Update app.py

app.py

@@ -3,6 +3,8 @@ import librosa
 from speechbrain.inference.classifiers import EncoderClassifier
 from pydub import AudioSegment
 import gradio as gr
+import uuid
+import os
 
 # Load model once
 classifier = EncoderClassifier.from_hparams(
@@ -11,15 +13,24 @@ classifier = EncoderClassifier.from_hparams(
 )
 
 def classify_accent(video):
+    # Generate unique filename
+    temp_wav = f"/tmp/{uuid.uuid4().hex}.wav"
+
+    # Convert to .wav
     audio = AudioSegment.from_file(video, format="mp4")
-    audio.export("output.wav", format="wav")
+    audio.export(temp_wav, format="wav")
 
-    waveform, sr = librosa.load("output.wav", sr=16000, mono=True)
+    # Load waveform
+    waveform, sr = librosa.load(temp_wav, sr=16000, mono=True)
     waveform_tensor = torch.tensor(waveform).unsqueeze(0)
 
+    # Predict
     prediction = classifier.classify_batch(waveform_tensor)
     _, score, _, text_lab = prediction
 
+    # Cleanup
+    os.remove(temp_wav)
+
     return f"Accent: {text_lab[0]} (Confidence: {score.item():.2f})"
 
 app = gr.Interface(
@@ -32,3 +43,4 @@ app = gr.Interface(
 
 if __name__ == "__main__":
     app.launch()
+