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/offline-ctc-fst-decoder.cc
+++ b/sherpa-onnx/csrc/offline-ctc-fst-decoder.cc
@@ -108,6 +108,9 @@ static OfflineCtcDecoderResult DecodeOne(kaldi_decoder::FasterDecoder *decoder,
    // -1 here since the input labels are incremented during graph
    // construction
    r.tokens.push_back(arc.ilabel - 1);
+    if (arc.olabel != 0) {
+      r.words.push_back(arc.olabel);
+    }

    r.timestamps.push_back(t);
    prev = arc.ilabel;