diff --git a/egs/commonvoice/ASR/pruned_transducer_stateless7/asr_datamodule.py b/egs/commonvoice/ASR/pruned_transducer_stateless7/asr_datamodule.py index 41009831c..bf0a3e245 100644 --- a/egs/commonvoice/ASR/pruned_transducer_stateless7/asr_datamodule.py +++ b/egs/commonvoice/ASR/pruned_transducer_stateless7/asr_datamodule.py @@ -409,6 +409,14 @@ class CommonVoiceAsrDataModule: self.args.cv_manifest_dir / f"cv-{self.args.language}_cuts_train.jsonl.gz" ) + @lru_cache() + def validated_cuts(self) -> CutSet: + logging.info("About to get validated cuts (with dev/test removed)") + return load_manifest_lazy( + self.args.cv_manifest_dir + / f"cv-{self.args.language}_cuts_validated.jsonl.gz" + ) + @lru_cache() def dev_cuts(self) -> CutSet: logging.info("About to get dev cuts")