Add Kotlin and Java API for KittenTTS (#2461)

@@ -309,12 +309,14 @@ jobs:
         run: |
           cd ./java-api-examples
 
+          ./run-non-streaming-tts-kitten-en.sh
           ./run-non-streaming-tts-kokoro-zh-en.sh
           ./run-non-streaming-tts-kokoro-en.sh
           ./run-non-streaming-tts-matcha-zh.sh
           ./run-non-streaming-tts-matcha-en.sh
           ls -lh
 
+          rm -rf kitten-nano-en-*
           rm -rf kokoro-multi-*
           rm -rf kokoro-en-*
 

@@ -144,3 +144,4 @@ sherpa-onnx-dolphin-base-ctc-multi-lang-int8-2025-04-02
 dict
 *.npz
 voices.bin
+kitten-nano-en-v0_1-fp16

@@ -40,7 +40,7 @@ public class NonStreamingTtsCoquiDe {
     String waveFilename = "tts-coqui-de.wav";
     audio.save(waveFilename);
     System.out.printf("-- elapsed : %.3f seconds\n", timeElapsedSeconds);
-    System.out.printf("-- audio duration: %.3f seconds\n", timeElapsedSeconds);
+    System.out.printf("-- audio duration: %.3f seconds\n", audioDuration);
     System.out.printf("-- real-time factor (RTF): %.3f\n", real_time_factor);
     System.out.printf("-- text: %s\n", text);
     System.out.printf("-- Saved to %s\n", waveFilename);

+// Copyright 2025 Xiaomi Corporation
+
+// This file shows how to use a KittenTTS English model
+// to convert text to speech
+import com.k2fsa.sherpa.onnx.*;
+
+public class NonStreamingTtsKittenEn {
+  public static void main(String[] args) {
+    // please visit
+    // https://k2-fsa.github.io/sherpa/onnx/tts/pretrained_models/kitten.html
+    // to download model files
+    String model = "./kitten-nano-en-v0_1-fp16/model.fp16.onnx";
+    String voices = "./kitten-nano-en-v0_1-fp16/voices.bin";
+    String tokens = "./kitten-nano-en-v0_1-fp16/tokens.txt";
+    String dataDir = "./kitten-nano-en-v0_1-fp16/espeak-ng-data";
+    String text =
+        "Today as always, men fall into two groups: slaves and free men. Whoever does not have"
+            + " two-thirds of his day for himself, is a slave, whatever he may be: a statesman, a"
+            + " businessman, an official, or a scholar.";
+
+    OfflineTtsKittenModelConfig kittenModelConfig =
+        OfflineTtsKittenModelConfig.builder()
+            .setModel(model)
+            .setVoices(voices)
+            .setTokens(tokens)
+            .setDataDir(dataDir)
+            .build();
+
+    OfflineTtsModelConfig modelConfig =
+        OfflineTtsModelConfig.builder()
+            .setKitten(kittenModelConfig)
+            .setNumThreads(2)
+            .setDebug(true)
+            .build();
+
+    OfflineTtsConfig config = OfflineTtsConfig.builder().setModel(modelConfig).build();
+    OfflineTts tts = new OfflineTts(config);
+
+    int sid = 7;
+    float speed = 1.0f;
+    long start = System.currentTimeMillis();
+    GeneratedAudio audio = tts.generate(text, sid, speed);
+    long stop = System.currentTimeMillis();
+
+    float timeElapsedSeconds = (stop - start) / 1000.0f;
+
+    float audioDuration = audio.getSamples().length / (float) audio.getSampleRate();
+    float real_time_factor = timeElapsedSeconds / audioDuration;
+
+    String waveFilename = "tts-kitten-en.wav";
+    audio.save(waveFilename);
+    System.out.printf("-- elapsed : %.3f seconds\n", timeElapsedSeconds);
+    System.out.printf("-- audio duration: %.3f seconds\n", audioDuration);
+    System.out.printf("-- real-time factor (RTF): %.3f\n", real_time_factor);
+    System.out.printf("-- text: %s\n", text);
+    System.out.printf("-- Saved to %s\n", waveFilename);
+
+    tts.release();
+  }
+}

@@ -50,7 +50,7 @@ public class NonStreamingTtsKokoroEn {
     String waveFilename = "tts-kokoro-en.wav";
     audio.save(waveFilename);
     System.out.printf("-- elapsed : %.3f seconds\n", timeElapsedSeconds);
-    System.out.printf("-- audio duration: %.3f seconds\n", timeElapsedSeconds);
+    System.out.printf("-- audio duration: %.3f seconds\n", audioDuration);
     System.out.printf("-- real-time factor (RTF): %.3f\n", real_time_factor);
     System.out.printf("-- text: %s\n", text);
     System.out.printf("-- Saved to %s\n", waveFilename);

@@ -54,7 +54,7 @@ public class NonStreamingTtsKokoroZhEn {
     String waveFilename = "tts-kokoro-zh-en.wav";
     audio.save(waveFilename);
     System.out.printf("-- elapsed : %.3f seconds\n", timeElapsedSeconds);
-    System.out.printf("-- audio duration: %.3f seconds\n", timeElapsedSeconds);
+    System.out.printf("-- audio duration: %.3f seconds\n", audioDuration);
     System.out.printf("-- real-time factor (RTF): %.3f\n", real_time_factor);
     System.out.printf("-- text: %s\n", text);
     System.out.printf("-- Saved to %s\n", waveFilename);

@@ -50,7 +50,7 @@ public class NonStreamingTtsMatchaEn {
     String waveFilename = "tts-matcha-en.wav";
     audio.save(waveFilename);
     System.out.printf("-- elapsed : %.3f seconds\n", timeElapsedSeconds);
-    System.out.printf("-- audio duration: %.3f seconds\n", timeElapsedSeconds);
+    System.out.printf("-- audio duration: %.3f seconds\n", audioDuration);
     System.out.printf("-- real-time factor (RTF): %.3f\n", real_time_factor);
     System.out.printf("-- text: %s\n", text);
     System.out.printf("-- Saved to %s\n", waveFilename);

@@ -56,7 +56,7 @@ public class NonStreamingTtsMatchaZh {
     String waveFilename = "tts-matcha-zh.wav";
     audio.save(waveFilename);
     System.out.printf("-- elapsed : %.3f seconds\n", timeElapsedSeconds);
-    System.out.printf("-- audio duration: %.3f seconds\n", timeElapsedSeconds);
+    System.out.printf("-- audio duration: %.3f seconds\n", audioDuration);
     System.out.printf("-- real-time factor (RTF): %.3f\n", real_time_factor);
     System.out.printf("-- text: %s\n", text);
     System.out.printf("-- Saved to %s\n", waveFilename);

@@ -48,7 +48,7 @@ public class NonStreamingTtsPiperEn {
     String waveFilename = "tts-piper-en.wav";
     audio.save(waveFilename);
     System.out.printf("-- elapsed : %.3f seconds\n", timeElapsedSeconds);
-    System.out.printf("-- audio duration: %.3f seconds\n", timeElapsedSeconds);
+    System.out.printf("-- audio duration: %.3f seconds\n", audioDuration);
     System.out.printf("-- real-time factor (RTF): %.3f\n", real_time_factor);
     System.out.printf("-- text: %s\n", text);
     System.out.printf("-- Saved to %s\n", waveFilename);

@@ -176,7 +176,7 @@ public class NonStreamingTtsPiperEn {
     String waveFilename = "tts-piper-en.wav";
     audio.save(waveFilename);
     System.out.printf("-- elapsed : %.3f seconds\n", timeElapsedSeconds);
-    System.out.printf("-- audio duration: %.3f seconds\n", timeElapsedSeconds);
+    System.out.printf("-- audio duration: %.3f seconds\n", audioDuration);
     System.out.printf("-- real-time factor (RTF): %.3f\n", real_time_factor);
     System.out.printf("-- text: %s\n", text);
     System.out.printf("-- Saved to %s\n", waveFilename);

@@ -54,7 +54,7 @@ public class NonStreamingTtsPiperEn {
     String waveFilename = "tts-vits-zh.wav";
     audio.save(waveFilename);
     System.out.printf("-- elapsed : %.3f seconds\n", timeElapsedSeconds);
-    System.out.printf("-- audio duration: %.3f seconds\n", timeElapsedSeconds);
+    System.out.printf("-- audio duration: %.3f seconds\n", audioDuration);
     System.out.printf("-- real-time factor (RTF): %.3f\n", real_time_factor);
     System.out.printf("-- text: %s\n", text);
     System.out.printf("-- Saved to %s\n", waveFilename);

+#!/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
+
+# please visit
+# https://k2-fsa.github.io/sherpa/onnx/tts/pretrained_models/kitten.html
+# to download more models
+
+if [ ! -f ./kitten-nano-en-v0_1-fp16/model.fp16.onnx ]; then
+  curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/kitten-nano-en-v0_1-fp16.tar.bz2
+  tar xf kitten-nano-en-v0_1-fp16.tar.bz2
+  rm kitten-nano-en-v0_1-fp16.tar.bz2
+fi
+
+java \
+  -Djava.library.path=$PWD/../build/lib \
+  -cp ../sherpa-onnx/java-api/build/sherpa-onnx.jar \
+  NonStreamingTtsKittenEn.java

@@ -140,6 +140,12 @@ function testTts() {
     rm kokoro-en-v0_19.tar.bz2
   fi
 
+  if [ ! -f ./kitten-nano-en-v0_1-fp16/model.fp16.onnx ]; then
+    curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/kitten-nano-en-v0_1-fp16.tar.bz2
+    tar xf kitten-nano-en-v0_1-fp16.tar.bz2
+    rm kitten-nano-en-v0_1-fp16.tar.bz2
+  fi
+
   out_filename=test_tts.jar
   kotlinc-jvm -include-runtime -d $out_filename \
     test_tts.kt \
@@ -477,7 +483,7 @@ function testOfflineNeMoCanary() {
   java -Djava.library.path=../build/lib -jar $out_filename
 }
 
-# testVersion
+testVersion
 
 testOfflineNeMoCanary
 testOfflineSenseVoiceWithHr

@@ -5,6 +5,7 @@ fun main() {
   testMatcha()
   testKokoroEn()
   testKokoroZhEn()
+  testKittenEn()
 }
 
 fun testKokoroZhEn() {
@@ -96,6 +97,27 @@ fun testVits() {
   println("Saved to test-en.wav")
 }
 
+fun testKittenEn() {
+  // see https://github.com/k2-fsa/sherpa-onnx/releases/tag/tts-models
+  var config = OfflineTtsConfig(
+    model=OfflineTtsModelConfig(
+      kitten=OfflineTtsKittenModelConfig(
+        model="./kitten-nano-en-v0_1-fp16/model.fp16.onnx",
+        voices="./kitten-nano-en-v0_1-fp16/voices.bin",
+        tokens="./kitten-nano-en-v0_1-fp16/tokens.txt",
+        dataDir="./kitten-nano-en-v0_1-fp16/espeak-ng-data",
+      ),
+      numThreads=2,
+      debug=true,
+    ),
+  )
+  val tts = OfflineTts(config=config)
+  val audio = tts.generateWithCallback(text="How are you doing today?", sid=7, callback=::callback)
+  audio.save(filename="test-kitten-en.wav")
+  tts.release()
+  println("Saved to test-kitten-en.wav")
+}
+
 /*
 1. Unzip test_tts.jar
 2.

@@ -43,6 +43,7 @@ java_files += OfflineRecognizerResult.java
 java_files += OfflineStream.java
 java_files += OfflineRecognizer.java
 
+java_files += OfflineTtsKittenModelConfig.java
 java_files += OfflineTtsKokoroModelConfig.java
 java_files += OfflineTtsMatchaModelConfig.java
 java_files += OfflineTtsVitsModelConfig.java

+// Copyright 2025 Xiaomi Corporation
+package com.k2fsa.sherpa.onnx;
+
+public class OfflineTtsKittenModelConfig {
+    private final String model;
+    private final String voices;
+    private final String tokens;
+    private final String dataDir;
+    private final float lengthScale;
+
+    private OfflineTtsKittenModelConfig(Builder builder) {
+        this.model = builder.model;
+        this.voices = builder.voices;
+        this.tokens = builder.tokens;
+        this.dataDir = builder.dataDir;
+        this.lengthScale = builder.lengthScale;
+    }
+
+    public static Builder builder() {
+        return new Builder();
+    }
+
+    public String getModel() {
+        return model;
+    }
+
+    public String getVoices() {
+        return voices;
+    }
+
+    public String getTokens() {
+        return tokens;
+    }
+
+    public String getDataDir() {
+        return dataDir;
+    }
+
+    public float getLengthScale() {
+        return lengthScale;
+    }
+
+
+    public static class Builder {
+        private String model = "";
+        private String voices = "";
+        private String tokens = "";
+        private String dataDir = "";
+        private float lengthScale = 1.0f;
+
+        public OfflineTtsKittenModelConfig build() {
+            return new OfflineTtsKittenModelConfig(this);
+        }
+
+        public Builder setModel(String model) {
+            this.model = model;
+            return this;
+        }
+
+        public Builder setVoices(String voices) {
+            this.voices = voices;
+            return this;
+        }
+
+        public Builder setTokens(String tokens) {
+            this.tokens = tokens;
+            return this;
+        }
+
+        public Builder setDataDir(String dataDir) {
+            this.dataDir = dataDir;
+            return this;
+        }
+
+        public Builder setLengthScale(float lengthScale) {
+            this.lengthScale = lengthScale;
+            return this;
+        }
+    }
+}
\ No newline at end of file

@@ -6,6 +6,7 @@ public class OfflineTtsModelConfig {
     private final OfflineTtsVitsModelConfig vits;
     private final OfflineTtsMatchaModelConfig matcha;
     private final OfflineTtsKokoroModelConfig kokoro;
+    private final OfflineTtsKittenModelConfig kitten;
     private final int numThreads;
     private final boolean debug;
     private final String provider;
@@ -14,6 +15,7 @@ public class OfflineTtsModelConfig {
         this.vits = builder.vits;
         this.matcha = builder.matcha;
         this.kokoro = builder.kokoro;
+        this.kitten = builder.kitten;
         this.numThreads = builder.numThreads;
         this.debug = builder.debug;
         this.provider = builder.provider;
@@ -35,10 +37,15 @@ public class OfflineTtsModelConfig {
         return kokoro;
     }
 
+    public OfflineTtsKittenModelConfig getKitten() {
+        return kitten;
+    }
+
     public static class Builder {
         private OfflineTtsVitsModelConfig vits = OfflineTtsVitsModelConfig.builder().build();
         private OfflineTtsMatchaModelConfig matcha = OfflineTtsMatchaModelConfig.builder().build();
         private OfflineTtsKokoroModelConfig kokoro = OfflineTtsKokoroModelConfig.builder().build();
+        private OfflineTtsKittenModelConfig kitten = OfflineTtsKittenModelConfig.builder().build();
         private int numThreads = 1;
         private boolean debug = true;
         private String provider = "cpu";
@@ -62,6 +69,11 @@ public class OfflineTtsModelConfig {
             return this;
         }
 
+        public Builder setKitten(OfflineTtsKittenModelConfig kitten) {
+            this.kitten = kitten;
+            return this;
+        }
+
         public Builder setNumThreads(int numThreads) {
             this.numThreads = numThreads;
             return this;
@@ -77,4 +89,4 @@ public class OfflineTtsModelConfig {
             return this;
         }
     }
-}
+}
\ No newline at end of file

@@ -166,6 +166,39 @@ static OfflineTtsConfig GetOfflineTtsConfig(JNIEnv *env, jobject config) {
   fid = env->GetFieldID(kokoro_cls, "lengthScale", "F");
   ans.model.kokoro.length_scale = env->GetFloatField(kokoro, fid);
 
+  // kitten
+  fid = env->GetFieldID(model_config_cls, "kitten",
+                        "Lcom/k2fsa/sherpa/onnx/OfflineTtsKittenModelConfig;");
+  jobject kitten = env->GetObjectField(model, fid);
+  jclass kitten_cls = env->GetObjectClass(kitten);
+
+  fid = env->GetFieldID(kitten_cls, "model", "Ljava/lang/String;");
+  s = (jstring)env->GetObjectField(kitten, fid);
+  p = env->GetStringUTFChars(s, nullptr);
+  ans.model.kitten.model = p;
+  env->ReleaseStringUTFChars(s, p);
+
+  fid = env->GetFieldID(kitten_cls, "voices", "Ljava/lang/String;");
+  s = (jstring)env->GetObjectField(kitten, fid);
+  p = env->GetStringUTFChars(s, nullptr);
+  ans.model.kitten.voices = p;
+  env->ReleaseStringUTFChars(s, p);
+
+  fid = env->GetFieldID(kitten_cls, "tokens", "Ljava/lang/String;");
+  s = (jstring)env->GetObjectField(kitten, fid);
+  p = env->GetStringUTFChars(s, nullptr);
+  ans.model.kitten.tokens = p;
+  env->ReleaseStringUTFChars(s, p);
+
+  fid = env->GetFieldID(kitten_cls, "dataDir", "Ljava/lang/String;");
+  s = (jstring)env->GetObjectField(kitten, fid);
+  p = env->GetStringUTFChars(s, nullptr);
+  ans.model.kitten.data_dir = p;
+  env->ReleaseStringUTFChars(s, p);
+
+  fid = env->GetFieldID(kitten_cls, "lengthScale", "F");
+  ans.model.kitten.length_scale = env->GetFloatField(kitten, fid);
+
   fid = env->GetFieldID(model_config_cls, "numThreads", "I");
   ans.model.num_threads = env->GetIntField(model, fid);
 

@@ -36,10 +36,19 @@ data class OfflineTtsKokoroModelConfig(
     var lengthScale: Float = 1.0f,
 )
 
+data class OfflineTtsKittenModelConfig(
+    var model: String = "",
+    var voices: String = "",
+    var tokens: String = "",
+    var dataDir: String = "",
+    var lengthScale: Float = 1.0f,
+)
+
 data class OfflineTtsModelConfig(
     var vits: OfflineTtsVitsModelConfig = OfflineTtsVitsModelConfig(),
     var matcha: OfflineTtsMatchaModelConfig = OfflineTtsMatchaModelConfig(),
     var kokoro: OfflineTtsKokoroModelConfig = OfflineTtsKokoroModelConfig(),
+    var kitten: OfflineTtsKittenModelConfig = OfflineTtsKittenModelConfig(),
     var numThreads: Int = 1,
     var debug: Boolean = false,
     var provider: String = "cpu",
@@ -189,13 +198,14 @@ fun getOfflineTtsConfig(
     modelName: String, // for VITS
     acousticModelName: String, // for Matcha
     vocoder: String, // for Matcha
-    voices: String, // for Kokoro
+    voices: String, // for Kokoro or kitten
     lexicon: String,
     dataDir: String,
     dictDir: String,
     ruleFsts: String,
     ruleFars: String,
-    numThreads: Int? = null
+    numThreads: Int? = null,
+    isKitten: Boolean = false
 ): OfflineTtsConfig {
     // For Matcha TTS, please set
     // acousticModelName, vocoder
@@ -203,13 +213,16 @@ fun getOfflineTtsConfig(
     // For Kokoro TTS, please set
     // modelName, voices
 
+    // For Kitten TTS, please set
+    // modelName, voices, isKitten
+
     // For VITS, please set
     // modelName
 
     val numberOfThreads = if (numThreads != null) {
         numThreads
     } else if (voices.isNotEmpty()) {
-        // for Kokoro TTS models, we use more threads
+        // for Kokoro and Kitten TTS models, we use more threads
         4
     } else {
         2
@@ -252,7 +265,7 @@ fun getOfflineTtsConfig(
         OfflineTtsMatchaModelConfig()
     }
 
-    val kokoro = if (voices.isNotEmpty()) {
+    val kokoro = if (voices.isNotEmpty() && !isKitten) {
         OfflineTtsKokoroModelConfig(
             model = "$modelDir/$modelName",
             voices = "$modelDir/$voices",
@@ -269,11 +282,23 @@ fun getOfflineTtsConfig(
         OfflineTtsKokoroModelConfig()
     }
 
+    val kitten = if (isKitten) {
+        OfflineTtsKittenModelConfig(
+            model = "$modelDir/$modelName",
+            voices = "$modelDir/$voices",
+            tokens = "$modelDir/tokens.txt",
+            dataDir = dataDir,
+        )
+    } else {
+        OfflineTtsKittenModelConfig()
+    }
+
     return OfflineTtsConfig(
         model = OfflineTtsModelConfig(
             vits = vits,
             matcha = matcha,
             kokoro = kokoro,
+            kitten = kitten,
             numThreads = numberOfThreads,
             debug = true,
             provider = "cpu",