fix asr infer.py

paddlespeech/cli/asr/infer.py

@@ -25,9 +25,6 @@ import librosa
 import numpy as np
 import paddle
 import soundfile
-from paddlespeech.audio.transform.transformation import Transformation
-from paddlespeech.s2t.frontend.featurizer.text_featurizer import TextFeaturizer
-from paddlespeech.s2t.utils.utility import UpdateConfig
 from yacs.config import CfgNode
 
 from ...utils.env import MODEL_HOME
@@ -37,6 +34,9 @@ from ..log import logger
 from ..utils import CLI_TIMER
 from ..utils import stats_wrapper
 from ..utils import timer_register
+from paddlespeech.audio.transform.transformation import Transformation
+from paddlespeech.s2t.frontend.featurizer.text_featurizer import TextFeaturizer
+from paddlespeech.s2t.utils.utility import UpdateConfig
 
 __all__ = ['ASRExecutor']
 
@@ -274,7 +274,7 @@ class ASRExecutor(BaseExecutor):
             # fbank
             audio = preprocessing(audio, **preprocess_args)
 
-            audio_len = paddle.to_tensor(audio.shape[0])
+            audio_len = paddle.to_tensor(audio.shape[0]).unsqueeze(axis=0)
             audio = paddle.to_tensor(audio, dtype='float32').unsqueeze(axis=0)
 
             self._inputs["audio"] = audio

paddlespeech/cli/ssl/infer.py

@@ -245,7 +245,7 @@ class SSLExecutor(BaseExecutor):
         # fbank
         audio = preprocessing(audio, **preprocess_args)
 
-        audio_len = paddle.to_tensor(audio.shape[0])
+        audio_len = paddle.to_tensor(audio.shape[0]).unsqueeze(axis=0)
         audio = paddle.to_tensor(audio, dtype='float32').unsqueeze(axis=0)
 
         self._inputs["audio"] = audio

paddlespeech/cli/whisper/infer.py

@@ -253,7 +253,7 @@ class WhisperExecutor(BaseExecutor):
         # fbank
         audio = log_mel_spectrogram(audio, resource_path=self.resource_path)
 
-        audio_len = paddle.to_tensor(audio.shape[0])
+        audio_len = paddle.to_tensor(audio.shape[0]).unsqueeze(axis=0)
 
         self._inputs["audio"] = audio
         self._inputs["audio_len"] = audio_len