From 16853c4310814cea6f31e79a27ea5f902ed75bad Mon Sep 17 00:00:00 2001
From: dohe0342 <kimdohe1070@gmail.com>
Date: Tue, 11 Apr 2023 15:47:17 +0900
Subject: [PATCH] from local

---
 .../.data2vec_audio.py.swp                    | Bin 40960 -> 40960 bytes
 .../.prompt_tuning.py.swp                     | Bin 86016 -> 86016 bytes
 .../data2vec_audio.py                         |   1 +
 3 files changed, 1 insertion(+)

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 2146c4daa43af67163ebc7b45a8688d7dcfef17b..05986257145cf4035c8de7dcc38a54b872460d72 100644
GIT binary patch
delta 112
zcmZoTz|?SnX~P5oM)%DV1&Z}~b66M{masA~FbFU(xK9>zcx~pu!N9-`#4p$x7#e}t
z9Eb&gm>-BAu`w__1me9wTnWTRK>Ql0dkO1iBgZ0^$$W0Io1eH&WST7NA-g%>V`4Y}
DTJ#wz

delta 122
zcmZoTz|?SnX~P5oMz_rq1&Z}~C$lgx9A{-<U=UznaGNaX@Y+m|gMon?h?lT4Fhl{d
zF%TO8F)I+W0PzDh28KC6916sIKn&6?$G+Lfv4~~z1ZOGE3LORoFqr(>Rd(}1*NIG<
IOFg>70qBz%s{jB1

diff --git a/egs/librispeech/ASR/pruned_transducer_stateless_d2v_v2/.prompt_tuning.py.swp b/egs/librispeech/ASR/pruned_transducer_stateless_d2v_v2/.prompt_tuning.py.swp
index 14bf7df31b6cc6e64ac2b2dd177215ac99156c3d..04617ff0889dbe50add6935998faae0c25ce440f 100644
GIT binary patch
delta 36
qcmZozz}m2YRV>LM%+puFQqO<^2m}}yESXGG+GaG0Z53mD<qrU+ZwWL2

delta 36
qcmZozz}m2YRV>LM%+puFQqO<^2m}}y>KIK^HcoF8+bYKR${zr$UkRcB

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 3a9228396..34ce40466 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
@@ -493,6 +493,7 @@ class Data2VecAudioModel(BaseFairseqModel):
             #features = torch.cat([features, prompt])
             prompt = prompt.expand((features.size()[0], prompt.size()[0], prompt.size()[1]))
             print(prompt.size())
+            print(features.size())
             features = torch.cat([prompt, features])
             print(features.size())
             prompt_padding_mask = torch.zeros(promt.size()).type(torch.BoolTensor).to(features.device)