Mansi Khamkar's repositories
Stutter-Therapy
Developed and trained Gated-CNN models to detect types of stutter in speech and SVM classifier to suggest new therapies to the user according to his stutter type and severity
Basic_GANs
Implemented and experimented with GAN, Conditional GAN and Deep-Convolutional GAN on various datasets for comparison, learning and demonstration purposes
LogisticReg
Implemented logistic regression from scratch to train on sign language digits dataset and titanic dataset using one-vs-one and one-vs-all algorithms
SiameseNet
Implemented and trained Siamese network for image classification using contrastive loss, triplet loss and regularized cross-entropy loss
AOS_reports
Reports on various topics in Advanced Operating Systems
AutoEncoder
Implemented and trained an Auto-encoder to reconstruct images, compared them with images reconstructed using PCA and conducted various experiments using PyTorch
BackPropagation
Implemented back-propagation algorithm on a neural network from scratch using Tanh and ReLU derivatives and performed experiments for learning purpose
CNN_classifier
Implemented and trained CNN using PyTorch, to classify images and performed transfer learning using VGG16
FIFA_Clustering
Performed KMeans, Agglomerative, Divisive, DBSCAN clustering on FIFA dataset along with outlier detection and cluster analysis
GradientDescent
Implemented gradient descent algorithm and its variants from scratch and visualized their results by training models, for comparison and learning purposes
KMeans_on_imgs
Implemented KMeans from scratch and trained it on Fashion-MNIST dataset by experimenting with initializaion methods like forgy, random partitions, kmeans++ and found the optimal number of clusters by implementing elbow & silhouette algorithms from scratch
mansikhamkar
IIITH - SSD Assignment 2 - Portfolio webpage
MLP_classifier
Implemented a fully functioning Multi-Layer-Perceptron classifier for image classification and conducted various experiments using PyTorch
multiway2P_mergesort
Implemented the two-phase merge sort algorithm to sort a large number of records
PHP_workshop
Simple PHP webpages created for teaching purposes
PySpark_MapReduce
Implemented programs for searching in a dataset using Spark architecture to enable distributed search
Social-Champ
Performed data retrieval from Facebook and Twitter feeds. Filtered and ranked social influencers, considering various aspects, who could help in promoting NGOs.
SortMerge_Join
Implemented the two-phase merge sort algorithm to sort and then join 2 large SQL relations