Research Tools: code, datasets, & models

This repository hosts samples that demonstrate how to use Trill, a high-performance one-pass in-memory streaming analytics engine from Microsoft Research. It can handle both real-time and offline data, and is based on a temporal data…

Our objective is to develop a library of efficient machine learning algorithms that can run on severely resource-constrained edge and endpoint IoT devices ranging from the Arduino to the Raspberry Pi. Please see our GitHub (opens in new…

The bistring library provides non-destructive versions of common string processing operations like normalization, case folding, and find/replace. Each bistring remembers the original string, and how its substrings map to substrings of the modified version.

Code on Github to reproduce data in Bias in Bios Paper

A high-resolution convolutional network backbone, better than ResNets. Many computer vision applications: ImageNet classification, semantic segmentation, object detection, and facial landmark detection.

The project is an official implementation of our CVPR2019 paper “Deep High-Resolution Representation Learning for Human Pose Estimation”

Deep Learning Based Real-Time Object Tracking

MSParS is a large-scale dataset for the open domain semantic parsing task. The whole dataset consists of 81,826 samples annotated by native English speakers. We randomly shuffle these samples and use 80% of them (63,826)…

The ELL models repository contains a number of pretrained models suited for different device footprints. Currently, there are pretrained models for image classification and audio keyword spotting.

Scripts to generate the knowledge-grounded conversation dataset. First used for DSTC7-Task2 challenge.