Deep-Learning-Experiments

Why we included this project

Self-taught learners get a coherent route through deep learning here, with each topic pairing a lecture note, a video where one exists, and a runnable PyTorch notebook. The sequence moves from supervised learning basics through CNNs, RNNs, transformers, and the newer Mamba architecture, and the small MNIST demos let you train a model quickly rather than just reading theory. A 2025 revision adds material like SAM2 segmentation, so the content tracks recent developments instead of freezing at an older snapshot. Instructors assembling a course and individuals who want to revisit a specific architecture before a real project will both find the experiments compact and easy to pick up.