DepthStream Accelerator: A TensorRT-optimized monocular depth estimation tool with ROS2 integration for C++. It offers high-speed, accurate depth perception, perfect for real-time applications in robotics, autonomous vehicles, and interactive 3D environments.

implementation of the ViT model with the help of torch.

Steel bar alignment detection system using Vision Transformers and KNN-based pseudo-labeling for real-time image classification.

Advanced Vision Transformer implementation that achieves 97.50% CIFAR-10 accuracy with minimal overfitting (2.07% gap) through strategic regularization. Features real-time inference, custom image uploads, confidence analysis, and demonstrates effective fine-tuning of pre-trained transformers for small datasets.