[Wav2Vec2] Fix dtype 64 bug (#13517)

* fix

* 2nd fix

src/transformers/models/speech_to_text/feature_extraction_speech_to_text.py

@@ -210,7 +210,7 @@ class Speech2TextFeatureExtractor(SequenceFeatureExtractor):
             raw_speech = [np.asarray(speech) for speech in raw_speech]
         elif not is_batched and not isinstance(raw_speech, np.ndarray):
             raw_speech = np.asarray(raw_speech)
-        elif isinstance(raw_speech, np.ndarray) and raw_speech.dtype is np.float64:
+        elif isinstance(raw_speech, np.ndarray) and raw_speech.dtype is np.dtype(np.float64):
             raw_speech = raw_speech.astype(np.float32)
 
         # always return batch

src/transformers/models/wav2vec2/feature_extraction_wav2vec2.py

@@ -207,10 +207,10 @@ class Wav2Vec2FeatureExtractor(SequenceFeatureExtractor):
         elif (
             not isinstance(input_values, np.ndarray)
             and isinstance(input_values[0], np.ndarray)
-            and input_values[0].dtype is np.float64
+            and input_values[0].dtype is np.dtype(np.float64)
         ):
             padded_inputs["input_values"] = [array.astype(np.float32) for array in input_values]
-        elif isinstance(input_values, np.ndarray) and input_values.dtype is np.float64:
+        elif isinstance(input_values, np.ndarray) and input_values.dtype is np.dtype(np.float64):
             padded_inputs["input_values"] = input_values.astype(np.float32)
 
         # convert attention_mask to correct format