From fd9690653e894a50cce74e50b7cc55aa0c8ca2a7 Mon Sep 17 00:00:00 2001 From: dohe0342 Date: Wed, 12 Apr 2023 14:45:58 +0900 Subject: [PATCH] from local --- .../.data2vec_audio.py.swp | Bin 40960 -> 40960 bytes .../data2vec_audio.py | 6 ++++-- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/egs/librispeech/ASR/pruned_transducer_stateless_d2v_v2/.data2vec_audio.py.swp b/egs/librispeech/ASR/pruned_transducer_stateless_d2v_v2/.data2vec_audio.py.swp index 09073357542def03d5ccbf48e33cab7fb2ab398b..f94443c1c17a98062855872246e2f4b8ac1db8fd 100644 GIT binary patch delta 194 zcmZoTz|?SnNi4}A%+puFQqO<^2m}}yzPOmBcpuy-Rwcmbwt1pJu|97K3j@OpR*01bx~1$442eLj55z1$%nZcW*cceD0`W#5&IMvMAifXOHDj}`V-Cw?eK(oO`L0}( z7r3%dj&T#+tmS6PI9b7$Wpcc`x}X9O6clCVm1qwZZ#82jT7oA+?&Lfdpk(^pk TV#NSb0RtsPl~$Wu-K!%2@ggsS delta 167 zcmZoTz|?SnNi4}A%+puFQqO<^2m}}y-q@I>RPWpj5!H$0;CQzFF5Xhh_2(S3Z6PC{Qja p%FHX#n0((=c5=Qe*W~>`3}tWL?;6NBS;Sp^vYi^w=1J~(5dfZCDdzwH diff --git a/egs/librispeech/ASR/pruned_transducer_stateless_d2v_v2/data2vec_audio.py b/egs/librispeech/ASR/pruned_transducer_stateless_d2v_v2/data2vec_audio.py index 7f0c5cf79..729fe6eef 100644 --- a/egs/librispeech/ASR/pruned_transducer_stateless_d2v_v2/data2vec_audio.py +++ b/egs/librispeech/ASR/pruned_transducer_stateless_d2v_v2/data2vec_audio.py @@ -494,8 +494,10 @@ class Data2VecAudioModel(BaseFairseqModel): prompt = prompt.expand((features.size()[0], prompt.size()[0], prompt.size()[1])) features = torch.cat([prompt, features], dim=1) prompt_padding_mask = torch.zeros(prompt.size()[0], prompt.size()[1]).type(torch.BoolTensor).to(features.device) - padding_mask = torch.cat([prompt_padding_mask, padding_mask], dim=1) - print(padding_mask.size()) + try: padding_mask = torch.cat([prompt_padding_mask, padding_mask], dim=1) + except: + print(prompt_padding_mask.size()) + print(padding_mask.size()) features = self.layer_norm(features)