DreamScene: Compositional and Object-Centric Text-to-3D Scene Generation with background image LLM-Generated Supervision for Egocentric Vision-based Legged Locomotion Policies An exploration of the capability of LLMs to be used as teacher policies for vision-based legged locomotion. NeRF and NeRS A simple implementation of neural radiance fields (NeRFs) and neural implicit surfaces (NeRS). 3D Perception Exploring the performance of models for pointcloud perception. Single View 3D An implementation of various deep learning models that predict voxel, pointcloud, and mesh representations of 3D objects from a single RGB image. Surface Reconstruction from Monocular Video An implementation of structure from motion on video captured by an iPhone camera. Text to 3D A simple implementation of text to 3D diffusion using score distillation sampling with NeRF and Gaussian splat representations. DeepDeepBlogger: An Academic Paper to Blogpost Converter An academic paper to blogpost converter using text-to-text transformers. Autonomous Car Control A longitudinal and lateral control system developed for an autonomous car. Pet Food Dispenser An automatic pet food dispenser powered by an Arduino.
