muhammadtahanasir/deepfoundry ? reverse-engineered prompt

Build me a polished GitHub style deep learning curriculum archive called DeepFoundry.

It should organize an 18 week learning path across 6 modules, from basic neural networks and backpropagation through CNNs, RNNs, self supervised learning, GANs, VAEs, and graph neural networks. I want each module to feel like a clear learning journey, with lesson notebooks or pages that explain the idea in simple language, show the math step by step, and include clean PyTorch examples that actually run.

Please make the repo easy to browse, with a strong README, a curriculum overview, and neat section pages for each module and week. Keep the style professional and educational, like something someone could use to study from or teach with. Include room for diagrams, code examples, and interview style review questions. If anything needs current best practice docs or references, look them up online.