Github repository for paper: VoxEval: Benchmarking the Knowledge Understanding Capabilities of End-to-End Spoken Language Models
Also check out our survey paper at Recent Advances in Speech Language Models: A Survey!
VoxEval is a novel speech question-answering benchmark specifically designed to assess SLMs' knowledge understanding through purely speech-based interactions.
Below are the three highlights of our VoxEval benchmark:
End-to-end speech-based evaluation:… See the full description on the dataset page:
https://huggingface.co/datasets/qqjz/VoxEval.
