Update app.py

app.py

@@ -85,8 +85,13 @@ if audio_file:
     # ================================
     # ✅ Fast Transcription Processing with Conformer
     # ================================
+    # Convert waveform into the required format
     inputs = processor(denoised_waveform.numpy(), sampling_rate=sr, return_tensors="pt", padding=True).to("cuda" if torch.cuda.is_available() else "cpu")
 
+    # Make sure the input has batch dimension (even if it's one example)
+    if len(inputs.input_values.shape) == 1:
+        inputs.input_values = inputs.input_values.unsqueeze(0)
+
     with torch.no_grad():
         logits = model(**inputs).logits