Repolode
Submit
rasbt

rasbt/deeplearning-models

A collection of Jupyter Notebook examples implementing many deep learning architectures and training tips in PyTorch and TensorFlow.

View on GitHub
Stars
17,632
Forks
4,090
Open issues
8
Last commit
last month

What is it?

This repository is a set of annotated Jupyter Notebooks that implement classic and modern deep learning models, from perceptrons and multilayer perceptrons to CNNs such as LeNet, AlexNet, VGG, ResNet and MobileNet, plus RNNs, LSTMs, transformers (DistilBERT fine-tuning), GANs, autoencoders and graph neural networks.

Beyond architectures, it covers practical topics: ordinal regression losses (CORAL, CORN), normalization layers, K-fold cross-validation, data augmentation with AutoAugment and TrivialAugment, transfer learning, gradient clipping, learning-rate schedules, custom data loaders, multi-GPU training with DataParallel and pipeline parallelism, and Lightning-based workflows.

Most examples are written in PyTorch, with a smaller set in TensorFlow 1.x and PyTorch Lightning variants for many models. Each notebook targets a common dataset such as MNIST, CIFAR-10, CelebA or IMDB.

Key features

  • 01Implementations of CNN, RNN, transformer, GAN, autoencoder and GNN architectures in Jupyter Notebooks.
  • 02Most examples in PyTorch, with TensorFlow 1.x and PyTorch Lightning variants for many models.
  • 03Ordinal regression methods such as CORAL, CORN and squared-error reformulation.
  • 04Practical training topics including data augmentation, K-fold validation, and gradient clipping.
  • 05Multi-GPU examples using DataParallel and pipeline parallelism.

Use cases

  • Learning how a specific architecture such as ResNet or DistilBERT is built and trained.
  • Finding a runnable reference for ordinal regression or a particular loss function.
  • Looking up data loading or multi-GPU training patterns for PyTorch projects.
  • Comparing PyTorch and TensorFlow implementations of the same model.

Quick start

  1. 1

    Open a PyTorch notebook linked from the README, such as basic-ml/perceptron.ipynb, in Jupyter

Things to know

  • The README gives no installation or dependency instructions.
  • Many entries have no description (marked TBD), so you must read each notebook to learn what it does.

In collections

Computer vision libraries & toolsNatural language processing tools & libraries

Topics

#deep-learning#pytorch#tensorflow#jupyter-notebook#computer-vision#nlp#neural-networks#tutorial

FAQ

What is deeplearning-models?
A collection of Jupyter Notebook examples implementing many deep learning architectures and training tips in PyTorch and TensorFlow. This repository is a set of annotated Jupyter Notebooks that implement classic and modern deep learning models, from perceptrons and multilayer perceptrons to CNNs such as LeNet, AlexNet, VGG, ResNet and MobileNet, plus RNNs, LSTMs, transformers (DistilBERT fine-tuning), GANs, autoencoders and graph neural networks.
Is deeplearning-models free to use? What license does it have?
deeplearning-models is open source under the MIT license. Check the license terms before using it in commercial products.
Is deeplearning-models still maintained?
Updated within 6 months. The last commit to deeplearning-models was on Sep 26, 2026.
What can I use deeplearning-models for?
Learning how a specific architecture such as ResNet or DistilBERT is built and trained. Finding a runnable reference for ordinal regression or a particular loss function. Looking up data loading or multi-GPU training patterns for PyTorch projects. Comparing PyTorch and TensorFlow implementations of the same model.
How do I install deeplearning-models?
Open a PyTorch notebook linked from the README, such as basic-ml/perceptron.ipynb, in Jupyter
Where can I get deeplearning-models?
The source code and installation instructions are on GitHub: https://github.com/rasbt/deeplearning-models

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects