概述
Common Voice 15 是由 Mozilla 收集的一个大规模多语言语音语料库,涵盖了 114 种语言,包含来自全球志愿者的数千小时经过验证的语音数据。
任务描述
主要特点
评估说明
en
zh-CN
fr
input_audio
属性
数据统计
各子集统计:
音频统计:
样例示例
子集: en
{ "input": [ { "id": "88959854", "content": [ { "text": "Please recognize the speech and only output the recognized content:" }, { "audio": "[BASE64_AUDIO: mp3, ~37.0KB]", "format": "mp3" } ] } ], "target": "Joe Keaton disapproved of films, and Buster also had reservations about the medium.", "id": 0, "group_id": 0, "subset_key": "en", "metadata": { "locale": "en", "path": "/home/tiger/.cache/huggingface/datasets/downloads/extracted/f54628fae82dd952031cdea3ec9c3d600c11d606e00cb8b3fd1b6ad500d7eb23/en_test_0/common_voice_en_27710027.mp3", "lang_id": "en" } }
提示模板
提示模板:
Please recognize the speech and only output the recognized content:
使用方法
使用 CLI
evalscope eval \ --model YOUR_MODEL \ --api-url OPENAI_API_COMPAT_URL \ --api-key EMPTY_TOKEN \ --datasets common_voice_15 \ --limit 10 # Remove this line for formal evaluation
使用 Python
from evalscope import run_task from evalscope.config import TaskConfig task_cfg = TaskConfig( model='YOUR_MODEL', api_url='OPENAI_API_COMPAT_URL', api_key='EMPTY_TOKEN', datasets=['common_voice_15'], dataset_args={ 'common_voice_15': { # subset_list: ['en', 'zh-CN', 'fr'] # optional, evaluate specific subsets } }, limit=10, # Remove this line for formal evaluation ) run_task(task_cfg=task_cfg)