Question 1

What are Speech Translation datasets used for?

Accepted Answer

Speech Translation datasets are collections of labelled or raw data used to train, fine-tune, and evaluate models on the speech translation task. This page lists 4 such datasets, each linking to its source and paper.

Question 2

Which Speech Translation dataset is best for benchmarking?

Accepted Answer

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

Question 3

How many Speech Translation datasets are there?

Accepted Answer

We catalog 4 Speech Translation datasets in one searchable directory.

Speech Translation Datasets

What languages do speech translation datasets cover?

Explore other dataset tasks

Frequently asked questions