Add Darija transcription and topic extraction app6

app.py

@@ -13,8 +13,9 @@ transcription_model = Wav2Vec2ForCTC.from_pretrained("boumehdi/wav2vec2-large-xl
 summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
 
 # Topic Classification Model (BERT for example)
-topic_model = BertForSequenceClassification.from_pretrained("your-finetuned-topic-model")
-topic_tokenizer = BertTokenizer.from_pretrained("your-finetuned-topic-model")
+topic_model = BertForSequenceClassification.from_pretrained("bert-base-uncased")  # Example model
+topic_tokenizer = BertTokenizer.from_pretrained("bert-base-uncased")
+
 
 # Function to resample audio to 16kHz if necessary
 def resample_audio(audio_path, target_sr=16000):