DLS1 - Deep Learning School (MIPT) notebooks

[🎓 DLS / MIPT] [🧠 Deep Learning] [📓 Notebooks] [🖼️ Computer Vision] [🧬 Generative]

A study project of homework notebooks completed during the Deep Learning School program at the Moscow Institute of Physics and Technology (MIPT, Faculty of Applied Mathematics and Informatics).

✨ Focus

Classical ML models, optimization, computer vision, and generative methods, all in practical notebook form.

📚 Notebooks

• notebooks/hw_1_game_of_thrones.ipynb - EDA and preprocessing, training baseline models, evaluation, and submission.csv.
• notebooks/hw_2_linear_models.ipynb - gradient descent, batching, logistic regression, L1/L2 and elastic-net regularization.
• notebooks/hw_3_kaggle.ipynb - Kaggle project: data exploration, feature processing, metric, and report.
• notebooks/hw_4_conv_cnn.ipynb - models from scratch: moons -> MNIST, dataloaders, training, and inference.
• notebooks/hw_5_simpsons_classification.ipynb - image classification (Simpsons), full ML pipeline from data to model.
• notebooks/hw_6_semantic_segmentation.ipynb - segmentation: IoU, BCE loss, SegNet, training, and inference.
• notebooks/hw_7_homework_detection.ipynb - detection: backbone/neck/FPN/head, label assignment, DIoU.
• notebooks/hw_8_autoencoders.ipynb - autoencoders and VAE: architectures, training, sampling, conditional VAE.
• notebooks/hw_9_gans_part_1.ipynb - GANs: data prep, training, generation, and 1-NN evaluation.
• notebooks/hw_9_gans_part_2.ipynb - StyleCLIP: image editing, CLIP/ID loss, latent optimization.

▶️ Run

Open notebooks in Jupyter or Colab. Dependencies and run steps are described inside each notebook (some require a GPU).