From 22af36a88e4b5fdab9012dce730c50fb0749d34c Mon Sep 17 00:00:00 2001 From: Yuekai Zhang Date: Thu, 7 Jul 2022 14:49:51 +0800 Subject: [PATCH] setting max duration --- egs/aishell2/ASR/pruned_transducer_stateless5/train.py | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/egs/aishell2/ASR/pruned_transducer_stateless5/train.py b/egs/aishell2/ASR/pruned_transducer_stateless5/train.py index 0d9bd0a6a..99a3596ef 100755 --- a/egs/aishell2/ASR/pruned_transducer_stateless5/train.py +++ b/egs/aishell2/ASR/pruned_transducer_stateless5/train.py @@ -957,15 +957,15 @@ def run(rank, world_size, args): train_cuts = aishell2.train_cuts() def remove_short_and_long_utt(c: Cut): - # Keep only utterances with duration between 1 second and 20 seconds + # Keep only utterances with duration between 1 second and 8 seconds # - # Caution: There is a reason to select 20.0 here. Please see + # Caution: There is a reason to select 8.0 here. Please see # ../local/display_manifest_statistics.py # # You should use ../local/display_manifest_statistics.py to get # an utterance duration distribution for your dataset to select # the threshold - return 1.0 <= c.duration <= 20.0 + return 1.0 <= c.duration <= 8.0 train_cuts = train_cuts.filter(remove_short_and_long_utt)