Deep-Learning

Why we included this project

Anyone learning deep learning in Python will find this repo a solid place to build real skills. It pairs working code with writeups for several concrete projects, including handwritten digit recognition, Discuz captcha recognition, garbage classification, and a UNet-based semantic segmentation series, so the theory in each lesson is backed by something you can actually run and train. The material is grouped by framework, letting you start with perceptrons and backpropagation or jump straight into the PyTorch segmentation tutorials. Code and explanations sit side by side, which suits a first model far better than reading theory alone. Note that the tutorials are written in Chinese, and because this is a curated collection of lessons rather than a library, treat it as a reference to adapt rather than production-ready software.