Dmitry Baranchuk

ivf-hnsw. Code for ECCV2018 paper: Revisiting the Inverted Indices for Billion-Scale Approximate Nearest Neighbors

memory-efficient-maml. Memory efficient MAML using gradient checkpointing

learning-to-route. Code for ICML2019 paper: Learning to Route in Similarity Graphs

VisualGenAI. Materials for the VisualGenAI course at YSDA2026

nns-meets-deep-rl. Code for paper: Towards Similarity Graphs Constructed by Deep Reinforcement Learning

human-pose-annotator. Web Service for Human Pose Annotation

fish-detection. Kaggle Competition NCFM (Top 1% solution)

bird-classification. Bird Classification using VGG16-pretrained model (Caffe and Keras implementations)

animal-segmentation. Animal Segmentation using Markov Random Field

height-estimation. Python

lstm-cuda. This is a c++ implementation of an LSTM Neural Network parallelized for a GPU using CUDA

bfs_swig. C

hnsw. C++

mae. PyTorch implementation of MAE https//arxiv.org/abs/2111.06377

search_hnsw_swig. C

multi-person-pose. Jupyter Notebook

facial-keypoints-detection. Facial Keypoints Detection using a simple custom CNN

human-pose-estimation. This repository implements a demo of the Human pose estimation via Convolutional Part Heatmap Regression paper.

pose-hg-demo. Code to test and use the model from "Stacked Hourglass Networks for Human Pose Estimation"