Data & Analytics

pytorch_active_learning is an open-source PyTorch library designed for active learning, accompanying the "Human-in-the-Loop Machine Learning" book. It offers a range of active learning methods, including Least Confidence, Margin of Confidence, Ratio of Confidence, and Entropy sampling. The library also supports more advanced techniques like Model-based Outlier sampling, Cluster-based sampling, and various forms of Active Transfer Learning. It is suitable for researchers and practitioners looking to experiment with and apply active learning strategies in computer vision and natural language processing, with a focus on real-world diversity to avoid bias. The code is stand-alone and can be easily integrated with existing PyTorch installations.

WenetSpeech offers a comprehensive 10000+ hour multi-domain Chinese corpus specifically designed for speech recognition tasks. This extensive dataset is compiled from YouTube and Podcast sources, utilizing both Optical Character Recognition (OCR) and Automatic Speech Recognition (ASR) techniques for labeling. To ensure high quality, the corpus undergoes a novel end-to-end label error detection method for validation and filtering. It categorizes data into High Label, Weak Label, and Unlabel sets, suitable for supervised, semi-supervised, or unsupervised training. The dataset also provides various training subsets (S, M, L) and evaluation sets (DEV, TEST_NET, TEST_MEETING) to support diverse ASR system development and benchmarking. Access to the dataset requires visiting the official website, agreeing to the license, and obtaining a password.

CLUENER2020 offers a PyTorch implementation of various models for Named Entity Recognition (NER), focusing on Chinese language tasks. It includes baseline code for the CLUENER2020 competition, featuring models like BiLSTM-CRF, BERT-base with Softmax/CRF/BiLSTM+CRF, and Roberta with Softmax/CRF/BiLSTM+CRF. The project utilizes the CLUENER2020 dataset, a Chinese fine-grained NER dataset derived from THUCNEWS, with 10 distinct categories such as organization, person name, and address. Users can configure model parameters and other hyperparameters, and the repository provides instructions for setting up the environment and running the models. It also includes pre-trained BERT and Roberta models for convenience.

SKY ENGINE AI revolutionizes computer vision with its Synthetic Data Cloud, offering deep learning solutions and synthetic data generation. The platform allows for the creation of robust, reliable synthetic datasets to train and validate accurate Vision AI models. Key features include 3D keypoints for pose estimation, semantic masks and bounding boxes for precision, depth maps for geometric context, and normal maps for surface orientation. It utilizes physically-based rendering and ray tracing to reduce the domain gap and offers sensor simulation for multimodal training. This platform is designed to accelerate AI development, improve model accuracy, and reduce costs for various industries.

FaceSymAI offers a free AI-powered facial symmetry checker that allows users to upload a photo and receive an analysis of their facial symmetry. The tool utilizes advanced image processing and AI algorithms to pinpoint facial features like eyes, nose, mouth, and overall facial structure. It then applies mathematical and statistical methods to evaluate symmetry across seven feature pairs: eyes, eyebrows, ears, nose, mouth, and cheeks. The platform emphasizes user privacy, stating that uploaded photos are not permanently stored and are used solely for analysis. FaceSymAI aims to make this service accessible to everyone, supporting its operations through advertising.

Luel is a two-sided marketplace designed to facilitate the exchange of high-quality AI training data, specifically focusing on video and audio content. It connects AI development teams seeking specific datasets with contributors who can provide the necessary video, audio, and image content. The platform ensures that all training data is curated, rights-cleared, and verified, making it suitable for commercial use. Enterprises can access a catalog of premium datasets or request custom data collection campaigns, benefiting from enterprise-grade quality and compliance. Contributors, on the other hand, can upload their content, such as cooking tutorials, voice samples, or product photos, to earn income with fast payouts and fair rates, without upfront costs.

DataDreamer is a powerful open-source Python library designed for prompting, synthetic data generation, and training workflows. It enables users to create and run complex, multi-step prompting workflows with major open-source or API-based LLMs. The library facilitates the generation of synthetic datasets for novel tasks or the augmentation of existing datasets using LLMs. Additionally, DataDreamer supports various model training processes, including fine-tuning, instruction-tuning, and distillation, on both existing and synthetic data. It emphasizes simplicity, efficiency through aggressive caching and resumability, and reproducibility, making it suitable for research-grade projects and easy sharing of workflows, datasets, and models.

FaceAge AI is an online tool that uses advanced AI to analyze uploaded photos and estimate various aspects of facial age, including facial age, eye age, skin age, and wrinkle age. Users can upload clear, front-facing photos in PNG, JPG, WEBP formats up to 10MB for instant analysis. The tool provides a detailed report on different age metrics and offers tips on how to potentially look younger. It emphasizes privacy, stating that photos are processed instantly and deleted immediately after analysis, with no storage or sharing. FaceAge AI is designed for anyone curious about their facial appearance and how daily habits might affect it, offering a fun and insightful way to understand aging signs.

Domain Specific Seed is a tool designed to streamline the creation of domain-specific datasets within the Hugging Face ecosystem. It automates the setup of essential resources, including dataset repositories and configuration spaces, making it easier for users to initiate new data projects. By providing a project name and Hugging Face user details, the tool facilitates the initial groundwork for data labeling and annotation tasks. This helps users quickly get started with building specialized datasets for various AI applications, leveraging the collaborative environment of Hugging Face.

Gretel.ai, now part of NVIDIA, offers advanced synthetic data generation capabilities, specifically tailored for agentic AI. The platform allows users to build Synthetic Data Generation (SDG) pipelines to support various AI applications, including conversational AI, benchmarks, and agentic AI workflows. It leverages NVIDIA's NeMo synthetic data tools, providing a robust solution for creating high-quality, privacy-preserving synthetic datasets. This is crucial for developing and testing AI models without relying on sensitive real-world data, accelerating AI development and deployment across diverse industries.

Raiinmaker specializes in providing high-quality data services for training and evaluating AI video models. The platform leverages a global network of over 300,000 human contributors across 190 countries to deliver real-time human feedback and natively captured video data. This data is ethically sourced, rights-cleared, and meta-data rich, ensuring compliance and scalability without legal risks. Raiinmaker offers custom data pipelines to meet specific model requirements, including objects, scenes, behaviors, and edge cases. It also provides real-time feedback loops for rapid iteration and improvement of AI models, supporting both LLMs with video-grounded context and next-gen vision models. The service includes detailed evaluation of generative AI video models through user feedback and task-based testing.

Vision Transformer is an AI tool hosted on Hugging Face Spaces, designed for image analysis tasks. It leverages the Vision Transformer model, which is known for its effectiveness in processing visual data. Users can explore capabilities such as image classification and object detection. The tool is currently experiencing a runtime error related to model loading, indicating it is under maintenance or experiencing high demand. Despite the current issue, it is intended to be a free resource, suitable for educational purposes and AI research, allowing users to experiment with advanced computer vision models.

IVARE specializes in developing and deploying proprietary Artificial Intelligence solutions tailored for both industrial and governmental sectors. The company's core offerings include advanced computer vision, facial recognition, biometrics, and LPR (License Plate Recognition) technologies. IVARE's solutions are designed to improve security, optimize industrial processes, and drive innovation across various domains such as public safety (Police 4.0), agribusiness, and smart cities. They provide real-time intelligence for analytics, predictions, and automation, helping organizations achieve greater efficiency, reduce costs, and scale operations. Notable applications include security for police and armed forces, industrial innovation, and smart solutions for condominiums and agribusiness.

mrc-for-flat-nested-ner is an open-source tool implementing a unified Machine Reading Comprehension (MRC) framework for named entity recognition (NER). Based on research presented at ACL 2020, this tool is designed for researchers and practitioners in natural language processing. It provides code, scripts, and data files for fine-tuning BERT models and treating NER as a sequence labeling task. The framework supports both flat and nested NER tasks, offering utilities to transform BMES NER annotations to MRC-format for flat NER and start-end NER annotations to MRC-format for nested NER. It leverages pytorch-lightning for training procedures and includes scripts for reproducing experimental results.

FaceShapeDetector is an online AI tool designed to accurately detect unique face shapes in seconds. Users can upload a well-lit, front-facing photo, and the AI analyzes key facial features like forehead width, cheekbone curve, jawline angle, and chin shape. The tool then compares these measurements to common shapes such as oval, round, square, heart-shaped, oblong, and diamond, providing a detailed report. This analysis helps users understand their unique facial structure, guiding them to choose the best hairstyles, makeup, and style options. It's perfect for personalized beauty and style decisions, offering a simple and fast process for both men and women.

OGB (Open Graph Benchmark) offers a comprehensive suite of benchmark datasets, data loaders, and evaluators specifically designed for graph machine learning. It supports a wide array of graph ML tasks, including predictions at the node, link, and graph levels, and covers diverse real-world applications. The platform provides datasets of varying scales, from those processable on a single GPU to large-scale graphs requiring advanced techniques. OGB's data loaders are fully compatible with leading graph deep learning frameworks like PyTorch Geometric and Deep Graph Library (DGL), offering automatic dataset downloading, standardized splits, and unified performance evaluation. This ensures reliable comparison of different methods and facilitates research in graph machine learning.

CollaNote is a top free note-taking app designed for students, creators, and planners, available on iPad, Mac, and iPhone. It offers an extensive range of features, including 25 pens and brushes for precise handwritten notes and sketches. The app integrates AI-powered tools for smart assistance, digital planners, and flashcards with audio and visuals to boost learning and retention. Users can choose from over 20 digital paper types, utilize templates and stickers, and organize their notes effectively. CollaNote aims to transform ideas into stunning notes, making smart note-taking easy and creative.

refinery is an open-source tool designed for data scientists to effectively manage and improve natural language data for NLP projects. It addresses common challenges such as insufficient labeled data, disorganized training data, and limited resources for annotation. The tool facilitates a data-centric approach to building NLP models, offering features like semi-automated labeling, identification of low-quality data subsets, and data monitoring. It integrates with state-of-the-art libraries like Hugging Face and spaCy, and supports neural search with Qdrant. refinery aims to make training data building a programmatic and enjoyable task, providing capabilities for extensive data management, monitoring, and team collaboration in its managed version.

GeoSpy is an AI-powered image intelligence tool designed for precise photo geolocation. It analyzes visual features within images, such as landmarks and architecture, to estimate their exact location. This tool is specifically built to support law enforcement, government agencies, and enterprise teams in their investigative efforts. GeoSpy provides users with accurate coordinates, detailed region descriptions, and interactive maps, making it an invaluable asset for OSINT investigations, journalism, and disaster response scenarios where rapid and accurate location identification is critical.

Kebab is a privacy-first mobile application designed to help users track and value their charitable donations, particularly non-cash contributions, for tax deductions. It leverages AI to provide fair market values based on real eBay sold listings, offering a significant advantage over tools using outdated valuation tables. The app is built to address the new 2026 tax rules, making accurate record-keeping more crucial than ever. Key features include barcode scanning, photo recognition for quick entry, and an industry-first 'Import from Photo' capability to digitize handwritten lists or thrift store receipts. Kebab ensures data privacy by storing all records directly on the user's device, and it supports various export formats like CSV, TXF for TurboTax/H&R Block Desktop, and pre-filled IRS Form 8283 for Premium users.

BodyFatEstimator.ai is an AI-powered tool designed to estimate body fat percentage directly from a user-uploaded photo. By analyzing visual cues such as body shape, fat distribution, and proportions, the AI provides a photo-based estimate. This tool is particularly useful for tracking changes in body composition over time, offering a simple and repeatable method without the need for tape measurements or complex formulas. Users can upload full-body photos, ideally taken with neutral posture, even lighting, and fitted clothing for optimal accuracy. While not a medical diagnostic tool like DEXA, it serves as an effective trend-tracking utility for fitness enthusiasts.

EfficientTAM is an AI tool designed for efficient object tracking within videos. Users can upload a video and then select specific points to initiate the segmentation and tracking process. The tool offers flexibility with two distinct tracking levels: coarse and fine, allowing for varying degrees of precision based on user needs. The output can be either detailed masks of the tracked objects or a fully masked video, making it suitable for various applications requiring object isolation and motion analysis. Built with Gradio, EfficientTAM provides an accessible interface for video analysis and is available under the Apache-2.0 license.

Documents To Synthetic QA is an AI tool designed to generate synthetic question-answer pairs from various document types, including text, Markdown, and PDF files. This tool is particularly useful for creating training data for question-answering models and augmenting existing datasets. Users can upload their documents, which are then processed into manageable chunks. The platform provides conformance and quality ratings for the generated QA pairs, ensuring high-quality output. This makes it an invaluable resource for researchers and educators who need to enhance their QA resources and build robust AI models.

GroundingDINO ⚔ OWL is an AI-powered object detection tool available as a Hugging Face Space. Users can upload an image and provide text queries to specify the objects they wish to find. The application then processes the image and highlights the identified objects, allowing for adjustments based on confidence thresholds. This tool is designed for tasks requiring precise object localization and identification within visual data. It is suitable for various applications, including research, development, and educational purposes, offering a straightforward interface for visual object recognition.