Minor Excerises

1. Use of PyTorch
2. Implemented Gradient Descent
3. Designing NN with Achitecture and Activation functions
4. Training NN with different techniques like regularisation, early stopping, dropout, ranmdom restart, etc.

Major Project 1: Developing a Handwritten Digits Classifier with PyTorch

• Used MLP to classify MINST data with accuracy 97%

Minor Excerises

1. Convolution Layer Visulization, Pooling Layer Visualization, CNNs for CIFAR Image Classification, Improving Performance using BatchgNorm and Image Augmentation and Exporting for Production
2. Transfer Learning
3. Linear Autoencoders, Convolutional Autoencoders, Denoising Autoencoders
4. Object Detection using RetinaNet, Semantic Segmentation using UNet

Major Project 2: Landmark Classification & Tagging for Social Media

• Created a CNN to classify landmarks from scratch.
• Used Transfer Learning with resnet18 to classify same landmarks.
• Deployed an app

Major Project 3: Text Translation and Sentiment Analysis using Transformers

Minor Excerises

1. MNIST GAN
2. DCGAN Generator Discriminator, Training, Frechet Inception Distance
3. CycleGAN
4. ProGAN, StyleGAN

Major Project 4: Face Generation Project

• Build a custom generative adversarial network to generate new images of faces.
• Used a dataset of high-resolution images of "celebrity" faces.