OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Evaluation & Observability

Sdgym

SDGym is a Python benchmarking framework for modeling and generating synthetic data, with support for comparing synthesizers, datasets and evaluation metrics.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Sdgym — OpenViST project profile in AI Evaluation & Observability.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use SDGym when you need a structured comparison of synthetic-data methods on public or custom datasets, including results for performance, memory and quality.

Who it is for

  • Data scientists comparing synthetic-data generation methods.
  • Machine-learning practitioners who want to include a custom synthesizer in an evaluation.
  • Teams evaluating synthetic-data performance, memory use, quality or privacy.

Documented uses

  • Compare selected synthesizers and baselines across datasets.
  • Measure performance and memory usage for synthetic-data methods.
  • Evaluate synthetic-data quality and privacy using available metrics.
  • Bring a custom synthesizer into a benchmark by defining its training and sampling logic.
Before you choose

Requirements & limits

The README documents installation with pip or conda and recommends a virtual environment.

  • The documented example benchmarks single tables.
  • The framework supports SDV synthesizers and baselines as well as custom synthesizers.
  • Datasets can come from the SDV public collection or be supplied as custom data, including private data stored locally or on Amazon S3.
  • Evaluation can include performance, memory usage, synthetic-data quality and privacy.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Compare selected single-table synthesizers

Benchmark SDV synthesizers and an SDGym baseline on publicly available datasets.

Before you start

  • SDGym installed in a Python environment.
  • Selected synthesizer names; the README example uses GaussianCopulaSynthesizer, CTGANSynthesizer and UniformSynthesizer.

Steps

  1. Define the selected SDV synthesizers and SDGym baseline in Python lists.

  2. Import sdgym.

  3. Call sdgym.benchmark_single_table with the combined synthesizer list.

Expected outcome

The example returns performance, memory and quality evaluation across the selected synthesizers on publicly available datasets.

Guide limitations

  • The documented example covers single-table benchmarking and does not provide a complete specification of benchmark procedures.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
18 October 2026
Original website title
sdv-dev/SDGym · Official homepage title · retained in its source language
How we review sourcesSuggest a correction