From 8677d83efce8a1eddb626280b1c3cd63469fd0dc Mon Sep 17 00:00:00 2001
From: ahadjawaid <94938815+ahadjawaid@users.noreply.github.com>
Date: Sun, 2 Feb 2025 23:49:42 -0600
Subject: [PATCH] Fix: Prepend 0 to tokenization to prevent word skipping for
 Kokoro. (#1787)

Addressed issue Skipping words #1777
---
 sherpa-onnx/csrc/piper-phonemize-lexicon.cc | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/sherpa-onnx/csrc/piper-phonemize-lexicon.cc b/sherpa-onnx/csrc/piper-phonemize-lexicon.cc
index ec312d8da..70ca37758 100644
--- a/sherpa-onnx/csrc/piper-phonemize-lexicon.cc
+++ b/sherpa-onnx/csrc/piper-phonemize-lexicon.cc
@@ -163,6 +163,8 @@ static std::vector<std::vector<int64_t>> PiperPhonemesToIdsKokoro(
   std::vector<int64_t> current;
   current.reserve(phonemes.size());
 
+  current.push_back(0);
+
   for (auto p : phonemes) {
     if (token2id.count(p)) {
       if (current.size() > max_len - 1) {