Add Kotlin and Java API for Moonshine models (#1474)

.github/workflows/apk-vad-asr.yaml

@@ -23,8 +23,8 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-latest]
-        total: ["5"]
-        index: ["0", "1", "2", "3", "4"]
+        total: ["10"]
+        index: ["0", "1", "2", "3", "4", "5", "6", "7", "8", "9"]
 
     steps:
       - uses: actions/checkout@v4
@@ -165,6 +165,7 @@ jobs:
 
             git clone https://csukuangfj:$HF_TOKEN@huggingface.co/csukuangfj/sherpa-onnx-apk huggingface
             cd huggingface
+            du -h -d1 .
             git fetch
             git pull
             git merge -m "merge remote" --ff origin main

.github/workflows/run-java-test.yaml

@@ -107,6 +107,31 @@ jobs:
             make -j4
             ls -lh lib
 
+      - name:  Run java test (Non-Streaming ASR)
+        shell: bash
+        run: |
+          cd ./java-api-examples
+
+          ./run-non-streaming-decode-file-moonshine.sh
+          rm -rf sherpa-onnx-moonshine-*
+
+          ./run-non-streaming-decode-file-sense-voice.sh
+          rm -rf sherpa-onnx-sense-voice-*
+
+          ./run-inverse-text-normalization-paraformer.sh
+
+          ./run-non-streaming-decode-file-paraformer.sh
+          rm -rf sherpa-onnx-paraformer-zh-*
+
+          ./run-non-streaming-decode-file-transducer.sh
+          rm -rf sherpa-onnx-zipformer-*
+
+          ./run-non-streaming-decode-file-whisper.sh
+          rm -rf sherpa-onnx-whisper-*
+
+          ./run-non-streaming-decode-file-nemo.sh
+          rm -rf sherpa-onnx-nemo-*
+
       - name:  Run java test (speaker diarization)
         shell: bash
         run: |
@@ -206,28 +231,6 @@ jobs:
           ./run-streaming-decode-file-transducer.sh
           rm -rf sherpa-onnx-streaming-*
 
-      - name:  Run java test (Non-Streaming ASR)
-        shell: bash
-        run: |
-          cd ./java-api-examples
-
-          ./run-non-streaming-decode-file-sense-voice.sh
-          rm -rf sherpa-onnx-sense-voice-*
-
-          ./run-inverse-text-normalization-paraformer.sh
-
-          ./run-non-streaming-decode-file-paraformer.sh
-          rm -rf sherpa-onnx-paraformer-zh-*
-
-          ./run-non-streaming-decode-file-transducer.sh
-          rm -rf sherpa-onnx-zipformer-*
-
-          ./run-non-streaming-decode-file-whisper.sh
-          rm -rf sherpa-onnx-whisper-*
-
-          ./run-non-streaming-decode-file-nemo.sh
-          rm -rf sherpa-onnx-nemo-*
-
       - name:  Run java test (Non-Streaming TTS)
         shell: bash
         run: |

.gitignore

@@ -121,3 +121,5 @@ sherpa-onnx-online-punct-en-2024-08-06
 *.mp4
 *.mp3
 sherpa-onnx-pyannote-segmentation-3-0
+sherpa-onnx-moonshine-tiny-en-int8
+sherpa-onnx-moonshine-base-en-int8

java-api-examples/NonStreamingDecodeFileMoonshine.java

@@ -0,0 +1,60 @@
+// Copyright 2024 Xiaomi Corporation
+
+// This file shows how to use an offline Moonshine,
+// i.e., non-streaming Moonshine model,
+// to decode files.
+import com.k2fsa.sherpa.onnx.*;
+
+public class NonStreamingDecodeFileMoonshine {
+  public static void main(String[] args) {
+    // please refer to
+    // https://k2-fsa.github.io/sherpa/onnx/moonshine/index.html
+    // to download model files
+
+    String preprocessor = "./sherpa-onnx-moonshine-tiny-en-int8/preprocess.onnx";
+    String encoder = "./sherpa-onnx-moonshine-tiny-en-int8/encode.int8.onnx";
+    String uncachedDecoder = "./sherpa-onnx-moonshine-tiny-en-int8/uncached_decode.int8.onnx";
+    String cachedDecoder = "./sherpa-onnx-moonshine-tiny-en-int8/cached_decode.int8.onnx";
+
+    String tokens = "./sherpa-onnx-moonshine-tiny-en-int8/tokens.txt";
+
+    String waveFilename = "./sherpa-onnx-moonshine-tiny-en-int8/test_wavs/0.wav";
+
+    WaveReader reader = new WaveReader(waveFilename);
+
+    OfflineMoonshineModelConfig moonshine =
+        OfflineMoonshineModelConfig.builder()
+            .setPreprocessor(preprocessor)
+            .setEncoder(encoder)
+            .setUncachedDecoder(uncachedDecoder)
+            .setCachedDecoder(cachedDecoder)
+            .build();
+
+    OfflineModelConfig modelConfig =
+        OfflineModelConfig.builder()
+            .setMoonshine(moonshine)
+            .setTokens(tokens)
+            .setNumThreads(1)
+            .setDebug(true)
+            .build();
+
+    OfflineRecognizerConfig config =
+        OfflineRecognizerConfig.builder()
+            .setOfflineModelConfig(modelConfig)
+            .setDecodingMethod("greedy_search")
+            .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();
+  }
+}

java-api-examples/VadFromMicWithNonStreamingMoonshine.java

@@ -0,0 +1,152 @@
+// Copyright 2024 Xiaomi Corporation
+
+// This file shows how to use a silero_vad model with a non-streaming
+// Moonshine tiny for speech recognition.
+
+import com.k2fsa.sherpa.onnx.*;
+import javax.sound.sampled.*;
+
+public class VadFromMicNonStreamingMoonshine {
+  private static final int sampleRate = 16000;
+  private static final int windowSize = 512;
+
+  public static Vad createVad() {
+    // please download ./silero_vad.onnx from
+    // https://github.com/k2-fsa/sherpa-onnx/releases/tag/asr-models
+    String model = "./silero_vad.onnx";
+    SileroVadModelConfig sileroVad =
+        SileroVadModelConfig.builder()
+            .setModel(model)
+            .setThreshold(0.5f)
+            .setMinSilenceDuration(0.25f)
+            .setMinSpeechDuration(0.5f)
+            .setWindowSize(windowSize)
+            .build();
+
+    VadModelConfig config =
+        VadModelConfig.builder()
+            .setSileroVadModelConfig(sileroVad)
+            .setSampleRate(sampleRate)
+            .setNumThreads(1)
+            .setDebug(true)
+            .setProvider("cpu")
+            .build();
+
+    return new Vad(config);
+  }
+
+  public static OfflineRecognizer createOfflineRecognizer() {
+    // please refer to
+    // https://k2-fsa.github.io/sherpa/onnx/moonshine/index.html
+    // to download model files
+
+    String preprocessor = "./sherpa-onnx-moonshine-tiny-en-int8/preprocess.onnx";
+    String encoder = "./sherpa-onnx-moonshine-tiny-en-int8/encode.int8.onnx";
+    String uncachedDecoder = "./sherpa-onnx-moonshine-tiny-en-int8/uncached_decode.int8.onnx";
+    String cachedDecoder = "./sherpa-onnx-moonshine-tiny-en-int8/cached_decode.int8.onnx";
+
+    String tokens = "./sherpa-onnx-moonshine-tiny-en-int8/tokens.txt";
+
+    OfflineMoonshineModelConfig moonshine =
+        OfflineMoonshineModelConfig.builder()
+            .setPreprocessor(preprocessor)
+            .setEncoder(encoder)
+            .setUncachedDecoder(uncachedDecoder)
+            .setCachedDecoder(cachedDecoder)
+            .build();
+
+    OfflineModelConfig modelConfig =
+        OfflineModelConfig.builder()
+            .setMoonshine(moonshine)
+            .setTokens(tokens)
+            .setNumThreads(1)
+            .setDebug(true)
+            .build();
+
+    OfflineRecognizerConfig config =
+        OfflineRecognizerConfig.builder()
+            .setOfflineModelConfig(modelConfig)
+            .setDecodingMethod("greedy_search")
+            .build();
+
+    return new OfflineRecognizer(config);
+  }
+
+  public static void main(String[] args) {
+    Vad vad = createVad();
+    OfflineRecognizer recognizer = createOfflineRecognizer();
+
+    // https://docs.oracle.com/javase/8/docs/api/javax/sound/sampled/AudioFormat.html
+    // Linear PCM, 16000Hz, 16-bit, 1 channel, signed, little endian
+    AudioFormat format = new AudioFormat(sampleRate, 16, 1, true, false);
+
+    // https://docs.oracle.com/javase/8/docs/api/javax/sound/sampled/DataLine.Info.html#Info-java.lang.Class-javax.sound.sampled.AudioFormat-int-
+    DataLine.Info info = new DataLine.Info(TargetDataLine.class, format);
+    TargetDataLine targetDataLine;
+    try {
+      targetDataLine = (TargetDataLine) AudioSystem.getLine(info);
+      targetDataLine.open(format);
+      targetDataLine.start();
+    } catch (LineUnavailableException e) {
+      System.out.println("Failed to open target data line: " + e.getMessage());
+      vad.release();
+      recognizer.release();
+      return;
+    }
+
+    boolean printed = false;
+    byte[] buffer = new byte[windowSize * 2];
+    float[] samples = new float[windowSize];
+
+    System.out.println("Started. Please speak");
+    boolean running = true;
+    while (targetDataLine.isOpen() && running) {
+      int n = targetDataLine.read(buffer, 0, buffer.length);
+      if (n <= 0) {
+        System.out.printf("Got %d bytes. Expected %d bytes.\n", n, buffer.length);
+        continue;
+      }
+      for (int i = 0; i != windowSize; ++i) {
+        short low = buffer[2 * i];
+        short high = buffer[2 * i + 1];
+        int s = (high << 8) + low;
+        samples[i] = (float) s / 32768;
+      }
+
+      vad.acceptWaveform(samples);
+      if (vad.isSpeechDetected() && !printed) {
+        System.out.println("Detected speech");
+        printed = true;
+      }
+
+      if (!vad.isSpeechDetected()) {
+        printed = false;
+      }
+
+      while (!vad.empty()) {
+        SpeechSegment segment = vad.front();
+        float startTime = segment.getStart() / (float) sampleRate;
+        float duration = segment.getSamples().length / (float) sampleRate;
+
+        OfflineStream stream = recognizer.createStream();
+        stream.acceptWaveform(segment.getSamples(), sampleRate);
+        recognizer.decode(stream);
+        String text = recognizer.getResult(stream).getText();
+        stream.release();
+
+        if (!text.isEmpty()) {
+          System.out.printf("%.3f--%.3f: %s\n", startTime, startTime + duration, text);
+        }
+
+        if (text.contains("exit the program")) {
+          running = false;
+        }
+
+        vad.pop();
+      }
+    }
+
+    vad.release();
+    recognizer.release();
+  }
+}

java-api-examples/run-non-streaming-decode-file-moonshine.sh

@@ -0,0 +1,37 @@
+#!/usr/bin/env bash
+
+set -ex
+
+if [[ ! -f ../build/lib/libsherpa-onnx-jni.dylib  && ! -f ../build/lib/libsherpa-onnx-jni.so ]]; then
+  mkdir -p ../build
+  pushd ../build
+  cmake \
+    -DSHERPA_ONNX_ENABLE_PYTHON=OFF \
+    -DSHERPA_ONNX_ENABLE_TESTS=OFF \
+    -DSHERPA_ONNX_ENABLE_CHECK=OFF \
+    -DBUILD_SHARED_LIBS=ON \
+    -DSHERPA_ONNX_ENABLE_PORTAUDIO=OFF \
+    -DSHERPA_ONNX_ENABLE_JNI=ON \
+    ..
+
+  make -j4
+  ls -lh lib
+  popd
+fi
+
+if [ ! -f ../sherpa-onnx/java-api/build/sherpa-onnx.jar ]; then
+  pushd ../sherpa-onnx/java-api
+  make
+  popd
+fi
+
+if [ ! -f ./sherpa-onnx-moonshine-tiny-en-int8/tokens.txt ]; then
+  curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-moonshine-tiny-en-int8.tar.bz2
+  tar xvf sherpa-onnx-moonshine-tiny-en-int8.tar.bz2
+  rm sherpa-onnx-moonshine-tiny-en-int8.tar.bz2
+fi
+
+java \
+  -Djava.library.path=$PWD/../build/lib \
+  -cp ../sherpa-onnx/java-api/build/sherpa-onnx.jar \
+  NonStreamingDecodeFileMoonshine.java

java-api-examples/run-vad-from-mic-non-streaming-moonshine.sh

@@ -0,0 +1,41 @@
+#!/usr/bin/env bash
+
+set -ex
+
+if [[ ! -f ../build/lib/libsherpa-onnx-jni.dylib  && ! -f ../build/lib/libsherpa-onnx-jni.so ]]; then
+  mkdir -p ../build
+  pushd ../build
+  cmake \
+    -DSHERPA_ONNX_ENABLE_PYTHON=OFF \
+    -DSHERPA_ONNX_ENABLE_TESTS=OFF \
+    -DSHERPA_ONNX_ENABLE_CHECK=OFF \
+    -DBUILD_SHARED_LIBS=ON \
+    -DSHERPA_ONNX_ENABLE_PORTAUDIO=OFF \
+    -DSHERPA_ONNX_ENABLE_JNI=ON \
+    ..
+
+  make -j4
+  ls -lh lib
+  popd
+fi
+
+if [ ! -f ../sherpa-onnx/java-api/build/sherpa-onnx.jar ]; then
+  pushd ../sherpa-onnx/java-api
+  make
+  popd
+fi
+
+if [ ! -f ./silero_vad.onnx ]; then
+  curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/silero_vad.onnx
+fi
+
+if [ ! -f ./sherpa-onnx-moonshine-tiny-en-int8/tokens.txt ]; then
+  curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-moonshine-tiny-en-int8.tar.bz2
+  tar xvf sherpa-onnx-moonshine-tiny-en-int8.tar.bz2
+  rm sherpa-onnx-moonshine-tiny-en-int8.tar.bz2
+fi
+
+java \
+  -Djava.library.path=$PWD/../build/lib \
+  -cp ../sherpa-onnx/java-api/build/sherpa-onnx.jar \
+  ./VadFromMicWithNonStreamingMoonshine.java

kotlin-api-examples/run.sh

@@ -168,6 +168,12 @@ function testSpokenLanguageIdentification() {
 }
 
 function testOfflineAsr() {
+  if [ ! -f ./sherpa-onnx-moonshine-tiny-en-int8/tokens.txt ]; then
+    curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-moonshine-tiny-en-int8.tar.bz2
+    tar xvf sherpa-onnx-moonshine-tiny-en-int8.tar.bz2
+    rm sherpa-onnx-moonshine-tiny-en-int8.tar.bz2
+  fi
+
   if [ ! -f ./sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/tokens.txt ]; then
     curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
     tar xvf sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2

kotlin-api-examples/test_offline_asr.kt

@@ -1,7 +1,7 @@
 package com.k2fsa.sherpa.onnx
 
 fun main() {
-  val types = arrayOf(0, 2, 5, 6, 15)
+  val types = arrayOf(0, 2, 5, 6, 15, 21)
   for (type in types) {
     test(type)
   }
@@ -16,6 +16,7 @@ fun test(type: Int) {
     5 -> "./sherpa-onnx-zipformer-multi-zh-hans-2023-9-2/test_wavs/1.wav"
     6 -> "./sherpa-onnx-nemo-ctc-en-citrinet-512/test_wavs/8k.wav"
     15 -> "./sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/test_wavs/zh.wav"
+    21 -> "./sherpa-onnx-moonshine-tiny-en-int8/test_wavs/0.wav"
     else -> null
   }
 

scripts/apk/generate-vad-asr-apk-script.py

@@ -369,6 +369,21 @@ def get_models():
 
             ls -lh
 
+            popd
+            """,
+        ),
+        Model(
+            model_name="sherpa-onnx-moonshine-tiny-en-int8",
+            idx=21,
+            lang="en",
+            short_name="moonshine_tiny_int8",
+            cmd="""
+            pushd $model_name
+
+            rm -rfv test_wavs
+
+            ls -lh
+
             popd
             """,
         ),

sherpa-onnx/java-api/Makefile

@@ -26,6 +26,7 @@ java_files += OnlineRecognizer.java
 java_files += OfflineTransducerModelConfig.java
 java_files += OfflineParaformerModelConfig.java
 java_files += OfflineWhisperModelConfig.java
+java_files += OfflineMoonshineModelConfig.java
 java_files += OfflineNemoEncDecCtcModelConfig.java
 java_files += OfflineSenseVoiceModelConfig.java
 java_files += OfflineModelConfig.java

sherpa-onnx/java-api/src/com/k2fsa/sherpa/onnx/OfflineModelConfig.java

@@ -6,6 +6,7 @@ public class OfflineModelConfig {
     private final OfflineTransducerModelConfig transducer;
     private final OfflineParaformerModelConfig paraformer;
     private final OfflineWhisperModelConfig whisper;
+    private final OfflineMoonshineModelConfig moonshine;
     private final OfflineNemoEncDecCtcModelConfig nemo;
     private final OfflineSenseVoiceModelConfig senseVoice;
     private final String teleSpeech;
@@ -22,6 +23,7 @@ public class OfflineModelConfig {
         this.transducer = builder.transducer;
         this.paraformer = builder.paraformer;
         this.whisper = builder.whisper;
+        this.moonshine = builder.moonshine;
         this.nemo = builder.nemo;
         this.senseVoice = builder.senseVoice;
         this.teleSpeech = builder.teleSpeech;
@@ -50,6 +52,10 @@ public class OfflineModelConfig {
         return whisper;
     }
 
+    public OfflineMoonshineModelConfig getMoonshine() {
+        return moonshine;
+    }
+
     public OfflineSenseVoiceModelConfig getSenseVoice() {
         return senseVoice;
     }
@@ -90,6 +96,7 @@ public class OfflineModelConfig {
         private OfflineParaformerModelConfig paraformer = OfflineParaformerModelConfig.builder().build();
         private OfflineTransducerModelConfig transducer = OfflineTransducerModelConfig.builder().build();
         private OfflineWhisperModelConfig whisper = OfflineWhisperModelConfig.builder().build();
+        private OfflineMoonshineModelConfig moonshine = OfflineMoonshineModelConfig.builder().build();
         private OfflineNemoEncDecCtcModelConfig nemo = OfflineNemoEncDecCtcModelConfig.builder().build();
         private OfflineSenseVoiceModelConfig senseVoice = OfflineSenseVoiceModelConfig.builder().build();
         private String teleSpeech = "";
@@ -135,6 +142,11 @@ public class OfflineModelConfig {
             return this;
         }
 
+        public Builder setMoonshine(OfflineMoonshineModelConfig moonshine) {
+            this.moonshine = moonshine;
+            return this;
+        }
+
         public Builder setTokens(String tokens) {
             this.tokens = tokens;
             return this;

sherpa-onnx/java-api/src/com/k2fsa/sherpa/onnx/OfflineMoonshineModelConfig.java

@@ -0,0 +1,70 @@
+// Copyright 2024 Xiaomi Corporation
+
+package com.k2fsa.sherpa.onnx;
+
+public class OfflineMoonshineModelConfig {
+    private final String preprocessor;
+    private final String encoder;
+    private final String uncachedDecoder;
+    private final String cachedDecoder;
+
+    private OfflineMoonshineModelConfig(Builder builder) {
+        this.preprocessor = builder.preprocessor;
+        this.encoder = builder.encoder;
+        this.uncachedDecoder = builder.uncachedDecoder;
+        this.cachedDecoder = builder.cachedDecoder;
+    }
+
+    public static Builder builder() {
+        return new Builder();
+    }
+
+    public String getPreprocessor() {
+        return preprocessor;
+    }
+
+    public String getEncoder() {
+        return encoder;
+    }
+
+    public String getUncachedDecoder() {
+        return uncachedDecoder;
+    }
+
+    public String getCachedDecoder() {
+        return cachedDecoder;
+    }
+
+    public static class Builder {
+        private String preprocessor = "";
+        private String encoder = "";
+        private String uncachedDecoder = "";
+        private String cachedDecoder = "";
+
+        public OfflineMoonshineModelConfig build() {
+            return new OfflineMoonshineModelConfig(this);
+        }
+
+        public Builder setPreprocessor(String preprocessor) {
+            this.preprocessor = preprocessor;
+            return this;
+        }
+
+        public Builder setEncoder(String encoder) {
+            this.encoder = encoder;
+            return this;
+        }
+
+        public Builder setUncachedDecoder(String uncachedDecoder) {
+            this.uncachedDecoder = uncachedDecoder;
+            return this;
+        }
+
+        public Builder setCachedDecoder(String cachedDecoder) {
+            this.cachedDecoder = cachedDecoder;
+            return this;
+        }
+    }
+
+
+}

sherpa-onnx/jni/offline-recognizer.cc

@@ -174,6 +174,39 @@ static OfflineRecognizerConfig GetOfflineConfig(JNIEnv *env, jobject config) {
   ans.model_config.whisper.tail_paddings =
       env->GetIntField(whisper_config, fid);
 
+  // moonshine
+  fid = env->GetFieldID(model_config_cls, "moonshine",
+                        "Lcom/k2fsa/sherpa/onnx/OfflineMoonshineModelConfig;");
+  jobject moonshine_config = env->GetObjectField(model_config, fid);
+  jclass moonshine_config_cls = env->GetObjectClass(moonshine_config);
+
+  fid = env->GetFieldID(moonshine_config_cls, "preprocessor",
+                        "Ljava/lang/String;");
+  s = (jstring)env->GetObjectField(moonshine_config, fid);
+  p = env->GetStringUTFChars(s, nullptr);
+  ans.model_config.moonshine.preprocessor = p;
+  env->ReleaseStringUTFChars(s, p);
+
+  fid = env->GetFieldID(moonshine_config_cls, "encoder", "Ljava/lang/String;");
+  s = (jstring)env->GetObjectField(moonshine_config, fid);
+  p = env->GetStringUTFChars(s, nullptr);
+  ans.model_config.moonshine.encoder = p;
+  env->ReleaseStringUTFChars(s, p);
+
+  fid = env->GetFieldID(moonshine_config_cls, "uncachedDecoder",
+                        "Ljava/lang/String;");
+  s = (jstring)env->GetObjectField(moonshine_config, fid);
+  p = env->GetStringUTFChars(s, nullptr);
+  ans.model_config.moonshine.uncached_decoder = p;
+  env->ReleaseStringUTFChars(s, p);
+
+  fid = env->GetFieldID(moonshine_config_cls, "cachedDecoder",
+                        "Ljava/lang/String;");
+  s = (jstring)env->GetObjectField(moonshine_config, fid);
+  p = env->GetStringUTFChars(s, nullptr);
+  ans.model_config.moonshine.cached_decoder = p;
+  env->ReleaseStringUTFChars(s, p);
+
   // sense voice
   fid = env->GetFieldID(model_config_cls, "senseVoice",
                         "Lcom/k2fsa/sherpa/onnx/OfflineSenseVoiceModelConfig;");

sherpa-onnx/kotlin-api/OfflineRecognizer.kt

@@ -33,6 +33,13 @@ data class OfflineWhisperModelConfig(
     var tailPaddings: Int = 1000, // Padding added at the end of the samples
 )
 
+data class OfflineMoonshineModelConfig(
+    var preprocessor: String = "",
+    var encoder: String = "",
+    var uncachedDecoder: String = "",
+    var cachedDecoder: String = "",
+)
+
 data class OfflineSenseVoiceModelConfig(
     var model: String = "",
     var language: String = "",
@@ -43,6 +50,7 @@ data class OfflineModelConfig(
     var transducer: OfflineTransducerModelConfig = OfflineTransducerModelConfig(),
     var paraformer: OfflineParaformerModelConfig = OfflineParaformerModelConfig(),
     var whisper: OfflineWhisperModelConfig = OfflineWhisperModelConfig(),
+    var moonshine: OfflineMoonshineModelConfig = OfflineMoonshineModelConfig(),
     var nemo: OfflineNemoEncDecCtcModelConfig = OfflineNemoEncDecCtcModelConfig(),
     var senseVoice: OfflineSenseVoiceModelConfig = OfflineSenseVoiceModelConfig(),
     var teleSpeech: String = "",
@@ -417,6 +425,19 @@ fun getOfflineModelConfig(type: Int): OfflineModelConfig? {
                 modelType = "nemo_transducer",
             )
         }
+
+        21 -> {
+            val modelDir = "sherpa-onnx-moonshine-tiny-en-int8"
+            return OfflineModelConfig(
+                moonshine = OfflineMoonshineModelConfig(
+                    preprocessor = "$modelDir/preprocess.onnx",
+                    encoder = "$modelDir/encode.int8.onnx",
+                    uncachedDecoder = "$modelDir/uncached_decode.int8.onnx",
+                    cachedDecoder = "$modelDir/cached_decode.int8.onnx",
+                ),
+                tokens = "$modelDir/tokens.txt",
+            )
+        }
     }
     return null
 }