We just published Deep Learning Crash Course (No Starch Press), a project based intro where you build modern models from scratch in PyTorch. It goes from basic neural nets all the way to CNNs, transformers, diffusion models, GNNs, plus RL and active learning. Free sample chapter here: https://nostarch.com/download/samples/DeepLearningCrashCours... and full free companion code here: https://github.com/DeepTrackAI/DeepLearningCrashCourse/ If you’re learning DL and want to build (not just read), this is for you. Looking forward to hearing what you think!