REaLTabFormer

Input—per 1M tokens

Output—per 1M tokens

Context—tokens

WeightsClosed

About

REaLTabFormer is ranked #11 of 23 in AI Synthetic Data Generators on Inferse. It runs on Linux, macOS, Self-hosted, Windows. There is a free plan.

Compared on AI Synthetic Data Generators

Deployment: self_hostedgithub.com
Relational data: Yesgithub.com
Unstructured data: Nogithub.com
Privacy-risk metrics: Yesgithub.com

Facts

Purpose: REaLTabFormer is a unified framework for synthesizing different types of tabular data.github.com · 1 Oct 2026
Relational generation: It uses a sequence-to-sequence model to generate synthetic relational datasets.github.com · 1 Oct 2026
Tabular model: Its non-relational tabular model uses GPT-2 and can model tabular data with independent observations out of the box.github.com · 1 Oct 2026
Installation: The package is installed from PyPI with pip install realtabformer.github.com · 1 Oct 2026
Python requirement: The current PyPI package requires Python 3.8 or newer.pypi.org · 1 Oct 2026
Operating systems: PyPI classifies the package as operating-system independent.pypi.org · 1 Oct 2026
Input format: Examples use pandas DataFrames as model input.github.com · 1 Oct 2026
Relational keys: Relational generation requires matching join-key columns in the parent and child tables.github.com · 1 Oct 2026
Stopping criterion: For non-relational tabular training, the model stops when the synthetic distribution is close to the real distribution.github.com · 1 Oct 2026
Validation: The framework provides observation validators, including a GeoValidator for filtering invalid synthetic samples.github.com · 1 Oct 2026
Privacy-oriented design: The paper says target masking is used to prevent data copying and the Qδ statistic with statistical bootstrapping is used to detect overfitting.arxiv.org · 1 Oct 2026
License: The package is distributed under the MIT License.pypi.org · 1 Oct 2026
Release: PyPI lists version 0.2.4 as released on January 4, 2026.pypi.org · 1 Oct 2026
Funding: The project acknowledges funding from the World Bank-UNHCR Joint Data Center on Forced Displacement.pypi.org · 1 Oct 2026
Purpose: REaLTabFormer is a framework for generating synthetic tabular and relational data with transformer models.github.com · 2 Oct 2026
Tabular model: For independent tabular observations, it uses GPT-2 and can model data out of the box.github.com · 2 Oct 2026
Relational model: A sequence-to-sequence model generates synthetic relational datasets.github.com · 2 Oct 2026
Sampling: The documented workflow fits a model, saves it locally, and samples synthetic data from it.github.com · 2 Oct 2026
Training behavior: For non-relational tabular models, training stops when the synthetic data distribution is close to the real data distribution.worldbank.github.io · 2 Oct 2026
Data validation: The framework provides an interface for observation validators that filter invalid synthetic samples, including a GeoValidator example.worldbank.github.io · 2 Oct 2026
Installation: The project is available through PyPI and documents installation with pip for Python 3.7 or later.github.com · 2 Oct 2026
License: The repository provides the software under the MIT License, which permits use, modification, distribution, sublicensing, and sale subject to its terms.github.com · 2 Oct 2026
Security reporting: The security policy asks users to report vulnerabilities by email rather than through public GitHub issues and says a response should arrive within 48 hours.github.com · 2 Oct 2026
Support: For vulnerability reports, the policy lists [email protected] and requests details that help reproduce and assess the issue.github.com · 2 Oct 2026
Documented audience: The project describes its use for projects or research and asks users to cite its research paper when using it.worldbank.github.io · 2 Oct 2026
Development context: The project acknowledges funding from the World Bank-UNHCR Joint Data Center on Forced Displacement for work involving responsible microdata access and synthetic population research.github.com · 2 Oct 2026