Update app.py

app.py

@@ -55,9 +55,11 @@ cosine_sim = torch.nn.CosineSimilarity(dim=-1)
 def similarity_fn(path1, path2):
     if not (path1 and path2):
         return '<b style="color:red">ERROR: Please record audio for *both* speakers!</b>'
-
-    wav1, _ = apply_effects_tensor(*torchaudio.load(path1), EFFECTS)
-    wav2, _ = apply_effects_tensor(*torchaudio.load(path2), EFFECTS)
+    
+    wav1, sr1 = librosa.load(path1, mono=True)
+    wav1, _ = apply_effects_tensor(torch.tensor(wav1).unsqueeze(0), sr1, EFFECTS)
+    wav2, sr2 = librosa.load(path2, mono=True)
+    wav2, _ = apply_effects_tensor(torch.tensor(wav2).unsqueeze(0), sr2, EFFECTS)
     print(wav1.shape, wav2.shape)
 
     input1 = feature_extractor(wav1.squeeze(0), return_tensors="pt", sampling_rate=16000).input_values.to(device)