Skip to content

Dataset · hlt-lab

LLM-powered voice assistant benchmark

hlt-lab/voicebench

English audio and text data for evaluating voice language models.

What's inside

The dataset contains between 10K and 100K English audio and text examples. Released under a permissive Apache 2.0 license as part of research on testing LLM-powered voice assistants.

What you can do with it

  • Comprehensive evaluation of end-to-end voice models and audio LLMs
  • Comparing speech comprehension and response generation quality in voice assistants
Languages
English
Size
10K–100K rows
Downloads this month
8.3K
Likes
18

How to load it

from datasets import load_dataset

ds = load_dataset("hlt-lab/voicebench")