Spaces:
Running
Running
| import gradio as gr | |
| from transformers import pipeline | |
| import soundfile as sf | |
| import os | |
| # --- Model Loading --- | |
| # We switched to 'superb/wav2vec2-base-superb-er' as it's a well-established and public model for emotion recognition. | |
| try: | |
| classifier = pipeline("audio-classification", model="superb/wav2vec2-base-superb-er") | |
| except Exception as e: | |
| # If there's an error during model loading, we can display it in the Gradio interface | |
| # This helps in debugging issues directly on the Hugging Face Space. | |
| def error_fn(audio_file): | |
| return {"error": f"Failed to load the model. Please check the logs. Error: {str(e)}"} | |
| classifier = None | |
| # --- Prediction Function --- | |
| def predict_emotion(audio_file): | |
| """ | |
| Predicts emotions from an audio file. | |
| Args: | |
| audio_file (str or tuple): Path to the audio file (from upload) or a tuple | |
| (samplerate, audio_array) from microphone input. | |
| Returns: | |
| dict: A dictionary of emotion labels and their probabilities. | |
| """ | |
| # Handle case where the model failed to load | |
| if classifier is None: | |
| return {"error": "The AI model could not be loaded. The application cannot start."} | |
| if audio_file is None: | |
| return {"error": "No audio input provided. Please upload a file or record."} | |
| # Gradio's Audio component can return a path to a temp file for file uploads, | |
| # or a tuple (samplerate, numpy_array) for microphone input. | |
| if isinstance(audio_file, str): | |
| # Handle file path (e.g., from file upload) | |
| audio_path = audio_file | |
| elif isinstance(audio_file, tuple): | |
| # Handle microphone input (samplerate, numpy_array) | |
| sample_rate, audio_array = audio_file | |
| # Save the numpy array to a temporary WAV file as the pipeline expects a file path or direct bytes | |
| temp_audio_path = "temp_audio_from_mic.wav" | |
| sf.write(temp_audio_path, audio_array, sample_rate) | |
| audio_path = temp_audio_path | |
| else: | |
| return {"error": f"Invalid audio input format: {type(audio_file)}"} | |
| try: | |
| # Perform inference | |
| results = classifier(audio_path, top_k=5) # top_k ensures we get all relevant emotion scores | |
| # Process results into a dictionary for better display | |
| emotion_scores = {item['label']: round(item['score'], 3) for item in results} | |
| return emotion_scores | |
| except Exception as e: | |
| return {"error": f"An error occurred during prediction: {str(e)}"} | |
| finally: | |
| # Clean up temporary file if created | |
| if 'temp_audio_path' in locals() and os.path.exists(temp_audio_path): | |
| os.remove(temp_audio_path) | |
| # --- Gradio Interface --- | |
| # Define the Gradio interface | |
| iface = gr.Interface( | |
| fn=predict_emotion, | |
| # THIS IS THE CORRECTED LINE: | |
| inputs=gr.Audio(sources=["microphone", "upload"], type="filepath", label="Upload Audio or Record with Microphone"), | |
| outputs=gr.Label(num_top_classes=5, label="Emotion Probabilities"), # This model has 4 emotions + 'no-emotion' | |
| title="AI Audio Emotion Detector", | |
| description="Upload an audio file or record your voice to detect emotions. This model is trained to recognize 'anger', 'happiness', 'neutral', 'sadness', and 'no-emotion'.", | |
| examples=[ | |
| # You can add example audio files to your Hugging Face Space and reference them here. | |
| # For now, we'll leave this empty. | |
| ] | |
| ) | |
| # Launch the Gradio app | |
| if __name__ == "__main__": | |
| iface.launch() |