PyTorch extensions for fast R&D prototyping and Kaggle farming

Collection of awesome test-time (domain/batch/instance) adaptation methods

Image Test Time Augmentation with PyTorch!

Code Implementation for EmbedSeg, an Instance Segmentation Method for Microscopy Images

My modified version of YoloV5 training, cross-validation and inference with Pseudo Labelling pytorch pipelines used in GWD Kaggle Competition

Wheat detection using Faster RCNN

Engage in a semantic segmentation challenge for land cover description using multimodal remote sensing earth observation data, delving into real-world scenarios with a dataset comprising 70,000+ aerial imagery patches and 50,000 Sentinel-2 satellite acquisitions.

My modified version of EfficientDet training, cross-validation and inference with Pseudo Labelling pytorch pipelines used in GWD Kaggle Competition

Source code for the CIKM 2024 paper "Post-Training Embedding Enhancement for Long-Tail Recommendation."

ARC-Test-Time-Training (ARC-TTT)

Object Detection and Bounding Box Prediction using YOLO5 and EfficientDet , Image Augmentations and Test Time Augmentations

Test-Time Augmentation library for Pytorch

Image recognition used to distinguish between bees and wasps in photographs

Test time augmentation with Tensorflow keras models for segmentation tasks. This package also enables creation of keras layers for GPU acceleration

This project explores workplace safety compliance in construction sites by evaluating a Convolutional Neural Network (CNN) for binary classification of images into "safe" (individuals wearing safety equipment) and "unsafe" (individuals not wearing safety equipment).

Test Time Augmentation for Deep Learning Inference

CVPR 2024 (Seattle, USA) - CLVision Workshop

A modern implementation of ECCV 2018 paper: "Hierarchical Relational Networks for Group Activity Recognition and Retrieval".

Just another ResNet implementation trained on CIFAR-100 dataset achieving 73% test accuracy

[72hr Speedrun] Using DL to classify real anime images vs. AI-generated anime images, achieving 97.28% accuracy with attention-enhanced MobileNetV2. Submitted as my final term report in VNU-UET's Machine Learning course, June 2025.