Label Studio is a multi-type data labeling and annotation tool with standardized output format

Data labeling react app that is backend agnostic and can be embedded into your applications — distributed as an NPM package

Label data using HuggingFace's transformers and automatically get a prediction service

A Streamlit component integrating Label Studio Frontend in Streamlit applications

ML backend for the Label Studio tool. The backend uses the YOLOv8 model for instance image segmentation or object detection.

Custom YOLOv8 backend for Label Studio

Full-fledged Data Exploration Tool for Label Studio

Fine tuning YoloV7 to detect white, red bloodcells and platelets to be used as backend in label studio for pre annotating

Exploring NLP weak supervision approaches to train text classification models. The project is also a prototype for a semi-automated text data labelling platform. Approaches: Snorkel and Zero-Shot Learning.

An AI-aided image segmentation ML-Module for Heartexlab/Label-Studio. Easy to deploy. Great to use.

Определение количества позиций товара на витрине по фотографиям. (label-studio, yolov5, torch, rabbitmq, pika, docker-compose)

Annotation assistent tool that uses CLIP to find described objects in dataset and label them

This small module connects Label Studio with Fonduer by creating a fonduer labeling function for gold labels from a label studio export. Documentation: https://irgroup.github.io/labelstudio-to-fonduer/

Developed a task automation program using API calls for Label Studio, reducing task assignment time from 3.41 hours to 10 seconds—a 99% reduction. The system managed approximately 24,000 tasks daily, with accurate logging in a file, demonstrating strong attention to detail and scalability.

Create ready-to-use Label Studio pre-populated JSON files from popular OCR formats.

Implementing Incremental Learning In Label Studio Using River ML Model

A simple python-script to augment an annotated dataset in JPG/XML Format as used by LabelIMG (https://github.com/tzutalin/labelImg) and now Label Studio. The script will rotate the images 4 times and mirror the resulting images and the annotations making for an expansion by the factor of 8.

Measure agreement between chart reviewers.

Explore and demo label-studio on OpenShift

datalabel is a UI-based data editing tool that makes it easy to create labeled text data in a dataframe. With datalabel, you can quickly and effortlessly edit your data without having to write any code. Its intuitive interface makes it ideal for both experienced data professionals and those new to data editing.