MulTaBench
Collection
Benchmarking multimodal tabular learning with text and images (NeurIPS 2026 Spotlight): leaderboard, 80 datasets and cached embeddings. • 84 items • Updated • 1
Error code: FileSystemError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Fruit No string | Color_K-Yellow_P_Green string | Fruit Grade string | Mass(kg) float64 |
|---|---|---|---|
images/1a.jpg | P | 2 | 0.5 |
images/2a.jpg | P | 2 | 0.5 |
images/3a.jpg | P | 2 | 0.35 |
images/4a.jpg | P | 2 | 0.45 |
images/5a.jpg | P | 2 | 0.5 |
images/6a.jpg | P | 2 | 0.4 |
images/7a.jpg | P | 2 | 0.7 |
images/8a.jpg | P | 2 | 0.45 |
images/9a.jpg | P | 2 | 0.3 |
images/10a.jpg | P | 2 | 0.3 |
images/11a.jpg | P | 2 | 0.4 |
images/12a.jpg | P | 2 | 0.35 |
images/13a.jpg | P | 2 | 0.3 |
images/14a.jpg | P | 2 | 0.5 |
images/15a.jpg | P | 2 | 0.5 |
images/16a.jpg | P | 2 | 0.3 |
images/17a.jpg | P | 2 | 0.35 |
images/18a.jpg | P | 2 | 0.35 |
images/19a.jpg | P | 2 | 0.5 |
images/20a.jpg | P | 2 | 0.4 |
images/21a.jpg | P | 2 | 0.4 |
images/22a.jpg | P | 2 | 0.3 |
images/23a.jpg | P | 2 | 0.35 |
images/24a.jpg | P | 2 | 0.4 |
images/25a.jpg | P | 2 | 0.35 |
images/26a.jpg | P | 2 | 0.3 |
images/27a.jpg | P | 2 | 0.35 |
images/28a.jpg | P | 2 | 0.35 |
images/29a.jpg | P | 2 | 0.4 |
images/30a.jpg | P | 2 | 0.35 |
images/31a.jpg | P | 2 | 0.4 |
images/32a.jpg | P | 2 | 0.25 |
images/33a.jpg | P | 2 | 0.35 |
images/34a.jpg | P | 2 | 0.3 |
images/35a.jpg | P | 2 | 0.3 |
images/36a.jpg | P | 2 | 0.4 |
images/37a.jpg | P | 2 | 0.45 |
images/38a.jpg | P | 2 | 0.4 |
images/39a.jpg | P | 2 | 0.35 |
images/40a.jpg | P | 2 | 0.4 |
images/41a.jpg | P | 2 | 0.35 |
images/42a.jpg | P | 2 | 0.5 |
images/43a.jpg | P | 2 | 0.4 |
images/44a.jpg | P | 2 | 0.3 |
images/45a.jpg | P | 2 | 0.55 |
images/46a.jpg | P | 2 | 0.35 |
images/47a.jpg | P | 2 | 0.45 |
images/48a.jpg | P | 2 | 0.45 |
images/49a.jpg | P | 2 | 0.35 |
images/50a.jpg | P | 2 | 0.5 |
images/51a.jpg | P | 2 | 0.5 |
images/52a.jpg | P | 2 | 0.35 |
images/53a.jpg | P | 2 | 0.4 |
images/54a.jpg | P | 2 | 0.45 |
images/55a.jpg | P | 2 | 0.3 |
images/56a.jpg | P | 2 | 0.5 |
images/57a.jpg | P | 2 | 0.35 |
images/58a.jpg | P | 2 | 0.45 |
images/59a.jpg | P | 2 | 0.45 |
images/60a.jpg | P | 2 | 0.4 |
images/61a.jpg | P | 2 | 0.5 |
images/62a.jpg | P | 2 | 0.5 |
images/63a.jpg | K | 2 | 0.45 |
images/64a.jpg | K | 2 | 0.5 |
images/65a.jpg | K | 2 | 0.45 |
images/66a.jpg | K | 2 | 0.5 |
images/67a.jpg | K | 2 | 0.45 |
images/68a.jpg | K | 2 | 0.45 |
images/69a.jpg | K | 2 | 0.4 |
images/70a.jpg | K | 2 | 0.4 |
images/71a.jpg | K | 2 | 0.5 |
images/72a.jpg | K | 2 | 0.45 |
images/73a.jpg | K | 2 | 0.45 |
images/74a.jpg | K | 2 | 0.55 |
images/75a.jpg | K | 2 | 0.6 |
images/76a.jpg | K | 2 | 0.3 |
images/77a.jpg | K | 2 | 0.4 |
images/78a.jpg | K | 2 | 0.3 |
images/79a.jpg | K | 2 | 0.3 |
images/80a.jpg | K | 2 | 0.5 |
images/81a.jpg | K | 2 | 0.51 |
images/82a.jpg | K | 2 | 0.52 |
images/83a.jpg | K | 2 | 0.65 |
images/84a.jpg | K | 2 | 0.41 |
images/85a.jpg | K | 2 | 0.53 |
images/86a.jpg | K | 2 | 0.6 |
images/87a.jpg | K | 2 | 0.45 |
images/88a.jpg | K | 2 | 0.45 |
images/89a.jpg | K | 2 | 0.58 |
images/90a.jpg | K | 2 | 0.44 |
images/91a.jpg | K | 2 | 0.38 |
images/92a.jpg | K | 2 | 0.52 |
images/93a.jpg | K | 2 | 0.58 |
images/94a.jpg | K | 2 | 0.57 |
images/95a.jpg | K | 2 | 0.49 |
images/96a.jpg | K | 2 | 0.53 |
images/97a.jpg | K | 2 | 0.55 |
images/98a.jpg | K | 2 | 0.61 |
images/99a.jpg | K | 2 | 0.67 |
images/100a.jpg | K | 2 | 0.54 |
Part of MulTaBench, a benchmark for multimodal tabular learning.
Paper: MulTaBench: Benchmarking Multimodal Tabular Learning with Text and Image · Leaderboard: multabench/leaderboard
Mass(kg)Load the benchmark splits with multabench.load_split("REG_IMAGE_MANGO_MASS", fold=0).
| Column | Type |
|---|---|
| Fruit No | image path (inside images-*.zip) |
| Color_K-Yellow_P_Green | categorical |
| Fruit Grade | categorical |
| Mass(kg) | float |
@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},
}