hitsz-ids / synthetic-data-generator

SDG is a specialized framework designed to generate high-quality structured tabular data.
Apache License 2.0
3.27k stars 545 forks source link

LaTable implementation for enhanced tabular data generation #195

Open jalr4ever opened 4 months ago

jalr4ever commented 4 months ago

🚅Search before asking

I have searched for issues similar to this one.

🚅Description

Based on the benchmark results presented in the paper [LaTable: Towards Large Tabular Models](https://arxiv.org/abs/2406.17673), LaTable demonstrates superior overall performance compared to CTGAN. Implementing LaTable in practical engineering settings could enable the processing of larger-scale tabular data and the generation of higher-quality tabular data.

This approach leverages deep learning techniques to improve both the efficiency and effectiveness of tabular data handling, potentially addressing the limitations seen with existing models like CTGAN.

image

ducdinhchu commented 2 months ago

good find!