543.543. Emotion Recognition from Audio

Emotion Recognition from Audio

In this problem, you will use a deep learning approach to classify audio clips into different emotions. You will need to use a pre-trained model and fine-tune it on a dataset of labeled audio clips.

**Example:** Given an audio clip, predict the emotion it represents.

**Constraints:** The input audio clip is a 10-second WAV file, and the model should output a probability distribution over 7 emotions: happy, sad, angry, fearful, surprised, disgusted, and neutral.

**Note:** You can use the Librosa library to load and preprocess the audio data, and the PyTorch library to build and train the model.