From bd81457222fb531d483695ce671200d6b7927dea Mon Sep 17 00:00:00 2001 From: Guo Liyong Date: Sat, 28 May 2022 01:37:51 +0800 Subject: [PATCH] adjust results format --- egs/librispeech/ASR/RESULTS-100hours.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/egs/librispeech/ASR/RESULTS-100hours.md b/egs/librispeech/ASR/RESULTS-100hours.md index c4c1e2073..5e7fd9a30 100644 --- a/egs/librispeech/ASR/RESULTS-100hours.md +++ b/egs/librispeech/ASR/RESULTS-100hours.md @@ -9,8 +9,11 @@ Related models/log/tensorboard is uploaded to: https://huggingface.co/GuoLiyong/stateless6_baseline_vs_disstillation Following results are otained by ./istillation_with_hubert.sh + The only differences is in pruned_transducer_stateless6/train.py. + For baseline: set enable_distillation=False + For distillation: set enable_distillation=True (the default) Decoding method is modified beam search.