Kartik Bali's repositories
YOLOV3_SSD
YOLOV3 vs SSD Object Detection on KITTI (tiny) dataset
PointNet_Semantic_Seg
Semantic segmentation on pointclouds using PointNet
DeepSDF-and-3D-EPN
3D shape reconstruction with DeepSDF and 3D EPN using the ShapeNet Dataset
Pedestrian-Distribution-using-VAE
Implemented and tested VAE for MNIST digit generation and then trained it to generate pedestrian distribution in a given space.
mlops-template
MLops template for deployment
DS-demo
Portfolio Project
2D-Sphere-Packing-using-Differential-Physics
Solution of 2D Sphere Packing Problem due to Overlapping Potentials using Differential Physics in PhiFlow
3D-Ray-Tracer
3D Ray Tracer App using SDL2 Media Layer
Topic-Modelling
Topic Modelling using BERTopic
LoRA-BERT
BERT_Sentiment_Classification
Vector-Field-Learning-using-RBFs
Task 3: Vector Function representing a dynamical system of points was learned using Radial Basis Functions. Task 5 : Taken's theorem was applied to create a time delay embedding. The PCA version of this curve was learnt by the RBFs to predict MI building utilization for longer duration.
heuristics
Copilot assisted algorithms and heuristics
Pedestrian_sim
A GUI based notebook for visualizing crowd simulation based on Cellular Automata Approach
MLCMS-exercises
Exercises for the course of Machine Learning in Crowd Modelling and Simulation @ TUM
Smoke-Reconstruction-using-Density-Priors
Trained a neural network that reconstructs smoke simulation from an only partially observed sequence of state,s which are the x- and y-projections of the smoke density over time, the full physical states.
1D-Kuramoto-Sivashinsky-Equation
Solution to 1D KS equation using Convolutional Neural Networks in Phiflow.
3DCNN-and-PointNet
Implementations of 3DCNN and PointNet Models for 3D Shape Classification and Segmentation
Aerial-Semantic-Segementation
Semantic Segmentation for 2D Aerial Drone Imagery using a Unet model
invertible-neural-networks
Tutorials for invertible neural networks.