Update asr.py

asr.py

@@ -20,15 +20,15 @@ def resample_to_16k(audio, orig_sr):
 
 def transcribe(audio):
     print(audio)
-    audio = librosa.load(audio, sr=16_000, mono=True)[0]
-    print("After loading: ",audio)
+    # audio = librosa.load(audio, sr=16_000, mono=True)[0]
+    # print("After loading: ",audio)
     sr,y = audio
     y = y.astype(np.float32)
     y /= np.max(np.abs(y))
     y_resampled = resample_to_16k(y, sr)
     print("Without using librosa to load:",y_resampled)
-    inputs = processor(audio, sampling_rate=16_000,return_tensors="pt")
-    # inputs = processor(y_resampled, sampling_rate=16_000,return_tensors="pt")
+    # inputs = processor(audio, sampling_rate=16_000,return_tensors="pt")
+    inputs = processor(y_resampled, sampling_rate=16_000,return_tensors="pt")
     with torch.no_grad():
         tr_start_time = time.time()
         outputs = model(**inputs).logits