.

README.md

@@ -62,7 +62,7 @@ python app.py
 2. Open your web browser and navigate to the provided URL (typically http://localhost:7860)
 
 3. Choose your input method:
-   - Upload an audio file
+   - Upload an audio file (supports .mp3, .wav, .ogg, .flac)
    - Enter lyrics text
 
 4. Adjust generation parameters:
@@ -75,7 +75,7 @@ python app.py
 ## Models Used
 
 - **Genre Classification**:
-  - Audio: `anton-l/wav2vec2-base-superb-gc`
+  - Audio: `superb/wav2vec2-base-superb-gc` (Pre-trained on music genre classification)
   - Text: `facebook/bart-large-mnli` (Zero-shot classification)
 - **Lyric Generation**: `gpt2-medium`
 

src/classifier.py

@@ -13,10 +13,10 @@ class MusicGenreClassifier:
             model="facebook/bart-large-mnli"
         )
         
-        # For audio classification, we'll use a pre-trained model
+        # For audio classification, we'll use a different pre-trained model
         self.audio_classifier = pipeline(
             "audio-classification",
-            model="anton-l/wav2vec2-base-superb-gc"
+            model="superb/wav2vec2-base-superb-gc"
         )
         
         self.genres = [
@@ -29,7 +29,11 @@ class MusicGenreClassifier:
         try:
             # Load audio using librosa (handles more formats)
             waveform, sample_rate = librosa.load(audio_path, sr=16000)
-            return torch.from_numpy(waveform)
+            # Convert to torch tensor and ensure proper shape
+            waveform = torch.from_numpy(waveform).float()
+            if len(waveform.shape) == 1:
+                waveform = waveform.unsqueeze(0)
+            return waveform
         except Exception as e:
             raise ValueError(f"Error processing audio file: {str(e)}")
 
@@ -37,8 +41,10 @@ class MusicGenreClassifier:
         """Classify genre from audio file."""
         try:
             waveform = self.process_audio(audio_path)
-            predictions = self.audio_classifier(waveform)
+            predictions = self.audio_classifier(waveform, top_k=1)
             # Get the top prediction
+            if isinstance(predictions, list):
+                predictions = predictions[0]
             top_pred = max(predictions, key=lambda x: x['score'])
             return top_pred['label'], top_pred['score']
         except Exception as e: