Regularization and Training Diagnostics

teach why Keras models fail and how to repair them carefully.

Units

1. Unit 06.00: 06.00 Underfitting, overfitting, and instability
2. Unit 06.01: 06.01 Dropout, weight decay, and early stopping
3. Unit 06.02: 06.02 Batch normalization and normalization layers
4. Unit 06.03: 06.03 Input scaling and preprocessing layers
5. Unit 06.04: 06.04 Reading learning curves and confusion matrices
6. Unit 06.05: 06.05 Project step: failed-training repair report

Module work

• Activity
• Knowledge check
• Summary