Dataset

A structured archive of datasets for data analysis and research, offering aggregated and cleaned data resources.

Deduplicated subset of vivym/midjourney-prompts (Hugging Face dataset). Prompts collected from the public Midjourney Discord channels.

Collect and upload conversations for chat-bot dataset.

A collaborative workspace for managers, engineers and data experts to develop AI features.

An open-source repository archiving The Economist articles from 2020, providing structured news data for analysis and research.

Extension for automatically collecting and sharing chatbot conversations for open-source use.

Get Accurate and Complete Datasets Ready-to-Use

Facial biometric datasets for AI projects.

AI dataset crowdsourcing platform

MNIST dataset in CSV format for machine learning and image classification tasks.

SQuAD is a reading comprehension dataset for training and evaluating question-answering AI models, with versions including answerable and unanswerable questions.

A public archive of Git repositories for dataset creation and analysis, supporting code research and open-source projects.

A public dataset for spam email classification, containing 4601 instances with 57 features, used for machine learning tasks.

A benchmark dataset for language modeling experiments with 1 billion words, providing training/test data and baseline model outputs for reproducible research.

A Chinese sentiment dataset for sentiment analysis, derived from human-computer conversation logs and segmented using Jieba.

A sentiment analysis tool and dataset from Stanford, using deep learning for semantic compositionality, integrated into CoreNLP.

A public dataset of Google Trends data in CSV format, available on GitHub for analysis and research.

A dataset of Amazon product co-purchasing networks and reviews from 2006, useful for research and analysis.

A dataset for Chinese abbreviation prediction, including negative full forms, released for academic research in NLP.

A Chinese reading comprehension dataset with cloze-style queries, including People Daily and Children's Fairy Tale data for NLP research.

CIFAR-10 and CIFAR-100 are labeled image datasets for computer vision and machine learning, with 60,000 32x32 color images in 10 or 100 classes.

A dataset from HetRec 2011 containing 92,800 artist listening records from 1,892 users for music recommendation research.

A Chinese names corpus with over 56 million entries, used for NLP tasks like name recognition and generation.

A large dataset for binary sentiment classification with 50,000 polar movie reviews, used for training and testing in NLP and machine learning.