DeepInterests

Why we included this project

DeepInterests is a Chinese-language deep learning course built around roughly thirty complete, runnable projects, and that emphasis on doing rather than just reading is what makes it useful. Beginners and intermediate practitioners can follow along as the code walks through image style transfer, GANs and WGANs, image captioning, seq2seq translation, object detection, and even a DQN-trained FlappyBird agent, all wired together with Python, TensorFlow 1.x, and Keras. Each project pairs working code with a written lesson in the included PDF and an online Zhihu column, so the material is easy to pick up on your own. The catch is that the code targets TensorFlow 1.9 and Keras 2.2 from 2018, so expect to treat it as a conceptual and architectural reference rather than something that runs on current frameworks.