multabench/core-text-reg-video-games-sales

REG_TEXT_VIDEO_GAMES_SALES

Part of MulTaBench, a benchmark for multimodal tabular learning.

Paper: MulTaBench: Benchmarking Multimodal Tabular Learning with Text and Image · Leaderboard: multabench/leaderboard

• Target: Global_Sales
• Rows: 16598
• Original source: https://www.kaggle.com/datasets/gregorut/videogamesales

Load the benchmark splits with multabench.load_split("REG_TEXT_VIDEO_GAMES_SALES", fold=0).

Column	Type
Name	string
Platform	categorical
Year	float
Genre	categorical
Publisher	string
Global_Sales	float

Citation

@misc{arazi2026multabenchbenchmarkingmultimodaltabular,
  title={MulTaBench: Benchmarking Multimodal Tabular Learning with Text and Image},
  author={Alan Arazi and Eilam Shapira and Shoham Grunblat and Mor Ventura and Elad Hoffer and Gioia Blayer and David Holzmüller and Lennart Purucker and Gaël Varoquaux and Frank Hutter and Roi Reichart},
  year={2026},
  eprint={2605.10616},
  archivePrefix={arXiv},
  primaryClass={cs.LG},
  url={https://arxiv.org/abs/2605.10616},
}