AI Model Compression (Quantization, Pruning and Knowledge Distillation)