Update app.py

app.py

@@ -136,52 +136,29 @@ def decode_emotion_prediction(prediction_tensor, label_encoder):
     
     return predicted_emotion, confidence
 
-# device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-# model = model.to(device)
-
 def predict(wave):
-    wave = preprocess_single_audio(wave)
-    le = LabelEncoder()
-    le.classes_ = np.array(['Angry', 'Disgusting', 'Fear', 'Happy', 'Neutral', 'Sad'])
-    wave = wave.unsqueeze(0)
-    with torch.no_grad():
-        prediction = model(wave)
-    predicted_emotion, confidence = decode_emotion_prediction(prediction, le)
-    return f"Predicted emotion: {predicted_emotion} (Confidence: {confidence:.2f})"
+    if wave is None or wave == '':
+        return "No audio input provided."
+    try:
+        wave = preprocess_single_audio(wave)
+        le = LabelEncoder()
+        le.classes_ = np.array(['Angry', 'Disgusting', 'Fear', 'Happy', 'Neutral', 'Sad'])
+        wave = wave.unsqueeze(0)
+        with torch.no_grad():
+            prediction = model(wave)
+        predicted_emotion, confidence = decode_emotion_prediction(prediction, le)
+        return f"Predicted emotion: {predicted_emotion} (Confidence: {confidence:.2f})"
+    except Exception as e:
+        return f'Error in processing audio: {str(e)}'
 
 # Gradio Interface
-# iface = gr.Interface(
-#     fn=predict,
-#     inputs=gr.Audio(sources="microphone", type="filepath"),
-#     outputs="text",
-#     live=True,
-#     title="Speech Emotion Recognition",
-#     description="Record your voice and get the predicted emotion."
-# )
-
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    gr.Markdown("# Emotion Recognition App")
-    gr.Markdown("Upload an audio file or record directly to get a prediction")
-    
-    with gr.Row():
-        audio_input = gr.Audio(sources="microphone", type="filepath")
-        audio_output = gr.Audio(label="Processed Audio")
-    
-    with gr.Row():
-        submit_btn = gr.Button("Get Prediction", variant="primary")
-        clear_btn = gr.Button("Clear")
-    
-    prediction_output = gr.Textbox(label="Prediction")
-
-    submit_btn.click(
-        fn=predict,
-        inputs=[audio_input],
-        outputs=[audio_output, prediction_output]
-    )
-
-    clear_btn.click(
-        fn=lambda: (None, None, ""),
-        outputs=[audio_input, audio_output, prediction_output]
-    )
-
-demo.launch()
+iface = gr.Interface(
+    fn=predict,
+    inputs=gr.Audio(sources="microphone", type="filepath"),
+    outputs="text",
+    live=True,
+    title="Speech Emotion Recognition",
+    description="Record your voice and get the predicted emotion."
+)
+
+iface.launch()