Engineering Datasets Hub
Curated, indexed open datasets from UCI Machine Learning Repository, Kaggle, and Hugging Face. Filter by domain, format, or architecture, and download directly for production training and benchmarks.
GSM8K (Grade School Math 8K Reasoning)
A dataset of 8,500 high-quality linguistically diverse grade school math word problems created by human problem writers. Essential benchmark for LLM chain-of-thought, reasoning steps, and mathematical problem-solving.
SQuAD 2.0 (Stanford Question Answering Dataset)
Stanford Question Answering Dataset (SQuAD) combines 100,000+ questions from SQuAD 1.1 with over 50,000 unanswerable questions written adversarially by crowdworkers to look similar to answerable ones.
MS COCO (Common Objects in Context)
Large-scale object detection, segmentation, and captioning dataset with 330,000 images, 1.5 million object instances, and 80 object categories. Foundational dataset for vision-language models and YOLO architectures.
Titanic: Machine Learning from Disaster
The definitive beginner-to-intermediate machine learning dataset. Use passenger data (name, age, fare, class, cabin) to build a predictive model predicting survival probabilities after the Titanic disaster.
IMDb Movie Reviews Large Sentiment Corpus
Large Movie Review Dataset consisting of 50,000 reviews for binary sentiment classification (25,000 training, 25,000 test). The standard baseline benchmark for text classification and fine-tuning encoder models.
UCI Iris Flower Dataset
The famous 1936 Fisher Iris dataset containing 150 instances of Iris setosa, versicolor, and virginica across 4 attributes (sepal length, sepal width, petal length, petal width). Ideal for clustering and multiclass validation.
Mozilla Common Voice Speech Corpus
Massive multilingual transcribed speech audio dataset with thousands of hours of speech across 100+ languages recorded by global volunteers. Standard dataset for Whisper, Wav2Vec2, and ASR model training.
House Prices: Advanced Regression Techniques
79 explanatory variables describing almost every aspect of residential homes in Ames, Iowa. The gold standard for feature engineering, gradient boosting (XGBoost, LightGBM, CatBoost), and stacking techniques.
Credit Card Fraud Detection Benchmark
Highly imbalanced dataset of European cardholder transactions in September 2013 containing 284,807 transactions with 492 frauds. Contains PCA-transformed numerical features (V1-V28). Perfect for anomaly detection.
UCI Breast Cancer Wisconsin (Diagnostic)
Features computed from a digitized image of a fine needle aspirate (FNA) of a breast mass describing characteristics of cell nuclei. Widely used for medical diagnosis and high-precision classification models.
UCI Adult Census Income Dataset
Predict whether income exceeds $50K/year based on census data. Extraction was done by Barry Becker from the 1994 Census database. Classic benchmark for tabular binary classification and algorithmic fairness.
UCI Wine Quality Dataset
Physicochemical laboratory tests on red and white Portuguese "Vinho Verde" wine variants with sensory quality scores. High-dimensional continuous feature regression and classification.
Global Crop Yields & Climate Telemetry (1990-2024)
Aggregated agricultural telemetry covering rain precipitations, temperature variations, pesticide usage, and crop productivity across 101 countries. Excellent for time-series forecasting and climate modeling.