From 88b7895adf03424497619b54cdd9a230e9216b5c Mon Sep 17 00:00:00 2001 From: Fangjun Kuang Date: Tue, 27 Dec 2022 13:59:55 +0800 Subject: [PATCH] fix librispeech.py in multi-dataset setup (#791) --- .../ASR/pruned_transducer_stateless3/librispeech.py | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless3/librispeech.py b/egs/librispeech/ASR/pruned_transducer_stateless3/librispeech.py index 6dba8e9fe..9f2cb6225 100644 --- a/egs/librispeech/ASR/pruned_transducer_stateless3/librispeech.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless3/librispeech.py @@ -72,3 +72,12 @@ class LibriSpeech: f = self.manifest_dir / "librispeech_cuts_dev-other.jsonl.gz" logging.info(f"About to get dev-other cuts from {f}") return load_manifest_lazy(f) + + def train_all_shuf_cuts(self) -> CutSet: + logging.info( + "About to get the shuffled train-clean-100, \ + train-clean-360 and train-other-500 cuts" + ) + return load_manifest_lazy( + self.manifest_dir / "librispeech_cuts_train-all-shuf.jsonl.gz" + )