This is an official implementation for "Swin Transformer: Hierarchical Vision Transformer using Shifted Windows" on Object Detection and Instance Segmentation.

Extract video features from raw videos using multiple GPUs. We support RAFT flow frames as well as S3D, I3D, R(2+1)D, VGGish, CLIP, and TIMM models.

1st solution for the Webly-supervised Fine-grained Recognition competition in https://www.cvmart.net/race/10412/base

Pytorch implementation of some vision transformers, trained on CIFAR-10.

provide some new architecture, channel pruning and quantization methods for yolov5

[IEEE TGRS 2025] Be the Change You Want to See: Revisiting Remote Sensing Change Detection Practices

[Brainlesion 2022] Official PyTorch Code for Hybrid Window Attention Based Transformer Architecture for Brain Tumor Segmentation: Solution for FeTS 2022 Task 2

Master: Meta Style Transformer for Controllable Zero-Shot and Few-Shot Artistic Style Transfer

Project for coding information into images via teganography and cryptography.

Swin backbone for UNet network for semantic segmentation

This repository presents a radiodosiomics framework for personalized [¹⁷⁷Lu]Lu-PSMA-617 RLT in mCRPC. It includes feature selection and ML models using clinical, radiomic, and dosiomic features, plus nnU-Net and Swin UNETR DL models with SSL to predict Monte Carlo–based dose rate maps.

A Swin Transformer equivalent for 3D sparse volumes

PyTorch Lightning wrapper to make training classifiers easier.

This Repo contains code for used for age and gender detection in CCTV videos

The initial stage of this work focuses on creating a segmented image from a given satellite image with different classes. In the second stage, these images should be converted to an ABM-friendly format. At the final stage, they can be used to simulate the real-world environment with related agents using a suitable Agent-based model.

Java Gui ile veritabanına bağlanan kullanıcı kaydı uygulaması

M.Sc Lab Code

This project is an implementation of the classic international Checkers game developed using Java and Swift. The game features a comprehensive set of functionalities that includes saving and loading game states, undoing moves, and restarting the game, providing players with a smooth and engaging experience.