first working version

Author: csukuangfj

sherpa-onnx/csrc/kokoro-multi-lang-lexicon.cc

@@ -260,7 +260,7 @@ class KokoroMultiLangLexicon::Impl {
 
   std::vector<std::vector<int32_t>> ConvertTextToTokenIDsWithEspeak(
       const std::string &text, const std::string &voice) const {
-    auto temp = ConvertTextToTokenIdsKokoro(
+    auto temp = ConvertTextToTokenIdsKokoroOrKitten(
         phoneme2id_, meta_data_.max_token_len, text, voice);
     std::vector<std::vector<int32_t>> ans;
     ans.reserve(temp.size());

sherpa-onnx/csrc/offline-tts-frontend.h

@@ -59,7 +59,7 @@ class OfflineTtsFrontend {
 void InitEspeak(const std::string &data_dir);
 
 // implementation in ./piper-phonemize-lexicon.cc
-std::vector<TokenIDs> ConvertTextToTokenIdsKokoro(
+std::vector<TokenIDs> ConvertTextToTokenIdsKokoroOrKitten(
     const std::unordered_map<char32_t, int32_t> &token2id,
     int32_t max_token_len, const std::string &text,
     const std::string &voice = "");

sherpa-onnx/csrc/offline-tts-impl.cc

@@ -16,6 +16,7 @@
 #include "rawfile/raw_file_manager.h"
 #endif
 
+#include "sherpa-onnx/csrc/offline-tts-kitten-impl.h"
 #include "sherpa-onnx/csrc/offline-tts-kokoro-impl.h"
 #include "sherpa-onnx/csrc/offline-tts-matcha-impl.h"
 #include "sherpa-onnx/csrc/offline-tts-vits-impl.h"
@@ -40,9 +41,15 @@ std::unique_ptr<OfflineTtsImpl> OfflineTtsImpl::Create(
     return std::make_unique<OfflineTtsVitsImpl>(config);
   } else if (!config.model.matcha.acoustic_model.empty()) {
     return std::make_unique<OfflineTtsMatchaImpl>(config);
+  } else if (!config.model.kokoro.model.empty()) {
+    return std::make_unique<OfflineTtsKokoroImpl>(config);
+  } else if (!config.model.kitten.model.empty()) {
+    return std::make_unique<OfflineTtsKittenImpl>(config);
   }
 
-  return std::make_unique<OfflineTtsKokoroImpl>(config);
+  SHERPA_ONNX_LOGE("Please provide a tts model.");
+
+  return {};
 }
 
 template <typename Manager>
@@ -52,9 +59,14 @@ std::unique_ptr<OfflineTtsImpl> OfflineTtsImpl::Create(
     return std::make_unique<OfflineTtsVitsImpl>(mgr, config);
   } else if (!config.model.matcha.acoustic_model.empty()) {
     return std::make_unique<OfflineTtsMatchaImpl>(mgr, config);
+  } else if (!config.model.kokoro.model.empty()) {
+    return std::make_unique<OfflineTtsKokoroImpl>(mgr, config);
+  } else if (!config.model.kitten.model.empty()) {
+    return std::make_unique<OfflineTtsKittenImpl>(mgr, config);
   }
 
-  return std::make_unique<OfflineTtsKokoroImpl>(mgr, config);
+  SHERPA_ONNX_LOGE("Please provide a tts model.");
+  return {};
 }
 
 #if __ANDROID_API__ >= 9