aditisingh02/deepfake-audio-detection ? reverse-engineered prompt

Build me a simple Streamlit app that can tell whether an uploaded voice recording is real or deepfake. It should let me upload a WAV file, analyze it with a trained CNN model, and then clearly show the result, the confidence score, and a few visualizations like the spectrogram and other audio feature plots so I can see what the model is looking at.

Also include the training and testing scripts so I can retrain the model on my own dataset and run predictions from the command line too. Use the dataset layout from the README with real and fake audio folders, and make the app work with the pre trained model if it’s available. Keep the interface clean and easy to use, with fast inference and a friendly message explaining the result. If you need to, look up current docs online for Streamlit or TensorFlow details.