This repository provides TensorFlow implementations of various deep learning architectures for CTR (Click-Through Rate) prediction in advertising. It includes models such as Wide&Deep, DeepFM, xDeepFM, Deep Interest Network, ESMM, AFM, FiBiNET, and Transformer-based architectures. The code uses tf.estimator API with tfrecord data format and tf.Dataset for efficient data loading, designed for industrial-scale applications.