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 591c7ff5c..d856a433d 100644
Binary files a/egs/librispeech/ASR/pruned_transducer_stateless_d2v_v2/.data2vec_audio.py.swp and b/egs/librispeech/ASR/pruned_transducer_stateless_d2v_v2/.data2vec_audio.py.swp differ
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 32b06f290..3f95fec84 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
@@ -496,7 +496,7 @@ class Data2VecAudioModel(BaseFairseqModel):
             print(features.size())
             features = torch.cat([prompt, features], dim=1)
             print(features.size())
-            prompt_padding_mask = torch.zeros(prompt.size()).type(torch.BoolTensor).to(features.device)
+            prompt_padding_mask = torch.zeros(prompt.size()[0], prompt.size()[1]).type(torch.BoolTensor).to(features.device)
             print(prompt_padding_mask.size())
             padding_mask = torch.cat([prompt_padding_mask, padding_mask])
             print(padding_mask.size())