Question 1

What are Multi-Modal datasets used for?

Accepted Answer

Multi-Modal datasets are collections of labelled or raw data used to train, fine-tune, and evaluate models on the multi-modal task. This page lists 5 such datasets, each linking to its source and paper.

Question 2

Which Multi-Modal dataset is best for benchmarking?

Accepted Answer

None of the listed Multi-Modal datasets are currently tracked as standard LLM benchmarks, but many are widely used for evaluation.

Question 3

How many Multi-Modal datasets are there?

Accepted Answer

We catalog 5 Multi-Modal datasets in one searchable directory.

Multi-Modal Datasets

What languages do multi-modal datasets cover?

Explore other dataset tasks

Frequently asked questions