Update app.py

app.py

@@ -34,7 +34,7 @@ from gpuinfo import GPUInfo
 #import csv
 import numpy as np
 import torch
-#import torchaudio
+import torchaudio
 import torchaudio.transforms as transforms
 
 from transformers import pipeline, AutoModel
@@ -69,9 +69,17 @@ def convert_to_wav(filepath):
 pipe = pipeline("automatic-speech-recognition", model="NbAiLab/nb-whisper-large", chunk_length_s=30, generate_kwargs={'task': 'transcribe', 'language': 'no'})
 
 @spaces.GPU()
-def transcribe_audio(audio_file, batch_size=16): # sample_rate=16000
-    sample_rate, samples = wavfile.read(audio_file)
-    waveform, sample_rate = torchaudio.load(audio_file) # avoids TypeError here?
+def transcribe_audio(audio_file, batch_size=16, sample_rate=16000): 
+    audio_file = filepath.read(audio_file) 
+    waveform, sample_rate = torchaudio.load(audio_file) 
+    
+    if samples.ndim > 1:
+        samples = samples[0, :]
+
+    # --waveform to ndnumpy array
+    samples = waveform.numpy()
+    
+    start_time = time.time()
         
     # --convert to mono
     if len(samples.shape) > 1: