Add LODR support to online and offline recognizers (#2026)

This PR integrates LODR (Level-Ordered Deterministic Rescoring) support from Icefall into both online and offline recognizers, enabling LODR for LM shallow fusion and LM rescore.

- Extended OnlineLMConfig and OfflineLMConfig to include lodr_fst, lodr_scale, and lodr_backoff_id.
- Implemented LodrFst and LodrStateCost classes and wired them into RNN LM scoring in both online and offline code paths.
- Updated Python bindings, CLI entry points, examples, and CI test scripts to accept and exercise the new LODR options.
This commit is contained in:
Askars Salimbajevs
2025-07-09 11:23:46 +03:00
committed by GitHub
parent 6122a678f5
commit f0960342ad
21 changed files with 613 additions and 14 deletions

View File

@@ -25,6 +25,7 @@ set(sources
jieba.cc
keyword-spotter-impl.cc
keyword-spotter.cc
lodr-fst.cc
offline-canary-model-config.cc
offline-canary-model.cc
offline-ctc-fst-decoder-config.cc