_link_libraries(streaming-paraformer-c-api sherpa-onnx-c-api)
add_executable(telespeech-c-api telespeech-c-api.c)
target_link_libraries(telespeech-c-api sherpa-onnx-c-api)
if(SHERPA_ONNX_HAS_ALSA)
add_subdirectory(./asr-microphone-example)
elseif((UNIX AND NOT APPLE) OR LINUX)

80
c-api-examples/paraformer-c-api.c Normal file
View File

@@ -0,0 +1,80 @@
// c-api-examples/paraformer-c-api.c
//
// Copyright (c) 2024 Xiaomi Corporation
//
// This file demonstrates how to use non-streaming Paraformer with sherpa-onnx's C API.
// clang-format off
//
// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-paraformer-zh-small-2024-03-09.tar.bz2
// tar xvf sherpa-onnx-paraformer-zh-small-2024-03-09.tar.bz2
// rm sherpa-onnx-paraformer-zh-small-2024-03-09.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-paraformer-zh-small-2024-03-09/test_wavs/0.wav";
const char *model_filename = "sherpa-onnx-paraformer-zh-small-2024-03-09/model.int8.onnx";
const char *tokens_filename = "sherpa-onnx-paraformer-zh-small-2024-03-09/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;
}
// Paraformer config
SherpaOnnxOfflineParaformerModelConfig paraformer_config;
memset(&paraformer_config, 0, sizeof(paraformer_config));
paraformer_config.model = model_filename;
// Offline model config
SherpaOnnxOfflineModelConfig offline_model_config;
memset(&offline_model_config, 0, sizeof(offline_model_config));
offline_model_config.debug = 1;
offline_model_config.num_threads = 1;
offline_model_config.provider = provider;
offline_model_config.tokens = tokens_filename;
offline_model_config.paraformer = paraformer_config;
// Recognizer config
SherpaOnnxOfflineRecognizerConfig recognizer_config;
memset(&recognizer_config, 0, sizeof(recognizer_config));
recognizer_config.decoding_method = "greedy_search";
recognizer_config.model_config = offline_model_config;
SherpaOnnxOfflineRecognizer *recognizer =
SherpaOnnxCreateOfflineRecognizer(&recognizer_config);
if (recognizer == NULL) {
fprintf(stderr, "Please check your config!\n");
SherpaOnnxFreeWave(wave);
return -1;
}
SherpaOnnxOfflineStream *stream = SherpaOnnxCreateOfflineStream(recognizer);
SherpaOnnxAcceptWaveformOffline(stream, wave->sample_rate, wave->samples,
wave->num_samples);
SherpaOnnxDecodeOfflineStream(recognizer, stream);
const SherpaOnnxOfflineRecognizerResult *result =
SherpaOnnxGetOfflineStreamResult(stream);
fprintf(stderr, "Decoded text: %s\n", result->text);
SherpaOnnxDestroyOfflineRecognizerResult(result);
SherpaOnnxDestroyOfflineStream(stream);
SherpaOnnxDestroyOfflineRecognizer(recognizer);
SherpaOnnxFreeWave(wave);
return 0;
}

136
c-api-examples/streaming-paraformer-c-api.c Normal file
View File

@@ -0,0 +1,136 @@
// c-api-examples/streaming-paraformer-c-api.c
//
// Copyright (c) 2024 Xiaomi Corporation
//
// This file demonstrates how to use streaming Paraformer with sherpa-onnx's C API.
// clang-format off
//
// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-paraformer-bilingual-zh-en.tar.bz2
// tar xvf sherpa-onnx-streaming-paraformer-bilingual-zh-en.tar.bz2
// rm sherpa-onnx-streaming-paraformer-bilingual-zh-en.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-paraformer-bilingual-zh-en/test_wavs/0.wav";
const char *encoder_filename = "sherpa-onnx-streaming-paraformer-bilingual-zh-en/encoder.int8.onnx";
const char *decoder_filename = "sherpa-onnx-streaming-paraformer-bilingual-zh-en/decoder.int8.onnx";
const char *tokens_filename = "sherpa-onnx-streaming-paraformer-bilingual-zh-en/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;
}
// Paraformer config
SherpaOnnxOnlineParaformerModelConfig paraformer_config;
memset(&paraformer_config, 0, sizeof(paraformer_config));
paraformer_config.encoder = encoder_filename;
paraformer_config.decoder = decoder_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.paraformer = paraformer_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;
SherpaOnnxOnlineRecognizer *recognizer =
SherpaOnnxCreateOnlineRecognizer(&recognizer_config);
if (recognizer == NULL) {
fprintf(stderr, "Please check your config!\n");
SherpaOnnxFreeWave(wave);
return -1;
}
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;
}

138
c-api-examples/streaming-zipformer-c-api.c Normal file
View File

@@ -0,0 +1,138 @@
// 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;
SherpaOnnxOnlineRecognizer *recognizer =
SherpaOnnxCreateOnlineRecognizer(&recognizer_config);
if (recognizer == NULL) {
fprintf(stderr, "Please check your config!\n");
SherpaOnnxFreeWave(wave);
return -1;
}
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;
}

74
c-api-examples/telespeech-c-api.c Normal file
View File

@@ -0,0 +1,74 @@
// c-api-examples/telespeech-c-api.c
//
// Copyright (c) 2024 Xiaomi Corporation
//
// This file demonstrates how to use TeleSpeech-ASR CTC model with sherpa-onnx's C API.
// clang-format off
//
// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04.tar.bz2
// tar xvf sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04.tar.bz2
// rm sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04.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-telespeech-ctc-int8-zh-2024-06-04/test_wavs/3-sichuan.wav";
const char *model_filename = "sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04/model.int8.onnx";
const char *tokens_filename = "sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04/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;
}
// Offline model config
SherpaOnnxOfflineModelConfig offline_model_config;
memset(&offline_model_config, 0, sizeof(offline_model_config));
offline_model_config.debug = 1;
offline_model_config.num_threads = 1;
offline_model_config.provider = provider;
offline_model_config.tokens = tokens_filename;
offline_model_config.telespeech_ctc = model_filename;
// Recognizer config
SherpaOnnxOfflineRecognizerConfig recognizer_config;
memset(&recognizer_config, 0, sizeof(recognizer_config));
recognizer_config.decoding_method = "greedy_search";
recognizer_config.model_config = offline_model_config;
SherpaOnnxOfflineRecognizer *recognizer =
SherpaOnnxCreateOfflineRecognizer(&recognizer_config);
if (recognizer == NULL) {
fprintf(stderr, "Please check your config!\n");
SherpaOnnxFreeWave(wave);
return -1;
}
SherpaOnnxOfflineStream *stream = SherpaOnnxCreateOfflineStream(recognizer);
SherpaOnnxAcceptWaveformOffline(stream, wave->sample_rate, wave->samples,
wave->num_samples);
SherpaOnnxDecodeOfflineStream(recognizer, stream);
const SherpaOnnxOfflineRecognizerResult *result =
SherpaOnnxGetOfflineStreamResult(stream);
fprintf(stderr, "Decoded text: %s\n", result->text);
SherpaOnnxDestroyOfflineRecognizerResult(result);
SherpaOnnxDestroyOfflineStream(stream);
SherpaOnnxDestroyOfflineRecognizer(recognizer);
SherpaOnnxFreeWave(wave);
return 0;
}

2
c-api-examples/whisper-c-api.c
View File

@@ -1,4 +1,4 @@
// c-api-examples/offline-stt-c-api.c
// c-api-examples/whisper-c-api.c
//
// Copyright (c) 2024 Xiaomi Corporation

85
c-api-examples/zipformer-c-api.c Normal file
View File

@@ -0,0 +1,85 @@
// c-api-examples/zipformer-c-api.c
//
// Copyright (c) 2024 Xiaomi Corporation
//
// This file demonstrates how to use non-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-zipformer-small-en-2023-06-26.tar.bz2
// tar xvf sherpa-onnx-zipformer-small-en-2023-06-26.tar.bz2
// rm sherpa-onnx-zipformer-small-en-2023-06-26.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-zipformer-small-en-2023-06-26/test_wavs/0.wav";
const char *encoder_filename = "sherpa-onnx-zipformer-small-en-2023-06-26/encoder-epoch-99-avg-1.onnx";
const char *decoder_filename = "sherpa-onnx-zipformer-small-en-2023-06-26/decoder-epoch-99-avg-1.onnx";
const char *joiner_filename = "sherpa-onnx-zipformer-small-en-2023-06-26/joiner-epoch-99-avg-1.onnx";
const char *tokens_filename = "sherpa-onnx-zipformer-small-en-2023-06-26/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
SherpaOnnxOfflineTransducerModelConfig zipformer_config;
memset(&zipformer_config, 0, sizeof(zipformer_config));
zipformer_config.encoder = encoder_filename;
zipformer_config.decoder = decoder_filename;
zipformer_config.joiner = joiner_filename;
// Offline model config
SherpaOnnxOfflineModelConfig offline_model_config;
memset(&offline_model_config, 0, sizeof(offline_model_config));
offline_model_config.debug = 1;
offline_model_config.num_threads = 1;
offline_model_config.provider = provider;
offline_model_config.tokens = tokens_filename;
offline_model_config.transducer = zipformer_config;
// Recognizer config
SherpaOnnxOfflineRecognizerConfig recognizer_config;
memset(&recognizer_config, 0, sizeof(recognizer_config));
recognizer_config.decoding_method = "greedy_search";
recognizer_config.model_config = offline_model_config;
SherpaOnnxOfflineRecognizer *recognizer =
SherpaOnnxCreateOfflineRecognizer(&recognizer_config);
if (recognizer == NULL) {
fprintf(stderr, "Please check your config!\n");
SherpaOnnxFreeWave(wave);
return -1;
}
SherpaOnnxOfflineStream *stream = SherpaOnnxCreateOfflineStream(recognizer);
SherpaOnnxAcceptWaveformOffline(stream, wave->sample_rate, wave->samples,
wave->num_samples);
SherpaOnnxDecodeOfflineStream(recognizer, stream);
const SherpaOnnxOfflineRecognizerResult *result =
SherpaOnnxGetOfflineStreamResult(stream);
fprintf(stderr, "Decoded text: %s\n", result->text);
SherpaOnnxDestroyOfflineRecognizerResult(result);
SherpaOnnxDestroyOfflineStream(stream);
SherpaOnnxDestroyOfflineRecognizer(recognizer);
SherpaOnnxFreeWave(wave);
return 0;
}