Support getting word IDs for CTC HLG decoding. (#978)

2024-06-06 14:22:39 +08:00
parent 69347ffc8f
commit 1a43d1e37f
13 changed files with 60 additions and 13 deletions
--- a/sherpa-onnx/csrc/online-recognizer.h
+++ b/sherpa-onnx/csrc/online-recognizer.h
@@ -47,6 +47,8 @@ struct OnlineRecognizerResult {
  /// log-domain scores from "hot-phrase" contextual boosting
  std::vector<float> context_scores;

+  std::vector<int32_t> words;
+
  /// ID of this segment
  /// When an endpoint is detected, it is incremented
  int32_t segment = 0;