Contents

Basics

• QuickStart [File | Notebook] - QuickStart gives general overview of Basics section.
• Tensors [File | Notebook] - Operations on tensors, numpy arrays and casting them to tensor or vice versa.
• Datasets and DataLoaders [File | Notebook] - Creating datasets and dataloaders.
• Transforms [File | Notebook] - Torchvision's augmentation methods and using them together.
• Build Neural Network [File | Notebook] - Building simple neural network from scratch.
• Optimizers, Model Save and Load [File | Notebook] - Creating optimizer, saving trained model and loading for inference.

Beginner

• Autograd and Freeze weights - Autograd: Automatic Differentiation.
• Neural Networks - To look deep into Neural Networks.

Intermediate

• Convolutional Neural Network - Classifying images of CIFAR10 using CNNs.
• Residual Neural Network - Using Residual Blocks to build a CNN for image classification.
• Recurrent Neural Network - Image classification using RNN networks.
• Long Short Term Memory - MNIST digit classifier using LSTM.
• Variational Auto Encoder - Reconstructing MNIST data samples using VAE.

Advanced

• Adversarial Generative Network - Deep Convolutional Generative Adversarial Network.
• Fast Neural Style Transfer - Transferring image with different styles.
• Super Resolution. Running it Using ONNX Runtime - Inference using ONNX Runtime.

Updating...

References

1. https://github.com/pytorch/examples
2. https://www.deeplearningwizard.com