Expose 'language' 'task' 'tailPaddings' in OfflineWhisperModelConfig (#…

…643) Co-authored-by: Gary <[email protected]>
k2-fsa · Mar 8, 2024 · ac43c2d · ac43c2d
1 parent 4b708e0
commit ac43c2d
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 0 deletions.
diff --git a/android/SherpaOnnx2Pass/app/src/main/java/com/k2fsa/sherpa/onnx/SherpaOnnx.kt b/android/SherpaOnnx2Pass/app/src/main/java/com/k2fsa/sherpa/onnx/SherpaOnnx.kt
@@ -75,6 +75,9 @@ data class OfflineParaformerModelConfig(
 data class OfflineWhisperModelConfig(
     var encoder: String = "",
     var decoder: String = "",
+    var language: String = "en", // Used with multilingual model
+    var task: String = "transcribe", // transcribe or translate
+    var tailPaddings: Int = 1000, // Padding added at the end of the samples
 )
 
 data class OfflineModelConfig(

diff --git a/sherpa-onnx/jni/jni.cc b/sherpa-onnx/jni/jni.cc
@@ -616,6 +616,22 @@ static OfflineRecognizerConfig GetOfflineConfig(JNIEnv *env, jobject config) {
   ans.model_config.whisper.decoder = p;
   env->ReleaseStringUTFChars(s, p);
 
+  fid = env->GetFieldID(whisper_config_cls, "language", "Ljava/lang/String;");
+  s = (jstring)env->GetObjectField(whisper_config, fid);
+  p = env->GetStringUTFChars(s, nullptr);
+  ans.model_config.whisper.language = p;
+  env->ReleaseStringUTFChars(s, p);
+
+  fid = env->GetFieldID(whisper_config_cls, "task", "Ljava/lang/String;");
+  s = (jstring)env->GetObjectField(whisper_config, fid);
+  p = env->GetStringUTFChars(s, nullptr);
+  ans.model_config.whisper.task = p;
+  env->ReleaseStringUTFChars(s, p);
+
+  fid = env->GetFieldID(whisper_config_cls, "tailPaddings", "I");
+  ans.model_config.whisper.tail_paddings = env->GetIntField(whisper_config,
+                                                            fid);
+
   return ans;
 }