ML-Projects

Project 0: Face Recognition and Classification

Project 1: Covid-19

• Data Visualization
• Prediction of Confirmed cases of various countries (just basic implementation)

Project 2: Handwritten digit recognition

• trained MNIST dataset on Neural Network with one hidden layer of 128 neurons
• using Loss Function: sparse categorical crossentropy, optimizer: Adam
• attaining an accuracy of ~98%

Project 3: Titanic Survival Classification

• develop the neural network from scratch
• Best accuracy of 80%

Project 4: Transfer Learning using MobileNet

• Image CLassification of Flowers
• used concept of Transfer Learning
• can attain accuracy of ~99%

Project 5: Image Classification using CNNs

• worked on fashion_mnist dataset
• implemented CNNs
• can attain accuracy of ~92.8%
• also tried to implement on CIFAR-10 dataset

Project 6: Face Verification (Siamese network)

• can be deployed as application for fae verification
• used Siamese network

Project 7: Sentiment Classification)

• used imdb reviews dataset
• implemented RNNs from scratch (vocabulary intialization, embedding, RNN architecture, training & updating parameters, Prediction)