0.13 Data Preprocessing

Описание: Get ready for a hands-on introduction to data preprocessing for both audio (signal) and image data, tailored for beginners in deep learning!

In this instructional video, you’ll learn:

Why data preprocessing is important for machine learning and deep learning projects.

Key steps for audio data: Downloading and cleaning raw speech files, extracting meaningful features (like Mel spectrograms and MFCCs), normalizing features, and using data augmentation (like time/frequency masking) to build more robust audio models.

Core image preprocessing concepts: Downloading and organizing images, resizing, creating datasets with PyTorch, normalization, and a wide variety of augmentation techniques (flipping, rotation, color jitter, and more).

Practical, code-based walkthroughs for each step, with visualizations that show how preprocessing and augmentation transform your data.

🔗 Materials

Link to slides and notebook used:
https://github.com/CMU-IDeeL/CMU-IDee...

👍 Stay Connected
Subscribe and hit the 🔔 to get notified when new Recitation‑0 videos premiere each weekday at 1 PM ET.