Add Example Notebook for Training Custom Watermarking Model on Public Dataset

[hastagAB]

Add a new example notebook that provides a comprehensive, step-by-step guide to training a watermarking model using AudioCraft and Dora. The example uses the Librispeech dataset and covers all essential steps, including:

• 🚀 Setting up the environment and installing dependencies (Torch, Torchaudio, Hydra, Dora, AudioCraft, etc.)
• 📂 Preparing the Librispeech dataset for training, including manifest creation in AudioCraft format.
• 📄 Creating YAML configurations for dataset and training management.
• 🏋️ Running a single-epoch training job using Dora to demonstrate the training process.
• 📊 Evaluating the trained model checkpoint for performance assessment.

This notebook aims to help users get started with training and evaluating watermarking models effectively, using the robust capabilities provided by AudioCraft and Dora.

🚫 The notebook was prepared on Google Colab which has resources limitations. his code is designed to run effectively on systems with higher computational resources.