KERNEL: ONLINE
3-DAY STREAK|350 XP (LVL 2)
CURATED CATALOG // 29 PINNED DATASET BLUEPRINTS & 20 LIBRARIES

DATASET EXPLORER & SKELETON DIRECTORY

Explore ready-to-train datasets with automated PyTorch DataLoaders, feature statistics, class distributions, and 7-step guided workspace pipelines.

SHOWING 29 OF 29 DATASETS
Tabular ML891 SAMPLES

Titanic: Machine Learning from Disaster

Predict passenger survival on the Titanic using passenger class, sex, age, ticket fare, cabin, and family size with PyTorch Multi-Layer Perceptron and ROC-AUC evaluation.

Target:Survived (0 or 1)
Features:12
Task Type:binary classification
#tabular#classification#pytorch#pandas
Computer Vision70,000 SAMPLES

MNIST Handwritten Digits

The standard computer vision benchmark dataset of handwritten digits (0-9) to master PyTorch Convolutional Neural Networks (Conv2d, MaxPool2d, BatchNorm2d).

Target:Digit (0-9)
Features:28x28 grayscale images
Task Type:multiclass classification
#vision#cnn#pytorch#torchvision
Computer Vision60,000 SAMPLES

CIFAR-10 Object Recognition

Natural RGB image recognition benchmark dataset for training deep residual networks (ResNet-18) and real-time image augmentation pipelines in PyTorch.

Target:10 Classes (airplane, car, bird, cat, deer, dog, frog, horse, ship, truck)
Features:32x32x3 RGB images
Task Type:multiclass classification
#vision#deep-learning#resnet#torchvision
Tabular ML20,640 SAMPLES

California Housing Prices

Predict median district housing values across California based on median income, housing age, average rooms, population, and latitude/longitude coordinates.

Target:MedHouseVal (Continuous $100k)
Features:8
Task Type:regression
#tabular#regression#pytorch#pandas
NLP / Sequential50,000 SAMPLES

IMDb Large Movie Review Dataset

Binary natural language sentiment classification benchmark with 25,000 training and 25,000 testing movie reviews using PyTorch word embeddings and Bidirectional LSTMs.

Target:Sentiment (Positive / Negative)
Features:Text reviews
Task Type:binary classification
#nlp#sentiment#lstm#pytorch
Industrial & Anomaly284,807 SAMPLES

Credit Card Fraud Detection

Highly imbalanced anonymized financial transactions requiring Focal Loss, weighted cross-entropy, Precision-Recall AUC tracking, and anomaly detection in PyTorch.

Target:Class (0: Legitimate, 1: Fraud 0.17%)
Features:30
Task Type:imbalanced classification
#tabular#fraud#imbalanced#focal-loss
Computer Vision70,000 SAMPLES

Fashion-MNIST Clothing Classification

A drop-in replacement for MNIST containing 10 categories of Zalando clothing articles with higher intra-class visual complexity and spatial textures.

Target:10 Apparel Classes (T-shirt, Trouser, Pullover, Dress, Coat, Sandal, Shirt, Sneaker, Bag, Ankle Boot)
Features:28x28 grayscale images
Task Type:multiclass classification
#vision#fashion#cnn#pytorch
Tabular ML6,497 SAMPLES

Wine Quality (Red & White)

Physicochemical laboratory tests (fixed acidity, volatile acidity, citric acid, residual sugar, chlorides, alcohol) to predict wine sensory quality scores.

Target:Quality Score (0-10)
Features:11
Task Type:regression
#tabular#regression#chemistry
Medical & Bio303 SAMPLES

Heart Disease UCI (Cleveland)

Clinical cardiology database to predict coronary artery disease presence from resting blood pressure, cholesterol, fasting blood sugar, and exercise ST-depression.

Target:Heart Disease Present (0 or 1)
Features:14
Task Type:binary classification
#medical#tabular#healthcare
Medical & Bio569 SAMPLES

Breast Cancer Wisconsin Diagnostic

Computed nuclear morphometry features from fine needle aspirate (FNA) biopsy images (radius, texture, perimeter, area, smoothness, concavity, symmetry).

Target:Diagnosis (Malignant vs Benign)
Features:30
Task Type:binary classification
#medical#binary-classification#pytorch
Tabular ML150 SAMPLES

Iris Flower Dataset

The classic multi-class pattern recognition dataset for learning PyTorch softmax distributions, one-hot target encoding, and multi-class decision boundaries.

Target:Species (Setosa, Versicolour, Virginica)
Features:4
Task Type:multiclass classification
#starter#multiclass#pytorch
NLP / Sequential21,578 SAMPLES

Reuters-21578 News Wire Topics

Multi-label text classification of financial and economic news articles with class imbalances using PyTorch tokenization and BCEWithLogitsLoss.

Target:Topics (Money, Grain, Crude, Earn, etc.)
Features:Financial News Articles
Task Type:multilabel classification
#nlp#multilabel#finance
Time Series17,379 SAMPLES

Bike Sharing Demand

Forecast hourly bike rental demand using weather, seasonality, holiday indicators, and cyclical temporal encodings (sin/cos hour transformations).

Target:Rental Count (Hourly)
Features:16
Task Type:time series_regression
#time-series#regression#forecasting
Tabular ML344 SAMPLES

Palmer Archipelago Penguins

Modern tabular dataset featuring culmen dimensions, flipper length, body mass, and island origins with missing value imputation and categorical encoding in PyTorch.

Target:Species (Adelie, Chinstrap, Gentoo)
Features:7
Task Type:multiclass classification
#tabular#eda#multiclass
NLP / Sequential5,574 SAMPLES

SMS Spam Collection

Detect fraudulent SMS spam messages using subword tokenization, vocabulary mapping, PyTorch EmbeddingBag, and imbalanced cross-entropy loss.

Target:Spam vs Ham
Features:Mobile SMS text messages
Task Type:binary classification
#nlp#text#binary-classification
Tabular ML53,940 SAMPLES

Diamonds Dataset

Predict the price of round cut diamonds based on carat weight, cut quality, color grade, clarity rating, and physical spatial dimensions (x, y, z).

Target:Price ($326 - $18,823)
Features:10
Task Type:regression
#tabular#regression#embeddings
Tabular ML45,211 SAMPLES

Bank Marketing (Direct Telemarketing)

Predict whether a retail banking client will subscribe a long-term deposit based on campaign contact history and macroeconomic financial indicators.

Target:Deposit Subscribed (Yes/No)
Features:17
Task Type:binary classification
#tabular#marketing#pytorch
Computer Vision20,000 SAMPLES

CelebA Facial Attributes Mini

Predict multiple facial binary attributes simultaneously using deep multi-task convolutional neural network architectures in PyTorch.

Target:40 Facial Attributes (Smiling, Eyeglasses, Young, etc.)
Features:Aligned Face Crops
Task Type:multilabel vision
#vision#multilabel#faces
Industrial & Anomaly1,941 SAMPLES

Steel Plates Faults (UCI)

Classify surface defects and structural faults in stainless steel plates using 27 geometric, thickness, and luminosity sensor measurements.

Target:7 Fault Types (Pastry, Z_Scratch, K_Scatch, Stains, Dirtiness, Bumps, Other_Faults)
Features:27
Task Type:multiclass classification
#industrial#quality-control#pytorch
Computer Vision5,000 SAMPLES

COCO Mini Object Detection & BBoxes

Hands-on PyTorch bounding box coordinate regression, Complete IoU (CIoU) loss calculation, and non-maximum suppression (NMS) evaluation.

Target:Classes + Bounding Box Coordinates [x, y, w, h]
Features:Natural Scene Images with Bounding Boxes
Task Type:object detection
#vision#detection#bounding-box
NLP / Sequential120,000 SAMPLES

AG News 4-Class Topic Classification

Academic benchmark for multi-class news article topic modeling with word tokenization, PyTorch BiLSTM, 1D-CNN, and Macro-F1 evaluation.

Target:Topic: World (0), Sports (1), Business (2), Sci/Tech (3)
Features:News article headlines & description text
Task Type:multiclass classification
#nlp#topics#lstm#pytorch
NLP / Sequential10,662 SAMPLES

Rotten Tomatoes Critic Reviews

Sentence-level movie review sentiment analysis with nuanced linguistic phrasing, idioms, and negation handling via PyTorch Recurrent Networks.

Target:Sentiment: Rotten (0) vs Fresh (1)
Features:Sentence-level film critic review snippets
Task Type:binary classification
#nlp#sentiment#gru#pytorch
NLP / Sequential20,000 SAMPLES

Emotion Tweet & Text Classifier

Multi-class affective state detection classifying subtle human psychological emotions from informal digital text using PyTorch LSTMs.

Target:6 Emotions: Sadness (0), Joy (1), Love (2), Anger (3), Fear (4), Surprise (5)
Features:Emotion-labeled social text snippets
Task Type:multiclass classification
#nlp#emotion#lstm#deep-learning
Computer Vision7,349 SAMPLES

Oxford-IIIT Pet Breed Classification

Fine-grained visual classification of 37 domestic dog and cat breeds with subtle inter-class visual similarities using transfer learning in PyTorch.

Target:37 Breeds: 25 Dogs, 12 Cats
Features:RGB pet photos with breed annotations
Task Type:multiclass classification
#vision#fine-grained#resnet#torchvision
Medical & Bio12,444 SAMPLES

Medical Blood Cell Microscope Vision

Automated clinical hematology cytology classification for white blood cell differential counts using deep convolutional neural networks in PyTorch.

Target:4 Leukocyte Types: Eosinophil, Lymphocyte, Monocyte, Neutrophil
Features:Microscopic white blood cell images (360x240 RGB)
Task Type:multiclass classification
#vision#medical#cytology#cnn
Computer Vision10,000 SAMPLES

Food-101 Mini Cuisine Image Classifier

Visual culinary dish recognition with color jitter, random crops, and transfer learning with MobileNetV2 in PyTorch.

Target:10 Dishes: Pizza, Burger, Sushi, Ramen, Steak, Tacos, Salad, Cake, Dumplings, Pasta
Features:Food & cuisine RGB photos (256x256)
Task Type:multiclass classification
#vision#food#transfer-learning#torchvision
Text Generation40,000 SAMPLES

Tiny Shakespeare Playwright Generator

Autoregressive sequential language modeling to generate dramatic Elizabethan dialogue using PyTorch Character/Word LSTMs and temperature sampling.

Target:Next Character / Token Prediction (Autoregressive)
Features:Sequential dramatic dialogue lines from Shakespeare plays
Task Type:text generation
#text-generation#rnn#lstm#nlp
Text Generation15,000 SAMPLES

Edgar Allan Poe Gothic Poetry Generator

Autoregressive language model synthesizing dark romantic poetry and gothic prose mimicking Poe's vocabulary and rhythmic cadence in PyTorch.

Target:Next Word Sequence (Autoregressive)
Features:Gothic stanzas, rhythmic cadences, and poetic lines
Task Type:text generation
#text-generation#poetry#lstm#nlp
Text Generation30,000 SAMPLES

Python Code & Docstring Auto-Complete

Causal code language model predicting Python AST expressions, docstrings, and function bodies with PyTorch Recurrent / Causal Transformer models.

Target:Next Code Token Sequence
Features:Python function signatures, docstrings, and implementation blocks
Task Type:text generation
#text-generation#code#transformer#gru