mirror of
https://github.com/k2-fsa/sherpa-onnx.git
synced 2026-11-10 21:11:03 +08:00
Fangjun Kuang
0b3647632f
Fixes #3731 espeak-ng and piper-phonemize are no longer built into sherpa-onnx. Text frontends are now: - **lexicon** — built in, no extra dependency - **phonemize** — phonemize outside sherpa-onnx (e.g. [piper-phonemize](https://github.com/csukuangfj/piper-phonemize)) and pass the result in. That library is GPL, so it stays optional: sherpa-onnx is GPL-free after this PR. ## API - `SherpaOnnxGenerationConfig` gains `phoneme_codepoints` and `tokens` (mutually exclusive alternatives to `text`) - `data_dir` on the TTS model configs is deprecated and ignored; the ABI slot is kept - all bindings updated: C/C++, Python, Java, Kotlin, Go, Rust, C#, Swift, Pascal, Dart/Flutter, node-addon, WASM, HarmonyOS
62 lines
2.1 KiB
Java
62 lines
2.1 KiB
Java
// Copyright 2026 Xiaomi Corporation
|
|
|
|
// This file shows how to use a non-streaming Zipformer CTC model with an HLG
|
|
// decoding graph via sherpa-onnx's Java API.
|
|
//
|
|
// The HLG graph is passed to the C API through CtcFstDecoderConfig.
|
|
import com.k2fsa.sherpa.onnx.*;
|
|
|
|
public class NonStreamingZipformerCtcHlg {
|
|
public static void main(String[] args) {
|
|
// please refer to
|
|
// https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-zipformer-ctc-en-2023-10-02.tar.bz2
|
|
// to download model files
|
|
String dir = "./sherpa-onnx-zipformer-ctc-en-2023-10-02";
|
|
String model = dir + "/model.int8.onnx";
|
|
String tokens = dir + "/tokens.txt";
|
|
String graph = dir + "/HLG.fst";
|
|
String waveFilename = dir + "/test_wavs/0.wav";
|
|
|
|
WaveReader reader = new WaveReader(waveFilename);
|
|
|
|
OfflineZipformerCtcModelConfig zipformerCtc =
|
|
OfflineZipformerCtcModelConfig.builder().setModel(model).build();
|
|
|
|
OfflineModelConfig modelConfig =
|
|
OfflineModelConfig.builder()
|
|
.setZipformerCtc(zipformerCtc)
|
|
.setTokens(tokens)
|
|
.setNumThreads(1)
|
|
.setDebug(true)
|
|
.build();
|
|
|
|
// This is what the example exercises: the graph and max_active reach the C
|
|
// API through OfflineCtcFstDecoderConfig.
|
|
OfflineCtcFstDecoderConfig ctcFstDecoderConfig =
|
|
OfflineCtcFstDecoderConfig.builder()
|
|
.setGraph(graph)
|
|
.setMaxActive(3000)
|
|
.build();
|
|
|
|
OfflineRecognizerConfig config =
|
|
OfflineRecognizerConfig.builder()
|
|
.setOfflineModelConfig(modelConfig)
|
|
.setDecodingMethod("greedy_search")
|
|
.setCtcFstDecoderConfig(ctcFstDecoderConfig)
|
|
.build();
|
|
|
|
OfflineRecognizer recognizer = new OfflineRecognizer(config);
|
|
OfflineStream stream = recognizer.createStream();
|
|
stream.acceptWaveform(reader.getSamples(), reader.getSampleRate());
|
|
|
|
recognizer.decode(stream);
|
|
|
|
String text = recognizer.getResult(stream).getText();
|
|
|
|
System.out.printf("filename:%s\nresult:%s\n", waveFilename, text);
|
|
|
|
stream.release();
|
|
recognizer.release();
|
|
}
|
|
}
|