Deep Learning from Scratch

This repository contains the code to build popular deep learning models from scratch using only tensors from PyTorch, and minimal usage of other libraries.

Models

• Convolutional Neural Network (CNN)
  • LeNet5
  • AlexNet
  • VGGNet
  • ResNet
  • Image Captioning
  • Show, Attend and Tell
• Generative Models
  • LSTM
  • GAN + DCGAN
  • VAEs
  • Diffusion Models
• Transformers
  • Vanilla Transformer
  • BERT
  • Vision Transformer (ViT)
  • Masked Autoencoders (MAE)
  • GPT